ABCG1 is required for pulmonary B-1 B cell and natural antibody homeostasis

Animals

All animals were bred and maintained at UCLA in temperature-controlled, pathogen-free conditions under a 12-hour light/dark cycle. Abcg1^−/− LacZ knock-in mice (Deltagen, (18, 31)) were backcrossed at least 10 times onto a C57BL/6 background. Control C57BL/6 mice (originally purchased from The Jackson Laboratory) were generated from Abcg1^+/− breeding. Mice were fed a chow diet, or a Western diet (Research Diets #D12079B, containing 21% fat and 0.2% cholesterol) where indicated. Mice expressing the green fluorescent protein (GFP) C57BL/6-Tg(CAG-EGFP)1Osb/J were purchased from the Jackson Laboratory (Strain #003291). The Institutional Animal Care and Research Advisory Committee at UCLA approved all experimental protocols.

Adoptive Transfer

Cells were isolated with Ab-tagged magnetic beads and Auto-MACS (Miltenyi Biotec). Peritoneal CD19⁺CD23⁻ B-1 B cells were isolated from C57BL/6-Tg(CAG-EGFP)1Osb/J mice by negative selection on a CD23⁺ column, followed by positive selection of CD19⁺ cells. Cell purity (>98%) was confirmed by FACS analysis using fluorochrome-labeled CD19, CD23 and CD5 Abs (eBioscience). Cell viability (>97%) was assessed by trypan blue exclusion. To obtain 10 × 10⁶ B-1 B cells, a pool of peritoneal cells from 20 donor mice was used, and 1 × 10⁶ B-1 B cells were adoptively transferred into 6 month old chow-fed wildtype and Abcg1^−/− mice.

Surfactant Isolation

Pulmonary surfactant was isolated from 6 month old wildtype and Abcg1^−/− mice by bronchoalveolar lavage as previously described (31). Briefly, tracheas were exposed and canulated before the lungs were flushed 3 times with 1 mL aliquots of BAL buffer (10 mmol/L Tris, 100 mmol/L NaCl, 0.2 mmol/L EGTA, pH 7.2). The aliquots were combined and centrifuged (200 × g, 5 min) to separate surfactant and cells.

Lipid Analyses

Cells, bronchoalveolar lavage fluid (BAL), or lung tissue were snap frozen in liquid nitrogen. Lung tissue was homogenized on ice in phosphate buffered saline. Cell suspensions, BAL fluid, or lung homogenates were subsequently subjected to a modified Bligh-Dyer lipid extraction (32) in the presence of lipid class internal standards including eicosanoic acid, 1-0-heptadecanoyl-sn-glycero-3-phosphocholine, 1,2-dieicosanoyl-sn-glycero-3-phosphocholine, and 1,2-ditetradecanoyl-sn-glycero-3-phosphoethanolamine (33). Fatty acids were converted to their pentafluorobenzyl esters and then were subsequently quantified using GC-MS with negative ion chemical ionization with methane as the reactant gas (34). For phospholipids, lipid extracts were diluted in methanol/chloroform (4/1, v/v) and molecular species were quantified using electrospray ionization mass spectrometry on a triple quadrupole instrument (Themo Fisher Quantum Ultra) employing shotgun lipidomics methodologies (35). Phosphatidylcholine molecular species were quantified as lithiated adducts in the positive ion mode using neutral loss scanning for 59.1 amu (collision energy = −28eV). Individual molecular species were quantified by comparing the ion intensities of the individual molecular species to that of the lipid class internal standard with additional corrections for type I and type II ¹³C isotope effects (35).

Flow Cytometry

Single-cell suspensions from lungs and spleen were depleted of red blood cells using hypotonic lysis. Cells were resuspended in phosphate buffered saline (PBS) with 0.2% bovine serum albumin (BSA) and 0.1% sodium azide (FACS buffer). Single cell suspensions were incubated for 15 min with anti-CD16/32 (Fc block), and stained for 30 min at 4°C. DAPI, anti-mouse CD3 (17A2), CD19 (1D3), CD11b (M1/70), CD5 (53-7.3), and IgM (II/41) were purchased from eBioscience. Cells were analyzed on LSRII (Becton Dickinson). More than 0.5 × 10⁵ cells were analyzed per sample, with dead cells excluded by DAPI positive staining. Surface marker analysis was performed using FlowJo software (Treestar Inc.). B cells (CD19⁺), B-1 B cells (CD19⁺, sIgM⁺, CD11b⁺), B-1a B cells (CD19⁺, sIgM⁺, CD11b⁺, CD5⁺), and B-1b B cells (CD19⁺, sIgM⁺, CD11b⁺, CD5⁻) were identified with the appropriate gating. For gating strategy and controls see Figure 1.

RNA isolation and analysis

Total RNA was isolated from 50 mg lung tissue (tissue weight determined after blotting of excess buffer) using Qiazol (Qiagen). Gene expression profiling of inflammatory markers was determined using GEArray mouse inflammatory cytokines and receptors microarray system (SABiosciences) as previously described (36). Total RNA (0.5 µg) was reverse transcribed with random hexamers using the Taqman Reverse Transcription Reagents Kit (Applied Biosystems). Taqman real-time quantitative PCR assays were performed using an Applied Biosystems 7900HT sequence detector. Real Time qPCR (RT-qPCR) assays were performed on a Lightcycler 480 (Roche). Results show averages of triplicate experiments normalized to GAPDH (Taqman) or 36B4 (RT-qPCR). Primer sequences are available on request.

Measurement of Ab titers

Total and specific antibody titers were determined by chemiluminescent enzyme immunoassays as previously described (37). In brief, capture antigens were coated on plates at 5 µg/mL in PBS overnight at 4°C (AB-12, CuOxLDL, IgM (Goat anti-mouse-IgM; Sigma), IgG (Goat anti-mouse-IgG; Sigma), IgA (Rat anti-mouse-IgA; Sigma), and MDA-LDL). Plates were blocked with 1% BSA in tris-buffered saline (TBS), and serially diluted antisera from individual mice was added. Plates were incubated for 1.5 h at room temperature. Bound plasma immunoglobulin isotype levels were detected with various anti-mouse immunoglobulin isotype-specific alkaline phosphatase (AP) conjugates using LumiPhos 530 (Lumigen, Southfield, MI, USA) solution, and a Dynex Luminometer (Dynex Technologies, Chantilly, VA, USA). Several secondary antibodies were used at dilutions of 1:30,000. These included alkaline phosphatase–labeled (AP-labeled) Goat-α-mouse-IgM–AP (µ-chain specific) (Sigma), Goat-α-mouse-IgG–AP (γ-chain specific) (Sigma), Goat-α-mouse-IgA-AP (α-chain specific) (Sigma). Specific controls were used for each specific antibody, and formal antibody dilution curves were determined as an initial study to identify the linear range of each antibody titer measurement. It was determined from these dilution curves that plasma samples could be optimally measured at 1:100 dilutions and lung samples at 1:10 dilutions, yielding concentrations within the linear detection range for each assay.

Where indicated antibodies were extracted from lung tissue by a method modified from Ylä-Herttuala et al. (38). Lungs were extensively perfused with physiological saline to ensure removal of all blood prior to analysis. Briefly, a small piece of lung (50 mg) was homogenized (Homogenizers POLYTRON®) on ice, in 0.5 ml PBS, pH 7.2, containing the following preservatives: 2.7 mmol/L EDTA, 2 mmol/L benzamidine, 1 mmol/L PMSF, 40 µmol/L elastatinal, 10 µmol/L probucol, 0.01% aprotinin, 0.008% chloramphenicol, 0.008% gentamicin and Protease Inhibitor Cocktail (1:100; Sigma P8340). Samples were incubated overnight at 4°C. The extract was collected by low-speed centrifugation at 4°C (30 min at 3000 RPM) and re-centrifuged for 15 min at 4,000 RPM at 4°C prior to assay.

Protein-lipid overlay analysis

Total lung lipids were isolated by Folch extraction as previously described (31) from 25 mg tissue. Total lung lipid extracts were spotted on to nitrocellulose membrane and allowed to dry. Membranes were either incubated in the presence of total lung protein extracts (15 mg tissue) or E06-IgM antibody (1:500) for 16 h at 4°C. Extracts containing antibodies were removed by extensive washing in PBS, and cross-reactivity detected using HRP-conjugated anti-mouse IgM (Invitrogen).

Immunohistochemistry

Frozen-embedded sections (lungs) from wildtype and Abcg1^−/− mice were fixed in 4% PFA, blocked with 5% goat serum, and stained with either HRP-conjugated anti-mouse IgM, IgG or IgA and detected with ECL. A Vectastain ABC-Alkaline phosphatase kit (Vector Laboratories) was used to visualize the antibody staining. Where indicated, slides were counter-stained with Harris Hematoxylin (Fisher Scientific). Frozen tissue sections of lungs from wildtype and Abcg1^−/− mice were also stained with antibodies that recognize CXCL13 (Genetex), oxPL (E06), B220 (B cell marker; BD Biosciences Clone RA3–6B2) and PCNA (proliferative marker; Genetex), followed by anti-mouse IgM AlexaFluor 488, anti-rat AlexaFluor 594 or anti-rabbit AlexaFluor 488 secondary antibodies (Molecular Probes, Life Sciences). Immunostaining of adjacent sections in the absence of primary antibody was used as a negative control.

TUNEL staining

The presence of apoptotic cells was assessed by terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay of frozen-embedded tissue sections or primary alveolar macrophages as previously described (39).

Statistics

Lipid parameters (cholesterol, oxysterols, phosphatidylcholine, oxidized phospholipids) were analyzed by two-way ANOVA, with genotype as one factor and lipid species as another. Where there was an effect of either genotype or lipid species with no apparent interaction, data were further analyzed by post hoc Bonferroni test to determine differential effects. Absolute cell numbers (determined using flow cytometry) were analyzed by unpaired Student t test. Antibody titers were analyzed by two-way ANOVA, with genotype as one factor and antigen (MDA-LDL, Cu-OxLDL, E06/T15) as another. Where there was an effect of either genotype or lipid species with no apparent interaction, data were further analyzed by post hoc Bonferroni test to determine differential effects. Ch25h, Cyp7b1, Cxcl13, Cxcr5, Gpr183, total IgM, and E06-IgM mRNA were analyzed by unpaired Student t test.

ABCG1 regulates pulmonary B cell homeostasis

To investigate the role of ABCG1 in B cell homeostasis and innate immunity, we examined specific immunological properties of 6 month old Abcg1^−/− mice. Flow cytometric analysis of the spleen demonstrated no significant difference in the number of B (Figure S1A) or T (Figure S1B) cells recovered from the spleens of wildtype and Abcg1^−/− mice. We previously observed gross lymphocytic infiltrates consisting predominantly of B cells in the lungs of Abcg1^−/− mice (36). Consistent with these observations, FACS analysis of the lungs revealed a significant increase in the B cell population (defined as CD19⁺) in the lungs of chow-fed Abcg1^−/− mice (Figure 2A). In contrast, there was no significant difference in the number of T cells (defined as CD3⁺) in the lung (Figure 2B).

B cells can be subdivided into B-1 and B-2 B cells. B-1 B cells are primarily localized to the peritoneal cavity and pleural space whereas B-2 cells are localized to the spleen (1, 4). FACS analysis of cells recovered from the lung or pleural cavity demonstrated that B-1 B cells (CD19⁺, sIgM⁺, CD11b⁺) were increased in Abcg1^−/−, compared to wildtype mice (Figure 2C, D). In contrast, there was no difference in the number of B-1 B cells present in the spleen or peritoneal cavity (PerC) (Figure 2E). Taken together, these data support the conclusion that a global loss of ABCG1 results in a niche-specific increase in B-1 B cells in the lungs and pleural space. Interestingly there was also a marked increase in the numbers of conventional B-2 cells in the lungs, but not the spleens, of Abcg1^−/− mice (Figure 2F). Since B-2 and B-1 cells are generally considered to be mediators of adaptive immunity (40) and innate immunity (6, 41) respectively, these data demonstrate that loss of ABCG1 has a broad impact on pulmonary B cell biology in mice.

Innate B-1 B cells can also be further divided into B-1a (sIgM⁺, CD11b⁺, CD5⁺) and B-1b (sIgM⁺, CD11b⁺, CD5⁻) cells. In the absence of infection, B-1a cells are the source of most serum IgM, are present in serous cavities, and are known to generate NAbs, which for example specifically bind to phosphocholine (PC) containing oxidized phospholipids (OxPL) (42). FACS analysis of cells recovered from the pleural cavity and lungs of wild type and Abcg1^−/− mice shows that there is a significant increase in B-1a cells recovered from Abcg1^−/− mice (Figure 2G, H). In contrast to B1-a cells, B-1b cells are thought to be the primary source of T cell-independent antibody production, and provide delayed but long-term protection against infectious pathogens (1). The finding that B-1b cells are also increased in both the lungs and pleural cavities of Abcg1^−/− mice, as compared to wildtype mice (Figure 2I), provides further support for the proposal that loss of ABCG1 affects multiple arcs of immunity. These effects are niche/lung-specific since there were was no significant difference in the numbers of B-1a (Figure S1C) or B-1b (Figure S1D) cells in the spleen or peritoneal cavity of wildtype and Abcg1^−/− mice.

Lipid-driven expansion of B-1 B cells in the lungs of Abcg1^−/− mice

To test the hypothesis that the lung-specific expansion of B cells is dependent upon lipid accumulation, we analyzed cells isolated from lungs, spleens, peritoneal and pleural cavities of chow-fed 12 week old wildtype and Abcg1^−/− mice. Importantly there was little/no lipid accumulation in the lungs of these young Abcg1^−/− mice (18). Consistent with our hypothesis, in the absence of pulmonary lipid accumulation, there was no significant difference in the number of B-1 B cells in the tissues of wildtype and Abcg1^−/− mice (Figure 3A). We previously demonstrated that the pulmonary lipidosis in Abcg1^−/− mice can be accelerated by feeding mice a Western diet (18). Consequently, 4 week old wildtype and Abcg1^−/− mice were fed a Western diet for 8 weeks to promote lipid deposition. FACS analysis of cells recovered from the pleural cavity and lungs shows that there was a significant increase in the number of B-1 B cells recovered from Abcg1^−/− mice, with no change observed in spleen and peritoneal cavity (Figure 3B, C).

It is possible that there are off-target effects from feeding a Western diet. Further, it is known that Abcg1^−/− mice >12 weeks old exhibit increasing pulmonary lipidosis on a chow diet (18). Therefore, to test the requirement of the pulmonary lipidosis for B-1 B cell expansion in Abcg1^−/− mice, we isolated B-1 B cells from 12 week old transgenic mice expressing green fluorescent protein (GFP) under the control of the chicken β-actin promoter. These GFP⁺ B-1 B cells were injected (1×10⁶ cells/mouse) intraperitoneally into 6 month old chow-fed wildtype and Abcg1^−/− mice. After 10 weeks reconstitution, FACS analysis of cells recovered from lung, spleen, pleural and peritoneal cavities shows that there were significantly increased numbers of GFP⁺ cells specifically in the lungs and pleural cavities of Abcg1^−/− mice, compared to wildtype mice (Figure 3D, E). In contrast, there was no difference in the number of GFP⁺ cells in either the spleens or peritoneal cavities of wildtype and Abcg1^−/− mice (Figure 3E).

Accumulation of oxysterols and oxidized phospholipids in the lungs of Abcg1^−/− mice

To better define the lipids that accumulate in the lungs of Abcg1^−/− mice, we utilized GC/MS and ESI-MS/MS to analyze lipids extracted under conditions that minimize autoxidation. Figure 4A–F shows for the first time that both the lungs and surfactant from 6 month old chow-fed Abcg1^−/− mice contain significantly increased levels of a number of enzymatically synthesized oxysterols (24-, 25- and 27-hydroxycholesterol), and autoxidation derivatives of cholesterol (7α-hydroxycholesterol, 7β-hydroxycholesterol, 3,5,6-triolcholesterol and 7-ketocholesterol), in addition to total cholesterol. Further, we demonstrate that the lungs of Abcg1^−/− mice contain 3- to 6-fold higher levels of numerous phosphatidylcholine species that differ in their content of different fatty acids at the sn-1 and/or sn-2 positions (Figure 4G). Phosphatidylcholines, especially di-palmitoyl phosphatidylcholine, represent the major phospholipids in the mammalian lung (43, 44). Thus, whole body loss of ABCG1 leads to a generalized increase in all lung phosphatidylcholines, rather than affecting the metabolism of specific species.

Given the increase in autoxidation products of cholesterol in the lungs and surfactant of Abcg1^−/− mice (Figure 4A, D), we used ESI-MS/MS to identify PC-containing oxidized phospholipids (OxPL), possible oxidation products of phosphatidylcholine. These analyses identified two OxPL, 1-palmitoyl-2-(5’-oxovaleroyl)-sn-glycero-3-phosphocholine (POVPC) and 1-palmitoyl-2-(9’-oxononanoyl)-sn-glycero-3-phosphocholine (PoxnoPC), which were increased 2–15 fold in the lungs of Abcg1^−/− mice (Figure 4H). Interestingly, we have previously reported that POVPC and PoxnoPC are able to serve as antigens for B-1 cells (42), a cell type that has a critical role in innate immunity (1, 4).

ABCG1 regulates B-1 B cell homing

Despite chronic inflammation in the lungs of Abcg1^−/− mice (18, 36), pathway analysis of inflammatory gene expression identified a signature of chemokines and cytokines, present in Abcg1^−/− mice compared to wildtype animals, reflective of B cell activation and homing (Table 1). Importantly, lipids are now widely accepted to act as both signaling molecules and mediators produced locally in response to various stimuli (45). Indeed, 25-hydroxycholesterol is a metabolite of cholesterol, produced and secreted by macrophages, that has potent and diverse effects on the immune system (46). Catabolism of 25-hydroxycholesterol by the enzyme CYP7B1 results in the formation of 7α,25-dihydroxycholesterol, a ligand for the G protein-coupled receptor (GPCR) GPR183/EBI2, expressed on the surface of B cells (47). The data of Figure 4B and (27) demonstrate that 25-hydroxycholesterol levels are significantly increased in the lungs and alveolar macrophages of Abcg1^−/− mice, respectively, suggesting there are enhanced B cell homing signals in the lungs of Abcg1^−/− mice. Indeed, mRNA levels of the enzymes cholesterol 25-hydroxylase (Ch25h) and Cyp7b1 are significantly increased in the lungs of Abcg1^−/− mice (Figure 5A).

Naïve B cells, expressing the CXCL13 receptor CXCR5, migrate in response to the chemokine CXCL13, and mice lacking CXCL13 fail to migrate B cells into the peritoneal and pleural cavities (48). The data of Figure 5B (left panel) show that in young 12-week old mice, that lack significant lung lipidosis there is no change in Cxcl13 mRNA expression. In contrast, Cxcl13 mRNA levels were increased in Abcg1^−/− mice fed a Western diet for 8 weeks to induce lipidosis (Figure 5B; right panel). Six month old chow-fed Abcg1^−/− mice also exhibit pulmonary lipidosis (18, 31). Compared to wildtype mice, the lungs of these older, chow-fed Abcg1^−/− mice also have elevated levels of Cxcl13 mRNA levels (Figure 5C). Additionally, B-1 B cells isolated from the pleural cavities of Abcg1^−/− mice expressed significantly increased levels of Gpr183/EBI2 and Cxcr5 mRNA compared to cells from wildtype mice (Figure S2A), which is in agreement with increased homing of B-1 B cells to the lungs and pleural cavity. Finally, immunohistochemical analysis of the lungs of 6 month old chow-fed mice demonstrates significantly increased levels of CXCL13 in the lungs of Abcg1^−/− mice, compared to wildtype animals (Figure 5D).

Increased immunoglobulins in the lungs of Abcg1^−/− mice

The data of Figures 2–5 suggest that the lungs of Abcg1^−/− mice may contain increased titers of antibodies that function in innate (IgM, IgA, or IgG3 from B-1 cells) and adaptive (IgG from B-2 cells) immunity. Consistent with this hypothesis, immunohistochemical analysis of the lungs and spleens of wildtype and Abcg1^−/− mice demonstrate that the lungs, but not spleens, of the knockout mice contain elevated levels of IgM, IgA and IgG, with the most positive staining observed adjacent to alveolar sacs (Figure 6A; arrows). Western blot analysis confirmed that the levels of IgM and IgA were markedly increased, and IgG slightly increased in the lungs of the knockout mice (Figure 6B). Further, titers of total IgM and IgA determined by ELISA were also increased in the lungs of Abcg1^−/− mice (Figure 6C). Thus, loss of ABCG1 results in increased levels of multiple immunoglobulins that function in both innate and adaptive immunity.

Loss of ABCG1 results in pulmonary accumulation of lipid antigens and specific IgM antibodies

We next performed a modification of the protein-lipid-overlay (PLO) technique (49) to determine if the lungs of Abcg1^−/− mice contained increased levels of lipid antigens and corresponding immunoglobulins. Total lipids were isolated from perfused lungs of 6 month old chow-fed wildtype and Abcg1^−/− mice, and spotted onto a nitrocellulose membrane. The membrane containing the bound lipids was then incubated with protein extracts from wildtype or Abcg1^−/− mice before addition of anti-IgM, -IgG or -IgA HRP-conjugated secondary antibodies for detection. Figure 6D (right panels) shows that the lungs of Abcg1^−/− mice, but not wildtype mice, contain IgM antibodies that bind to an endogenous lipid antigen(s) present in the lungs of knockout mice. This observation is consistent with the presence of elevated levels of POVPC and PoxnoPC (Figure 4H), known to be antigens for B-1 B cells, in the lungs of Abcg1^−/− mice. In contrast, the lungs of wildtype and knockout mice contain both IgA and IgM antibodies that bind lipid antigens present only in the lungs of Abcg1^−/− mice (Figure 6D).

Increased antibody titers to oxidation-specific epitopes (OSE) in the lungs of Abcg1^−/− mice

The increase in specific oxidized sterols and phospholipids (Figure 4) in the lungs of Abcg1^−/− mice suggested an enhanced oxidative environment and the generation of oxidation-specific epitopes (OSE). Consequently, we performed chemiluminescent ELISAs, using various antigens, to specifically determine whether the lungs of Abcg1^−/− mice contained antibodies binding to OSE. Figure 7A–C shows that titers of IgM, IgA and IgG to MDA-LDL and Cu-OxLDL were increased 2.5- to 6-fold in the lungs of knockout mice. In addition, using the anti-idiotypic antibody AB1–2, which binds specifically to the E06/T15 idiotype that binds OxPL (50), we measured titers of the specific E06 IgM and T15 IgA NAbs. There were markedly increased titers of both E06 IgM (Figure 7B) and T15 IgA (Figure 7C) in the lungs of Abcg1^−/− mice. We also noted markedly increased titers of IgA to the classic B-1 cell antigen α1,3-dextran in the lungs of Abcg1^−/− mice. The latter finding suggests a generalized increase in B-1 cell derived IgA antibodies, which are typically mucosal, but a preferential increase in IgM NAbs to OSE. Further, the lungs of Abcg1^−/− mice exhibited a 2-fold increase in total IgM mRNA but a 6-fold increase in E06-IgM-specific mRNA (Figure 7D), consistent with the idea that there are increased levels of E06-secreting B-1a B cells in the Abcg1^−/− lungs. These data indicate that the lungs of Abcg1^−/− mice contain increased levels of B-1a B cells (Figures 2, 3), oxidized lipid antigens (Figure 4H), IgA and IgM NAbs (Figures 6, 7) that are part of innate immunity, and IgGs that are likely part of adaptive immunity (Figures 6, 7).

Elevated plasma titers of natural and adaptive antibodies in Abcg1^−/− mice

To assess whether loss of ABCG1 leads to a more global expansion of antibodies involved in innate and/or adaptive immunity, we used chemiluminescent ELISAs to measure total and OSE antibody titers in the plasma of wildtype and Abcg1^−/− mice. Table 2 shows that plasma from Abcg1^−/− mice contained increased titers of total IgG and IgM, increased titers of IgA and IgG to MDA-LDL, and increased IgM titers to OxLDL and α-1,3-dextran, as well as increased titers of the IgM E06 to OxPL. Interestingly, although IgA are predominantly serosal antibodies, there were also increased plasma titers of IgA to MDA-LDL in Abcg1^−/− mice.

Selective stimulation and increased proliferation of B-1a B cells in the lungs of Abcg1^−/− mice

Together, the data of Figures 2–7 and Table 2 demonstrate that loss of ABCG1 results in increased titers of NAbs in the lungs and plasma of Abcg1^−/− mice, and to a niche-specific pulmonary expansion of B cells and especially B-1 B cells. However, we cannot rule out the possibility that in addition to increased homing of B-1 B cells to the lungs and pleural cavity (Figure 5, Table 1), there may also be local expansion of B-1 B cells. To directly test whether a localized increase of B-1a B cells might be related to a selective stimulation of these clones with their respective antigens, (e.g. the presence of OxPL in the lungs leading to enhanced secretion of IgM E06 and IgA T15), we performed immunohistochemical analyses of lung sections to determine the presence of OxPL that are recognized and bound by E06. In agreement with the increased titers of E06 NAbs (Figure 7) and increased OxPL antigens (Figure 4H), we observed positive E06 staining in the lungs of 6 month old chow-fed Abcg1^−/− mice compared to wildtype mice (Figure 8A). We also determined whether there was increased proliferation of B cells in situ in the lung using proliferative and B cell marker co-staining on lung tissue sections from wildtype and Abcg1^−/− mice that were injected with GFP⁺ B-1 B cells (as in Figure 3). First, consistent with the data of Figure 2 and (36) we observed increased staining of B-1 B cells in the lungs of Abcg1^−/− mice compared to wildtype mice, using GFP expression levels (Figure 8B, left panel). We also observed increased proliferation of B-1 B cells, indicated by the increased staining and co-localization of the proliferative marker PCNA with GFP (Figure 8B, right panel merged image). Finally, we also observed that cells expressing PCNA also express IgM, a marker of B-1 B cells (Figure S2B). Taken together, these data demonstrate that loss of ABCG1 and tissue sterol homeostasis profoundly affects both adaptive and innate immune responses.