Peroxisome proliferator-activated receptors (PPARs) are members of the nuclear hormone receptor superfamily. PPARγ, a ligand-activated transcription factor, has important anti-inflammatory and antiproliferative functions, and it has been associated with diseases including diabetes, scarring, and atherosclerosis, among others. PPARγ is expressed in most bone marrow-derived cells and influences their function. PPARγ ligands can stimulate human B cell differentiation and promote Ab production. A knowledge gap is that the role of PPARγ in B cells under physiological conditions is not known. We developed a new B cell-specific PPARγ (B-PPARγ) knockout mouse and explored the role of PPARγ during both the primary and secondary immune response. In this article, we show that PPARγ deficiency in B cells decreases germinal center B cells and plasma cell development, as well as the levels of circulating Ag-specific Abs during a primary challenge. Inability to generate germinal center B cells and plasma cells is correlated to decreased MHC class II expression and decreased Bcl-6 and Blimp-1 levels. Furthermore, B-PPARγ–deficient mice have an impaired memory response, characterized by low titers of Ag-specific Abs and low numbers of Ag-experienced, Ab-secreting cells. However, B-PPARγ–deficient mice have no differences in B cell population distribution within primary or secondary lymphoid organs during development. This is the first report, to our knowledge, to show that, under physiological conditions, PPARγ expression in B cells is required for an efficient B cell-mediated immune response as it regulates B cell differentiation and Ab production.
Peroxisome proliferator-activated receptors (PPARs) are members of the nuclear hormone receptor superfamily (1). These ligand-activated transcription factors are divided into three subtypes: PPARα, PPARβ/δ, and PPARγ (2). PPAR signaling is activated by natural and synthetic ligands. Of interest, PPARγ can be activated by the endogenous PG 15-deoxy-Δ12,14 PG J2, or by the synthetic antidiabetic thiazolidinediones (3–5). PPARγ is generally considered an anti-inflammatory and antiproliferative transcription factor (2). PPARγ and its ligands have been implicated in diseases such as diabetes, scarring, and atherosclerosis, among others (6–8).
PPARγ plays an important role in driving adipogenesis and lipid metabolism, and in dampening inflammation (2, 9, 10). PPARγ is also involved in regulating aspects of the immune system. T cells, B cells, macrophages, and dendritic cells express PPARγ (11–14). PPARγ is involved in monocyte and dendritic cell differentiation (12, 13, 15). Furthermore, in dendritic cells, PPARγ signaling downregulates IL-12 production (12, 13). Similarly, in T cells, PPARγ ligands decrease IL-12 and IFN-γ production, as well as cell survival (16–18).
B cells play an important role during both the innate and adaptive immune response. After initial Ag encounter, experienced B cells differentiate into Ab-secreting cells or to memory cells, ensuring an efficient response upon Ag re-encounter (19). Thus, mounting a strong but controlled humoral response is crucial for establishing long-term immune protection (20). We previously demonstrated that normal and malignant B cells express PPARγ (14). Using human B cells, we determined that PPARγ plays a role in regulating B cell function, particularly in promoting Ab production and B cell differentiation toward a plasma cell in vitro (21). In addition, malignant B-lineage cancer cells lines, with PPARγ deliberately overexpressed, have decreased proliferation and enhanced apoptosis (22–24).
Despite its role in B cell function in vitro, the role of PPARγ in B cell biology under physiological conditions is not known. Much of the animal work has been limited by technical challenges including the inability to generate complete PPARγ-deficient animals, because they are embryonically lethal (9). An alternative approach, which used global PPARγ haploinsufficient mice, has shown that a systemic reduction in PPARγ levels increases B cell proliferation and the Ag-specific immune response (25). However, even though that study showed the involvement of PPARγ in the humoral response, the experimental model used suffers from the fact that all cells were PPARγ haploinsufficient, not just B cells. Therefore, to further explore the role of PPARγ in B cell function under physiological conditions, we generated a new B cell-specific PPARγ (B-PPARγ) knockout mouse and analyzed the role of PPARγ during both the primary and secondary immune response.
Materials and Methods
Mice generation and breeding
Mice carrying a recombinant Pparγ gene with loxP sites flanking exon 2 were a gift from Dr. Frank Gonzalez (National Institutes of Health, Bethesda, MD) (26). Strain C.129P2-Cd19tm1(cre)Cgn/J, which contains the Cre recombinase under control of the Cd19 promoter on the BALB/C genetic background, was purchased from The Jackson Laboratory (Bar Harbor, ME). These strains were crossed to generate Cd19-Cre+/− Pparγfl/− heterozygous mice, which were then backcrossed to C57BL/6J for five generations. Sibling crosses were then used to generate Cd19-Cre+/− Pparγfl/fl mice, and this genotype was maintained by sibling and cousin mating using male Cd19-Cre+/− Pparγfl/fl and female Cd19-Cre−/− Pparγfl/fl mice. The progeny are either Cd19-Cre+/− Pparγfl/fl (B cell PPARγ knockout) or Cd19-Cre−/− Pparγfl/fl (normal B cell littermate controls). Progeny were genotyped by a commercial service (Transnetyx, Cordova, TN) using PCR oligos that span the junction of the Cd19 promoter and the Cre coding sequence, and the junction of Pparγ intron 1 and the loxP site.
Because Cre is knocked into the CD19 locus, Cd19-Cre+/− mice have only one functional copy of CD19. To control for Cd19 copy number effects, we bred Cd19-Cre+/− males to C57BL/6J females, and the resulting Cd19-Cre+/− Pparγwt/wt offspring were used as controls for some experiments. B cells and Ab titers in naive mice were analyzed as in Figs. 1 and 3 and Supplemental Fig. 2, and the OVA immune response was analyzed as in Figs. 4 and 5 and Supplemental Fig. 3. In both cases, the results were similar to experiments performed using Cd19-Cre−/− Pparγfl/fl controls (data not shown).
|B Cell Subset .||Phenotypea .|
|Pro B cell||B220lowCD43+sIgM−|
|Pre B cells||B220lowCD43−sIgM−|
|Immature B cells||B220highsIgM+|
|Transitional B cells||B220+CD93+sIgM+|
|Follicular B cells||B220+CD23+IgDhighIgMlow|
|Marginal zone precursor||B220+IgMhighCD23highCD21high|
|B Cell Subset .||Phenotypea .|
|Pro B cell||B220lowCD43+sIgM−|
|Pre B cells||B220lowCD43−sIgM−|
|Immature B cells||B220highsIgM+|
|Transitional B cells||B220+CD93+sIgM+|
|Follicular B cells||B220+CD23+IgDhighIgMlow|
|Marginal zone precursor||B220+IgMhighCD23highCD21high|
All cells are live.
To confirm B cell deletion of PPARγ, we extracted DNA from highly purified B cells obtained by FACS and subjected to quantitative PCR analysis. Excision of exon 2 in Cre+ mice was confirmed using the sense 5′-GTAGAACCTGCATCTCCACC-3′ and antisense 5′-CTTGCATCCTTCACAAGCATG-3′ primers; exon 1 was used as a control and amplified with sense 5′-CATGGTTGACACAGAGATGC-3′ and antisense 5′-GTGTGGAGCAGAAATGCTGG-3′ primers. Lack of PPARγ protein expression was further confirmed by Western blot analysis. Purified B cells were lysed with Nuclear Extract Kit (Active Motif, Carlsbad, CA). Ten micrograms of nuclear extract were loaded onto gradient SDS-PAGE gels (Pierce/Thermo Fisher Scientific, Rockford, IL). Western blots were probed using anti-PPARγ Ab (D69; Cell Signaling Technology, Beverly, MA) and anti-actin Ab (Calbiochem/EMD Chemicals, Gibbstown, NJ). HRP-conjugated goat anti-rabbit or goat anti-mouse Abs (Jackson ImmunoResearch) were used. Western blots were visualized with ECL (PerkinElmer Life Sciences).
To assess the efficiency of Cre recombination, we used the reporter strain B6.129X1-Gt(ROSA)26Sortm1(EYFP)Cos/J (R26-Stop-enhanced yellow fluorescent protein [EYFP]; The Jackson Laboratory). This strain contains a gene expressing EYFP, preceded by a floxed STOP sequence, inserted in the Gt(ROSA)26Sor locus (27). When crossed with a strain expressing Cre recombinase, the STOP sequence is deleted, allowing expression of EYFP in the Cre-expressing cells or tissues. Cd19-Cre+ males were bred to R26-Stop-EYFP+/+ females, and Cre+/EYFP+/− offspring were identified by tail snip DNA PCR as described earlier. Peripheral blood mononucleated cells were harvested and analyzed by flow cytometry.
Mouse B lymphocyte isolation and in vitro culture conditions
Spleens were harvested, processed into a single-cell suspension, and B cells were isolated using CD19 magnetic beads (Miltenyi, Auburn, CA). The purity was >98% as determined by CD19 surface staining. B cells (1 × 106 cells/ml, unless otherwise specified) were cultured in RPMI 1640 media (Invitrogen Life Technologies) supplemented with 5% FBS, 50 μM 2-ME (Eastman Kodak, Rochester, NY), 10 mM HEPES (U.S. Biochemical, Cleveland, OH), 2 mM l-glutamine (Invitrogen Life Technologies, Carlsbad, CA), 50 μg/ml gentamicin (Invitrogen Life Technologies, Carlsbad, CA). B cells were stimulated with LPS (1 μg/ml, E. coli 055:B5; Sigma) or anti-mouse CD40 (5 μg/ml, HM40-3; BD Pharmingen) plus IL-4 (25 ng/ml; eBioscience).
Activated B cells (105 cells/ml) were cultured in triplicate using 96-well round-bottom plates. Cell cultures were pulsed with [3H]thymidine (1 μCi/well) for 24 h before harvest. [3H]thymidine incorporation was measured by scintillation spectroscopy using a Topcount Luminometer (PerkinElmer, Boston, MA).
IgM and IgG ELISA
Sera from mice or supernatants from activated B cells were collected for total IgM, IgG, or IgA quantification by ELISA kits as specified by the manufacturer (Bethyl Laboratories, Montgomery, TX). OVA-specific Abs were measured using precoated OVA (10 μg/ml) plates and mouse-specific Ab ELISA kit (Bethyl Laboratories, Montgomery, TX).
IgM- and IgG-specific ELISPOT assay
Cells were incubated in OVA-coated ELISPOT plates (Millipore, Billerica, MA) for 5 h at 37°C. Alkaline phosphatase-conjugated goat anti-mouse IgM, IgG, or IgA Abs (Southern Biotech, Birmingham, AL) were used as recommended by the manufacturer. ELISPOT plates were developed with Vector AP substrate kit III (Vector Laboratories, Burlingame, CA) and quantified using the CTL plate reader and ImmunoSpot software (Cellular Technologies, Shaker Heights, OH).
In vivo experiments
Primary mouse immunization was done using 10 μg OVA absorbed on CFA (1:1 ratio) by i.p. injection and samples collected 2 wk after injection. Secondary stimulation was done i.p. with 10 μg OVA (in PBS) 10 wk later and samples collected 2 wk after immunization. Cd19-Cre−/− Pparγfl/fl or Cd19-Cre+/− Pparγwt/wt animals were used as controls. Primary footpad immunizations were done using 25 μg OVA adsorbed on CFA (1:1 ratio), 20 μl final volume. Two weeks after immunization, popliteal lymph nodes were collected for analysis.
Single-cell suspensions were prepared from mouse bone marrow, blood, spleen, and lymph nodes, and stained with a mixture of fluorochrome-conjugated anti-mouse mAbs: CD93 (clone AA4.1l eBioscience), CD1d (clone 1B1; BD Pharmingen), CD24 (clone M1/69; eBioscience), CD5 (clone 53-7.3; BioLegend), IgM (clone II/41; eBioscience), CD23 (clone B3B4; BioLegend), CD21(clone 7E9; BioLegend), IgD (clone 11.26c.2a; BioLegend), CD19 (clone 6D5; BioLegend), B220 (clone RA3-6B2; eBioscience), CD43 (clone 1B11; BioLegend), CD25 (clone PC61.5; eBioscience), CD3 (clone 17A2; BioLegend), FasL 9 (clone MFL3; BioLegend), CD138 (clone 281-2; BD Pharmingen), CD62L (clone MEL-14; eBioscience), CD95 (clone Jo2; BD Pharmingen), CD86 (clone GL-1; BioLegend), GL7 (clone GL7; eBioscience), CD44 (clone IM7; eBioscience), I-A/I-E (clone M5/114.15.2; BD Pharmingen), CD80 (clone 16-10A1; eBioscience), CXCR5 (clone 2G8; BD Pharmingen), IgG (eBioscience), IgM (clone II/41; eBioscience), CD3 (clone 145-2C11; BioLegend), CD4 (clone GK1.5; BioLegend), CD69 (clone H1.2F3; BioLegend), CD138 (clone 281-2; BD Pharmingen) followed by staining with fluorochrome-conjugated streptavidin (Invitrogen) if biotin-conjugated Ab was included in the staining panel. All the samples were stained for dead cell exclusion using Live/Dead fixable violet dead cell staining kit (Invitrogen). Samples were run on a 12-color LSRII cytometer (BD Pharmingen) and analyzed by FlowJo software (Tree Star, Ashland, OR).
Total RNA was isolated using miRNAeasy Mini Kit (Qiagen, Valencia, CA). iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA) was used to reverse transcribe RNA to cDNA. Steady-state levels of Bcl-6, Blimp-1, and GAPDH RNA were assessed by real-time PCR using iQ SYBR Green Supermix (Bio-Rad). Primers used for Bcl-6 sense, 5′-AGACGCACAGTGACAAACCATACAA-3′ and antisense, 5′-GCTCCACAAATGTTACAGCGATAGG-3′; for Blimp-1 were sense, 5′-TTCTTGTGTGGTATTGTCGGGACTT-3′ and antisense, 5′-TTGGGGACACTCTTTGGGTAGAGTT-3′; and for GAPDH sense, 5′-AGCCTCGTCCCGTAGACAAA-3′ and antisense, 5′-CCTTGACTGTGCCGTTGAAT-3′ (28). Results were analyzed using Bio-Rad iCycler software. Bcl-6 and Blimp-1 mRNA steady-state levels were normalized to GAPDH.
Freshly dissected spleens were embedded in optimum cutting temperature compound (Sakura) and stored at −80°C. Cryostat sections (10 μm) were prepared from frozen tissues. Sections were fixed in cold acetone and rehydrated in PBS, blocked with 5% whole-rat serum, and stained with allophycocyanin–anti-CD3ε (clone 145-2C11; BioLegend) and PE–anti-B220 (clone RA3-6B2; eBioscience). Images were acquired using a fluorescence microscope connected to a monochrome CCD digital camera and were analyzed using ImagePro Plus software (MediaCybernetics, Bethesda, MD).
Data are expressed as mean ± SEM. Significance was determined by one-way ANOVA with a Tukey’s posttest, or a two-tailed unpaired Student t test where applicable. Statistical analyses were done using GraphPad Prism 5.0 (GraphPad Software, La Jolla, CA.). Probability values of p ≤ 0.05 were considered statistically significant.
Generation of a mouse Cre/loxP recombination system to study PPARγ deficiency in the B cell lineage
Global PPARγ knockout animals are embryonically lethal, limiting in vivo studies on PPARγ function. Therefore, we developed a new B-PPARγ knockout mouse using transgenic mice in which Cre recombinase expression is under control of the CD19 promoter, which is expressed only in the B cell lineage (29, 30). The efficiency of Cre-mediated recombination in B cells using an EYFP reporter was >90% (Supplemental Fig. 1A). The CD19-Cre mice were bred with a targeted strain in which LoxP sites are positioned flanking exon 2 of the PPARγ gene (26) to generate conditional knockouts in which only B cells harbor a Cre-dependent deletion of PPARγ (B-PPARγ–deficient mice). In these Cre+ B cells, exon 2 is excised, thus rendering the PPARγ receptor nonfunctional (Supplemental Fig. 1B). Efficiency of exon 2 deletion in highly purified B cells was near 100%, determined by PCR analysis of genomic DNA (Supplemental Fig. 1C). Knockout of PPARγ expression in B cells from B-PPARγ–deficient mice was further confirmed by Western blot analysis using purified B cells (Supplemental Fig. 1D).
Loss of PPARγ in B cell lineage does not affect B cell development in the bone marrow or spleen
Considering the potential importance of PPARγ in B cell function, we first evaluated any possible deficiencies in B cell subpopulations. The total number of splenocytes, as well as the distribution of splenic CD19+ B cells in the B-PPARγ–deficient mice, were measured and compared with control mice (Fig. 1A, 1B). Naive B-PPARγ–deficient mice have a normal count of total splenocytes and CD19+ cells.
Early B cell development takes place in the bone marrow, whereas late stages of differentiation can take place in secondary lymphoid organs. We performed a comprehensive evaluation of B cell subpopulations (as defined in Table I and Supplemental Fig. 2A, 2B) in B-PPARγ–deficient mice (29, 31, 32). Pro-B cell, pre-B cell, immature B cell, and transitional B cell population frequencies do not change in the bone marrow of B-PPARγ–deficient mice compared with control animals (Fig. 1C–E). In addition, follicular, marginal zone B, and B1 cell subpopulations in spleen and lymph node were also found to be unchanged (Supplemental Fig. 2C–H).
PPARγ-deficient B cells have decreased proliferation and IgG production in vitro
We have previously shown that PPARγ signaling is important for human B cell Ab production in vitro (21). We therefore evaluated the primary B cell response under in vitro conditions. Ab production was measured in highly purified B cells isolated from B-PPARγ–deficient and control mouse spleens and treated with LPS (Fig. 2). Interestingly, LPS-treated B cells from B-PPARγ–deficient mice produced similar levels of IgM, but significantly lower levels of IgG (Fig. 2A, 2B).
PPARγ affects normal and malignant B cell differentiation and proliferation (21, 33). Considering that B cells from PPARγ-deficient mice produced less IgG upon activation, we investigated whether there were any differences in cell proliferation. Accordingly, LPS-activated, purified PPARγ-deficient B cells had decreased proliferation compared with control cells (Fig. 2C).
B-PPARγ–deficient mice have an impaired Ag-specific Ab response
Next, we measured Ab concentration in sera of naive B-PPARγ–deficient animals. Both B-PPARγ–deficient and control groups showed similar total IgM, IgG, and IgA serum Ab titers (Fig. 3A–C).
Because the observed decrease in Ab levels and proliferation were seen under stimulating culture conditions (Fig. 2), we next asked whether B-PPARγ–deficient animals would also have an impaired Ab-mediated immune response upon Ag exposure. To test our hypothesis, B-PPARγ–deficient and control mice were challenged using the OVA Ag model. The primary immune response was measured 2 wk after immunization. Interestingly, B-PPARγ–deficient mice had an impaired Ab-mediated immune response, seen by a significantly lower amount of both Ag-specific IgM and IgG (Fig. 4A, 4B). This defect cannot be ascribed to either structural disruption in lymphoid organ architecture in B-PPARγ–deficient mice (Fig. 4C–H) or deficiencies in development of peripheral B cell populations (Fig. 1, Supplemental Fig. 3).
B-PPARγ–deficient mice have decreased Ag-specific Ab titers and Ab-producing cells during memory response
Adaptive immunity is mediated, in part, by plasma and memory cells. These fully differentiated cells are responsible for the quick Ab-mediated immune response upon Ag re-encounter. To further study the impact of B-PPARγ deficiency on the adaptive immune response, mice were rechallenged with OVA 10 wk after initial immunization. Two weeks after Ag re-exposure, the Ab memory response was analyzed and measured (Fig. 5A, 5B). Control mice had an enhanced memory response and higher Ag-specific Ab titers. Again, B-PPARγ–deficient mice had considerably lower Ag-specific serum IgM and IgG levels. IgA levels were also measured during the primary and secondary response; however, IgA levels were below detection.
The B cell memory response was further analyzed by measuring the number of spleen Ag-specific Ab-producing cells present in reimmunized B-PPARγ–deficient mice (Fig. 5C–F). Correlating with the decreased serum Ab titers, B cells from immunized B-PPARγ–deficient animals did not respond to Ag re-exposure. In fact, the number of Ag-specific Ab-secreting cells was comparable with that of nonimmunized animals. This significant defect was true for IgM, IgG, and IgA isotypes (Fig. 5D–F).
B-PPARγ–deficient mice have decreased germinal center B cell and plasma cell populations
It is possible that the decreased Ab response observed is a consequence of impaired B cell differentiation after Ag presentation. To test this hypothesis, we challenged B-PPARγ–deficient mice by delivery of OVA Ag into the footpad. B and T cells were analyzed in the immunized and nonimmunized contralateral popliteal lymph nodes 2 wk after the OVA challenge. Distribution of germinal center (GC) B cell, plasma cell, and follicular Th (Tfh) cell populations were analyzed (Fig. 6). Immunized B-PPARγ–deficient mice had a significant decrease in the percentage of GC B cells, similar to those of the nonimmunized control group (Fig. 6A, 6B). Furthermore, the plasma cell population was also absent in the popliteal lymph nodes of immunized B-PPARγ–deficient mice (Fig. 6C, 6D). Tfh cells, which are crucial in the GC reaction (34), had a normal distribution in B-PPARγ–deficient mice compared with control (Fig. 6E, 6F).
PPARγ-deficient B cells do not upregulate MHC class II surface expression and have decreased Bcl-6 and Blimp-1 expression
Given the differences in GC B cell and plasma cell differentiation, we further analyzed B cell activation in vitro. Purified splenic B cells from control and B-PPARγ–deficient mice were stimulated with anti-CD40 and IL-4, a stimulation resembling T cell-dependent activation in vivo. Similar to LPS treatment (Fig. 2), PPARγ-deficient B cells stimulated with anti-CD40 plus IL-4 produced significantly lower levels of IgM and IgG compared with B cells from control mice (Fig. 7A, 7B).
B cell–T cell interactions within secondary lymphoid organs are necessary for B cell activation and initiation of the Ag-specific immune response (35). Upon stimulation, B cells increase expression of activation markers including MHC class II, CD80, CD86, and CD69. However, stimulated PPARγ-deficient B cells did not upregulate MHC class II (Fig. 7C, 7D). CD80, CD86, and CD69 expression levels in PPARγ-deficient B cells were similar to those of the control group (Supplemental Fig. 4).
To further confirm the mechanisms responsible for the decreased B cell activation and differentiation seen in vivo and in vitro, we measured Bcl-6 and Blimp-1 expression. Bcl-6 and Blimp-1 are key transcription factors involved in B cell differentiation. Bcl-6 is necessary for GC B cell development (36, 37), whereas Blimp-1 regulates plasma cell differentiation (38, 39). Bcl-6 and Blimp-1 steady-state mRNA levels were significantly lower in PPARγ-deficient B cells compared with control (Fig. 7E, 7F).
PPARγ is a widely expressed transcription factor that influences many areas of biology. In this article, we show evidence demonstrating a physiological role of PPARγ in B cells. Using a newly developed B-PPARγ knockout mouse model, we have shown that PPARγ expression in B cells is necessary for optimal humoral immune responses. A B-PPARγ–deficient mouse model provides a novel and direct system to study the role of PPARγ as a regulator of Ab production. In this model, mice lacking PPARγ expression in B cells are shown to have decreased proliferation and IgG production in vitro, as well as low levels of circulating Ag-specific Abs during a primary response. Furthermore, PPARγ-deficient mice have an impaired immune memory response, characterized by low titers of Ag-specific Abs and low numbers of Ag-experienced, Ab-secreting cells. Our results denote the importance of PPARγ expression in B cells during both the primary and secondary immune response.
PPARγ is involved in key B cell processes, including differentiation and Ab production (21). Interestingly, our results show that B-PPARγ–deficient mice have normal basal Ab levels, as well as a normal percentage of B cells present in the spleen. Detailed analysis of B cell development in primary and secondary lymphoid organs showed neither differences in the population distribution of the PPARγ knockout animals, nor changes in their spleen follicle architecture. We conclude that the decreased primary humoral response is not due to changes in B cell development.
Upon Ag encounter, B cells activate, proliferate, and differentiate. Ag-experienced B cells will differentiate into memory or plasma cells, which produce Ag-specific Abs during the adaptive immune response (19). Our results show that the number of Ag-specific, Ab-secreting cells in immunized B-PPARγ–deficient mice is equivalent to background levels seen in the nonimmunized group. The lack of Ag-specific, Ab-secreting cells in immunized B-PPARγ–deficient mice is true for all Ab isotypes measured. These findings correlate to the decreased serum Ab titers during the primary and secondary immune response. Interestingly, immunized animals have a normal distribution of follicular, transitional, marginal zone, and B1 cell populations, confirming that PPARγ is not required for early B cell development.
Notably, the Cre gene was introduced into the CD19 locus by homologous recombination, so that Cd19-Cre−/− Pparγfl/fl mice, used in this study as normal B cell littermate controls, have two functional copies of the CD19 gene, in contrast with the Cd19-Cre+/− Pparγfl/fl mice (B-PPARγ deficient), which have one functional CD19 locus. To control for CD19 copy number effects, some experiments were repeated using Cd19-Cre+/− Pparγwt/wt mice as controls. B cell number and Ab titer in naive mice were analyzed as in Figs. 1 and 3 and Supplemental Fig. 2, and the OVA immune response was analyzed as in Figs. 4 and 5 and Supplemental Fig. 3. In both cases, the results were similar to experiments performed using Cd19-Cre−/− Pparγfl/fl controls (data not shown). We conclude that deletion of one copy of the CD19 gene has no effect on B cell development or function in these mice.
Initial B cell priming takes place within lymphoid organs, leading to the generation of GCs (35). The GC reaction is necessary for B cell somatic hypermutation, affinity maturation, and class-switch recombination (35). Even though early B cell development in B-PPARγ–deficient mice is normal, our results show that PPARγ is required for B cell activation during Ag presentation. Our analysis has revealed that B-PPARγ–deficient mice generate Tfh cells upon Ag stimulation. However, B-PPARγ–deficient mice are unable to develop GC B cells, as well as differentiated plasma cells. We have not directly surveyed the memory cell populations. However, absence of GC B cell and plasma cell populations during the primary immune response and lack of Ag-specific, Ab-producing cells during both primary and secondary immunizations makes it very likely that B-PPARγ–deficient mice also have decreased numbers of memory B cells.
Furthermore, PPARγ-deficient B cells do not upregulate MHC class II upon activation. Decreased MHC class II expression could: 1) disrupt the immunological synapse between Ag-primed T cells and naive B cells, thus preventing GC B cell development; and 2) decrease B cell Ag presentation, consequently decreasing the Ag-specific immune response. Further in vivo investigation of B cell activation is required.
At a molecular level, we demonstrate that PPARγ deficiency in B cells decreases Bcl-6 and Blimp-1 expression, which are crucial during GC B cell and plasma cell development (36–39). There are both direct and indirect mechanisms that can explain the effects of PPARγ on Bcl-6 and Blimp-1. Signaling through CD40 and IL-4R activates NF-κB signaling. Bcl-6 expression can be downregulated via NF-κB (40, 41). Furthermore, PPARγ is known to inhibit NF-κB translocation to the nucleus (42). Thus, it is possible that PPARγ regulates Bcl-6 expression through NF-κB signaling, which initiates B cell activation and Blimp-1–mediated plasma cell differentiation. Alternatively, Bcl-6 and Blimp-1 contain multiple PPAR response elements within their 3′-untranslated region with high PPARγ binding efficiency. Therefore, it is also possible for PPARγ to directly regulate Bcl-6 and Blimp-1.
In this article, we provide conclusive evidence that PPARγ has a physiological role during the primary and secondary immune response, using a novel B-PPARγ–deficient mouse model. PPARγ has been associated with multiple diseases including autoimmune disorders, which involve B cell function (25, 43, 44). The in vivo study of PPARγ and PPARγ ligands has been limited because of challenges in the development of animal models. Our novel B-PPARγ–deficient mouse model will serve as a powerful tool for future studies on the role of PPARγ, as well as PPARγ ligand-dependent and -independent signaling mechanisms in multiple disease models.
This work was supported by National Institutes of Health Grants DE011390 (to R.P.P.), ES01247 (to R.P.P.), HL75432 (to P.J.S.), T32 DE007202, and T32 HL66988.
The online version of this article contains supplemental material.
The authors have no financial conflicts of interest.