Report

Impaired B and T Cell Antigen Receptor Signaling in p110δ PI 3-Kinase Mutant Mice

See allHide authors and affiliations

Science  09 Aug 2002:
Vol. 297, Issue 5583, pp. 1031-1034
DOI: 10.1126/science.1073560

Abstract

Class IA phosphoinositide 3-kinases (PI3Ks) are a family of p85/p110 heterodimeric lipid kinases that generate second messenger signals downstream of tyrosine kinases, thereby controlling cell metabolism, growth, proliferation, differentiation, motility, and survival. Mammals express three class IA catalytic subunits: p110α, p110β, and p110δ. It is unclear to what extent these p110 isoforms have overlapping or distinct biological roles. Mice expressing a catalytically inactive form of p110δ (p110δD910A) were generated by gene targeting. Antigen receptor signaling in B and T cells was impaired and immune responses in vivo were attenuated in p110δ mutant mice. They also developed inflammatory bowel disease. These results reveal a selective role for p110δ in immunity.

There is increasing evidence for an important role for class IA PI3Ks in regulation of the immune system (1–4). Mice lacking the p85α regulatory subunit show impaired B cell development and activation but normal T cell activation, whereas T cell–restricted deletion of the gene for the phosphoinositide 3-phosphatase PTEN, a negative regulator of PI3K signaling, results in a lethal lymphoproliferative disease (5–7). To date, the specific role for each of the three class IA PI3K catalytic subunits in lymphocyte signaling has not been determined. p110δ is expressed predominantly in leukocytes (8, 9), which indicates that it may play a unique role in immune signaling.

The study of class IA PI3Ks is complicated by the heterodimeric nature of these proteins, and altering the expression of one subunit can affect the expression of the others (10). In cells from p85α knockout mice, p85β expression was up-regulated, and expression of p110α, p110β, and p110δ was down-regulated (5, 6, 11). In contrast, p85α was overexpressed in p110α knockout embryos, which died after 9.5 days of gestation (12). In this study, p110δ was inactivated by point mutation instead of by deletion to prevent changes in the expression levels of the other PI3K catalytic and regulatory subunits (see fig. S1 for details of the targeting strategy). Homozygous mutant mice (p110δD910A/D910A) were born at normal Mendelian ratios, were fertile, and did not show gross anatomical or behavioral abnormalities (13).

Expression of the mutated p110δ protein was equivalent to that of the wild-type (WT) protein, as was the expression of the other PI3K subunits (Fig. 1A). p110δ lipid kinase activity was completely abrogated in p110δD910A/D910Amice, with no alteration in the kinase activities of p110α and p110β (Fig. 1B). Total class IA PI3K activity was reduced 30 to 50% in thymocytes as well as in mature B and T cells (13).

Figure 1

Impaired PI3K signaling in B and T cells. (A) Immunoblots of class IA PI3K subunits expressed in thymocytes. As a control for equal protein loading, the filter was probed with anti-β-actin. (B) p110 subunits were immunoprecipitated from thymocyte lysates with isoform-specific antibodies, and the associated lipid kinase activity was assayed with phosphatidylinositol-4,5-bisphosphate as a substrate. Average results from duplicate measurements are represented as arbitrary units (a.u.). No PI3K activity was detected in p110δ immunoprecipitates from p110δD910A/D910Amice. Solid bars, WT; open bars, p110δD910A/D910A. (C) Purified B cells were treated with or without soluble anti-IgM (αIgM) F(ab')2 (10 μg/ml). Purified T cells were treated with or without anti-CD3 (αCD3) (10 μg/ml) bound to polystyrene beads. Immunoblots of total cell lysates with antibodies to Akt/PKB [total or Ser(P)473] or to Erk [total or Tyr(P)204] are shown. (D) Ca2+ flux was measured in B cells stimulated with αIgM F(ab')2 (10 μg/ml) and in T cells stimulated with αCD3 (5 μg/ml) and anti-hamster immunoglobulin (10 μg/ml) (24). Black, WT; gray, p110δD910A/D910A.

Antibodies to immunoglobulin M and to CD3 (anti-IgM and anti-CD3) were used to stimulate the B and T cell antigen receptors (BcR and TcR), respectively, and phosphorylation of the PI3K target Akt/protein kinase B (PKB) was used as an indirect measure of PI3K activity. In both cases, the induction of phosphoserine-473 [Ser(P)473] Akt/PKB was compromised in p110δD910A/D910A cells (Fig. 1C). Phosphorylation of the mitogen-activated protein kinase Erk was also reduced in p110δD910A/D910A B and T cells (Fig. 1C), which is consistent with the observation that, depending on signal strength, Erk activation can be regulated by PI3Ks (14). Ca2+ flux in response to anti-IgM crosslinking and to anti-CD3 crosslinking was attenuated in p110δD910A/D910A mice (Fig. 1D), in accord with an important role for PI3Ks upstream of Tec family kinase–mediated activation of phospholipase Cγ (PLCγ) and Ca2+signaling by antigen receptors (3, 15). These results implicate p110δ as an important mediator of antigen receptor signals in B and T cells.

The bone marrow of p110δD910A/D910A mice had reduced numbers of B220+IgM+ B cell progenitors (Fig. 2A). More specifically, the ratios of pre- to pro-B cells were altered: in WT mice there were almost four times as many CD43IgM+ pre-B cells as CD43+IgM pro-B cells, whereas the p110δD910A/D910A mice had equivalent numbers of pre- and pro-B cells (Fig. 2A). The spleens of p110δD910A/D910Amice had, on average, 50% fewer cells than the spleens of WT mice, with a more severe reduction of B cells than T cells (13). Analysis of the B220+ population in the spleen indicated similar levels of IgM+IgD+ follicular B cells; however, the fraction of CD23CD21+IgM+ marginal zone B cells was reduced in p110δWT/D910A spleens and was undetectable in the p110δD910A/D910A spleens (Fig. 2B). CD5+IgM+ peritoneal B-1 cells were readily detected in WT mice, were detected to a lesser extent in p110δWT/D910A littermates, and were almost undetectable in p110δD910A/D910A mice (Fig. 2C). These results show that p110δ plays an important role in the development and differentiation of B cells.

Figure 2

Impaired B and T cell maturation. Cells from bone marrow (A), spleen (B), peritoneal washes (C), thymus (D), and lymph nodes (E) were analyzed by flow cytometry with the indicated antibodies (24). The percentage of gated cells in particular quadrant is indicated in (A to D). The percentage of CD44lo cells is indicated in (E).

T cell development in the thymus appeared to be normal as judged by CD4 versus CD8 profiles (Fig. 2D). The lymph nodes contained normal ratios of CD4+ and CD8+ T cells, but the expression of CD44, an activation marker that is also used to distinguish CD44lo naı̈ve T cells from CD44hi memory T cells (16), was considerably reduced in T cells from p110δD910A/D910A mice (Fig. 2E). These results indicate that p110δ plays a role in the differentiation or survival of effector and, possibly, memory T cells. Peripheral blood cell counts appeared normal, with the exception of a twofold increase in the numbers of neutrophils (table S1).

Proliferation of purified B cells by anti-IgM cross-linking was almost completely abrogated (Fig. 3A). In contrast, proliferation in response to anti-CD40, interleukin 4 (IL-4), or lipopolysaccharide (LPS) was impaired but not abrogated (Fig. 3A). Stimulation with anti-CD40 in combination with IL-4 was also impaired (13). Proliferation of purified CD4+ T cells in response to anti-CD3 was reduced in p110δD910A/D910A mice (Fig. 3B). In contrast, proliferation in response to costimulation with anti-CD3 and anti-CD28 was normal, or even enhanced, in p110δD910A/D910A T cells (Fig. 3B). Stimulation of IL-2 production by anti-CD3 and anti-CD28 was normal (Fig. 3C), as was the capacity of phorbol 12,13-dibutyrate (PdBu)–stimulated T cells to proliferate in response to IL-2 (Fig. 3B). To examine the response of T cells to a physiological ligand, we crossed the p110δD910A/D910A mice with DO11-10 mice, which express a TcR specific for an ovalbumin peptide [Ova(323–339)] (17). T cells from p110δD910A/D910ADO11-10 mice exhibited diminished proliferative responses to the Ova peptide presented by WT antigen presenting cells (APCs) (Fig. 3D). IL-2 production was also attenuated in response to peptide stimulation, especially at elevated peptide concentrations (Fig. 3E). These results indicate that p110δ plays an important role in T cell activation but that the requirement for p110δ can be overcome by stimulation through CD28, a receptor that can enhance proliferation and IL-2 production independently of PI3K (18). The functional interaction between B and T cells depends on integrins that stabilize cell-cell adhesion. However, anti-CD3–induced cell adhesion to fibronectin via β1 integrins and to ICAM-1 via β2 integrins was unaffected by the p110δD910A mutation (Fig. 3F). TcR signaling is thought to be amplified by CD28-dependent accumulation of lipid rafts, membrane structures enriched in signaling proteins, at the interface between the T cell and the APC (19). The recruitment of lipid rafts, as induced by stimulating T cells with beads coated with anti-CD3 and anti-CD28 (19, 20), was about 60% less frequent in p110δD910A/D910A T cells than in WT T cells (Fig. 3, G and H). Therefore, although raft recruitment appears to be dispensable for T cell activation by antibodies, it may play a critical role during T cell activation by peptide–major histocompatibility complex ligands presented by APCs.

Figure 3

Impaired B and T cell responses in vitro. (A) Purified spleen B cells were stimulated with anti-IgM (αIgM) F(ab')2 (10 μg/ml), αCD40 (10 μg/ml), IL-4 (20 ng/ml), or LPS (10 μg/ml). Proliferation was measured by [3H]thymidine incorporation during the last 6 hours of a 48-hour culture. Similar results were obtained after 24 and 72 hours (13). Bars are as in Fig. 1. (B) Purified lymph node CD4+ T cells were stimulated with polystyrene beads coated with αCD3 (0.1, 1, or 10 μg/ml) with or without αCD28 (10 μg/ml), and proliferation was measured as described in (A). Proliferation was also measured in response to PdBu (50 ng/ml) and IL-2 (10 ng/ml) and in response to PdBu (50 ng/ml) and ionomycin (iono) (1 μg/ml). (C) CD4+ T cells were stimulated with polystyrene beads coated with αCD3 (1 μg/ml) with or without αCD28 (10 μg/ml), and supernatant was harvested after 24 hours for analysis of IL-2 production by enzyme-linked immunosorbent assay (ELISA). (D) CD4+ T cells from WT and p110δD910A/D910Amice that had been crossed onto the DO11-10 background were stimulated with mitomycin C–treated Balb/c spleen B cells and the indicated concentrations of Ova323–339 peptide. Proliferation was measured by [3H]thymidine incorporation during the last 6 hours of a 48-hour culture. Solid symbols, WT; open symbols, p110δD910A/D910A (E) CD4+ DO11-10 WT and p110δD910A/D910A T cells were stimulated as in (D). Supernatants were harvested at 40 hours and assayed for IL-2 content by ELISA. (F) Adhesion of T cells stimulated with anti-CD3 to plate-bound fibronectin and ICAM-1 is expressed as fold increase in adhesion of anti-CD3–stimulated versus nonstimulated T cells. (G) Purified spleen and lymph node T cells were stimulated for 25 min with polystyrene beads coated with αCD3, αCD28, or αCD3 and αCD28. (Left) Bead (dotted circle) with a WT T cell, scored as positive (presenting raft aggregation toward the bead). (Right) p110δD910A/D910A T cell/bead complex, scored as negative. Graph represents the percentage of cells presenting raft aggregation toward the bead. The values presented for stimulation by αCD3 and αCD28 are the average of six independent experiments and of two experiments for the other conditions. (H) Cells were incubated with αCD3 and αCD28–coated beads, fixed after the indicated incubation times, and analyzed as described in (G). All results presented are from three or four measurements ±SD, except where noted (24).

The p110δD910A/D910A mice had somewhat lower serum immunoglobulin levels than WT littermates, the IgM and IgA isotypes being most consistently reduced (Fig. 4A). To examine in vivo humoral immune responses, we immunized mice with TNP-KLH [(2,4,6- trinitrophenyl)–keyhole limpet hemocyanin] and measured relative TNP-specific antibody titers in the serum on day 12. The total titer of TNP-specific antibodies made by the p110δD910A/D910A mice was about 4% of that from the WT mice, with a more profound reduction of the IgM, IgG1, and IgG3 isotypes than IgG2a and IgG2b (Fig. 4B). Consistent with these findings, the spleens of TNP-KLH–immunized WT mice contained numerous germinal centers (GCs), whereas formation of these was largely absent in the spleens of TNP-KLH–immunized p110δD910A/D910Amice (Fig. 4C). These results demonstrate that T cell–dependent humoral immune responses are deficient in p110δD910A/D910A mice. To measure T cell–independent antibody responses, we immunized mice with TNP-Ficoll and measured relative TNP-specific antibody titers on day 7. Consistent with the impaired capacity of p110δD910A/D910A B cells to signal through the BcR, T cell–independent humoral responses were strongly impaired (Fig. 4D).

Figure 4

Impaired in vivo immune response and IBD. (A) Total immunoglobulin levels in serum from four WT and three p110δD910A/D910A mice. Each dot represents an individual mouse. (A, B, and D) Average median effective concentration of serum titers is indicated by a horizontal bar. Statistically significant differences are indicated by * (P < 0.05) or ** (P < 0.01), as determined by the two-tailed Student's t test. (B) Anti-TNP titers from four WT and four p110δD910A/D910A mice immunized with TNP-KLH with alum used as an adjuvant. Serum was obtained from tail bleeds on day 12 after immunization. (C) GC formation in spleens from TNP-KLH immunized mice as detected by peanut agglutinin (PNA) staining. (D) Anti-TNP titers from four WT and three p110δD910A/D910A mice immunized with TNP-Ficoll. (E) Histological examination of hematoxylin and eosin–stained sections of mesenteric lymph nodes (i and ii), Peyer's patches (PPs) (iii and iv), and cecum (v to viii) of WT and p110δD910A/D910A mice. The mesenteric lymph node (i) and the PPs (iii) of WT mice have secondary follicles with prominent GCs (star). GCs are conspicuously absent from the follicles (star) of the mesenteric lymph node (ii) and PPs (iv) of p110δD910A/D910A mice. (v and vii) Normal intestinal mucosa of the cecum of a control mouse with short crypts (arrow) and sparsely cellular lamina propria. (vi) Cecum of p110δD910A/D910A mice have thickened mucosa with hyperplasic glands and mixed inflammatory infiltration of the lamina propria. (vii) Higher magnification of the cecum of a p110δD910A/D910A mouse showing focal erosion of mucosa with neutrophilic inflammation (arrowhead) and a regenerating gland (arrow). (i to vi) Bar, 300 μm; (vii and viii) bar, 50 μm (24).

Histological examination revealed lymphoid hypoplasia and lack of GCs in the spleen, lymph nodes, and Peyer's patches (PPs) in p110δD910A/D910A mice (Fig. 4E, i to iv) (13). All other organs appeared to be normal, with the interesting exception of the large intestine. The p110δD910A/D910A mice developed a mild inflammatory bowel disease (IBD), which was segmental and mostly limited to the cecal and rectal sections of the large intestine. The lesions were characterized by mucosal hyperplasia, crypt abscesses, and mixed leukocyte infiltrates, associated with regenerative changes of the glandular epithelium (Fig. 4E, v to viii). Other gene-targeted mice with T cell defects develop IBD, which may reflect an important role for regulatory T cells in maintaining tolerance to gut flora by secreting anti-inflammatory cytokines such as IL-10 and transforming growth factor–β (21). IBD in humans, which includes Crohn's disease and ulcerative colitis, can result from inherited mutations in disease susceptibility genes interacting with environmental factors such as the intestinal bacterial flora. The human p110δ gene (PIK3CD) maps to the IBD7 susceptibility locus on chromosome 1p36 (22, 23). Given the IBD phenotype of p110δD910A/D910A mice,PIK3CD should be further investigated as a candidate human IBD susceptibility gene.

p110δ plays a unique role in antigen receptor signaling that is not readily compensated for by p110α or p110β. The selective attenuation of immune function in p110δD910A/D910A mice suggests that a specific inhibitor of p110δ could effectively suppress B and T cell–mediated autoimmunity and possibly B and T cell transformation.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1073560/DC1

Materials and Methods

Fig. S1

Table S1

  • * These authors contributed equally to this report.

  • Present address: Roslin Institute, Roslin, Midlothian, EH25 9PS, UK.

  • To whom correspondence should be addressed. E-mail: bartvanh{at}ludwig.ucl.ac.uk

REFERENCES AND NOTES

View Abstract

Navigate This Article