Function of PI3Kγ in Thymocyte Development, T Cell Activation, and Neutrophil Migration

See allHide authors and affiliations

Science  11 Feb 2000:
Vol. 287, Issue 5455, pp. 1040-1046
DOI: 10.1126/science.287.5455.1040


Phosphoinositide 3-kinases (PI3Ks) regulate fundamental cellular responses such as proliferation, apoptosis, cell motility, and adhesion. Viable gene-targeted mice lacking the p110 catalytic subunit of PI3Kγ were generated. We show that PI3Kγ controls thymocyte survival and activation of mature T cells but has no role in the development or function of B cells. PI3Kγ-deficient neutrophils exhibited severe defects in migration and respiratory burst in response to heterotrimeric GTP-binding protein (G protein)–coupled receptor (GPCR) agonists and chemotactic agents. PI3Kγ links GPCR stimulation to the formation of phosphatidylinositol 3,4,5-triphosphate and the activation of protein kinase B, ribosomal protein S6 kinase, and extracellular signal-regulated kinases 1 and 2. Thus, PI3Kγ regulates thymocyte development, T cell activation, neutrophil migration, and the oxidative burst.

PI3Ks constitute a family of evolutionarily conserved lipid kinases that regulate a vast array of fundamental cellular responses, including proliferation, transformation, protection from apoptosis, superoxide production, cell migration, and adhesion (1). These responses result from the activation of membrane-trafficking proteins and enzymes such as the phosphoinositide-dependent kinases (PDKs), protein kinase B (PKB), and S6 kinases (S6Ks) by the key second-messenger phosphatidylinositol 3,4,5-triphosphate [PtdIns (3,4,5)P3] (2). PtdIns (3,4,5)P3 is generated from phosphatidylinositol 4,5-bisphosphates through phosphorylation at the 3′ position by PI3Ks that are linked to tyrosine kinase–based receptors or G proteins. PI3Kγ is activated in response to GPCRs and can be directly activated by G protein βγ subunits (Gβγ) in vitro (1). The catalytic subunit of PI3Kγ (p110γ) associates with a p101 regulatory subunit but not with the p85 family proteins that regulate other PI3K proteins (3).

To define the physiological roles of PI3Kγ, we disrupted the p110γ catalytic PI3Kγ subunit and generated PI3Kγ null mice (4). Homozygous PI3Kγ −/−mice were born at the expected Mendelian ratio, appear healthy, and are fertile. Hematopoietic lineages were examined in peripheral blood and bone marrow of PI3Kγ +/− andPI3Kγ −/− littermates. No significant differences in basophil, erythrocyte, or platelet numbers were observed, and all white blood cells exhibited normal morphology on smears stained with Wright-Giemsa or myeloperoxidase. However, PI3Kγ deficiency led to increases in neutrophil, monocyte, and eosinophil populations (Table 1). Total and relative numbers of Gr1+Mac1+ myeloid cells were also significantly increased in the spleen, but not in bone marrow, ofPI3Kγ −/− mice (Tables 1 and2). Thymocyte numbers were significantly reduced in PI3Kγ −/− mice (Table 3). The reduction in total thymocyte numbers was also apparent when PI3Kγ −/−rag1 −/− chimeric mice created by blastocyst complementation were compared to PI3Kγ +/−rag1 −/− chimeric mice, showing that the developmental defect was intrinsic to thymocytes. Thus,PI3Kγ −/− mice exhibit two gross phenotypes: reduced numbers of thymocytes and increased numbers of myeloid cells in the spleen and blood.

Table 1

Blood cell numbers inPI3Kγ −/− mice. Seven- to 12-week-oldPI3Kγ +/− (n = 5) andPI3Kγ −/− (n = 5) littermate mice were used. Bold numbers indicate statistically significant differences between PI3Kγ +/− andPI3Kγ −/− mice (Mann-Whitney Utest; P < 0.05). Values are given as the mean ± SEM. WBC, white blood cells.

View this table:
Table 2

Cell populations in PI3Kγ-deficient mice. Seven- to 12-week-old PI3Kγ +/− andPI3Kγ −/− littermate mice were used. Total cells from thymi (n = 7), spleens (n = 7), lymph nodes (n = 3), and bone marrows (n = 3) were stained with antibodies against the indicated proteins. Populations were determined by FACScan. Bold numbers indicate statistically significant differences betweenPI3Kγ +/− andPI3Kγ −/− mice (Mann-Whitney Utest; P < 0.05). Values are given as the mean ± SEM.

View this table:
Table 3

Numbers of lymphoid cells inPI3Kγ −/− mice. Thymi (n = 12), spleens (n = 5), and all lymph nodes (n= 5) from 4- to 7-week-old PI3Kγ +/− andPI3Kγ −/− littermate mice or 6-week-oldPI3Kγ +/− chimeric andPI3Kγ −/− chimeric mice (n = 4 for all organs) were analyzed. Bold numbers indicate statistically significant differences between PI3Kγ +/− andPI3Kγ −/− mice or betweenPI3Kγ +/−rag1 −/−and PI3Kγ −/−rag1 −/− chimeric mice (Mann-WhitneyU test; P < 0.05). Values are given as the mean ± SEM.

View this table:

In vitro studies indicate that, like PI3Kγ, type-IA PI3Ks can also be activated in response to GPCR stimulation (5). We examined whether PI3Kγ is essential for generating PtdIns (3,4,5)P3 after GPCR stimulation (6). Receptors of exogenous and endogenous chemoattractants, including the peptide N-formyl-Met-Leu-Phe (fMLP) and the complement component C5a are coupled to G Protein (Gi). No PtdIns (3,4,5)P3accumulation could be detected in PI3Kγ-deficient neutrophils treated with the GPCR agonists fMLP or C5a (Fig. 1A). PI3Ks and PtdIns (3,4,5)P3 have been implicated in Ca2+mobilization, phosphotyrosine signaling, and the activation of kinases such as PKB, S6Ks, and extracellular signal-regulated kinases 1 and 2 (ERK1 and ERK2, respectively) (1, 2, 7). Stimulation of PI3Kγ +/− neutrophils withfMLP or C5a increased kinase activity of PKB (Fig. 1B) and ERK1 and ERK2 (Fig. 1C). In contrast, activation of PKB (Fig. 1B) and phosphorylation of PKB at Ser473 (Fig. 1D) and Thr308 induced by fMLP or C5a were completely abrogated in PI3Kγ −/−neutrophils. Phosphorylation of S6K is regulated by PDKs and PKB, both of which are activated by PtdIns (3,4,5)P3. No S6K phosphorylation was observed inPI3Kγ −/− cells (Fig. 1E). Activation (Fig. 1C) and phosphorylation (Fig. 1D) of ERK1 and ERK2 inPI3Kγ −/− neutrophils treated withfMLP or C5a were decreased to amounts similar to those observed in wortmannin-treated control cells. Ca2+mobilization (Fig. 1F), activation of p38 kinase, and overall tyrosine phosphorylation induced by fMLP or C5a treatment occurred to the same extent and with the same kinetics inPI3Kγ +/− andPI3Kγ −/− neutrophils. Activation of PKB in response to the tyrosine kinase-coupled receptor agonist granulocyte-macrophage colony-stimulating factor was comparable inPI3Kγ +/− andPI3Kγ −/− neutrophils, indicating that the lack of PKB activation in PI3Kγ −/− cells in response to GPCR agonists is not due to an intrinsic defect of PKB activation.

Figure 1

PI3Kγ regulates PtdIns (3,4,5)P3 production and activation of PKB and ERKs. (A) Failure of PI3Kγ −/−neutrophils to produce PtdIns (3,4,5)P3. Cells labeled with32P were incubated with or without wortmannin (wort, 100 nM) for 10 min before stimulation with fMLP (8 μM) or C5a (100 ng/ml). Left panel, thin-layer chromatography. Ori, origin. Right panel, PtdIns (3,4,5)P3 quantitation. (Bthrough E) Activation of PKB [(B) and (D)], ERK1 and ERK2 [(C) and (D)], and S6K (E). Neutrophils were incubated with or without wortmannin (wort, 100 nM) before stimulation withfMLP (8 μM) or C5a (100 ng/ml). PKB and ERK1 and ERK2 activation were determined in triplicate by in vitro kinase assays and with phospho-specific antibodies against PKB (Ser473), ERK1 and ERK2 (Thr202/Tyr204), and S6K (Thr421/Ser424). (F) Ca2+ mobilization. Indo-1–loaded neutrophils were stimulated with 8 μM fMLP, and Ca2+ flux was monitored in real time by FACS. Arrows indicate the time of fMLP addition. (G) GPCR agonist–specific activation of PKB and ERK1 and ERK2. Serum-starved BMMCs were treated with thrombin (Thr, 5 units/ml), SCF (100 ng/ml), or IL-3 (5 ng/ml). Phosphorylation of PKB and ERK1 and ERK2 was monitored by phospho-specific antibodies. (H) In vitro PI3K activity. BMMCs were left untreated (none), or activated with SCF (100 ng/ml) or thrombin (Thr, 5 units/ml) for 1 min. Total phosphorylated proteins were immunoprecipitated with an antibody to phosphotyrosine, and the associated PI3K activity was determined. Baseline p85α/PI3K activity was similar inPI3Kγ +/− andPI3Kγ −/− BMMCs. Upper panel, thin-layer chromatography; lower panel, quantitation of PtdIns (3,4,5)P3. Error bars in (A), (B), (C), and (H) indicate ±SD.

To extend the role of PI3Kγ in GPCR signaling to a different cell type, we established bone marrow mast cell (BMMC) lines. BMMCs respond to engagement of the G protein–coupled thrombin receptor, the tyrosine kinase–based receptors for c-Kit, interleukin-3 (IL-3), and immunoglobulin E (IgE) (8). Stimulation ofPI3Kγ +/− BMMCs with thrombin, stem cell factor (SCF or c-Kit ligand), IL-3, and IgE led to phosphorylation of PKB and ERK1 and ERK2 (Fig. 1G). Phosphorylation of ERK1 and ERK2 and PKB induced by the GPCR agonist thrombin was abolished inPI3Kγ −/− BMMCs (Fig. 1G). Thrombin did not activate type-IA PI3Ks in BMMCs assessed by PI3K activity present in antiphosphotyrosine immunoprecipitates (Fig. 1H). SCF-stimulated activation of type-IA PI3Ks occurred equally well inPI3Kγ −/− and control BMMC. These data show that PI3Kγ is the essential lipid kinase linking GPCRs to PtdIns (3,4,5)P3 accumulation and activation of PKB and S6K.

To determine the role of PI3Kγ in the inflammatory response, we examined neutrophil migration in experimental models of peritonitis in vivo. Accumulation of neutrophils in the peritoneal cavities was significantly reduced in both casein- and Listeria monocytogenes–treated PI3Kγ −/− mice in comparison to control animals (Fig. 2A). We further tested purified neutrophils for in vitro chemotaxis in response to GPCR agonists. fMLP- and C5a-induced chemotaxis of neutrophils from PI3Kγ −/− mice were decreased by 70% in comparison to neutrophils from heterozygous littermates (Fig. 2B). PI3Kγ −/−neutrophils adhered to fibronectin-coated surfaces as tightly asPI3Kγ +/− cells (9), suggesting that the decreased chemotaxis in PI3Kγ −/−neutrophils is due to impaired motility and not altered adhesion. These in vivo and in vitro results identify PI3Kγ as a critical link between GPCR stimulation and the chemotactic response.

Figure 2

Neutrophil functions. (A) Impaired accumulation of PI3Kγ −/−neutrophils in response to casein-induced peritonitis andListeria infection. Mean (error bars, ±SEM) neutrophil numbers in the peritoneal cavities of 12 (casein) and 8 littermates (Listeria) were determined after intraperitoneal injection of casein or intraperitoneal Listeria infection. Baseline numbers of peritoneal neutrophils are shown (none). For all figures, statistical differences between PI3Kγ −/− andPI3Kγ +/− groups were determined with the Student's t test; *P < 0.05; **P < 0.01. (B) Decreased neutrophil chemotaxis. Freshly isolated neutrophils were tested for the ability to migrate in response to 8 μM fMLP. Basal migration of PI3Kγ +/− neutrophils is shown as 100%. Values represent the mean (error bars, ±SEM) (n = 8). (C) Oxidative burst. Neutrophils were isolated from the bone marrow and incubated for 15 min with fMLP (8 μM), C5a (100 ng/ml), or PMA (0.5 μM). Mean values of O2 production (O2 produced over 5 min by 106cells) are shown. (D) Normal FcγR-mediated phagocytosis. Neutrophils were incubated with IgG-coated beads at 0° or 37°C. Representative FACScan profiles of ingested IgG beads are shown. The response was abolished by wortmannin.

In addition to chemotaxis, it has been reported that PI3Ks link stimulation of GPCRs to respiratory burst (superoxide anion O2 production) (10). GPCR-induced respiratory burst was decreased in freshly isolated bone marrow neutrophils from PI3Kγ −/− mice (Fig. 2C). The direct protein kinase C–activator phorbol 12-myristate 13-acetate (PMA) induced superoxide anion formation in bothPI3Kγ −/− andPI3Kγ +/− neutrophils. Thus, GPCR-triggered respiratory burst depends on PI3Kγ. However, the requirement of PI3Kγ for oxidative burst is not absolute because primed neutrophils isolated from the peritoneal cavity still produce O2 in response to C5a and fMLP.PI3Kγ −/− neutrophils were able to ingest IgG-coated particles as readily as PI3Kγ+/−cells (Fig. 2D), suggesting that PI3Kγ does not act downstream of these tyrosine kinase–based receptors. The kinetics and extents of apoptosis were comparable among freshly isolated total myeloid cells and neutrophils from the spleens ofPI3Kγ +/− andPI3Kγ −/− mice. Thus, in neutrophils, PI3Kγ links GPCRs to cell migration and superoxide formation.

In addition to neutrophil defects,PI3Kγ −/− mice exhibit a reduction in thymic cellularity (Table 3). Thymocytes undergo defined stages of development, from CD4CD8 double-negative (DN) precursors to CD4+CD8+ double-positive (DP) immature thymocytes, and finally, they mature to single-positive (SP) CD4+ or CD8+ T cells (11). The proportions of mature CD4+ and CD8+ SP thymocytes were equal in PI3Kγ −/−,PI3Kγ +/−, andPI3Kγ −/−rag1 −/− mice. There was a slight decrease in the proportion of DP cells and an increase in DN cells (Table 2). No differences were found in the amount of αβ T cell receptor (TCRαβ), CD3, CD4, CD8, CD28, CD45, TCRVβ subclasses, and CD95 expressed on the surface of SP and DP thymocytes. The maturation of DN precursor populations (defined by surface expression of CD44 and CD25), maturation of DP thymocytes to SP thymocytes (defined by expression of CD69, CD44, heat stable antigen, CD5, and H2-Kb), and proliferation of DN precursors and mature thymocytes were similar amongPI3Kγ+/+, PI3Kγ +/−, andPI3Kγ −/− mice (12). These data suggest that early thymocyte development and progression from DP to mature thymocytes do not require PI3Kγ.

It is estimated that 90 to 95% of DP thymocytes are lost to apoptotic “death by neglect” due to the expression of nonfunctional antigen receptors (11). To examine the impact of PI3Kγ deficiency on thymocyte survival, we evaluated DP thymocyte apoptosis to treatment with antibodies to CD3ɛ and Fas, γ irradiation, or dexamethasone (13). No differences in the kinetics or extents of apoptosis were observed between mutant and control mice (Fig. 3A). Human patients with mutations in adenosine deaminase exhibit severe combined immunodeficiency disease and a defect in thymocyte maturation (14). Adenosine is naturally present in thymi in vivo, and the adenosine A2A receptor is expressed on thymocytes. The adenosine A2A receptor is a GPCR that can activate adenylate cyclase and PI3Kγ (15). Stimulation of PI3Kγ +/− thymocytes with adenosine A2A receptor agonistsN6-[4-[[[4-[[[(2-aminoethyl)amino]carbonyl]methyl]-anilino]carbonyl] methyl]phenyl]adenosine (ADAC) or 2-p-(2-carboxyethyl)phenethylamino-5′-N-ethylcarboxamidoadenosine hydrochloride (CGS) caused a small amount of cell death (Fig. 3A). However, PI3Kγ −/− thymocytes exhibited much more apoptosis after stimulation with either adenosine receptor agonist (Fig. 3A). Treatment of thymocytes with ADAC or CGS with antibody to CD3ɛ (anti-CD3ɛ) slightly increased the death ofPI3Kγ +/− cells but substantially increased cell death of PI3Kγ −/− thymocytes (Fig. 3B). We further investigated the influence of PI3Kγ on thymocyte apoptosis in vivo by injecting control and mutant mice with monoclonal antibodies (mAbs) to CD3ɛ, which results in apoptosis of DP thymocytes (16). In PI3Kγ −/− mice, not only was the total number of thymocytes decreased, but also the specific depletion of the DP population was enhanced in comparison to that in PI3Kγ +/− mice (Fig. 3, C and D). These results indicate that PI3Kγ has a role in the maintenance of homeostasis of the thymus by regulating TCR- and GPCR-induced thymocyte apoptosis.

Figure 3

PI3Kγ regulates thymocyte survival (A) Increased susceptibility ofPI3Kγ −/− thymocytes to adenosine receptor–mediated apoptosis. Freshly isolated thymocytes were stimulated for 20 hours with dexamethasone (0.1 μM), γ irradiation (5 Gy), anti-CD95 (FAS, 1 μg/ml), immobilized anti-CD3ɛ(0.1 μg/ml), or the adenosine receptor agonists ADAC (10 μM) or CGS (10 μM). Mean percentages (error bars, ±SD) of viable DP thymocytes are indicated (14). Spontaneous apoptosis was comparable among nonstimulated PI3Kγ +/− andPI3Kγ −/− thymocytes. (B) Adenosine receptor agonists increase anti-CD3–mediated cell death. Thymocytes were cultured for 20 hours in medium alone (–) or in the presence of (+) immobilized anti-CD3ɛ (0.1 μg/ml), ADAC (10 μM), CGS (10 μM), or anti-CD3ɛ plus adenosine analogs. Percent viability of DP thymocytes (14) was normalized to the percentage of viable DP cells in untreated cultures (100%). (C) Increased anti-CD3–mediated cell death in vivo. Total thymocytes were isolated from mice 48 hours after intraperitoneal injection with phosphate-buffered saline (PBS) (control) or mAb to CD3ɛ (50 μg/200 μl PBS). Cells were stained with anti-CD4-PE, anti-CD8-FITC, and 7-AAD and analyzed by FACS. Percentages of different thymocyte subpopulations in gated live cells are shown within quadrants. Numbers above panels indicate total numbers of viable DP cells (mean ± SD) (n = 6 mice per group). (D) Time course of anti-CD3–mediated cell death in vivo. Total thymocytes were isolated from mice at 0, 24, or 48 hours after intraperitoneal injection with anti-CD3ɛ (50 μg/200 μl PBS). Mean percentages (error bars, ±SD) of DP thymocytes that remained viable from four (24 hours) and six (48 hours) independent experiments are shown.

Gene targeting of the PI3K subunit p85α leads to a block in early B cell development and defective proliferation of mature B cells in response to IL-4, CD40, and B cell receptor (BCR) (anti-IgM) stimulation (17). Development, proliferation, and IL-2 production of T cells appeared normal inp85α −/− mice, although inhibition of PI3K with wortmannin partially inhibits T cell proliferation and cytokine production after TCR and CD28 stimulation.PI3Kγ −/− andPI3Kγ −/−rag1 −/− chimeric mice displayed normal numbers and differentiation of B220+CD25+, B220+CD25, B220+CD43+, B220+CD43, B220+sIgM+, and CD19+sIgM+sIgD+ B cells in the bone marrow (sIg, surface Ig); conventional CD19+sIgM+sIgD+CD23+ B cells in peripheral lymphoid organs; and B1 (CD5+IgM+) B cells in the peritoneal cavity (Tables 2 and 3). Proliferation of PI3Kγ −/−B cells in response to a mAb to IgM, the F(ab′)2 fragment of the antibody to IgM, anti-CD40 (Fig. 4A), IL-4, or lipopolysaccharide was similar to that of PI3Kγ +/− B cells (18). These data show that PI3Kγ has no apparent role in B cell development or in BCR- or CD40-mediated B cell proliferation.

Figure 4

Defective T cell activation inPI3Kγ −/− mice. (A) B cell activation. Purified spleen B cells (1 × 105 per well) were incubated for 36 hours in medium alone (none), anti-IgM (20 μg/ml), anti-IgM F(ab′)2 [F(ab′)2, 15 μg/ml], anti-CD40 (5 μg/ml), or anti-IgM F(ab′)2[F(ab′)2, 15 μg/ml) plus anti-CD40 (5 μg/ml). Mean values of [3H]-thymidine uptake (error bars, ±SD) are shown. Similar results were observed for various seeding numbers and at earlier and later time points of activation. (B) Proliferation and (C) IL-2, and (D) IFN-γ production. T cells were purified from lymph nodes and activated with anti-CD3ɛ (0.1 or 0.5 μg/ml), anti-CD3ɛ(0.1 μg/ml) plus anti-CD28 (20 ng/ml), Con A (2 μg/ml), and PMA (10 ng/ml) plus Ca2+ ionophore (100 ng/ml) (PMA+Iono). Culture supernatants were analyzed after 24 hours for cytokine production by ELISA. Proliferation was determined at 48 hours. Values are the means (error bars, ±SEM) of triplicate cultures. Phenotypes and activation status of input T and B cells populations were similar between PI3Kγ +/− andPI3Kγ −/− mice. (E) Footpad-swelling reaction. Mice were inoculated with 2000 PFU of LCMV, and swelling was assessed daily. Mean values (error bars, ±SEM) from five PI3Kγ +/− and fivePI3Kγ −/− littermates are shown. (F) Cytotoxicity. Mice were inoculated with LCMV as in (E). Spleen cells were harvested 10 days later, and cytotoxicity was measured using EL-4 target cells pulsed with the LCMV-specific peptide NP118-126 or the nonspecific peptide AV. (G) Responses to the T cell–dependent hapten NIP-OVA. Serum IgG1 titers were determined 7, 14, and 21 days after immunization. Arbitrary units of optical density (at 405 nm) of NIP-specific IgG1 titers are shown for individual mice.

In contrast to B cells, thymic development was altered inPI3Kγ −/− mice. Lymph nodes ofPI3Kγ −/− mice contained normal numbers and ratios of CD4+ and CD8+ T cells. However, CD4+ T cells, but not CD8+ T cells, were reduced in PI3Kγ −/− spleens (Table 2). The amounts of TCRαβ, CD3, CD4, CD8, CD28, CD45, CD44, LFA-1, CD25, and CD69 on the surface of both splenic and lymph node CD4+ and CD8+ T cells were comparable inPI3Kγ +/− andPI3Kγ −/− mice. WhereasPI3Kγ +/− andPI3Kγ −/− T cells proliferated equally well in response to stimulation by PMA/Ca2+ ionophore,PI3Kγ −/− T cells showed impaired proliferation in response to anti-CD3ɛ or concanavalin A (Con A) stimulation (Fig. 4B) (18). The engagement of the costimulatory CD28 receptor increased TCR/CD3–mediated proliferation in both PI3Kγ +/− andPI3Kγ −/− T cells and rescued the proliferation defect of PI3Kγ −/− T cells (Fig. 4B). Nevertheless, PI3Kγ −/− T cells produced lower amounts of IL-2 (Fig. 4C) and interferon-γ (IFN-γ) (Fig. 4D) in response to treatment with anti-CD3ɛ and anti-CD28 or Con A. The efficacy of splenic antigen-presenting cells to induce proliferation and IL-2 production was comparable betweenPI3Kγ +/− andPI3Kγ −/− mice with mixed lymphocyte reactions (18). The functional defect in cytokine production can also be observed in PI3Kγ −/− T cells treated with PMA/Ca2+ ionophore, a stimulus that bypasses the initial TCR signal (Fig. 4, C and D). Moreover, TCR-mediated Ca2+ flux, tyrosine phosphorylation, and activation of tyrosine kinases were comparable amongPI3Kγ +/− andPI3Kγ −/− T cells. Thus, it appears that PI3Kγ does not act downstream of the TCR but regulates a second signal through GPCRs.

To examine the role of PI3Kγ in T cell responses in vivo, we injected PI3Kγ+/− andPI3Kγ−/− mice in the footpad with lymphocytic choriomeningitis virus (LCMV) (19). The early phase (days 6 to 8 after infection) of this footpad-swelling reaction is mediated by CD8+ cytotoxic T lymphocytes (CTLs), whereas the later phase depends on CD4+ T cells (20). Whereas PI3Kγ+/− mice developed an effective LCMV-induced footpad-swelling reaction starting from day 6 after infection, footpad swelling was reduced inPI3Kγ −/− mice (Fig. 4E). T cells recovered from the spleens of both PI3Kγ +/− andPI3Kγ −/− mice generated a normal primary cytotoxic response to LCMV peptides (Fig. 4F).PI3Kγ +/− andPI3Kγ −/− mice were further immunized with the T helper cell–dependent hapten nitroiodophenylacetic acid conjugated to ovalbumin (NIP-OVA). WhereasPI3Kγ+/− mice exhibited high titers of antibodies to NIP-specific IgG1, antibodies to NIP were reduced inPI3Kγ −/− mice (Fig. 4G). Thus, PI3Kγ is required to generate effective CD8+ T cell–dependent antiviral responses and functional T helper cell–dependent responses to hapten antigens in vivo.

PI3Ks link surface receptors to phosphatidylinositol-regulated signaling pathways. We report the generation of viable mice lacking the p110γ catalytic PI3K subunit (PI3Kγ −/−).PI3Kγ −/− mice display two principal phenotypes: accumulation of neutrophils and reduced thymic cellularity. Neutrophils had defects in migration and oxidative burst in response to agonists that trigger GPCRs. PI3Kγ mediates PtdIns (3,4,5)P3 production and activation of PKB and S6K in response to the GPCR agonists C5a, fMLP, and thrombin in neutrophils and mast cells. In addition, PI3Kγ is critical for GPCR-triggered activation of ERK1 and ERK2. In contrast to mutation of p85α (the regulatory subunit of PI3Kα, PI3Kβ, and PI3Kδ, which leads to developmental and functional defects in B cells, but not in T cells), deletion of PI3Kγ had no effect on B lymphocytes. Instead, PI3Kγ regulates proliferation and cytokine production of T lymphocytes. Moreover, PI3Kγ expression is required for an effective T cell–dependent footpad-swelling reaction following viral challenge and functional T helper cell–dependent responses to hapten antigens in vivo. In thymocytes, PI3Kγ provides a developmental survival signal.

  • * To whom correspondence should be addressed. E-mail: jpenning{at}


Stay Connected to Science

Navigate This Article