PI3-kinase p110α-null mice are embryonic lethal (24). So, to assess the role of p110α PI3-kinase in T cell function, mice with conditional deletion of the pik3ca gene in T cells were generated by crossing CD4-Cre mice and mice with a floxed pik3ca gene (p110α^flox/flox) (24). CD4-Cre^+/−/p110α^flox/flox mice will be hereinafter referred to as p110α^−/−ΔT, whereas CD4-Cre^−/−/p110α^flox/flox littermates will be termed wild type (WT). PI3-kinase p110α was efficiently removed from peripheral CD4⁺ and CD8⁺ T lymphocytes of p110α^−/−ΔT mice; however, the PI3-kinase p110δ subunit or other proteins like CD4, or Erk were unaffected (see Figure S1 in Supplementary Material, and data not shown). Most subpopulations in the peripheral lymphoid organs of WT and p110α^−/−ΔT mice were not significantly changed, including the percentage of total T (CD3⁺) cells, CD8 T lymphocytes, B lymphocytes, γδ, and NKT lymphocytes, or NK cells (see Figure S1 in Supplementary Material, and data not shown). However, p110α^−/−ΔT mice showed a slightly lower number of spleen cells and a lower percentage of CD4⁺ cells (Figure S1 in Supplementary Material), even though the percentage of naive and memory or Treg cells within the CD4⁺ T cell population was not significantly different. Analysis of thymus cells indicated that this was not due to a deficient development of mature CD4⁺ T lymphocytes (Figure S1 in Supplementary Material).

Next, the effect of PI3-kinase p110α removal on the activation of naive CD4⁺ T lymphocytes in vitro was determined. Secretion of IL-10 and, particularly, IFN-γ were significantly enhanced in p110α^−/−ΔT cells activated with anti-CD3 plus anti-CD28 antibodies, as compared to WT littermates (Figure 1A). In contrast, IL-2 secretion or proliferation was not significantly changed (Figure 1A, and data not shown). The levels of the IFN-γ master transcription factor T-bet were also significantly enhanced in activated CD4⁺ T cells of p110α^−/−ΔT mice (Figure 1B). Induction of T-bet expression in CD4⁺ T lymphocytes depends on the activity of MAP kinases like P38 and, particularly, Erk, as revealed using specific inhibitors (Figure 1C). Consequently, the impact of p110α removal in early MAP kinase activation was analyzed (Figure 1D). As expected, activation of naive CD4⁺ T with anti-CD3 plus anti-CD28 induced Tyr phosphorylation of specific substrates, and some of them showed enhanced phosphorylation in p110α^−/−ΔT cell lysates. Furthermore, Erk activation was clearly higher in p110α^−/−ΔT cells than in WT cells. In unstimulated WT cells, the basal phosphorylation of P38 was not significantly changed upon anti-CD3 plus anti-CD28 stimulation. In p110α^−/−ΔT cells, the basal level of P38 activation was higher than in WT cells, and was enhanced by CD3 plus CD28 stimuli (Figure 1D). Interestingly, T cells lacking p110α show enhanced levels of phosphorylation of the PI3K target Akt. This suggests that other PI3K catalytic subunits like p110δ can advantageously replace p110α concerning PI3K activation. Taken together, these data indicate that p110α removal enhances early activation signals in CD4⁺ T lymphocytes, leading to enhanced MAPK activity, T-bet induction, and eventually to higher IFN-γ secretion.

Naive CD4⁺ T cells were then differentiated into effector T cell populations “in vitro,” as under neutral activation conditions, i.e., in the absence of added cytokines, the effect of p110α-deficiency on cytokines other than IL-2, IL-10, or IFN-γ could not be determined.

In differentiated Tfh cells, the production of IL-21 (Tfh) was significantly enhanced in p110α-deficient cells (Figure 2A). Then, ICOS-induced elongation of Tfh cells was analyzed. Tfh express ICOS at high levels, and it is known that ICOS interaction with ICOS-ligand expressed by bystander B cells is needed for Tfh cells to reach germinal centers. This phenomenon is dependent on PI3-kinase activity, so that ICOS-induced, PI3-kinase-dependent elongation of ICOS⁺ T cells might be involved in germinal center formation in vivo. Our previous data using specific inhibitors indicated that, in Th2 cell lines, ICOS-induced elongation is largely dependent on the activity of the p110α PI3K subunits (20). Hence, anti-ICOS antibody was used to compare the elongation of “in vitro” differentiated Tfh cells of p110α^−/−ΔT or WT genotype. As shown in Figure 2B, the contribution of p110α to this phenomenon was confirmed by the significantly lower elongation of p110α^−/−ΔT Tfh cells. Indeed, the elongation of Tfh cells from WT, but not p110α^−/−ΔT mice was significantly inhibited by a p110α-specific PI3K inhibitor (A66, 1 µM). In both cases, a p110δ-specific inhibitor (IC87114, 5 µM) blocked ICOS-induced elongation.

Our results using Th1 T cells differentiated “in vitro” indicated that IL-2 and IFN-γ secretion as well as T-bet expression were enhanced in cells lacking PI3K p110α (Figure S2A in Supplementary Material), in agreement with the results obtained under neutral conditions (Figure 1). In differentiated Th17, the production of IL-17A was also enhanced in p110α-deficient cells Figure S2A in Supplementary Material.

Last, the differentiation of naive CD4⁺ T cells into iTreg was checked. No significant differences were observed in the percentage of CD4⁺Foxp3⁺ (Figure 3), of Foxp3⁺ cells expressing the inhibitory checkpoint marker PD-1 (Figure 3), or of Foxp3⁺ cells expressing ICOS (data not shown). Furthermore, similar amounts of IL-10 were secreted on a per-cell basis (Figure 3B). In contrast, the number of cells recovered from p110α^−/−ΔT cultures was significantly lower than from WT cultures, and so were the number of iTreg cells (Figure 3C). Overall, this might result in a weaker Treg activity in p110α^−/−ΔT mice.

Then, the role of PI3K p110α deficiency on the activation and differentiation “in vitro” of T CD8⁺ lymphocytes was also addressed. Naive CD8⁺ T cells were cultured for 72 h with anti-CD3 and anti-CD28. IFN-γ levels were much higher in p110α^−/−ΔT cell cultures, in agreement with the results observed in CD4⁺ T cells. As shown in Figure 4A, the levels of secreted TNF-α were clearly enhanced in activated p110α^−/−ΔT cells, whereas IL-2 secretion was not significantly different in WT or p110α^−/−ΔT cells.

Class I PI3-kinases control the differentiation of TCR-activated CD8⁺ in the presence of IL-2. So, naive CD8⁺ T cells were activated for 72 h with anti-CD3 and anti-CD28, washed, and expanded in the presence of IL-2 for further 72 h. The expanded p110α^−/−ΔT cells also showed enhanced effector mechanisms. These included the expression of molecules involved in cytotoxic mechanisms like CD107a (LAMP-1) and Granzyme B (Figures 4B,C), or IFN-γ and TNF-α secretion (Figure S2B in Supplementary Material).

The impact of T cell PI3K p110α deficiency “in vivo” was also analyzed in a model of immune response to the protein antigen keyhole limpet hemocyanin (KLH). Mice were first immunized s.c. with KLH in Freund’s adjuvant (KLH/FCA). Cytokine release upon antigen re-stimulation “in vitro” of cells from draining lymph nodes was then determined (Figure 5A). The levels of IFN-γ in cell cultures from p110α^−/−ΔT mice were significantly higher than those of WT mice. No significant differences were observed in the other cytokines examined (IL-2, IL-10, IL-17A, and TNF-α), even though the levels of the inflammatory cytokines IL-17A and TNF-α were higher in p110α^−/−ΔT cells (Figure 5A). No significant differences were found in the levels of KLH-specific antibodies of the isotypes examined (Figure S3A in Supplementary Material).

Mice were also immunized i.p. with KLH and Alum as adjuvant (KLH/Alum). In these conditions, the level of anti-KLH-specific antibodies of the IgG1 and IgG2b isotypes was significantly higher in the sera of p110α^−/−ΔT mice (Figure 5B), as was the percentage of germinal center B lymphocytes and T lymphocytes with a Tfh cell phenotype in the spleen cells from immunized mice (Figure 5C). Spleen cells were re-stimulated “in vitro” with KLH and the cytokines released determined (Figure 5D; Figure S3B in Supplementary Material). Significantly enhanced secretion of the Th2 cytokine IL-4 fitted with the higher titer of IgG1 anti-KLH antibodies, but no significant differences were observed in IL-10 and IFN-γ levels (Figure 5D) or in IL-2, IL-17A, and TNF-α levels (Figure S3B in Supplementary Material).

IFN-γ is a major mediator in antitumor adaptive responses. Since T cells from p110α^−/−ΔT mice had enhanced effector mechanisms including IFN-γ secretion, and might have altered Treg proliferation, the effect of PI3K p110α deletion in T cells on the growth of s.c. injected B16.F10 melanoma was examined. Tumor growth was significantly lower in p110α^−/−ΔT mice, as shown by tumor size or by the survival rate (Figure 6A). Furthermore, spleen cells from tumor-bearing mice were analyzed by flow cytometry and activated “in vitro” with melanoma proteins. Activated cells from p110α^−/−ΔT mice secreted significantly higher amounts of IFN-γ; however, this difference was not found in other cytokines like IL-2 and TNF-α (Figure 6B). Interestingly, the percentage of CD8⁺ T cells was significantly enhanced in tumor-bearing p110α^−/−ΔT mice (Figure 6C). Furthermore, CD4⁺Foxp3⁺ Treg cells (CD4⁺Foxp3⁺) were significantly diminished in p110α^−/−ΔT mice, and the level of Foxp3 in these cells was lower than in their WT counterparts (Figure 6C). The fraction of CD4⁺ and CD8⁺ T lymphocytes expressing the inhibitory checkpoint marker PD-1 was also lower in the spleen of p110α^−/−ΔT mice, yet the difference was not significant. In summary, our data show enhanced T-cell-mediated mechanisms of tumor rejection in p110α^−/−ΔT mice.

Mice used in this study were C57BL/6J, CD4-Cre [(43) Tg(Cd4-cre)1Cwi; MGI:2386448, backcrossed for nine generations with C57BL/6NTac mice], and Pik3ca^flox, p110α^flox/flox described in Ref. (24). They were bred under specific pathogen-free conditions in the animal care facility of the Centro de Investigaciones Biológicas from stock purchased from Charles River (C57BL/6J, p110α^flox/flox), or the European Mouse Mutant Archive Repository (CD4-Cre), and used at 8–12 weeks of age. CD4-Cre and p110α^flox/flox mice were crossed to obtain CD4-Cre^+/− p110α^flox/flox (p110α^−/−ΔT) mice, and these crossed with CD4-Cre^−/− p110α^flox/flox (WT) mice to obtain p110α^−/−ΔT and WT littermates used in the experiments. All experimental procedures were performed according to established institutional and national guidelines under project licenses PROEX 181/15 (JMR, CIB) and PROEX 330/15 (PP, ISCIII) from the Consejeria de Medio Ambiente y Ordenación del Territorio de la Comunidad de Madrid. Mice were genotyped for Cre, p110α, p110αflox, or Cre-mediated removal of floxed p110α using the oligonucleotides and conditions described in Ref. (24).

Antibodies used were monoclonal rat anti-mouse CD3ε [YCD3-1 (44)]; rat anti-mouse CD4 [GK1.5 (45)]; rat anti-mouse CD8 [53–6.72 (46)]; syrian hamster anti-mouse CD28 [37.51 (47)]; rat anti-mouse CD44 (IM-7.8.1); rat anti-mouse CD62L (MEL-14); armenian hamster anti- ICOS (CD278) [C398.4A (48, 49)]; anti-IL-4 (11B11); and anti-IFN-γ (XMG.1). They were prepared in-house as protein A- or protein G-purified preparations from hybridoma supernatants, or purchased from eBioscience (37.51). Where needed, they were coupled with biotin, FITC, or DyLight₆₄₇ in our laboratory using standard methods or as recommended by the manufacturer. Anti-γδ-FITC was from BD Biosciences, RM-5 anti-CD4-FITC, -PE, -PE-Cy7, and -APC conjugates, PE conjugates of anti-mouse CD8a, anti-CD19, anti-CD25, anti-NK1.1, anti-T-bet (4B10), and anti-Foxp3 (FJK-16s) antibodies; APC conjugates of anti-CD62L-, anti- ICOS- (C398.4A), anti-PD-1- (J43) antibodies were from eBioscience. Anti-mouse CXCR5-Biotin (2G8) was from BD-Pharmingen. Anti-CD107a (LAMP-1) (eBio1D4B) Alexa Fluor 488 conjugate and anti-Granzyme B-PE (NGZB-PE) were from eBioscience; anti-Granzyme A-FITC (3G8.5-FITC) was from Santa Cruz Biotechnology.

Anti-phosphotyrosine antibody (PY-20) was from GE Healthcare. Polyclonal rabbit anti-ZAP-70 and anti-CD4 antibodies were obtained and purified as described previously (12, 50). Rabbit antibody specific for dually phosphorylated ERK (Anti-Active MAPK ref. V6671) was from Promega; rabbit anti-phospho-p38 MAPK (Thr180/Tyr182) (#9211) and monoclonal rabbit anti-phospho-Akt(Ser473) (D9E, #4060) antibodies were from Cell Signaling Technology. Conjugates of Streptavidin with AlexaFluor647 (Molecular Probes), PE (Southern Biotech) and PE-Cy7 (BD-Pharmingen) were also used. Peanut agglutinin FITC conjugate was from Vector.

Horseradish peroxidase-coupled polyclonal goat antibodies specific for mouse IgM, IgG1, IgG2b, IgG2c, and IgG3 immunoglobulin were from Southern Biotech.

Rabbit polyclonal anti-ERK antibody was from Upstate Biotechnology. Rabbit anti-PI3K p110δ (H-219) was from Santa Cruz Biotechnology; the rabbit anti-p110α monoclonal antibody C73F8 was from Cell Signaling Technology.

Recombinant human IL-2 and mouse IL-4 and IL-6 were from Preprotech; recombinant human TGF-β was from R&D. PI3Kα inhibitor (A66) was from Selleckchem; the PI3K PI3Kδ Inhibitor IC87114 was from Symansis Pty. The P38 inhibitor SB203580 was from Sigma; the MEK1 inhibitor U0126 was from Calbiochem.

Spleen or thymus cell suspensions were prepared in culture medium (Click’s medium supplemented with 10% heat inactivated fetal bovine serum) and passed through 30-µm filters to remove clumps. Naive CD4⁺ (CD4⁺CD62L⁺) T cells were isolated using mouse naive CD4⁺ T cell isolation kit II (130-090-860, Miltenyi Biotech, Bergisch Gladbach, Germany). CD8⁺CD62L⁺ T cells were obtained with the CD8a T Cell Isolation Kit (130-104-075) and anti-CD62L beads (Miltenyi Biotech). Where required, CD4⁺ T cells were purified to >98% CD4⁺ T cells by sorting using a FACS Vantage-DIVA flow cytometer (B-D Bioscience).

CD4⁺CD62L⁺ or CD8⁺CD62L⁺ T cells were cultured at 0.5–1 × 10⁶ cells/ml in 24-well culture plates (Costar 3524, 1 ml/well) or 96-well culture plates (Costar 3590, 0.2 ml/well) previously coated with anti-CD3 (YCD3-1, 5–10 µg/ml) plus soluble anti-CD28 (2.5 µg/ml), as described in the Section “Results.” At 72 h, supernatants were taken for cytokine determination. Differentiation of naive CD4⁺ T cells into distinct functional effector subsets was performed by activation with anti-CD3 and anti-CD28 antibodies plus cytokines and antibodies as follows: for Th1 differentiation, 20 ng/ml IL-12 and 5 µg/ml anti-IL-4 (11B11); for Th2 differentiation, 10 ng/ml IL-4 and 10 µg/ml anti-IFN-γ (XMG.1); for Th17 differentiation, 20 ng/ml IL-6, 5 ng/ml human TGF-β1 plus anti-IL-4 and anti-IFN-γ; for Tfh differentiation, cells were cultured in the presence of anti-IL-4 and anti-IFN-γ antibodies plus 20 ng/ml IL-6. Treg differentiation from naive CD4⁺ T cells was achieved with anti-CD3 (YCD3-1, 5 µg/ml) plus huIL-2 (20 ng/ml), 5 ng/ml TGF-β1, 5 µg/ml anti-IL-4, and 10 µg/ml anti-IFN-γ. At 48 h, cells were stained for Foxp3 with the PE anti-mouse/rat FoxP3 staining set (eBioscience), according to the manufacturer’s instructions.

Where needed, CD8⁺CD62L⁺ T cells were first activated for 72 h with anti-CD3 plus anti-CD28, washed, adjusted to 15 × 10⁴ cells/ml in culture medium and then expanded for further 96 h in the presence of 20 ng/ml huIL-2.

Naive CD4⁺ cells were first differentiated into Tfh cells that express high levels of ICOS. Elongation assays were performed in 24-well culture plates containing glass coverslips previously coated with anti-ICOS antibody or Poly-l-Lysine (10 µg/ml in PBS), as described (20). After washing with PBS, the wells received 2 × 10⁵ cells in 0.5 ml of PBS, 10 mM HEPES, 0.1% glucose, pH 7.2. The plates were briefly spun and then incubated for 60 min at 37°C, washed twice with PBS, and fixed with 4% formaldehyde 5 min at 37°C. After washing with PBS, the coverslips were mounted on Vectastain (Vector). The cells were analyzed in an Axioplan Universal microscope (Carl Zeiss, Jena, Germany) and a Leica DFC 350 FX CCD at a 40× magnification to acquire images. Cell elongation was expressed as the Axis Ratio of the major and minor axis of the ellipse fitting the cell using the shape descriptors of the ImageJ 1.51j8 public domain software (National Institutes of Health).

Cytokines in culture supernatants (IL-2, IL-4, IL-10, IL-17A, IL-21, IFN-γ, and TNF-α) were determined with Ready-SET-Go! kits (eBioscience).

Phosphorylated Tyr in proteins and activated Erk and P38 MAP kinases were determined in lysates of activated cells. Freshly isolated CD4⁺CD62L⁺ T cells from the spleen of WT or p110α^−/−ΔT mice were washed with PBS and activated for 20 min at 37°C in serum-free medium by mixing the cells at 50 × 10⁶/ml with an equal number of latex beads coated with anti-CD3 or control antibodies (5 µg/ml) in the presence of anti-CD28 antibodies (5 µg/ml), as described in Ref. (12). Separation of cell lysates by SDS-PAGE, transfer to PVDF membranes, and immunoblot was performed as described previously in detail (12, 51). ImageJ 1.51j8 was used for densitometry analysis of immunoblots.

For staining of the transcription factors Foxp3 and T-bet, or Granzymes A and B, the cells were cultured for the times specified in each case. After washing, they were fixed, permeabilized, and stained with fluorochrome-conjugated antibodies using the Transcription Factor Staining Buffer Set (Affimetrix/eBioscience), according to the manufacturer’s instructions. Staining of CD107a was performed as described in detail in Ref. (52). Flow cytometry analysis of surface and intracellular stained cells was performed in a FC-500 flow cytometer (Beckman Coulter, Brea, SA, USA).

Keyhole limpet hemocyanin (Sigma) was injected at 100 μg/mouse. It was either emulsified v:v in Freund’s complete adjuvant (FCA, Sigma-Aldrich) and injected s.c. in two sites in the base of the tail, or mixed with Alum (Imject Alum, Thermo Scientific) and injected i.p. Mice were bled and sacrificed on day 7 (FCA) or day 10 (Alum) after immunization. Cells from inguinal lymph nodes (FCA) or the spleen (Alum) of individual mice were analyzed by flow cytometry, or cultured in round-bottom 5 ml polystyrene tubes at 5 × 10⁶ cells/ml plus 100 µg/ml KLH in 1 ml of culture medium. Culture supernatants were taken at 96 h to determine KLH-induced cytokine secretion. The titer of anti-KLH-specific antibodies of the IgM, IgG1, IgG2b, IgG2c, and IgG3 subclasses were determined in the serum of individual mice, as described previously (51).

Mouse melanoma B16.F10 cells from a mycoplasma-free stock were maintained by serial passage in culture medium. Primary B16.F10 tumors were induced by s.c. inoculation of 10⁵ B16.F10 melanoma cells in one flank of mice. Tumor size was periodically measured with calipers by experimenters blind to phenotype. Mice were sacrificed when tumor diameter reached 10 mm, or on day 21 after inoculation, and spleen was obtained.

Spleen cells from individual mice were analyzed by flow cytometry, or cultured (5 × 10⁶ cells/culture) in 1 ml of culture medium in round-bottom 5 ml polystyrene tubes containing 100 µg/ml of protein extract from B16.F10 cell lysates. Supernatants were collected after 6 days of culture and cytokine content determined by ELISA.

The statistical analyses were performed with GraphPad Prism 6.05 using the Student’s t-test, or the log-rank Mantel–Cox test. Significant differences between data are indicated by asterisks (*p < 0.05, **p < 0.01, ***p < 0.001).