PI3Kγ Activates Integrin α₄ and Promotes Immune Suppressive Myeloid Cell Polarization during Tumor Progression

Chronic inflammation promotes tumor progression by stimulating angiogenesis and suppressing the antitumor immune response (1–3). CD11b⁺Gr1⁺ immunosuppressive myeloid cells [myeloid-derived suppressor cells (MDSCs)] accumulate in the blood, spleen, lymph nodes, bone marrow, and tumor microenvironments in tumor-bearing animals and cancer patients (4–8). These cells, as well as CD11b⁺Gr1⁻ tissue resident macrophages, inhibit innate and adaptive immunity, thereby promoting tumor immune escape (1–8). These immunosuppressive cells inhibit maturation of antigen-presenting dendritic cells (DCs), suppress T-cell activation and NK-cell cytotoxicity and induce regulatory T-cell (Treg) development, leading to dysfunctional cell-mediated antitumor immunity through the release of immunosuppressive factors such as IL10, TGFβ1, IL6, VEGF, and ROS (9–10).

Experimental evidence suggests that therapeutic strategies targeting MDSCs may be beneficial in cancer therapy (11–15). Studies show that modulation of recruitment, differentiation, and expansion of myeloid cells could promote antitumor immunity and block tumor progression (11–15). Strategies targeting the mCSF1 receptor (CSF1R), the CCL2 receptor (CCR2), or PI3Kγ reduce tumor growth and inflammation in animal models and are currently undergoing testing in cancer patients (11–14). We previously found that CD11b⁺Gr1⁺ monocyte and granulocyte trafficking to tumors from circulation depends on PI3Kγ-mediated activation of integrin α₄β₁, a receptor for endothelial cell VCAM (14–18). As the most prevalent PI3K isoform in normal myeloid cells, PI3Kγ is also required for neutrophil motility and trafficking during inflammatory disease (19–20). We found that numerous cytokines and chemokines activate PI3Kγ, including SDF-1, VEGF-A, IL4, IL6, and CSF1 (17). Once activated, PI3Kγ stimulates BTK, PLCγ, RAPGEF, Rap1a, RIAM, and paxillin-dependent integrin α₄ activation (14–18). Inhibition of PI3Kγ, BTK, PLCγ, RAPGEF, Rap1a, RIAM, or integrin α₄β₁ suppresses monocyte and granulocyte accumulation in tumors and inhibits tumor growth (14–18).

PI3Kγ not only controls myeloid cell trafficking during tumor growth and inflammation, but also controls myeloid cell polarization by regulating immune suppressive gene expression in tumor-associated macrophages (TAMs), monocytes, and granulocytes (21–22). Our studies showed that PI3Kγ stimulates mTor, S6Kα, and C/EBPβ-mediated anti-inflammatory gene expression and inhibits NFκB-mediated proinflammatory gene expression, thereby promoting expression of immune suppressive factors, such as IL10, TGFβ, and Arginase, and inhibiting expression of IL12, IFNγ, and Nos2 (21–22). PI3Kγ activity thereby suppresses T-cell immunity and stimulates tumor growth (21). Because our prior studies showed that PI3Kγ signaling activates integrin α₄β₁ to promote myeloid cell adhesion and invasion, we speculated that integrin α₄β₁ might also regulate myeloid cell polarization downstream of PI3Kγ and thereby affect the antitumor immune response.

We show here that integrin α₄β₁ promotes immune suppressive myeloid cell polarization, inhibits antitumor immunity, and stimulates tumor growth. Genetic or pharmacologic blockade of integrin α₄β₁ or its activator PI3Kγ inhibits immunosuppressive myeloid cell polarization, stimulates DC maturation, and restores antitumor T cell–mediated immunity in mouse models of cancer. These studies indicate that blockade of the PI3Kγ-integrin α₄ pathway can suppress tumor growth.

Lewis lung carcinoma cells (LLC) were obtained from the ATCC and cultured in Dulbecco's Modified Medium (DMEM) supplemented with 10% FBS, 2 mmol/L glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin. Pancreatic adenocarcinoma cells (Panc02) were obtained from the David Cheresh Laboratory, University of California, San Diego, and were cultured in RPMI (GIBCO) supplemented with L-glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin and 10% FBS. Cultures were maintained in a humidified incubator at 37°C in an atmosphere containing 5% CO_2. Both cell lines were verified by RT-PCR or RNA sequencing, in vitro morphological and biochemical criteria, and in vivo tumor assays in syngeneic mice.

Bone marrow–derived cells (BMDC) were aseptically harvested from 6- to 8-week-old female mice by flushing leg bones of euthanized mice with phosphate buffered saline (PBS), 0.5% BSA, 2 mmol/L EDTA, incubating in red cell lysis buffer (155 mmol/L NH4Cl, 1 mmol/L NaHCO₃ and 0.1 mmol/L EDTA) and centrifuging over Histopaque 1083. Purified mononuclear cells were cultured in RPMI + 20% serum + 50 ng/mL M-CSF (PeproTech). Bone marrow–derived macrophages were polarized with either IFNγ (20 ng/mL, Peprotech) plus LPS (100 ng/mL, Sigma) or LPS alone for 24 hours or IL4 (20 ng/mL, Peprotech) for 24 to 48 hours. Total RNA was harvested from macrophages using the RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions.

Animal studies were approved by the Institutional Animal Care and Use Committee (IACUC), University of California, San Diego: 5 × 10⁵ LLC cells were injected subcutaneously into syngeneic (C57Bl/6J) 6- to 8-week-old wild-type (WT), integrin α4Y991A, or PI3Kγ^–/− (p110γ^–/−) mice (n = 8–10). Tumors dimensions were recorded and excised at 14 to 21 days. Tumors were cryopreserved in O.C.T., solubilized for RNA purification or collagenase-digested for flow-cytometric analysis of immune cell infiltration as detailed below. Alternatively, orthotopic Panc02 pancreatic tumor were initiated by implanting 1 × 10⁶ Panc02 pancreatic carcinoma cells into the pancreas of syngeneic mice (n = 8–10). The abdominal cavities of immunocompetent C57Bl/6J mice, integrin α4Y991A mutant, and PI3Kγ^–/− mice were opened and the tails of the pancreas were exteriorized. One million Panc02 cells were injected into the pancreatic tail, the pancreas was placed back into the abdominal cavity, and the incision was closed. Pancreas were excised and cryopreserved after 5 weeks. Tumor weight and immune cell infiltration were quantified as described.

Studies with blocking mAb to integrin α₄:C57Bl/6J mice were subcutaneously implanted on day 1 with 5 × 10⁵ LLC cells. Mice were treated every third day with intraperitoneal (i.p.) injections of mAb PS/2, an α₄ function-blocking antibody, or isotype-matched control rat IgG2b, at a dose of 200 μg/mouse (10 mg/kg) in a 100 μL volume (n = 8 per group). Tumors were harvested at 14 to 21 days, weighed and further analyzed by quantitative RT-PCR, flow cytometry, and immunohistochemistry.

IL10 blocking studies: C57Bl/6J mice were subcutaneously implanted on day 1 with 5 × 10⁵ LLC cells. Mice were treated on day 7 and day 11 with i.p. injections of function-blocking anti-IL10 (JES052A5, R&D Systems) or isotype-matched (rat IgG1) control antibodies at doses of 200 μg/mouse (n = 6 per group). Tumors were harvested at 14 days, weighed, and further analyzed by quantitative RT-PCR, flow cytometry, and immunohistochemistry. PI3Kγ inhibitor studies: C57Bl/6J mice were subcutaneously implanted on day 1 with 5 × 10⁵ LLC. Mice were treated by i.p. injection with 2.5 mg/kg of PI3Kγ inhibitor (TG100-115) or with a chemically similar inert control (n = 10) twice daily for 14 days for a total daily dose of 5 mg/kg. Tumor volumes and weights, as well as myeloid cell densities were measured.

Bone marrow cells were aseptically harvested from 6- to 8-week-old female mice by flushing leg bones of euthanized mice with phosphate buffered saline (PBS) containing 0.5% BSA and 2 mmol/L EDTA, incubating cells in red cell lysis buffer and centrifuging over Histopaque 1083. Approximately 5 × 10⁷ bone marrow cells were purified by gradient centrifugation from the femurs and tibias of a single mouse. Two million cells were intravenously injected into tail veins of each lethally irradiated (1,000 rad) 6-week-old syngeneic recipient mouse. After 4 weeks of recovery, tumor cells were injected in BM-transplanted animals. LLC (n = 8, 3 experiments) tumor growth in C57BL/6 and α4Y991A mice transplanted with BM from α4Y991A or WT were compared as described above.

Tumors were isolated, minced, and digested to single-cell suspension for 1 hours at 37°C in 5 mL of Hanks Balanced Salt Solution (HBSS, Invitrogen) containing 1 mg/mL collagenase type IV (Sigma), 0.1 mg/mL hyaluronidase (Sigma) and 20 U/mL DNase type IV (Sigma). Cell suspensions were filtered through a 70-μm cell strainer and then incubated with different antibodies to perform flow cytometry.

Tumor-infiltrating immune cells were incubated with Fc-blocking reagent (anti CD16/CD32, BD Biosciences), followed by CD11b-APC (M1/70, BD Biosciences), Gr1-FITC (RB6-8C5, BD Biosciences), CD11c-APC (HL3, BD Biosciences), MHC II-FITC (I-Ab, AF6-120.1, BD Biosciences), CD3-APC (145-2C11, eBioscience), CD4-FITC (GK1.5, eBioscience), CD8a-APC (53-6.7, eBioscience) along with isotype-matched control. In vitro cultures DCs were stained with CD11b-APC, Gr1-FITC, CD11c-APC, MHC II-FITC, CD80-FITC (16-10A1, eBioscience), and CD86-FITC (GL1, eBioscience).

Total tumor RNA was prepared with TRIzol Reagent (Invitrogen) according to the manufacturer's instructions. In some experiments, total isolated CD11b⁺ or CD90.2⁺ cells RNA was isolated using RNeasy Mini Kit (Qiagen). cDNA synthesis was performed from 0.5 to 1 μg RNA with SuperScript III First-Strand Synthesis Kit (Invitrogen). Quantitative PCR was performed using primers for Gapdh, Arg1, Il10, Il6, Vegfa, Tgfb, Il12, or Ifng (QuantiTect Primer Assay). qPCR analysis was performed using a Smart Cycler System (Cepheid).

Tumor samples were collected and progressively frozen in cold 2-methylbutane solution. Sections (6 μm) were fixed in 100% cold acetone, blocked with 8% normal goat serum for 2 hours, and incubated with the appropriate primary antibodies (rat anti-mouse CD11b, hamster anti-mouse CD11c, rat anti-mouse CD8a or rat anti-mouse CD11b, BD Biosciences) for 2 hours at room temperature. Sections were washed 3 times with PBS and incubated with goat anti-rat secondary antibodies coupled to Alexa488 or 597 for CD11b, CD8, and CD4 studies or with Cy5-goat anti-hamster IgG for CD11c studies. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI). The slides were washed and mounted in DAKO fluorescent mounting medium. The detection of apoptotic cells was performed using a TUNEL-assay (ApoTag Fluorescein In Situ Apoptosis Detection Kit, Promega) in accordance with the manufacturer's instructions. Immunofluorescence images were collected on a Nikon microscope (Eclipse TE2000-U) and pixels of fluorescence or area of staining was quantified using MetaMorph Software.

Myeloid CD11b⁺ cells and CD90.2⁺ (Thy1.2) T cells were isolated from tumor cell suspensions using CD11b or CD90.2 microbeads (Miltenyi Biotech), respectively. The purity of the sorted cells was typically >80% as assessed by FACS analysis. In some experiments, tumor-infiltrating lymphocytes (TILs) were stained with CD4-FITC (GK1.5, eBioscience) and CD8-APC (53-6.7, eBioscience).

Thy1.2⁺ T cells (effector cells) in tumor were purified from LLC tumor-bearing WT and α4Y991A mice and then coincubated with LLC tumor cells (target cells) at 2.5:1, 5:1, and 10:1 ratios (2 × 10³ LLC tumor cells per well) for 6 hours. Target cell killing was assayed by collecting the supernatants from each well for measurement of the lactate dehydrogenase release (Cytotox96 NonRadioactive Cytotoxicity Assay kit, Promega).

CD90.2⁺ T cells were first isolated from the spleens of WT, α4Y991A and PI3Kγ^–/− mice bearing tumors using the MACS system and then mixed with viable LLC tumor cells at ratio 1:2. The mixture was injected into the flanks of naïve mice. Tumor growth, intratumoral apoptosis, and necrosis were investigated after 10 days.

Tumors were collected and homogenized in RIPA lysis buffer (137 mmol/L NaCl, 20 mmol/L Tris-HCL pH8, 10% Glycerol, 1% Triton X-100, 0.1% SDS and protease inhibitor cocktail). Protein extracts were clarified by centrifugation at 12,000 ×g for 10 minutes at 4°C, and clarified protein samples were used for IL10 ELISA quantification (R&D Systems).

Results were expressed as mean ± SEM. Data comparing differences between two groups were assessed using unpaired Student t test. One-way ANOVA with post hoc testing was used for multiple comparisons. A value of P < 0.05 was considered significant.

Tumor inflammation promotes the accumulation of CD11b⁺Gr1⁺ MSDCs and CD11b⁺Gr1⁻ TAMs that inhibit antitumor immunity, thereby favoring tumor progression (4–8). Recruitment of circulating Gr1⁺ myeloid cells into tumors is regulated by PI3Kγ-mediated integrin α₄β₁ activation (14–18). PI3Kγ also promotes TAM polarization by activating an mTor, S6Kα, and C/EBPβ-dependent transcriptional pathway, leading to expression of immune suppressive factors, such as Arginase, IL10, and TGFβ, that promote tumor immune suppression (21–22). As integrin α₄β₁ is downstream of PI3Kγ, we speculated that integrin α₄β₁ might also regulate immune suppressive transcription in myeloid cells.

To investigate whether integrin α₄ regulates immune suppressive myeloid cell polarization in tumors, we first compared tumor growth and inflammation in two different models of tumor growth, the syngeneic subcutaneous LLC and orthotopic Panc02 pancreatic adenocarcinoma (PDAC) models in WT, PI3Kγ^–/−, or α4Y991A knock-in mice and in mice treated with inhibitors of integrin α₄ or PI3Kγ^–/− (Fig. 1). Integrin α4Y991A mice express a knock-in point mutation in the α₄ cytoplasmic tail that inhibits integrin activation, which is controlled by PI3Kγ signaling in myeloid cells (17). LLC and Panc02 tumors implanted in integrin α4Y991A and PI3Kγ^–/− mice were both significantly smaller than tumors implanted in control-treated mice (P < 0.001), whether analyzed at endpoint or over time (Fig. 1A–C). Growth of lung LLC tumors was similarly inhibited in animals treated with inhibitors of α₄ or PI3Kγ: PS2, a function-blocking antibody (anti-α₄₎ or TG100-115, a PI3Kγ inhibitor; Fig. 1A; Supplementary Fig. S1). Flow cytometric analyses of LLC or Panc02 tumors implanted in WT, integrin α4Y991A, and PI3Kγ^–/− mice revealed significant (P < 0.01) reductions in infiltration of total CD11b⁺Gr1⁺ myeloid cells (Fig. 1D–F), as well as in CD11b⁺Gr1^lo and CD11b⁺Gr1^hi (P < 0.05) myeloid cell subpopulations but not of CD11bGr1⁻ cells (Fig. 1G) in PI3Kγ^–/− and α4Y991A animals compared with WT animals. These populations were further defined by marker analysis as CD11b⁺Gr1^loLy6G⁻F4/80^loCD206⁻CD11c⁻ monocytes, CD11b⁺Gr1^hiLy6G⁺F4/80⁻CD206⁻CD11c⁻ granulocytes and CD11b⁺Gr1⁻Ly6G⁻F4/80^hiCD206⁺CD11c⁺ macrophages (Fig. 1H). Similar results were observed in tumor-bearing mice treated with the anti-α4 or with the PI3Kγ inhibitor TG100-115 (Fig. 1D–G). These data show that loss of PI3Kγ or integrin α₄ activity affects trafficking of Gr1⁺ but not Gr1⁻ myeloid cells.

Tumors from PI3Kγ^–/− and α4Y991A animals not only exhibited decreased CD11b⁺ cell infiltration but also increased CD11c⁺ DC infiltration in tumors and lymph nodes compared with control animals (Fig. 2A–C). Similar results were observed in α4Y991A mice transplanted with α4Y991A bone marrow but not in mice transplanted with WT bone marrow (Fig. 2D–F). The percentages of mature MHCII-expressing DCs (CD11c⁺MHCII⁺) increased in α4Y991A and PI3Kγ^–/− Panc02 and LLC tumors compared with WT tumors (P < 0.01) and in tumors from animals that were treated with anti-α4 or the PI3Kγ inhibitor TG100-115 (Fig. 2G–I). The ratio of MHCII⁺ mature DC to MHCII⁻ immature DC was 2-fold higher in tumors from these mice compared with control mice (P < 0.001; Fig. 2J and K). In addition, levels of the costimulatory protein CD80 were significantly (P < 0.05) higher in DCs in tumors from PI3Kγ^–/− and α4Y991A animals (Fig. 2L). Together, these results indicate that PI3Kγ and α₄ integrin inhibit DC recruitment and maturation in association with tumor inflammation and growth and that inhibitory targeting of these molecules can stimulate DC recruitment and activation.

PI3Kγ promotes immune suppressive transcription in TAMs and MDSCs, thereby contributing to tumor immune escape (18, 21–22). Therefore, as integrin α₄ is downstream of PI3Kγ, we investigated the role of integrin α4 in myeloid cell polarization. We used RT-PCR to assess the expression of pro- and anti-inflammatory factors in LLC and Panc02 tumors from WT, α4Y991A, and PI3Kγ^–/− mutant mice and in LLC tumors from anti-α₄- and TG100-115–treated mice. We observed increased expression of mRNAs encoding immunostimulatory cytokines, such as IL12b and IFNγ, and decreased expression of mRNAs encoding immune suppressive and proangiogenic factors, such as IL10, IL6, Arg1, and VEGFA, in LLC and Panc02 tumors from α4Y991A and PI3Kγ^–/− mice (Fig. 3A and B). Treatment of LLC tumors with anti-α₄ or the PI3Kγ inhibitor (TG100-115) had similar effects on inflammatory factor expression (Fig. 3C and D). These gene expression changes were also observed in tumor-associated CD11b⁺ myeloid cells purified from α4Y991A and PI3Kγ^–/− mice (Fig. 3E).

To determine whether integrin α₄β₁ like PI3Kγ controls macrophage polarization, we also performed gene expression analyses on in vitro cultured, IL4-stimulated macrophages derived from bone marrow of WT and α4Y991A mice. We found that loss of integrin activity in α₄Y991A macrophages suppressed expression of Il10, Arg1, and Tgfb1 and stimulated expression of mRNAs encoding immune stimulatory factors such as IL12 and IFNγ in vitro (Fig. 3F). As IL4-stimulated macrophages exhibit gene expression patterns similar to those of tumor derived macrophages (21), our results indicate that integrin α₄β₁ plays a role in the PI3Kγ signaling cascade that regulates myeloid cell polarization. Taken together, these results suggest that integrin α₄ regulates immunosuppressive myeloid cell transcription downstream of PI3Kγ during tumor inflammation.

IL10, an immune suppressive factor expressed by Tams, MDSCs and other immune cells, impairs DC maturation and effector T-cell function (10). We observed that CD11b⁺ myeloid cells are the principal source of Il10 expression in LLC tumors (Fig. 4A). IL10 protein and mRNA expression was suppressed in α4Y991A MDSCs and in tumors from α4Y991A BM transplanted animals (Fig. 4B and C). As integrin α₄ activity promotes IL10 expression, we further elucidated the role of IL10 in the control of DC maturation during tumor growth. Mice bearing LLC tumors were treated over the course of 10 days with anti-IL10. Although this treatment did not affect tumor growth, it did increase the total percentage of intratumoral mature MHCII⁺ DC compared with treatment with control isotype-matched antibody (Fig. 4D–F). Mice treated with blocking mAbs to IL10 exhibited increased expression of Il12b, Ifng, and Cd8a mRNA (Fig. 4G–I) as well as increased infiltration of CD8⁺ T cells as detected by IHC (Fig. 4J and K). Together, these data suggest that the integrin α₄ regulated production of IL10 by CD11b⁺ myeloid cells prevents DC maturation and inhibits antitumor T-cell responses.

To promote an effective antitumor immune response, tumor-specific T cells must be present in sufficient numbers to kill their targets. A direct correlation exists between the number of TILs and a favorable clinical outcome (23–26). Furthermore, the functional status of TILs has been correlated with a favorable prognosis in human malignancies (25–26). To further understand the mechanisms underlying the immune suppression mediated by inhibition of the PI3Kγ–integrin α₄ pathway, we evaluated whether blockade of this pathway alters both the number and the activation state of the TILs. Histological examination of tumors revealed reduced numbers of CD4⁺ T cells and increased numbers of CD8⁺ T cells in tumors from both α4Y991A and PI3Kγ^–/− mice when compared with WT mice (Fig. 5A and B). Additionally, flow cytometric analysis of TILs isolated from tumors showed that numbers of CD4⁺ T cells were reduced, whereas numbers of CD8⁺ T cells were increased, in α4Y991A and PI3Kγ^–/− mice compared with WT mice (Fig. 5C and D). These observations were supported by increases in total Cd8a mRNA in tumors from both α4Y991A and PI3Kγ^–/− mice when compared with WT mice (Fig. 5E). TILs from α4Y991A and PI3Kγ^–/− mice expressed more Ifng and less Tgfb1 and Il10 mRNAs than WT TILs, suggesting that inhibition of myeloid cell PI3Kγ or integrin α₄β₁ resulted in increased recruitment and priming of CD8⁺ T cells in tumors (Fig. 5F). To determine whether myeloid cell PI3Kγ or integrin α₄β₁ promote T cell–mediated tumor cell killing, TILs isolated from WT, α4Y991A, or PI3Kγ^–/− mice were cocultured with LLC cells in a direct ex vivo CTL assay. TILs from α4Y991A or PI3Kγ^–/− mutant mice exhibited a 2- to 3-fold increase in LLC tumor cell cytotoxicity over TILs from WT mice (Fig. 5G).

To determine whether PI3Kγ and α4 integrin also suppress T cell–mediated cytotoxicity in vivo, we mixed tumor-infiltrating T lymphocytes isolated from tumors of WT, α4Y991A, or PI3Kγ^–/− mice with LLC tumor cells and injected the mixture into the flanks of C57Bl/6J-naïve mice. When LLC tumor cells were mixed with α4Y991A and PI3Kγ^–/− T cells, tumor dimensions, and weights were reduced compared with tumors derived from LLC cells mixed with WT T cells (P < 0.001; Fig. 6A and B). Tumors derived from LLC cells mixed α4Y991A and PI3Kγ^–/− T cells exhibited increased intratumoral cell death, as detected by in situ TUNEL assay (Fig. 6C and D) and H&E staining (Fig. 6E and F) in mice injected with α4Y991A and PI3Kγ^–/− T cells compared with mice receiving WT T cells. Together, these studies show that integrin α₄β₁, like PI3Kγ, suppresses T cell–mediated tumor cell cytoxicity and that inhibiting the PI3Kγ–integrin α₄ pathway can stimulate T-cell cytotoxicity.

These studies indicate that myeloid cell PI3Kγ and integrin α₄ regulate myeloid cell polarization to control antitumor T-cell activation. In summary, these data show that blocking the PI3Kγ–integrin α₄ pathway in myeloid cells can restore antitumor immunity by modulating the tumor microenvironment and promoting tumor cell killing by T cells. We show here that integrin α₄β₁ regulates myeloid cell polarization and antitumor immunity and that inhibition of this pathway could be a useful therapeutic approach to stimulate antitumor immunity.

Trafficking of myeloid cells to sites of tumor inflammation depends on integrin α₄ and PI3Kγ (14–18, 21–22). Here, we show that integrin α₄β₁ controls myeloid cell polarization during tumor progression. Our studies demonstrate that integrin α₄, like PI3Kγ, not only regulates myeloid cell trafficking to tumors but also promotes immune suppressive myeloid cell polarization, inhibition of antitumor immunity and stimulation of tumor growth. PI3Kγ stimulates integrin α4 activation in a manner dependent on BTK, PLCγ, RAPGEF, Rap1a, RIAM, and paxillin (14–18). PI3Kγ activates BTK to promote immune suppressive myeloid cell polarization by inducing expression of IL10, TGFβ, and Arginase, which are dependent on mTor, S6Kα, or C/EBPβ, and inhibiting expression of IL12, IFNγ, and Nos2 (18, 21, 22). C/EBPβ controls immune suppression in tumor-infiltrating myeloid cells (27). Our present studies show that a PI3Kγ–integrin α₄ pathway inhibits antitumor immunity through stimulation of CD11b⁺Gr1⁺ MDSC cell recruitment and immunosuppressive myeloid cell polarization. Although it is not clear whether integrin α₄ directly participates in PI3Kγ-mediated activation of mTor, S6Kα, or C/EBPβ, our results indicate that the PI3Kγ–integrin α₄ pathway modulates tumor inflammation and immunosuppression.

Several lines of evidence presented here indicate that blockade of this pathway regulates antitumor immune responses. First, inhibition of the PI3Kγ–integrin α₄ pathway is associated with reduction of CD11b⁺Gr1⁺ MDSC infiltration and immunosuppressive factors (IL10, IL6, TGFβ, or Arginase) in the tumor microenvironment and lymph nodes of tumor-bearing animals. Blocking this pathway reduced expression of immunosuppressive factors in tumor-infiltrating CD11b⁺ myeloid cells.

Second, disruption of the PI3Kγ–integrin α₄ pathway increased the recruitment of mature DCs into the tumor microenvironment and into the draining lymph nodes. Defects in DCs caused by the immunosuppressive tumor microenvironment result in the accumulation of immature DCs. These immature DCs can promote tumor angiogenesis (28–36) and suppress T-cell responses (10), thus favoring tumor progression. We observed that blockade of a PI3Kγ–integrin α₄ pathway could facilitate CD11c⁺ DC infiltration and maturation. DC differentiation and maturation depends on factors produced by the microenvironment, such as IL10. Our results showed that PI3Kγ and integrin α₄ regulate IL10 production by myeloid cells and thereby regulate DC differentiation and maturation.

Third, blockade of the PI3Kγ–integrin α₄ pathway increased expression of immunostimulatory cytokines (IL12β, IFNγ), reduced CD4⁺ T-cell infiltration and promoted CD8⁺ T-cell accumulation into the tumor microenvironment. Furthermore, inhibition of this pathway promoted T cell–mediated tumor cytotoxicity and tumor rejection. We did not observe fewer T cells, but rather more T cells, in tumors from α4Y991A and PI3Kγ^–/− mice. Our studies show that integrin α₄ does not control T-cell trafficking to tumors. Indeed, PI3Kγ deficiency increases CD8⁺ cell differentiation and Th1 cytokine production (IL12, IFNγ) but does not directly affect Th1/Th2 T-cell polarization (22).

Finally, the PI3Kγ–integrin α₄ pathway also affects tumor progression by promoting angiogenesis (17). Myeloid cells can support tumor angiogenesis and growth through the production of proangiogenic factors (28–32). We found that blockade of the PI3Kγ–integrin α₄ pathway reduced expression of proangiogenic factors and tumor angiogenesis. Inhibition of this pathway increased myeloid cell expression of IL12, an antiangiogenic Th1 cytokine (33, 34). Therefore, inhibition of PI3Kγ and integrin α4 controls both tumor angiogenesis and immunity.

Inhibition of the PI3Kγ-α₄ integrin pathway could also influence development of regulatory T cells (Treg), which maintain tumor immune tolerance (34). In fact, we found that TGFβ, a cytokine that induces the generation of Treg (35), was downregulated in T cells from PI3Kγ and α₄ integrin mutant mice, suggesting that disruption of this pathway reduces the generation of Tregs in the tumor microenvironment.

The tumor microenvironment is a barrier to immune cell function. Our results indicate that blocking the PI3Kγ–α₄ integrin pathway can inhibit the immunosuppressive tumor microenvironment and promote antitumor immunity. Our results support the development of strategies based on the inhibition of the PI3Kγ–integrin α₄ pathway to control tumor immune suppression.

No potential conflicts of interest were disclosed.

Conception and design: P. Foubert, J.A. Varner

Development of methodology: P. Foubert, M.M. Kaneda, J.A. Varner

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P. Foubert, M.M. Kaneda

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P. Foubert, M.M. Kaneda, J.A. Varner

Writing, review, and/or revision of the manuscript: P. Foubert, J.A. Varner

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.A. Varner

Study supervision: J.A. Varner

We thank the UCSD Moores Cancer Center for their Flow Cytometry and Histology Shared Resources and for technical assistance.

This research was supported by fellowships from the “Fondation pour la Recherche Medical,” the “Fondation Bettencourt-Schueller” and the Pancreatic Cancer Action Network-AACR to P. Foubert, and NIH grants R01CA83133 and R01CA167426 J.A. Varner. The authors declare no competing financial interests.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.