Adenomatous polyposis coli (Apc) is a cell polarity regulator and a tumor suppressor associated with familial adenomatous polyposis and colorectal cancer. Apc involvement in T lymphocyte functions and antitumor immunity remains poorly understood. Investigating Apc-depleted human CD8 T cells and CD8 T cells from ApcMin/+ mutant mice, we found that Apc regulates actin and microtubule cytoskeleton remodeling at the immunological synapse, controlling synapse morphology and stability and lytic granule dynamics, including targeting and fusion at the synapse. Ultimately, Apc tunes cytotoxic T cell activity, leading to tumor cell killing. Furthermore, Apc modulates early TCR signaling and nuclear translocation of the NFAT transcription factor with mild consequences on the expression of some differentiation markers. In contrast, no differences in the production of effector cytokines were observed. These results, together with our previous findings on Apc function in regulatory T cells, indicate that Apc mutations may cause a dual damage, first unbalancing epithelial cell differentiation and growth driving epithelial neoplasms and, second, impairing T cell–mediated antitumor immunity at several levels.
TCRs recognize peptide Ags associated with MHC molecules on APCs. Ag interaction induces early TCR signaling, leading to the T cell polarization toward APCs and the formation of the immunological synapse at the T cell Ag-presenting cell interface. Immunological synapses are key to control T cell activation leading to T cell growth, differentiation and cytokine production. They also control T cell effector functions such as polarized secretion of cytokines by Th and regulatory T cells (Tregs) and lytic granules by CTLs (1).
Immunological synapse generation and function depend on the orchestrated action of the actin and microtubule cytoskeleton and of intracellular vesicle traffic that polarize at the T cell side of the immunological synapse. This drives the targeting and dynamic clustering of T cell Ag receptors and signaling molecules to optimally control T cell activation (2). Furthermore, the Golgi and lytic granule intracellular traffic reorient toward the synapse, delivering Th cytokines directly to Ag-presenting B cells (3–5) or lytic granules to infected or tumor target cells. Polarized secretion of lytic granules depends on the fine concomitant dynamic remodeling of actin and microtubule cytoskeleton at the immunological synapse (6–9).
Cell polarity complexes are key to ensure stable cell polarity (10) as well as induced polarization in migrating cells (11). Scribble, Dlg1, and PKCζ polarity regulators were shown to control lymphocyte migration, immunological synapse formation, and T cell activation (12–17). The polarity regulator and tumor suppressor adenomatous polyposis coli (Apc) is known for its association with familial adenomatous polyposis, human colorectal tumors, and intestinal carcinomas in mice (18–21). Apc interacts with a variety of proteins, including transcription factor regulators as β-catenin, polarity regulators, such as Dlg1 or Scribble, cytoskeleton regulators, such as Cdc42 or EB1, nuclear pore and nuclear transport proteins, and apoptosis- or mitosis-related proteins (22, 23).
Apc mutations alter intestinal epithelium differentiation and induce tumor progression in colorectal cancer patients and in mouse models (18–22, 24). The role of Apc in immune responses, in particular against tumors, is much less explored. However, altered intestinal immune homeostasis and control of inflammation by Tregs was reported in Apc mutant mice (25–28). We have previously unveiled that Apc underpins various molecular mechanisms controlling CD4 T cell functions (29). These include microtubule network organization and centrosome polarization at the immunological synapse and NFAT-driven cytokine gene activation. Interestingly, Apc regulates NFATc2 nuclear translocation upon T cell activation. Furthermore, NFATc2 forms microclusters associated with microtubules and needs microtubules to translocate to the nucleus upon T cell activation. Finally, in ApcMin/+ mutant mice, lamina propria Tregs display lower capacity to differentiate and produce cytokines, mainly IL-10 (29). This cytokine is central to regulation of intestinal inflammation and adenocarcinoma progression (30). These findings suggested that Apc mutations may also affect immune cell functions and, particularly, the anti-inflammatory capacity of Tregs, which could contribute to tumor development in patients carrying Apc mutations.
NFAT is involved in cytotoxic T cell differentiation and function (31). Furthermore, microtubule-mediated lytic granule polarization and release at cytotoxic T cell synapses is crucial for efficient tumor cell killing (7). Therefore, we hypothesized that Apc deficiency might also affect cytotoxic T cell functions, reducing their capacity to eliminate tumor cells and contributing to tumor escape in Apc-dependent polyposis patients. To investigate this, we studied Apc-silenced human primary CD8 T cells and CD8 T cells from ApcMin/+ mice. These heterozygous mutant mice have been largely used as an animal model to investigate the molecular bases of Apc-mediated intestinal polyposis and carcinoma (19).
In this study, we show that Apc is involved in microtubule and actin cytoskeleton remodeling at the immunological synapse of CD8 T cells. Furthermore, CD8 T cells from ApcMin/+ mice displayed reduced T cell Ag receptor-mediated Erk and Akt kinase activation and NFAT nuclear translocation. However, mild or no effects were observed on some CD8 differentiation markers or production of cytokines. Importantly, synapse stability, lytic granule dynamics and fusion at the synapse, as well as cytotoxic activity against tumor cells were reduced in ex vivo–differentiated CTLs from ApcMin/+ mice or from Apc-silenced human CD8 T cells. These results reveal a novel role of Apc in modulating cytotoxic T cell responses, with potential consequences in antitumor immunity.
Materials and Methods
Human cell isolation, small interfering RNA transfection, lentiviral infection, CTL generation, and tumor target cell culture
Human peripheral blood T cells from healthy volunteers were obtained from the French Blood Bank Organization (Etablissement Français du Sang) or through the Institut Pasteur Biological Resources Core Facility, Clinical Investigation and Access to BioResources (ICAReB) (NSF 96-900 certified, from sampling to distribution, reference BB-0033-00062/ICAReB platform/Institut Pasteur, Paris, France/BBMRI AO203/1 distribution/access: 2016, May 19th, [BIORESOURCE]), under CoSImmGEn protocol approved by the Committee of Protection of Persons, Ile de France-1 (No 2010-dec-12483). Informed consent was obtained from all donors.
PBMCs were isolated by centrifugation through Ficoll-Hypaque. CD8 T cells were isolated from PBMCs using the MACS CD8 T Cell Isolation Kit (Miltenyi Biotec) and maintained in human CD8 medium: RPMI 1640 plus GlutaMAX-I (Life Technologies) supplemented with 10% FBS, 1 mM sodium pyruvate and nonessential amino acids, 10 mM HEPES, 1% penicillin–streptomycin (v/v).
To generate CTLs, freshly isolated CD8 T cells were stimulated 2 d with coated anti-CD3 (10 μg/ml; UCHT1 produced by A. Alcover), soluble anti-CD28 (7 μg/ml; Beckman Coulter), and recombinant human IL-2 (100 U/ml; PeproTech) in human CD8 medium. Cells were infected for 24 h with lentiviruses coding for control or Apc-specific short hairpin RNAs (shRNAs) (10% v/v) in human CD8 medium with IL-2, in which FBS was replaced by human serum and then selected for 3 d with puromycin (3.9 μg/ml). The percentage of infected CTLs was 70–85%, as assessed by GFP expression by FACS.
For short interfering RNA (siRNA) experiments, the following small dsRNA oligonucleotide sequence was used for Apc depletion (siRNA for Apc depletion [siApc]): 5′-GAGAAUACGUCCACACCUU-3′ (GE Healthcare), as we previously described (29). As siRNA control (siCtrl), the sequence used was 5′-UAGCGACUAAACACAUCAA-3′ (siGENOME Non-Targeting siRNA #1; GE Healthcare). A total of 1 × 107 freshly isolated CD8 T cells were transfected with 1 nmol siCtrl or siApc using the Human T Cell Nucleofector Kit and the program U-14 on an Amaxa Nucleofector II (Lonza). Cells were then harvested in human CD8 medium without penicillin–streptomycin and used 72 h after transfection.
For shRNA experiments, lentiviruses were produced by HEK293T cells transfected with the transient calcium phosphate DNA precipitation technique. Cells were transfected with pCMV-deltaR8-2 and pCMV-env-VSV, together with a pLKO.1-puro-CMV-tGFP lentiviral vector expressing or not (as negative control) as shRNA–targeting Apc (5′-GACTGTCCTTTCACCATATTT-3′) (Sigma-Aldrich). Forty-eight hours later, supernatant was recovered and concentrated 40× by ultracentrifugation (26,000 rpm, 1.5 h, 4°C). Lentiviruses stocks were stored at −80°C.
The P815 mouse mastocytoma cell line was used as the tumor target cell. P815 cells were maintained in DMEM supplemented with 10% FBS and 1% penicillin–streptomycin (v/v). To make stimulatory target cells, P815 were pulsed with anti-human or mouse CD3 Abs, as described below.
ApcMin/+ and wild-type mice, lymphocyte isolation, and cell culture
Heterozygous C57BL/6J-ApcMin (ApcMin/+) mice and wild-type controls were purchased from The Jackson Laboratory. Mice were housed and bred under pathogen-free conditions in the Central Animal Facility of the Institut Pasteur. The protocols used had been approved by the Ethical Committee for Animal Experimentation of the Institut Pasteur and by the French Ministry of Research. Wild-type and ApcMin/+ male and female littermates were sacrificed at 8–14 wk of age. The age, gender, and number of mice analyzed per type of experiment are depicted in Supplemental Table I. Individual mice were analyzed separately. We did not observe any significant gender effect in any of the experiments performed. Data from male and female mice were pooled for each experimental condition.
Spleen and lymph nodes were homogenized through a 70-μm filter. To generate CTLs, CD8 T cells were purified using the MACS CD8a T Cell Isolation Kit (Miltenyi Biotec) and stimulated 2 d with coated anti-CD3 (5 μg/ml; 145-2C11; eBioscience), soluble anti-CD28 (2 μg/ml; eBioscience), and IL-2 (10 ng/ml; Miltenyi Biotec). Cells were then cultured for 5–6 d in mouse CD8 medium: DMEM plus GlutaMAX-I (Life Technologies) supplemented with 10% FBS, 1 mM sodium pyruvate and nonessential amino acids, 50 μM 2-ME, 10 mM HEPES, and 1% penicillin–streptomycin (v/v), in the presence of IL-2 (10 ng/ml; Miltenyi Biotec).
For Apc detection by Western blot, total cell extracts were obtained by lysing cells for 5 min on ice in lysis buffer (50 mM Tris [pH 7.4], 100 mM NaCl, 0.5% Nonidet P-40, 5 mM EDTA, 5 mM EGTA, 40 mM 2-ME, 10 mM NaF, 10 mM Na4P2O7, 2 mM orthovanadate, and protease inhibitor mixture). Cell lysates were cleared by spinning at 20,800 × g for 20 min at 4°C. An equal amount of protein content was measured using the Bio-Rad Protein Assay (Bio-Rad Laboratories) was loaded in NuPAGE 3–8% Tris-Acetate gels (Thermo Fisher Scientific). Proteins were transferred onto nitrocellulose membranes (LI-COR Biosciences) for 4 h using a Bio-Rad Mini Trans-Blot system in buffer containing 50 mM Tris, 380 mM glycine, 20% ethanol, and 0.1% SDS. Membranes were saturated with blocking buffer (Rockland Immunochemicals) and incubated with anti-Apc (2 μg/ml; ALi 12–28; Abcam) and anti-ZAP70 (50 ng/ml; BD Biosciences) or anti-PLCγ1 (1/1000; Cell Signaling Technology) overnight at 4°C. They were washed and then incubated with specific secondary Abs conjugated with Alexa Fluor 680 or DyLight 800 (Thermo Fisher Scientific) for 45 min. Near-infrared fluorescence was imaged and quantified using the Odyssey Classic Near-Infrared Imaging System (LI-COR Biosciences), and the Apc band intensity was normalized to control ZAP70 or PLCγ1 band.
For Apc detection by retrotranscription quantitative PCR, total RNA was extracted using the RNeasy Plus Micro Kit (QIAGEN) following the manufacturer’s instructions. cDNA was obtained from 200 ng of total RNA using an iScript cDNA Synthesis Kit (Bio-Rad Laboratories). FastStart Universal SYBR Green PCR Master Mix (Roche) and an ABI PRISM 7900HT Sequence Detection system (Applied Biosystems) were used to quantify gene products. Quantitative PCR were performed in triplicates. Quantity values were calculated by the relative standard curve method and normalized to the mRNA expression of the RLP13a housekeeping gene. The primer sequences used were as follows: Apc forward, 5′-CCAACAAGGCTACGCTATGC-3′ and reverse, 5′-TACATCTGCTCGCCAAGACA-3′; RLP13a forward, 5′-CATAGGAAGCTGGGAGCAAG-3′ and reverse, 5′-GCCCTCCAATCAGTCTTCTG-3′
Confocal microscopy, image posttreatment, and quantitative image analysis
Coverslips were washed with HCl–ethanol 70% and coated with poly-l-lysine 0.002% in water (Sigma-Aldrich). For flat pseudosynapse formation, coverslips were further coated with anti-CD3 (10 μg/ml; UCHT1; BioLegend) overnight at 4°C. Coverslips were washed and blocked for 30 min at 37°C with human CD8 medium. Human T cells were plated on coverslips for 5 min at 37°C and fixed with 4% paraformaldehyde for 13 min at room temperature. For microtubule detection, cells were also incubated 20 min at −20°C in ice-cold methanol. For CTL–target cell immunological synapse formation, P815 cells, previously coated with anti-CD3 (20 μg/ml; OKT3 produced by A. Alcover) for 45 min at 4°C, were plated on poly-l-lysine–coated coverslips for 15 min at 37°C. Coverslips were carefully washed with PBS, and human CTLs were added at a 1:1 CTL/target ratio and incubated for 15 min at 37°C. Cells were then fixed with 4% paraformaldehyde for 13 min at room temperature. Coverslips were washed in PBS and incubated 1 h in PBS with 1% BSA (v/v) to prevent unspecific binding. Cells were then incubated 1 h at room temperature with PBS, 1% BSA, 0.1% Triton X-100 and anti-Apc (1/300; gift of I. Näthke, University of Dundee), anti–β-tubulin (6.6 μg/ml; Merck Millipore), anti-pericentrin (1/100; Abcam), and anti-perforin (10 μg/ml; BD Biosciences). Coverslips were then incubated with the corresponding fluorescent-coupled secondary Ab and Texas Red-X Phalloidin (1/100; Invitrogen) for 45 min at room temperature. After three washes in PBS with 1% BSA, coverslips were mounted on microscope slides using ProLong Gold Antifade Mountant with DAPI (Life Technologies).
Confocal images were acquired with an LSM 700 confocal microscope (Zeiss) using the Plan Apochromat 63×/1.40 numerical aperture objective. Optical confocal sections were acquired with ZEN software (Zeiss) by intercalating green and red laser excitation to minimize channel cross-talk. All analyses were performed using Fiji software (32). For Apc quantification and F-actin analysis, optical sections were acquired at 1-μm intervals. Apc fluorescent intensity was measured on the total cell. Formation of the actin ring and phalloidin fluorescence intensity analyses were performed on one confocal section at the cell–coverslip contact. F-actin intensity profiles were assessed along a line drawn across each cell image (as shown in Fig. 2D). Ranking of each cell in a category (presence of F-actin ring plus central clearance, intermediate; low, F-actin ring plus clearance) was decided after observing actin profiles on three to four different angles. For microtubule pattern and synapses analyses, optical sections were acquired at 0.2-μm intervals, and images were treated by deconvolution with the Huygens Professional software (Scientific Volume Imaging). Microtubule pattern analysis was performed on projection of four confocal sections at the cell–coverslip contact, and phenotypes were classified by blinded image observation by three different investigators. Synapse morphology and cytotoxic granule localization were analyzed in three-dimensional (3D) projections and classified as described for microtubule patterns. Centrosome localization was estimated by measuring the distance between the anti-pericentrin staining and the CTL plasma membrane at the center area of the immunological synapse. Quantification of cytotoxic granules per cell was obtained using the TrackMate plugin for ImageJ developed by J.-Y. Tinevez (Photonic BioImaging, Unit of Technology and Service, Institut Pasteur) (33).
Rupture force assay of cell–cell interactions in laminar flow chamber
Glass surface (Rogo-Sampaic) of the flow chamber (Slide I0.1, Ibidi, Germany) was washed using five chamber volumes (CV) of sulfuric acid (2 M; Merck-Sigma), then five CV of hydrogen peroxide (33%; Merck-Sigma), and rinsed with 10 CV of water, then 10 CV of PBS. Bovine fibronectin (10 μg/ml in PBS; Merck-Sigma) was incubated 1 h at room temperature and rinsed with 20 CV of PBS. The chamber was equilibrated with DMEM (Life Technologies) at 37°C, then loaded with P815 cells for 15 min at 37°C. One CV of anti-CD3 (20 μg/ml final; OKT3 produced by A. Alcover) in DMEM was carefully added, and cells were incubated for 15 min at 37°C. Unbound cells and Abs were cleared by flowing five CV of RPMI 1640 (37°C; Lonza). One CV of human CTLs in RPMI 1640–1% FBS (Lonza) was injected inside the chamber and incubated for 10 min at 37°C. A flow rate of PBS, increasing from 0 to 38.4 ml/min, was applied through the chamber for 115 s using a syringe pump (SP210iW, World Precision Instruments) controlled by a computer and synchronized with image acquisition (three images per s, 1100 × 840 μm) using an inverted transmission microscope (Axio Observer D1; Zeiss) with a 10×/0.3 numerical aperture objective controlled by MicroManager (34).
Image series were analyzed using ImageJ software (32) and the plugin Cell Counter. P815 and CTLs were distinguished by their size, shape, and nucleus/cell ratio. The flow rate value at the cell–cell rupture event was used to compute the dragging force on the released cell according to its size (mean diameter of 8 μm), shape (roughly spherical), and density (mean value 1.20 kg/l). Calibration of dragging force was performed from the sedimentation rate of cells in the chamber (measured PBS density 1.0034 kg/l and dynamic viscosity 0.6998 mPa per s at 37°C) (35) and the theoretical flow speed versus wall distance according to Poiseuille solution to Navier–Stokes formalism for a Reynolds number below 10 characterizing a laminar flow. The number of cells loaded in the chamber and located in the field of view may vary from one experiment to another. Moreover, the CTLs initially retained as conjugates were always lower for Apc-silenced cells, indicating initial adhesion impairment. To give the same weight to each of the four experiments performed on different cell numbers in the overall whisker-boxes, shCtrl (n = 3745) and shApc (n = 1091) cells released per force interval were scaled to a total number of 250 cells per experiment, 1000 cells after adding the four sets.
Live-cell total internal reflection fluorescence microscopy
Glass-bottom microwell dishes (MatTek) were washed with HCl–ethanol 70%. They were coated with poly-l-lysine 0.002% in water (Sigma-Aldrich) for 30 min at room temperature and then with anti-CD3 (10 μg/ml; UCHT1; BioLegend) overnight at 4°C. Human CTLs were incubated with LysoTracker Deep Red (0.1 μM; Molecular Probes by Life Technologies) for 45 min at 37°C to label cytotoxic granules. Cells were washed and resuspended in human CD8 medium. A total of 1 × 105 cells were dropped in a microwell, and once cells were seeded, images were acquired in live-cell total internal reflection fluorescence (TIRF) plan every 150 ms for 4 min. TIRF images were acquired with an LSM 780 Elyra PS.1 confocal microscope (Zeiss) using an α Plan Apo 100×/1.46 numerical aperture oil immersion objective. Images were analyzed using the TrackMate plugin for ImageJ software (33). A fluorescence threshold above 2× the mean granule fluorescence was set to select strong fluorescence events more likely closer to the plasma membrane.
At day 6–8 after initial stimulation, human or mouse CTLs were resuspended in RPMI 1640 plus GlutaMAX-I (Life Technologies) 3% FCS, 10 mM HEPES, and mixed at the indicated ratios with P815 target cells previously coated with anti-human-CD3 (6 μg/ml; UCHT1; BioLegend) or anti-mouse-CD3 (6 μg/ml; 145-2C11; eBioscience) for 45 min at 4°C. Mixed CTLs and target cells were incubated 4 h at 37°C. The percentage of target cell lysis was measured using the CytoTox 96 Non-Radioactive Cytotoxicity Assay (Promega) following manufacturer’s instructions. Absorbance at 490 nm was measured using a PR2100 Microplate Reader (Bio-Rad Laboratories).
At day 6–8 after initial stimulation, human or mouse CTLs were mixed at a 1:1 ratio with P815 target cells previously coated with the indicated anti-CD3 concentration (for human: OKT3, produced by A. Alcover; for mice: 145-2C11; eBioscience) in presence of anti-CD107a-PE (for human: 1/90, clone H4A3; BioLegend; for mice 2.5 μg/ml, clone 1D4B; BD Biosciences), and incubated for 3 h at 37°C. Alternatively, CTLs were stimulated for the indicated time with coated anti-CD3 (10 μg/ml; clone OKT3 produced by A. Alcover) and soluble anti-CD28 (7 μg/ml; Beckman Coulter). Cells were stained with either anti-human CD8a-allophycocyanin-Cy7 (1/25; BioLegend) or anti-mouse CD8a-PerCP-Cy5.5 (1 μg/ml; BD Biosciences). Events were acquired on a MACSQuant Analyzer (Miltenyi Biotec), and analysis was performed using FlowJo 10 software (FlowJo). All samples were gated on forward and side scatter and for singlets.
Analysis of protein phosphorylation
Mouse T cells were incubated with 20 μg/ml of biotin-conjugated anti-CD3 (145-2C11; eBioscience) and anti-CD28 (eBioscience) Abs for 30 min at 4°C. Cells were washed, resuspend in Opti-MEM plus GlutaMAX-I (Life Technologies), and placed 1 min at 37°C. Streptavidin (10 μg/ml; Sigma-Aldrich) was added, and cells were incubated at 37°C for the indicated times. Ice-cold PBS containing 2 mM orthovanadate and 0.05% sodium azide was added to stop cell stimulation. Cells were lysed in ice-cold lauryl–β-maltoside buffer (20 mM Tris [pH 7.4], 150 mM NaCl, 0.25% lauryl-β-maltoside, 50 mM NaF, 10 mM Na4P2O7, 1 mM EGTA, 2 mM orthovanadate, 1 mM MgCl2, and protease inhibitor mixture) for 10 min on ice. Cell lysates were cleared by spinning at 20,800 × g for 10 min at 4°C. Equal amounts of protein content measured using the Bio-Rad Protein Assay (Bio-Rad Laboratories) was loaded in NuPAGE 4–12% Bis-Tris gels (Thermo Fisher Scientific). Protein transfer was performed using the Trans-Blot Turbo system (Bio-Rad Laboratories). Membranes were saturated with blocking buffer (Rockland Immunochemicals) and incubated overnight at 4°C or 2 h at room temperature with anti–phospho-Akt (p-Ser473; 1/1000; Cell Signaling Technology), anti–phospho-Erk1/2 (p-Thr202/Tyr204; 1/1000; Cell Signaling Technology), anti-GAPDH (6.6 μg/ml; Calbiochem). Membranes were then incubated with specific secondary Abs conjugated with Alexa Fluor 680 or DyLight 800 (Thermo Fisher Scientific) for 45 min at room temperature. Near-infrared fluorescence was recorded and quantified using an Odyssey Classic Near-Infrared scanner (LI-COR Biosciences). Band intensity was normalized to control GAPDH. For a pooled analysis of several experiments, normalized intensities were then divided by the mean normalized intensity of the same experiment.
Nuclear NFAT detection
Nuclear NFAT was analyzed by Western blot analysis of cells fractionated for nucleus and cytoplasm separation. Mouse T cells were activated as for analysis of protein phosphorylation, for the indicated times. Cells were lysed in ice-cold low-salt lysis buffer (10 mM KCl, 10 mM HEPES, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 50 mM NaF, 10 mM sodium pyrophosphate [Na4P2O7], and a protease inhibitor mixture) for 15 min on ice. Detergent Nonidet P-40 (Sigma-Aldrich) was added (0.45% v/v final), and cells were vortexed for 10 s to break cytoplasmic membranes without altering nuclear membranes. Lysates were centrifuged at 20,800 × g for 30 min at 4°C. Supernatants corresponding to cytosolic fractions were recovered in ice-cold tubes. Pellets containing nuclei were washed once in PBS and resuspended in ice-cold high-salt buffer (20 mM Tris–HCl [pH 8], 1% SDS, 2 mM EDTA, and protease inhibitor mixture), solubilized by sonication (3 × 10 s at 60% power) using a Vibracell 72434 (Bioblock Scientific), then centrifuged at 20,800 × g for 30 min. Supernatants were recovered as nuclear fractions. An equal amount of protein content was measured using the Bio-Rad Protein Assay (Bio-Rad Laboratories) was loaded in NuPAGE 3–8% Tris-Acetate gels (Thermo Fisher Scientific). Protein transfer was performed in a Bio-Rad Mini Trans-Blot system. Membranes were saturated with blocking buffer (Rockland Immunochemicals) and incubated with anti-NFATc2 (NFAT1) (0.25 μg/ml; BD Biosciences), anti-SLP76 (1/1000; Thermo Fisher Scientific) or anti-Lamin B1 (1/2000; Abcam) overnight at 4°C. Then, they were incubated with specific secondary Abs conjugated with Alexa Fluor 680 or DyLight 800 (Thermo Fisher Scientific) for 45 min. Near-infrared fluorescence was imaged and quantified using the Odyssey Classic scanner as described above. NFAT band intensity was normalized to SLP76 cytoplasmic control or Lamin B1 nuclear control. For a pooled analysis of several experiments, normalized intensities were then divided by the mean normalized intensity of the same experiment.
Mouse T cell differentiation and proliferation assays
Mouse cells from 12-wk-old mice freshly isolated from the spleen and lymph nodes were stimulated with concanavalin A (ConA; 5 μg/ml; Sigma-Aldrich) in mouse CD8 medium with IL-2 (10 ng/ml; Miltenyi Biotec) for 2 d and then cultured four additional days with IL-2. At day 0, 2, 4, and 6, cells were incubated with anti-CD16/32 (10 μg/ml; BioLegend) to block Fc receptors and then stained with Fixable Viability Stain 450 (25 ng/ml; BD Biosciences), anti-CD3-FITC (5 μg/ml; 145-2C11, BD Biosciences), anti-CD4-allophycocyanin-Vio770 (1.5 μg/ml; Miltenyi Biotec), anti-CD8-PerCP-Cy5.5 (1 μg/ml; BD Biosciences), anti-CD25-PE-Cy7 (2 μg/ml; BD Biosciences), anti-granzyme B-A647 (1/50; BioLegend), anti-CD44-allophycocyanin (2 μg/ml; eBioscience), and anti-CD62L-RPE (2 μg/ml; eBioscience). Events were acquired on a MACSQuant Analyzer (Miltenyi Biotec), and analysis was performed using FlowJo 10 software (FlowJo). All samples were gated on forward and side scatter for singlets, and for live cells, samples were stained using Fixable Viability Stain 450 (250 ng/ml; BD Biosciences).
Proliferation was calculated by counting the cells at day 0, 2, 4, and 6 upon stimulation, as described above for the differentiation assays. Cell numbers were referred to the cell counts at day 0.
Detection of cytokine production
IL-2 was removed from CTL culture, and 20 h later, cytokine production was assessed by ELISA and flow cytometry as follows: for detection by ELISA, 1 × 3.105 human or mouse CTLs were restimulated at 37°C for 6 h with PMA (50 ng/ml; Sigma-Aldrich) and calcium ionophore A23187 (500 ng/ml; Sigma-Aldrich) or 20 h with coated anti-CD3 (for human, 5 μg/ml, UCHT1; BioLegend; for mice, 145-2C11, 5 μg/ml; eBioscience) and soluble anti-CD28 (for human, 5 μg/ml; Beckman Coulter; for mice, 2 μg/ml; eBioscience). Secreted IL-2, TNF-α, and IFN-γ levels were measured from culture supernatants with specific DuoSet ELISA Kits (R&D Systems) following manufacturer’s instructions.
For detection by flow cytometry, 1 × 105 mouse CTLs were restimulated as described above. Two hours after the beginning of restimulation, brefeldin A (10 μg/ml; Sigma-Aldrich) was added. Cells were fixed with 2% paraformaldehyde for 15 min at room temperature, incubated with anti-CD16/32 (10 μg/ml; BioLegend), and subsequently stained with anti-IL-2-RPE (4 μg/ml; eBioscience), anti-TNF-α-FITC (1.2 μg/ml; Miltenyi Biotec) and anti-IFN-γ-PE-Cy7 (8 μg/ml; BioLegend) in presence of 0.05% saponin. Events were acquired on a MACSQuant Analyzer (Miltenyi Biotec), and analysis was performed using FlowJo 10 software (FlowJo). All samples were gated on forward and side scatter for singlets, and for live cells, they were stained using Fixable Viability Stain 450 (250 ng/ml; BD Biosciences).
Statistical analyses were carried out using Prism Software (GraphPad). Error bars in plots represent the mean ± SEM. The p values are represented as follows: ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, NS p ≥ 0.05. The type of test used is mentioned in each figure legend.
Apc is expressed in human CD8 T cells
We first investigated the expression of Apc in primary human CD8 T cells. Immunoblotting on cell lysates revealed similar levels in CD4 and CD8 T cells of a protein band consistent with the 311 kDa full-length protein previously described (24, 29) (Fig. 1A). This band disappeared upon siRNA silencing in CD8 T cells (Fig. 1B). Apc depletion was similarly detected by quantitative RT-PCR upon siRNA transfection in human resting CD8 T cells or shRNA retroviral transduction in differentiated CTLs (Fig. 1C). We next analyzed Apc subcellular localization. Both resting and differentiated CD8 T cells showed a punctate pattern of Apc associated with microtubules and aligned at the edges of immunological synapses formed on anti-CD3–coated coverslips (arrowheads) (Fig. 1D, 1E). This Apc pattern was consistent with our findings in CD4 T cells (29) and with previous reports in other cell types (36, 37). The specks pattern was strongly diminished in Apc-silenced T cells (Fig. 1F), further supporting the specificity of the Apc immunofluorescence pattern.
Apc regulates microtubule and actin cytoskeleton remodeling at the immunological synapse
Apc was shown to regulate microtubule network organization in polarized migrating cells (38–40) as well as actin polymerization (41). Furthermore, Apc coordinates actin polymerization and microtubules at focal adhesions in migrating cells (42). In both CD4 and CD8 T cells, actin and microtubule cytoskeletons cooperate to build functional immunological synapses (2, 8, 14). Indeed, in CD8 T cells, microtubule polarization toward target cells combined with actin polymerization, followed by rapid clearance from the center of the synapse, appear to be key for cytolytic granule release and target cell lysis (6, 9, 43). Importantly, we previously showed that Apc silencing in CD4 T cells results in microtubule network disorganization at the immunological synapse (29).
Therefore, we investigated whether Apc controls cytoskeleton remodeling in CD8 T cell immunological synapses. Apc silencing altered microtubule network organization at the immunological pseudosynapses formed by either resting or differentiated CD8 T cells on anti-CD3–coated coverslips. Microtubule patterns were more frequently radially organized in control cells, reaching the synapse periphery and with visible microtubule organizing center. In contrast, Apc-silenced cells more often displayed randomly organized microtubules (Fig. 2A–C). In addition, we observed that actin remodeling at the synapse was also impaired (Fig. 2D–F). Actin could accumulate at the synapse, and no differences between control and Apc-silenced cells were observed with regard to the total amount of F-actin at the contact site (Fig. 2E). However, control cells often formed F-actin rings at the periphery of the synapse, excluding actin from the center (Fig. 2D, left panel, and Fig. 2F). In contrast, Apc-silenced cells less frequently displayed this actin pattern, more often forming synapses with a less defined actin ring and lower clearance from the center (Fig. 2D, center and right panels, and Fig. 2F).
Apc conditions CTL centrosome polarization, immunological synapse shape, symmetry, and stability
Our observations on cytoskeleton remodeling prompted us to investigate a larger number of features of immunological synapses formed between human CTLs and anti-CD3–coated P815 tumor target cells, a model of Ab-mediated T cell cytotoxicity. Following the examples of synapses shown in Fig. 3A, we analyzed the following: 1) CTL centrosome polarization, assessed by the distance of the centrosome (arrows) to the center of the synapse; 2) synapse symmetry, assessed by the relative aligned position of the centrosome and the T cell and target cell nuclei; and 3) synapse shape, assessed by the presence of large membrane extensions (arrowheads). Fig. 3A (left panel) provides an example of a symmetrical CTL–target cell conjugate, with a polarized centrosome (arrow) and no large membrane extensions. The right panel is an example of an asymmetrical conjugate, with the centrosome poorly polarized and displaying large cell extensions (arrowheads) and a more irregular shape. We observed that Apc silencing did not alter the number of conjugates (data not shown) but impaired T cell centrosome polarization toward the synapse and T cell symmetry and increased the presence of large cell extensions (Fig. 3B–D). These features may reflect or be the cause of the inability of Apc-silenced cells to make stable interactions with target cells.
To obtain further insight into the effect of Apc silencing on the stability of CTL–target cell conjugates, we measured the rupture force of cell–cell interactions between CTLs and anti-CD3–coated P815 cells immobilized in a microscope laminar flow chamber undergoing increasing flow rate. A laminar shear flow rate was applied, generating dragging forces from 0 to 5 nN on individual cells with an increasing slope of 34 pN/s. CTLs interacting with immobilized P815 were stretched by this dragging force until rupture of the contact and release of the T cells into the stream (Fig. 3E; Supplemental Video 1). CTL–P815 cell rupture forces ranged from 0 to 4.5 nN, with half of the values between 0.5 and 2.5 nN, suggesting a large variety of synapse formation kinetics or complexity. Median rupture forces were 1.52 and 0.67 nN for control and Apc-silenced cells, respectively, indicating that the interacting force was decreased by loss of Apc expression (Fig. 3F). Interestingly, cumulative plots of cell–cell rupture events versus dragging force show linear distributions for control cells and hyperbolic distributions for Apc-silenced cells (Fig. 3G). Of note is that, although equal cellular input of shCtrl and shApc cells was applied to the chamber, the number of CTLs forming initial conjugates with P815 cells before applying flow pressure was lower in shApc-treated cells in most experiments (n in Fig. 3F legend). This further reflects the lower capacity of Apc-silenced cells to initially stabilize conjugates with P815 target cells. To give the same weight to each of the four experiments performed on different cell numbers in the combined data whisker-boxes, shCtrl (n = 3745) and shApc (n = 1091) cells released per force interval were scaled to a total number of 250 cells per experiment, 1000 cells after adding the four sets.
Therefore, Apc silencing affects centrosome polarization, immunological synapse shape and symmetry, and the stability of CTL–tumor target cell interactions.
Apc controls lytic granule dynamics, targeting, and fusion at the immunological synapse
Perturbation of actin and microtubule remodeling at the immunological synapse, T cell centrosome polarization, and the stability of CTL–target cell interactions of Apc-silenced cells (Figs. 2, 3) might result in reduced accessibility of lytic granules to the immunological synapse, reducing the capacity of CTLs to kill target cells (6, 7, 9).
To investigate this, we first analyzed lytic granule dynamics at the immunological synapse. Human control or Apc-silenced CTLs were labeled with LysoTracker and set on anti-CD3–coated coverslips to obtain flat immunological synapses. Granule dynamics were then followed by live-cell TIRF microscopy. We measured the following events occurring at the TIRF plan during 4 min recording: 1) the total number of bright fluorescence spots as a readout of granules targeted to the synapse; 2) the granule movement by tracking individual fluorescent particles at the synapse; and 3) the number of granules generating a fluorescence burst as a readout of granule fusion with the plasma membrane (6, 9). Representative images are shown in Fig. 4E and Supplemental Video 2. Quantification showed that Apc-silenced cells displayed reduced targeting events (Fig. 4A). Furthermore, lytic granules in Apc-silenced cells remained for longer times at the synapse, and their speed was lower than in control cells (Fig. 4B). However, the length of displacement of granules was similar in control and Apc-silenced CTLs (Fig. 4B, 4C). Importantly, Apc-silenced cells displayed fewer fusion events (Fig. 4D, 4E, arrowheads, and Supplemental Video 2). Therefore, loss of Apc expression alters lytic granule dynamics, targeting, and fusion at the immunological synapse.
We next studied immunological synapses formed between CTLs and anti-CD3–coated P815 tumor cells upon 15 min of CTL–target cell contact. Subcellular localization of lytic granules was assessed by anti-perforin Ab staining. Interestingly, control cells displayed a low number of randomly distributed lytic granules (Fig. 4F, left panel, and Fig. 4G), whereas Apc-silenced cells displayed a higher number of granules that were more grouped and polarized toward the synapse (Fig. 4F, right panel, and Fig. 4G).
These data suggest that in Apc-silenced cells lytic granules can group close to the centrosome and reorient toward the synapse; however, they are less efficiently targeted and fused to the synapse plasma membrane. As a consequence, a higher number of lytic granules is retained in CTLs during their interaction with target cells.
Apc regulates CTL killing of tumor target cells
We next asked whether the defects in cytotoxic granule dynamics observed in Apc-silenced cells could impair their capacity to kill tumor target cells. To this end, we used conventional longer incubation assays to measure degranulation and tumor target cell killing. Both human Apc-silenced CTLs and mouse ApcMin/+ CTLs were less efficient in killing anti-CD3–coated P815 tumor target cells, as assessed by a lactate dehydrogenase–release colorimetric assay (Fig. 5A, 5B). However, no significant differences were observed in the capacity of these human and mouse CTLs to degranulate, as assessed by measuring cell surface expression of the LAMP1 luminal epitope CD107a upon CTL contact with anti-CD3–coated P815 tumor target cells for 3 h (44) (Fig. 5C, 5D).
To investigate the apparent contradiction between a less efficient lytic granule fusion at the immunological synapse of Apc-silenced cells and no alteration in LAMP1 expression, we stimulated the cells under the same conditions used for granule fusion detection, and we shortened the kinetics. No significant changes in LAMP1 expression between control and Apc-silenced CTLs were observed under these experimental conditions either (Supplemental Fig. 1). This dichotomy has been reported by others (45) and may account for LAMP1+ vesicle release in an unpolarized manner in response to calcium signals despite a less efficient lytic granule delivery at the CTL–target cell interface and reduced killing.
Altogether, the above described data indicate that Apc is necessary for efficient microtubule and actin cytoskeleton remodeling conditioning, lytic granule targeting and fusion at the immunological synapse, and ultimately, efficient killing of tumor target cells.
Apc modulates T cell signaling and NFAT nuclear translocation
We had shown in CD4 T cells that the polarity regulators Dlg1 and Apc were involved in microtubule network organization at the immunological synapse, conditioning NFAT transcriptional activation (14, 29, 46). We therefore investigated whether Apc modulation of cytoskeleton remodeling could affect CD8 T cell activation. To this end, we analyzed the capacity of ex vivo–differentiated CTLs from control and ApcMin/+ mutant mice to respond to anti-CD3 plus anti-CD28 stimulation. ApcMin/+ CTLs exhibited a mild but significantly reduced capacity to activate Erk and Akt serine-threonine kinases as assessed by the phosphorylation of regulatory residues of these kinases (Fig. 6A–C). In contrast, the upstream protein tyrosine kinase ZAP70 was more variably activated, and no significant differences between control and ApcMin/+ cells could be appreciated (data not shown).
Furthermore, NFATc2 nuclear translocation in response to anti-CD3 plus anti-CD28 stimulation was reduced in ApcMin/+ CTLs (Fig. 6D–I).
Apc defects mildly alter ex vivo differentiation of CD8 T cells
We then investigated whether Apc mutation was leading to defects in long-term T cell survival affecting T cell numbers in vivo. To this end, T cells from the spleen and lymph nodes were phenotyped and counted just upon the mice’s euthanasia. ApcMin/+ mice did not show evident defects because they presented similar T cell numbers at 8 wk of age and even higher T cell numbers at 10, 12, and 14 wk of age (Fig. 7). This T cell number increase coincided with the increasing development of intestinal adenomatous polyps in ApcMin/+ mice.
Next, we analyzed whether Apc defects could be associated with impaired CD8 T cell proliferation and differentiation upon in vitro T cell stimulation. Splenocytes and lymph node cells were stimulated with ConA and IL-2, and the proliferation and expression of the activation/differentiation molecules CD25, granzyme B, CD44, and CD62L was assessed at 2, 4, and 6 d upon activation. Expression of CD44 and CD62L adhesion molecules in CD8 T cells from ApcMin/+ mice appeared mildly affected, mainly at day 2, and granzyme B was affected at day 6. CD62L and granzyme B expression tended to be lower, but the high variability among the animals made differences between wild-type and ApcMin/+ statistically NS (Fig. 8A–D, 8G, 8H). In contrast, no differences were observed in CD25 expression (Fig. 8E, 8F). Interestingly, differences in CD44 and CD62L expression were more readily observed in CD4 T cells, with more pronounced and longer effects, and the level and kinetics of CD62L expression in CD4 T cells were different from in CD8 T cells (Supplemental Fig. 2A–F). However, no significant differences in the proliferative capacity were found within the same experimental setting (Supplemental Fig. 2G–I).
These findings show that CD8 T cells from ApcMin/+ mice present mildly reduced early T cell activation events, such as Erk and Akt activation, and NFAT nuclear translocation as well as mild changes in CD44, CD62L, and granzyme proteins expressed during differentiation ex vivo.
Apc defects do not alter ex vivo CTL cytokine production
Because NFAT drives the transcription of a variety of T cell cytokines, we investigated the capacity of Apc-silenced human CTLs and ApcMin/+ CTLs to produce effector cytokines. Despite the reduced capacity of ApcMin/+ CTLs to induce NFAT nuclear translocation, we did not observe significant differences in the production of IL-2, IFN-γ, or TNF-α by differentiated CTLs restimulated with anti-CD3 plus anti-CD28, either by intracellular staining and FACS analysis or by ELISA of cell culture supernatants (Fig. 9A, 9B). Similar results were found when assessing cytokine production in cell culture supernatants of control and Apc-silenced human CTLs by ELISA (Fig. 9C).
These findings indicate that Apc defects in both human and mouse CTLs do not affect their capacity to produce effector CD8 cytokines under our ex vivo stimulation conditions.
Altogether, the data shown in this study unveil a novel role of Apc in cytotoxic T cell effector functions. Apc-silenced human peripheral blood CD8 T cells allowed us to reveal the importance of Apc in cytoskeleton remodeling at the immunological synapse. Both microtubules and actin reorganization appeared affected. Thus, the radial organization of microtubules and centrosome polarization were perturbed, and F-actin ring formation at the periphery of the synapse and F-actin exclusion from the center of the synapse were impaired in Apc-defective T cells. Additionally, consistent with an impact on actin cytoskeleton dynamics (47), synapse stability, shape, and symmetry were also altered in Apc-deficient cells.
Apc regulation of microtubule organization in CD8 T cells is consistent with the described role of Apc as a polarity regulator in other cell systems (23). Apc performs this function in association with other regulators of cell polarity, including Cdc42, Scribble, Par6, PKCζ, and Dlg1, as shown, for instance, in migrating astrocytes (40, 48). Furthermore, we have shown before that Dlg1 and Apc control microtubule patterns, centrosome polarization, and signaling microcluster dynamics at the immunological synapse of CD4 T cells (14, 29). Although less extensively studied, Apc may also regulate actin polymerization in synergy with its partner, the formin family protein mDia (41). In addition, Apc regulates actin-microtubule cross-talk at focal adhesions of migrating cells (42). Interestingly, molecular partners of Apc, such as Cdc42 (40) and formins (41), were shown to control actin remodeling and centrosome polarization at the immunological synapse (49, 50). Therefore, Apc may collaborate with these cytoskeleton regulators to collectively reorganize both cytoskeleton components at the CTL immunological synapse, thus ensuring their interplay, which is crucial for properly tuning T cell functions.
Concomitant centrosome polarization and F-actin exclusion from the center of the immunological synapse has been shown to be key for lytic granule targeting and fusion at the synapse and efficient cytotoxic function against infected or tumor target cells (6, 7, 9, 43). Therefore, impairment of both microtubule organization and centrosome polarization, together with F-actin reorganization at the synapse in Apc-defective CTLs, may account for their altered lytic granule dynamics, targeting, and fusion at the synapse observed by live-cell TIRF microscopy. Apc-silenced cells could still concentrate lytic granules in the centrosomal area and polarize them to a certain extent toward the immunological synapse, suggesting that there is not a complete failure in granule transport on microtubules. Granules may be able to move toward the centrosome and microtubules minus end, a movement mediated by dynein motors. In contrast, granules do not properly reach the immunological synapse plasma membrane. Disorganization of microtubule network at the immunological synapse likely prevents microtubules from reaching the plasma membrane and, as a consequence, precludes centrosome closing up to the synapse, which facilitates lytic granule docking and fusion (7). Kinesin-mediated transport to the microtubules plus end may also help granules to reach the synapse plasma membrane (51, 52). As a likely consequence of inefficient lytic granule dynamics, targeting, and fusion, Apc-defective CTLs killed tumor target cells less efficiently than control cells. However, no difference in degranulation as assessed by increase in LAMP1 (CD107a) cell surface expression was detected. The apparent contradiction between the live-cell TIRF experiments and LAMP1 expression may be due to the distinct sensitivities of the two assays to discriminate differences. In particular, the fact that the LAMP1 vesicular compartment is larger than the lytic granule compartment and may fuse in a nonpolarized manner in response to normal calcium influx in Apc-silenced cells (29) could explain this dichotomy of findings. Of note, lack of direct correlation between degranulation and cytotoxicity has been reported in other experimental setups (45, 53).
Interestingly, our findings share a number of features with those previously reported in CTLs from Wiscott–Aldrich Syndrome patients, who carry mutations in the gene encoding the WAS protein (WASP), a key regulator of actin cytoskeleton dynamics. Thus, in vitro–derived CTLs from WAS patients had a lower capacity to kill tumor target cells, whereas they display equivalent degranulation as assessed by LAMP1 expression. Furthermore, although WAS CTLs could reorient their centrosome and lytic granules to the target cells, lytic granule distance to the immune synapse was higher than that of healthy donor cells. Finally, WAS CTLs displayed impaired synaptic ring formation, distorted less symmetric CTL–target cell synapses with lower stability, and delayed lethal hit (45, 54). However, WAS CTLs displayed defects in cytokine production, contrary to the Apc-defective CTLs described in this study (45).
Although significant, the differences in tumor cell killing activity between Apc-defective CTLs and control cells were modest. This may be due to the fact that ApcMin/+ mice are heterozygous (homozygous mutation is embryonically lethal) (55) and may express residual levels of Apc protein. Likewise, shRNA silencing in human primary CTLs is not fully efficient. Hence, remaining Apc protein could account for the observed function. Furthermore, other polarity regulators acting together with Apc may also compensate Apc defects. Finally, the T cell activation conditions we used for CTL generation in vitro may compensate, in part, for cytotoxicity defects because of Apc by increasing the expression of other proteins. This effect was described for some forms of familial hemophagocytic lymphohistiocytosis because of syntaxin 11 or Munc 18-2 mutations, in which the addition of IL-2 used to induce T cell proliferation in vitro was partially compensating for the loss of effector function (56, 57). Importantly, subtle differences in CTL cytotoxicity using in vitro assays may be predictive of more serious cytotoxic defects in vivo, leading to poor immune responses (58). Therefore, the small differences in cytotoxicity we observed could have consequences in long-term antitumor immunity in patients.
We have previously shown in CD4 T cells that Apc silencing altered NFAT nuclear translocation and reduced NFAT-driven transcriptional activation, leading to impaired IL-2 gene expression. Furthermore, CD4 T cells from ApcMin/+ mice had reduced capacity to produce IL-2, IL-4, and IFN-γ. Finally, Tregs in these mutant mice showed impaired differentiation and production of the anti-inflammatory cytokine IL-10 (29). We show in this study that CTLs from ApcMin/+ mice also had an impaired capacity to translocate NFATc2 to the nucleus. Interestingly, NFATc2 formed microclusters associated with microtubules in CD8 T cells (data not shown), as we previously showed in CD4 T cells (29). The total number of CD3+ lymphocytes, CD3+CD4+ or CD3+CD8+ was not reduced in ApcMin/+ mice but even increased starting from 10 wk of age and stayed enhanced at 12 and 14 wk of age, at which time signs of polyposis become evident. This suggests that long-term survival in vivo of CD4 or CD8 lymphocytes is not impaired, but T cell stimulation due to intestinal inflammation at the appearance of polyposis may increase T cell numbers. Interestingly, the analysis of the expression of several differentiation markers (e.g., CD25, granzyme B, CD44, and CD62L) did not show activation of freshly isolated T cells from ApcMin/+ mice. In addition, ex vivo activation with ConA and IL-2 revealed only mildly altered CD44 and CD62L expression at day 2 and of granzyme B at day 6 in ApcMin/+ mice. Interestingly, under the same stimulation conditions, the differences in expression of CD44 and CD62L between ApcMin/+ and control T cells were more pronounced and significant in the CD4 T cell population. This was not accompanied by differences in the proliferative capacity of CD4 and CD8 T cells under the same stimulatory conditions. This is, in part, in contrast with our previous data on lamina propria CD4 T cells that displayed impaired proliferation capacity at suboptimal stimulation (29). The differences may be due to the distinct tissue origin of T cells, lamina propria in our previous publication and spleen and lymph nodes in this study, or they may be due to the different stimulatory conditions used, different doses of CD3/CD28 on lamina propria cells in our previous work versus ConA in this study. In addition, Apc-silenced human and ApcMin/+ mouse ex vivo–differentiated CTLs produced an equal amount of IL-2, TNF-α, and IFN-γ after restimulation with CD3/CD28 Abs. Therefore, in comparison with our previous work (29), CD8 T cells appear to be less sensitive to Apc defects than CD4 T cells when performing their activation programs leading to differentiation and production of cytokines. Furthermore, the intestinal environment in ApcMin/+ mice could also influence these responses.
Altogether, the data presented in this study show that Apc is involved in the regulation of CTL effector functions that drive target cell killing. Hence, Apc defects impair both actin and microtubule cytoskeleton remodeling at the immunological synapse, centrosome polarization and lytic granule dynamics, targeting, and fusion at the synapse. However, CD8 T cell differentiation and cytokine production appear much less dependent on Apc compared with what has been observed in CD4 T cells (29). Thus, Apc defects may not significantly alter the acquisition of effector functions in CD8 T cells but impair the CTL capacities to kill tumor cells and could therefore have an impact in long-term antitumor immune responses in familial adenomatous polyposis patients, allowing easier tumor escape.
We thank Drs. F. Sepulveda and D. Scott-Algara for scientific advice, J.Y. Tinevez, A. Salles, J. Fernandes and the Unit of Technology and Service Photonic BioImaging facility, Institut Pasteur for technical support and the ICAReB Biological Resources core facility team, Institut Pasteur for providing primary T cell samples. We are grateful to I. Näthke for Abs.
This work was supported by grants from La Ligue Nationale contre le Cancer, Equipe Labellisée 2018, and institutional grants from the Institut Pasteur and INSERM. M.J. was supported by a Ligue Nationale contre le Cancer Doctoral Fellowship and the Institut Pasteur. M.M. is a scholar of the Pasteur Paris University International Doctoral Program, supported by the Institut Pasteur and the European Union Horizon 2020 Research and Innovation Programme under the Marie Sklodowska-Curie grant agreement 665807 (COFUND-PASTEURDOC). The Unit of Technology and Service Photonic BioImaging core facility is supported by the French National Research Agency France BioImaging (ANR-10–INSB–04, Investments for the Future).
Abbreviations used in this article:
adenomatous polyposis coli
Clinical Investigation and Access to BioResources
short hairpin RNA
siRNA for Apc depletion
small interfering RNA
total internal reflection fluorescence
regulatory T cell.
The online version of this article contains supplemental material.
The authors have no financial conflicts of interest.