β-Adrenergic receptor (β-AR) signaling exerts protumoral effects by acting directly on tumor cells and angiogenesis. In addition, β-AR expression on immune cells affects their ability to mount antitumor immune responses. However, how β-AR signaling impinges antitumor immune responses is still unclear. Using a mouse model of vaccine-based immunotherapy, we showed that propranolol, a nonselective β-blocker, strongly improved the efficacy of an antitumor STxBE7 vaccine by enhancing the frequency of CD8+ T lymphocytes infiltrating the tumor (TIL). However, propranolol had no effect on the reactivity of CD8+ TILs, a result further strengthened by ex vivo experiments showing that these cells were insensitive to adrenaline- or noradrenaline-induced AR signaling. In contrast, naïve CD8+ T-cell activation was strongly inhibited by β-AR signaling, and the beneficial effect of propranolol mainly occurred during CD8+ T-cell priming in the tumor-draining lymph node. We also demonstrated that the differential sensitivity of naïve CD8+ T cells and CD8+ TILs to β-AR signaling was linked to a strong downregulation of β2-AR expression related to their activation status, since in vitro–activated CD8+ T cells behaved similarly to CD8+ TILs. These results revealed that β-AR signaling suppresses the initial priming phase of antitumor CD8+ T-cell responses, providing a rationale to use clinically available β-blockers in patients to improve cancer immunotherapies.
Multiple immunosuppressive mechanisms within the tumor microenvironment (TME) alter the establishment of effective antitumor immune responses; such as, regulatory CD4+ T cells (Tregs; ref. 1), physical barriers imposed by the extracellular matrix (2), soluble mediators [e.g., prostaglandin E2 (PGE2), adenosine; refs. 3, 4], and enhanced expression of inhibitory immune checkpoints (5). Therefore, current immune-based therapies, such as anticancer vaccines, adoptive immunotherapy, and immune-checkpoint blockade, aim to induce robust T-cell responses, while limiting immune-suppressive mechanisms (6). Despite promising clinical results in various types of cancer, only a fraction of patients responds to immunotherapies, yet combinations that target immunosuppressive molecules within the TME, such as PGE2 and adenosine, may improve their efficiency (4, 7).
Over the last decade, the nervous sympathetic system, specifically adrenergic receptor (AR) signaling, has emerged as a component in cancer progression (8). The nervous sympathetic system depends on the basal release of endogenous catecholamines, adrenaline and noradrenaline, from the adrenal medulla and peripheral neurons, which is increased during psychological stress. Behavioral factors (e.g., anxiety, depression) can influence the incidence and development of lung, breast, lymphoid, and hematopoietic cancers (9, 10). Studies in various cancers have shown that patients taking β-AR antagonists, referred to as β-blockers, have a significantly better overall survival than the untreated ones (11). In mouse tumor models, β-AR signaling favors tumor growth and metastatic invasion by acting on tumor cells, angiogenesis, and the tumor stroma (12–14). Growing evidence suggests that adrenergic signals can also exert their suppressive effects on antitumor immune responses by acting directly on immune cells (15).
Although immune cells express both α- and β-ARs, β2-AR is the most highly expressed AR subtype, particularly in T and B cells (16). β2-ARs belong to the G protein–coupled receptor (GPCR) family and are coupled to Gαs protein (17). In the canonical pathway, β2-AR signaling activates downstream transcription factors in a cAMP-protein kinase A (PKA)–dependent manner (17). β2-AR signaling modulates the functions of natural killer (NK) cells, innate lymphoid cells, dendritic cells (DC), macrophages, B cells, and T cells (18). In the context of cancer, β-AR signaling promotes tumor growth and metastasis formation through M2 polarization of tumor-associated macrophages (TAM; refs. 19, 20), infiltration of immunosuppressive cells such as Tregs (21) and suppression of NK cell activity (22, 23). β-AR signaling prevents the infiltration of CD8+ T cells into tumors and inhibits their activity either directly (24–26) or indirectly by affecting DC maturation and antigen presentation (27) and by enhancing the generation and the suppressive activity of myeloid-derived suppressor cells (MDSC; ref. 28).
We investigated how β-AR signaling could modify the priming and/or the effector phase of an antitumor CD8+ T-cell response. To address this question, we used a murine model of vaccine-based tumor immunotherapy, the STxBE7 vaccine (composed of the nontoxic β-subunit of Shiga toxin coupled to HPV16 derived-E7 peptide) plus IFNα, which elicits E7-specific CD8+ T cells that can induce tumor regression (29, 30). We demonstrated that propranolol, a nonselective β-blocker, improved the antitumor efficacy of the STxBE7 vaccine by increasing the number of tumor-infiltrating CD8+ T cells (CD8+ TILs). However, we did not observe any effect of propranolol on the reactivity of CD8+ TILs, consistent with our observations that CD8+ TILs were insensitive ex vivo to adrenaline/noradrenaline treatment. In contrast, we found that naïve CD8+ T cells were sensitive to β-AR signaling, and that the beneficial effect of β-blockers on STxBE7 vaccine efficacy mainly occurred during the initial priming phase of CD8+ T cells in the tumor-draining lymph node (TDLN). Finally, we showed that this differential sensitivity of naïve CD8+ T cells and CD8+ TILs was related to a downregulation of β2-AR expression in CD8+ TILs, directly related to their activation status. Taken together, these data identified that the T-cell priming phase is a key target to prevent β-AR signaling inhibition of antitumor CD8+ T-cell responses.
Materials and Methods
C57BL/6J and OT-IxSJL mice were purchased from Charles Rivers and maintained in the Cochin Institute Specific-Pathogen-Free animal facility. β2-AR−/− mice were kindly provided by Sophie Ugolini (Centre d'Immunologie de Marseille-Luminy, Marseille, France). The specificity of OT-IxSLJ mice was confirmed by tetramer staining and the genotype of β2-AR−/− mice was confirmed by real-time PCR (TransnetXY). All experiments were performed in accordance with the Federation of European Laboratory Animal Science Associations. All procedures were approved by the Paris-Descartes Ethical Committee for Animal Experimentation (approval 03506.02).
L-(−)Epinephrine-(+)-bitartrate (adrenaline) and L-(−)norepinephrine-(+)-bitartrate (noradrenaline) were purchased from Calbiochem, prostaglandin E2 (PGE2) from Tocris Bioscience, and (±)-propranolol hydrochloride from Sigma-Aldrich.
Tumor model and in vivo treatments
TC1 (tissue culture 1) tumor cells expressing HPV16-E6 and HPV16-E7 proteins were kindly provided by Dr. Eric Tartour (Hôpital Européen Georges Pompidou, Paris, France) in 2011 and have not been genetically authenticated by our laboratory. Cell lines were not tested for Mycoplasma. TC1 tumor cells were cultured in RPMI-1640 GlutaMax (Gibco) supplemented with 10% fetal calf serum (FCS; GE Healthcare), penicillin 50 U/mL, 50 μg/mL streptomycin and 1 mmol/L sodium pyruvate (all from Gibco). Once thawed, cells were maintained in culture for 4 days and passed twice prior to use.
Eight-week-old C57BL/6J mice were subcutaneously (s.c.) injected on the right flank with 1 × 105 E7 expressing-TC1 tumor cells in 100 μL PBS. When tumors reached a volume of 100 mm3 (∼10 days after implantation), mice were primed (d0) with a peritumoral injection of 20 μg of STxBE7 vaccine (29), composed of the nontoxic β-subunit of Shiga toxin coupled to HPV16-derived–E7 peptide, and 6 × 105 U of IFNα in a total volume of 200 μL PBS (29, 30). The STxBE7 vaccine was provided by Dr. Eric Tartour (Hôpital Européen Georges Pompidou, Paris, France) and Dr. Ludger Johannes (Institut Curie, Paris, France). The nonvaccinated mice were injected with PBS as control. The next day, vaccinated mice received an additional peritumoral injection of 6 × 105 U of IFNα in a total volume of 100 μL PBS. IFNα was a kind gift from Dr. Agnès Le Bon (Institut Cochin, Paris, France; ref. 31). Mice were treated daily with propranolol (drinking water, 0.5 mg/mL) commencing 1 day prior to vaccination (d-1). Drinking water was changed two times per week. For some experiments, propranolol treatment was given to nonvaccinated mice 9 days after tumor implantation and to vaccinated mice 4 days after the vaccine prime.
For in vivo CD8+ T-cell depletion, 500 μg of anti-CD8a (clone 53–6.7, Bio X Cell, #BE0004-1) or rat IgG2a isotype control (clone 2A3, Bio X Cell, #BE0089) was injected intraperitoneally (i.p.) at days −2 and −1 prior to vaccination followed by injection of 200 μg every 3 days (all diluted in 100 μL PBS). Depletion was confirmed using flow cytometry.
Tumors were measured by calipers every 3 days, and the tumor volumes were estimated by tumor volume = width × thickness × length/2. For flow cytometry, transcriptomic analysis and immunofluorescence, tumors were harvested 10 days after priming as described below.
Tissue single-cell suspensions
Fresh TC1 tumors were cut into 2- to 3-mm pieces and were digested for 45 minutes at 37°C with DNase I (100 μg/mL, Roche) and collagenase (1 mg/mL, Roche) in RPMI-1640 GlutaMax under rotating agitation. Lymph nodes (LN) and spleens were processed into single-cell suspensions by mechanical disruption (crushed by the plunger of a 5-mL syringe) directly through a 40-μm nylon cell strainer in PBS with 10% FCS. Blood samples were collected into heparinized Eppendorf tubes by retro-orbital plexus puncture. Red blood cells were removed using an RBC lysis buffer (eBioscience) for 2 to 3 minutes at room temperature (RT). Cell suspensions were passed through a 40-μm nylon cell strainer (Corning), washed, and resuspended in PBS 2% FCS 1 mmol/L EDTA (Gibco).
T-cell isolation and culture
T cells from LNs of wild-type mice were purified using an EasySep T-cell negative selection kit (STEMCELL Technologies, #19851) as per the manufacturer's instructions. Purity after isolation kit was >95%.
CD8+ T cells from wild-type mice and β2-AR−/− mice and OT-I cells from OT-IxSJL mice were purified from spleen and LNs using a negative isolation kit (Dynabeads Untouched Mouse CD8 cells Kit, Invitrogen, #114717D) as per the manufacturer's instructions. Purity after isolation kit was >95%. The CD8+ T cell–enriched fraction was labeled with anti-CD8a, anti-CD44, and anti-CD62L, and naïve CD8+ T cells (CD8+CD44negCD62L+) were sorted on the ARIA III (BD Biosciences). Purity was routinely tested after cell sorting and was >99%.
Activated CD8+ T cells were obtained from CD8+ T cells activated for 7 days with plate-bound anti-CD3ϵ (2 μg/mL, clone 145-2C11, BD Biosciences, #01081D) in the presence of soluble anti-CD28 (2 μg/mL, clone 37.51, BD Biosciences, #553294) in complete RPMI medium (RPMI-1640 GlutaMax including 10% FCS, penicillin 50 U/mL, 50 μg/mL streptomycin, 1 mmol/L sodium pyruvate, hepes 0.01 mol/L, 1× nonessential amino acids (all from Gibco), and 0.05 mmol/L β-mercaptoethanol (Sigma-Aldrich). Nunc MaxiSorp 96-well flat bottom plates (Invitrogen, # 44-2404-21) were bounded with anti-CD3ϵ overnight at 4°C and washed twice in PBS prior to use. At days 3 and 5 of culture, cells were 1:2 split in complete RPMI medium supplemented with IL2 (20 U/mL).
For CD8+ TIL isolation, cell suspensions from TC1 tumors from STxBE7-vaccinated mice (10 days after vaccination) were resuspended in 40% Percoll (GE Healthcare) layered on top of 80% Percoll, and centrifuged at 325 × g for 25 minutes without the brake. Cells at the Percoll interface were collected, labeled with anti-CD45, anti-CD4, anti-CD8a, anti-CD11b, and CD8+ TILs (CD45+CD11bnegCD4negCD8+) were sorted on the ARIA III. Purity was routinely tested after cell sorting and was >99%. CD8+ TILs were either used immediately (RNA and protein extractions) or allowed to recover for 24 hours in complete RPMI medium supplemented with IL2 (20 U/mL).
After washing in PBS, cell suspensions (2–3 × 106) were stained in 96-well round-bottom plates with the Live/Dead fixable dead cell stain kit (Invitrogen) for 20 minutes on ice, to gate out dead cells, according to the manufacturer's indications. Cells were then washed in staining buffer (PBS 2% FCS 1 mmol/L EDTA), and Fc receptors were blocked with mouse Fc block (5 μg/mL, anti-CD16/CD32, clone 2.4G2, BD Biosciences, #553142) diluted in staining buffer for 20 minutes on ice. After washing in staining buffer, cells were labeled for surface markers with fluorescently conjugated mouse antibodies diluted in staining buffer for 20 minutes on ice. Antibodies used are described in Supplementary Table S1. For detection of E7-specific CD8+ T cells or OVA-specific OT-I cells, cells were respectively stained for 30 minutes on ice with a Db-E7 dextramer (H-2Db—RAHYNIVTF) or a Kb-S8L dextramer (H-2Kb—SIINFEKL), both purchased from Immudex. After staining, cells were washed in staining buffer, fixed in 2% paraformaldehyde (Electron Microscopy Sciences) for 15 minutes at RT, and resuspended in staining buffer.
For intracellular staining, cells were fixed, permeabilized, and stained using Foxp3/Transcription staining buffer set (eBioscience) or Cytofix/Cytoperm solution (BD Biosciences) as per the manufacturer's instructions. Cells were labeled overnight at 4°C with the fluorescently conjugated mouse intracellular antibodies described in Supplementary Table S1. For intracellular IFNγ staining, cells were stimulated in vitro for 4 hours at 37°C with either 5 μg/mL E7-peptide (kindly provided by Dr. Eric Tartour) or Dynabeads Mouse T-activator CD3/CD28 (1 bead for 2 CD8+ T cells, Thermo Fisher) or 0.5 μg/mL PMA + 2.5 μmol/L ionomycin (Calbiochem) in the presence of GolgiPlug (1/1000, BD Biosciences).
Cells were acquired on Fortessa flow cytometer (BD Biosciences) and analyzed using FlowJo v10 software (FlowJo LLC).
Calcium influx measurements
T cells, CD8+ TILs, and in vitro–activated CD8+ T cells (2 × 106/mL) were loaded for 30 minutes at 37°C in complete RPMI medium with the membrane-permeable fluorescent Ca2+ indicator dye Indo-1 AM (1 μmol/L, Invitrogen, #I1223), washed and labeled with anti-CD4, anti-CD8a, anti-CD62L and anti-CD44 for 20 minutes on ice. After washing in complete RPMI medium, T cells (5 × 106), CD8+ TILs (2 × 105) and in vitro–activated CD8+ T cells (2 × 106) were rested for 5 minutes and then pretreated with either adrenaline, noradrenaline, or PGE2 at the indicated dose for 10 minutes at 37°C. In some experiments, purified T cells were pretreated with 10 μmol/L propranolol for 30 minutes at 37°C prior to 10 μmol/L adrenaline or noradrenaline treatment. Changes in intracellular Ca2+ (iCa2+) concentration over the time (∼0 to 6–8 minutes) were monitored with the Indo-1 iCa2+-free (violet) and Indo-1 iCa2+-bound (blue) channels on the BD LSR II flow cytometer at 37°C. The calcium baseline level was measured for 30 seconds before stimulation with hamster anti-CD3ϵ (2.5 μg/mL, 145-2C11, BD Biosciences) and then for 1 minute before cross-linking with secondary goat anti-hamster IgG (H + L; 20 μg/mL, Invitrogen, #31115). Peaks of iCa2+ mobilization in naïve CD8+ T cells (CD4negCD8+CD44negCD62L+), CD8+ TILs (CD4negCD8+CD44neg), and in vitro–activated CD8+ T cells (CD4negCD8+CD44neg) were quantified as the ratio of Indo-1 iCa2+-free/Indo-1 iCa2+-bound and analyzed using FlowJo v10 software (FlowJo LLC).
Proliferation and survival assay
Naïve CD8+ T cells (4 × 106/mL) were washed twice in serum-free RPMI, labeled with 5 μmol/L Celltrace Violet (CTV, Invitrogen, #C34557) for 20 minutes at 37°C according to the manufacturer's instructions. Cells (5 × 104) were then stimulated with plate-bound anti-CD3ϵ (2 μg/mL, clone 145-2C11, BD Biosciences, #01081D) in the presence of soluble anti-CD28 (2 μg/mL, clone 37.51, BD Biosciences, #553294) and cultured in complete RPMI medium in the presence of adrenaline or noradrenaline at the indicated dose. Similar doses of adrenaline or noradrenaline were added in the cell culture at 24 hours, and proliferation was assessed by flow cytometry after 3 days. In some experiments, IL2 (20 U/mL) was added to the culture. The average number of CD8+ T-cell divisions was calculated as follows. First, we estimated the CTV dilution factor (f) due to stimulation: f = CTV mean fluorescence intensity (MFI) in the absence of stimulation divided by CTV MFI in the presence of stimulation. Then, as the intracellular amount of CTV is halved during each cell cycle, the average number of cell divisions (A) was calculated with the following formula: A = LOG2(f) (32). Proliferation percentage was calculated as (the average number of CD8+ T-cell divisions of treated CD8+ T cells)/(the average number of CD8+ T-cell divisions of untreated CD8+ T cells) × 100.
Sorted CD8+ TILs or activated CD8+ T cells (5 × 104) were cultured in complete RPMI medium with or without IL2 (20 U/mL) in the presence of 10 μmol/L adrenaline or noradrenaline for 24 hours. Cell populations were enumerated using the Accuri C6 flow cytometer (BD Biosciences) at days 0 and 1. Proliferation rates were quantified as the ratio of absolute cell number at day 1/absolute cell number at day 0.
To assess effects of adrenaline and noradrenaline on CD8+ T-cell viability, naïve CD8+ T cells (5 × 104) were cultured with IL7 (10 ng/mL, PeproTech) and CD8+ TILs (5 × 104) were cultured with or without IL2 (20 U/mL). Survival was evaluated with Live/Dead staining at 72 hours for naïve CD8+ T cells and at 24 hours for CD8+ TILs on Fortessa flow cytometer (BD Biosciences).
Generation of bone marrow–derived DCs
Bone marrow cells were collected from femurs of wild-type C57BL/6J mice. The bone marrow was flushed out with a syringe (25-G needle) into a sterile Petri dish with serum-free RPMI. Then, the cells were centrifuged, depleted of red blood cells using RBC lysis buffer (eBioscience), and cultured at 1 × 107 cells per well in tissue culture–treated 6-well plates in 4 mL of complete RPMI medium with 20 ng/mL GM-CSF (Bio-Legend). Half of the medium was removed at day 2 and warmed (37°C) complete RPMI medium with 40 ng/mL GM-CSF was added. The culture medium was entirely discarded at day 4 and replaced by warmed complete RPMI medium with 20 ng/mL GM-CSF. On day 6, nonadherent cells in the culture supernatant and loosely adherent cells were harvested by gently washing with PBS and were then pooled and used for experiments.
Bone marrow–derived DC maturation and coculture with T cells
Immature bone marrow–derived DCs (BMDC) were stimulated either with 2.5 μg/mL of the STxBOVA vaccine plus 105 U/mL of IFNα or with 125 μg/mL OVA protein (InvivoGen) plus 10 μg/mL Quil-A (InvivoGen) in the presence of 10 μmol/L adrenaline or noradrenaline for 18 hours. The STxBOVA vaccine was provided by Dr. Eric Tartour (Hôpital Européen Georges Pompidou, Paris, France) and Dr. Ludger Johannes (Institut Curie, Paris, France). For DC maturation analysis, BMDCs were collected, and Fc receptors were blocked with mouse Fc block diluted in staining buffer for 20 minutes on ice. After washing, cells were then stained with anti-CD11b, anti-CD11c, anti-MHC class II, anti-CD80, and anti-CD86 diluted in staining buffer for 20 minutes on ice. For DC/OT-I cocultures, 1 × 104 stimulated BMDCs were washed and cultured in complete RMPI medium with 5 × 104 purified OT-I cells labeled with 5 μmol/L CTV. OT-I cell proliferation was assessed by flow cytometry after 3 days as described above.
Immunization model with OVA protein
Eight-week-old C57BL/6J mice received adoptive transfer of 1 × 106 CTV-labeled naïve OT-I cells in 100 μL PBS (d-1), and 1 day later (d0) were s.c. immunized with a mixture of 50 μg OVA protein (InvivoGen) and 20 μg Quil-A adjuvant (InvivoGen) diluted in 100 μL PBS. Control mice received 100 μL PBS. Naïve OT-I cells were purified and labeled with CTV as described above. Propranolol treatment (drinking water, 0.5 mg/mL) was initiated 2 hours after adoptive transfer and continued for the duration of the experiment. Proliferation of transferred CTV-labeled OT-Is was evaluated in the LN draining the immunization site at day 2 and frequency of OT-I cells in the blood was measured at day 3 by flow cytometry as described above.
IL2 and IFNγ cytokine production was measured in culture supernatants collected 24 hours after stimulation of naïve CD8+ T cells or CD8+ TILs or activated CD8+ T cells (5 × 104) with plate-bound anti-CD3ϵ (2 μg/mL, clone 145-2C11, BD Biosciences, #01081D) in the presence of soluble anti-CD28 (2 μg/mL, clone 37.51, BD Biosciences, #553294) and of 10 μmol/L AR agonists. The ELISA tests were performed using IL2 or IFNγ ELISA Max kit (Bio-Legend, #431004 or #430804), according to the manufacturer's indications. ELISA plates were read with CLARIOstar (BMG Labtech) plate reader and analyzed using MARS Data Analysis Software (BMG Labtech).
RNA extraction and real-time qRT-PCR
Total RNA from whole tumor (day 10 after vaccination) or purified CD8+ T cells (naïve, activated, or TILs) were extracted using the RNeasy Mini Kit (Qiagen, #74104) according to the manufacturer's instructions and were quantified using Nanodrop ND-1000 Spectrophotometer. Gene expression was analyzed by RT-qPCR with the LightCycler 480 Real-Time PCR system (Roche Life Science). For cDNA synthesis, 0.2 to 1 μg of purified RNA was reverse-transcribed using the Advantage RT-for-PCR kit (Applied Clontech). PCR amplification reactions were carried out in a total volume of 10 μL using LightCycler 480 SYBR Green I Master (Roche Life Science), forward and reverse primers (10 μmol/L each, Eurofins Genomics) and 2 μL cDNA. Primers sequences used are listed in Supplementary Table S2. GAPDH was used as a housekeeping gene to normalize mRNA expression. Relative expression was calculated by the 2−ΔΔCt method to compare expression levels of the same transcript in different samples or 2−ΔCt to compare expression levels of several transcripts in the same sample. All the measures were performed in triplicate and validated when the difference in threshold cycle (Ct) between two measures was <0.3. Raw Ct values were calculated using the LightCycler 480 v1.5.0 SP3 software.
Western blotting analysis
Purified CD8+ T cells (naïve, activated, or TILs) were lysed for 45 minutes on ice in RIPA lysis buffer (Sigma-Aldrich) supplemented with a cocktail of 1× protease inhibitors (Complete Protease Inhibitor Cocktail, Roche, #11697498001) and 1× phosphatases (PhosSTOP, Roche, #4906845001). Lysates were collected after centrifugation at 14,000 rpm for 30 minutes at 4°C. Protein concentration in supernatants was quantified with GeneQuant Pro RNA/DNA Calculator Spectrophotometer using a Bradford protein assay (Bio-Rad, #5000006), according to the manufacturer's recommandations. Lysates were diluted in 5× Laemmli buffer [500 mmol/L Tris (Sigma-Aldrich), 10% SDS (Sigma-Aldrich), 10% glycerol (Sigma-Aldrich), 0.1 M DTT (Merck), 10% bromophenol blue (Sigma-Aldrich)] and then boiled for 5 minutes at 95°C. Next, 15 μg of proteins from each sample was loaded on 12% SDS-PAGE. After protein separation, samples were transferred on a 0.45-μm PVDF membrane (GE Healthcare Life Sciences), followed by blocking in T-TBS [0.1% Tween 20 (Sigma-Aldrich), 10% TBS 10×] containing 5% nonfat dry milk for 1 hour at RT. The membrane was incubated overnight at 4°C with the following primary antibodies diluted in T-TBS 5% nonfat dry milk: anti-β2-AR (0.5 μg/mL, rabbit polyclonal, Invitrogen, #PA5-14118) and anti-β-actin (0.2 μg/mL, clone AC-74, Sigma, #A2228). The membrane was washed with T-TBS and probed with the following relevant secondary antibodies: goat anti-rabbit or anti-mouse IgG (H + L) coupled to HRP (1/5,000, Jackson ImmunoResearch Laboratories, #111-035-144 or #11-035-146) in T-TBS for 1 hour at RT. After washing, the revelation was performed with an ECL kit (GE Healthcare Life Sciences) using a camera (Fusion FX7, Vilbert Lourmat). Protein bands were quantified with ImageJ software.
TDLNs from STxBE7-vaccinated mice were fixed for 2 hours at RT in a periodate–lysine–paraformaldehyde solution [0.05 mol/L phosphate buffer containing 0.1 mol/L L-lysine (pH 7.4; Sigma-Aldrich), 2 mg/mL NaIO4 (Sigma-Aldrich), and 10 mg/mL paraformaldehyde (EM Grade)]. Tissue slices were prepared as previously described (2). Briefly, fixed TDLNs were embedded in 5% low-gelling-temperature agarose (type VII-A; Sigma-Aldrich) prepared in PBS. Ganglionic slices (250 μm) were cut with a vibratome (VT 1000S; Leica) in a bath of ice-cold PBS. Immunostaining was performed by first blocking Fc receptors with mouse Fc block for 1 hour at RT. After washing in PBS 0.3% Triton X-100 (Sigma-Aldrich) 1% BSA (Sigma-Aldrich) 5% donkey serum (Jackson ImmunoResearch Laboratories), slices were stained for sympathetic nerves with tyrosine hydroxylase primary antibody (1/1,000, rabbit polyclonal, Merck Millipore, #AB152) overnight at 4°C. Slices were then washed and stained with PE-conjugated CD3ϵ antibody (1/50, clone 500A2, eBioscience, #12-0033-81) for 30 minutes at RT. After washing, immunodetection was performed using secondary Alexa Fluor 647-conjugated donkey anti-rabbit IgG (H + L; 1/300, polyclonal, Invitrogen, #A-31573) for 30 minutes at RT. All antibodies were diluted in PBS 0.3% Triton X-100 1% BSA 5% donkey serum. Images were obtained with a spinning disk confocal microscope (Leica) equipped with a CoolSnap HQ2 camera and a 25× water immersion objective (25×/0.95 N.A.). All images were acquired with MetaMorph 7 imaging software (Molecular Devices) and analyzed with ImageJ software.
Data were analyzed with GraphPad Prism6 Software. For comparison of two nonparametric data sets, Mann–Whitney U test was used. For multigroup comparisons, one-way ANOVA or two-way ANOVA tests with post hoc testing using Tukey multiple comparison were used. P values <0.05 were considered statistically significant (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001). All data are depicted as mean ± SEM.
Blocking β-AR signaling improved STxBE7 vaccine–induced tumor regression
To elucidate the effects of β-AR signaling on CD8+ T cell–dependent antitumor immune responses, we assessed the effects of the pan-β-AR antagonist propranolol on TC1 tumor–bearing mice vaccinated with the STxBE7 vaccine plus IFNα. C57BL/6J mice were transplanted with E7-expressing TC1 tumor cells and vaccinated when tumor nodules reached 100 mm3 at day 10. Mice were treated with propranolol daily, starting 1 day prior to STxBE7 vaccination (Fig. 1A). We designed a suboptimal vaccination protocol in which a minority of tumors completely regressed (≈30%) and the others either progressed (≈10%), stabilized (≈10%), or partially regressed (≈50%; Fig. 1B–D). Such a variable outcome is appropriate for evaluating if propranolol treatment can modulate the efficiency of the STxBE7 vaccine–induced antitumor response. Propranolol improved the efficiency of STxBE7 vaccine (Fig. 1B–D). Indeed, all the tumors from the propranolol-treated mice regressed, with 74% of the mice being tumor free at the sacrifice 22 days after the vaccination (Fig. 1D). In a separate experiment, mice with complete tumor regression remained tumor free up to 32 days (Supplementary Fig. S1). Treatment of unvaccinated control mice with propranolol had no effect on tumor growth, suggesting that propranolol mainly acts on the antitumor immune response induced by the STxBE7 vaccine (Fig. 1B and C). These results demonstrated that blocking β-AR signaling is able to improve the efficacy of an antitumor response induced by a vaccine-based immunotherapy.
Blocking β-AR signaling increased the frequency of CD8+ TILs without affecting their reactivity
To investigate why we observed better tumor regression after propranolol treatment, immune intratumoral infiltrates were analyzed 10 days after vaccination by multiparametric flow cytometry. The gating strategies of the different lymphoid and myeloid cell populations are shown in Supplementary Fig. S2A and S2B. STxBE7 vaccination led to an increased infiltration of CD45+ immune cells into tumors (Fig. 2A) and all lymphoid and myeloid cell subsets, except conventional DCs (cDC) which decreased (Supplementary Fig. S3A). Specifically, among intratumoral CD45+ immune cells, the frequency of CD8+ T cells, NK cells, monocytes increased whereas the frequency of Tregs and TAMs decreased after STxBE7 vaccination (Supplementary Fig. S3B). In STxBE7-vaccinated mice, propranolol significantly increased the abundance of CD45+ immune cells (Fig. 2A) and CD8+ T cells, including the E7-specific ones (Fig. 2B; Supplementary Fig. S3A and S3B). Moreover, in vivo depletion of CD8+ T cells completely abrogated the STxBE7 vaccine plus propranolol-induced regression, demonstrating that propranolol mainly acts on the CD8+ T cell–dependent antitumor immune response induced by the STxBE7 vaccine (Fig. 2C).
We next wondered whether the activity of CD8+ TILs was improved by β-AR signaling blockade. STxBE7 vaccination led to an increase in the activation markers CD44 and PD-1, granzyme B production, proliferation (Ki-67), and IFNγ production after restimulation of CD8+ TILs (Fig. 2D and E). Propranolol treatment did not further affect the status of CD8+ TILs from STxBE7-vaccinated mice (Fig. 2D and E). These results showed that propranolol treatment improved the efficacy of an antitumor vaccination by increasing the frequency of the CD8+ TILs without affecting their activation status or their ability to produce IFNγ.
Propranolol treatment improved CD8+ T-cell priming in the TDLN
To gain further insight into the potential mechanism underlying the increase in CD8+ TIL frequency induced by propranolol treatment, we first examined the transcripts of various chemokines in the whole tumor. All chemokines tested, including the two major T-cell chemoattractants CXCL9 and CXCL10 (33), were increased after STxBE7 vaccination, without any additive effect of propranolol (Supplementary Fig. S4). This suggested that the effect of propranolol on the enhanced CD8+ T-cell infiltration within the tumor is not chemokine-dependent.
We hypothesized that the increase in CD8+ TILs observed with propranolol may have been due to better antigen priming in the TDLN. Indeed, the frequency of E7-specific CD8+ T cells in the TDLN was increased in propranolol-treated mice (Fig. 3A). Moreover, TDLN exhibited a sympathetic innervation (positive for tyrosine hydroxylase) in close contact with T cells (Fig. 3B), providing anatomic evidence for a possible effect of β-AR signaling on the activation of naïve CD8+ T cells. To directly test this hypothesis, we designed a vaccination protocol in which β-AR blockade by propranolol was absent during the first 4 days following the STxBE7 vaccination, corresponding to the initial priming phase (Fig. 3C). In these conditions, the beneficial antitumor effect of propranolol on the STxBE7 vaccination was completely lost (Fig. 3D and E). Delayed propranolol treatment also failed to increase the frequency of CD45+ cells and CD8+ T cells in tumors (Fig. 3F). Together, these results demonstrated that the main effect of propranolol on the STxBE7 vaccine–induced tumor regression occurs during the priming phase in the TDLN by increasing the generation of tumor antigen–specific CD8+ T cells.
AR signaling inhibited CD8+ T-cell priming following vaccination without affecting DCs
The aforementioned findings suggest that blockade of β-AR signaling improved the priming of the naïve CD8+ T cells in the TDLN. To reinforce this conclusion, we next evaluated the effect of propranolol treatment in an in vivo antigen vaccination model in the absence of tumor (Fig. 4A). CTV-labeled naïve OT-I cells were adoptively transferred into mice treated with propranolol. One day after transfer, mice were immunized with OVA protein plus Quil-A adjuvant. At day 2, propranolol increased the proliferation of adoptively transferred OT-I cells in the LN (Fig. 4B). Moreover, their frequency in blood was also higher in propranolol-treated mice at day 3 (Fig. 4C).
To explain the effects on CD8+ T-cell priming after propranolol treatment in both the vaccine-induced tumor regression model and OVA protein immunization, we first speculated that a suppressive effect of β-AR signaling on DC maturation and antigen presentation could be involved. BMDCs were stimulated for 18 hours with the STxBE7 vaccine fused with OVA (instead of E7, STxBOVA vaccine) plus IFNα or with OVA protein plus Quil-A, in the presence or absence of the natural AR agonists, adrenaline and noradrenaline. Stimulation of DCs with STxBOVA vaccine induced their maturation, as reflected by an increased expression of MHC class II, CD80, and CD86. None of these maturation markers was significantly altered by AR agonists (Fig. 4D). Importantly, when these pulsed BMDCs were cocultured with OVA-specific naïve CD8+ OT-I cells, we did not observe any treatment-induced differences in OT-I proliferation, demonstrating that the antigen presentation capacity of DCs was not affected by AR signaling (Fig. 4E). Similar results were obtained with BMDCs pulsed with OVA protein plus Quil-A adjuvant (Supplementary Fig. S5). Thus, β-AR signaling inhibited the priming of naïve CD8+ T cells in LNs without affecting the maturation and antigenic presentation capacities of DCs.
Naïve CD8+ T cells were susceptible to β2-AR signaling
We next wondered if blockading β-AR signaling could directly act on naïve CD8+ T cells to improve their priming. To test this hypothesis, we evaluated the effect of adrenaline or noradrenaline treatment on naïve CD8+ T cells in vitro. The early calcium signaling response following CD3 stimulation was first analyzed. Both AR agonists inhibited the anti–CD3-induced calcium response of wild-type naïve CD8+ T cells in a dose-dependent manner, whereas they had no effect in naïve CD8+ T cells lacking β2-AR (β2-AR−/−) or treated with propranolol (Fig. 5A; Supplementary Fig. S6A and S6B). Accordingly, naïve CD8+ T cells showed predominant expression of β2-AR (Supplementary Fig. S6C). We next assessed the consequences of AR agonists' effect on T-cell proliferation and cytokine production. The proliferation of naïve CD8+ T cells at day 3 was reduced by AR agonists in a dose-dependent manner whereas naïve β2-AR−/− CD8+ T cells remained insensitive to the same treatment (Fig. 5B; Supplementary Fig. S6D). This inhibition was not due to an effect on naïve CD8+ T-cell viability (Supplementary Fig. S6E). We observed that proliferation could be rescued by adding IL2 in the culture medium (Fig. 5C), suggesting some inhibition of T-cell IL2 production, a hypothesis confirmed by the low level of IL2 found in supernatants 24 hours after stimulation in the presence of AR agonists (Fig. 5D). In addition, IFNγ secretion was also suppressed (Fig. 5E). Nevertheless, neither IL2 nor IFNγ secretion was affected by AR agonists in naïve β2-AR−/− CD8+ T cells (Fig. 5D and E). Together, these results demonstrated that naïve CD8+ T-cell activation can be strongly inhibited by β2-AR signaling, a phenomenon that likely affects their priming in the TDLN.
CD8+ T-cell sensitivity to β-AR signaling was related to activation status and β2-AR expression
Given the effects of β-AR signaling on naïve CD8+ T-cell activation, the lack of any obvious effect of propranolol treatment on CD8+ TIL activation and on their ability to secrete IFNγ was surprising (Fig. 2D and E). We thus investigated how purified E7-specific CD8+ TILs could react to adrenaline and noradrenaline ex vivo. These cells exhibited an activated/effector phenotype (CD44+CD62Lneg; Supplementary Fig. S7A). Contrary to what was observed with naïve CD8+ T cells, the anti–CD3-induced calcium responses of CD8+ TILs were not affected by AR agonists (Fig. 6A). In contrast, PGE2, another cAMP-elevating GPCR ligand previously described to inhibit T-cell functions (34), strongly inhibited CD8+ TIL response to anti-CD3 stimulation, eliminating the involvement of a defect in cAMP signaling (Fig. 6A). In addition, AR agonists had no effect on the proliferation and survival of CD8+ TILs cultured with or without IL2 (Fig. 6B; Supplementary Fig. S7B). This was consistent with the fact that AR agonists did not affect IL2 production by CD8+ TILs restimulated with anti-CD3/CD28 (Fig. 6C). Similarly, the production of IFNγ by CD8+ TILs restimulated with anti-CD3/CD28 was insensitive to AR agonists (Fig. 6D). This was confirmed by IFNγ intracellular staining experiments following restimulation of CD8+ TILs with either anti-CD3/CD28 or PMA/ionomycin in the presence of adrenaline or noradrenaline for 4 hours (Supplementary Fig. S7C). Absence of responsiveness of CD8+ TILs to AR agonists in vitro supported the results obtained in vivo showing the absence of propranolol treatments' effect on the activity of CD8+ TILs in STxBE7-vaccinated mice (Fig. 2D and E).
We next wondered whether the differential sensitivity of naïve CD8+ T cells and CD8+ TILs to AR agonists could be related to their activation status. To test this hypothesis, we generated activated CD8+ T cells in vitro by stimulating them with anti-CD3/CD28 for 3 days followed by a 4-day culture with IL2 and evaluated their sensitivity to adrenaline and noradrenaline using the same read-outs used with CD8+ TILs. Similarly, AR agonists had no effect on their anti–CD3-induced calcium responses (Fig. 6E), proliferation (Supplementary Fig. S7D) or IL2/IFNγ production (Supplementary Fig. S7E and S7F). These results strongly suggest that activation of CD8+ T cells rendered them insensitive to AR signaling, and that the refractory state observed with CD8+ TILs was likely a consequence of their activation status. In order to explain the difference between naïve CD8+ T cells and CD8+ TILs/activated CD8+ T cells to AR signaling, we finally compared their β-AR expression level. As for naïve CD8+ T cells, CD8+ TILs and activated CD8+ T cells predominantly express β2-ARs (Supplementary Fig. S7G and S7H). However, this expression was much lower in CD8+ TILs and in activated CD8+ T cells, as compared with naïve CD8+ T cells, both at mRNA (Fig. 6F) and protein levels (Fig. 6G). Together, these results demonstrated that the differential sensitivity of naïve CD8+ T cells and CD8+ TILs to AR signaling was related to their activation status, most likely due to the downregulation of the β2-AR expression driven by activation.
In this study, we showed that blocking β-AR signaling by propranolol markedly enhanced the antitumor efficacy of an STxBE7-based vaccine. This effect was due to an increase in tumor-infiltrating CD8+ T cells. Propranolol improved the efficacy of STxBE7 vaccination by directly enhancing the priming of naïve CD8+ T cells against tumor antigen in the TDLN. We did not observe any influence of propranolol on the reactivity of already primed CD8+ TILs, a result reinforced by data showing that CD8+ TILs were insensitive to AR signaling, contrary to naïve CD8+ T cells. Finally, the lack of response of CD8+ TILs to β-AR signaling was likely related to their activated phenotype, characterized by a downregulation of β2-AR expression.
Over the last decade, several studies have revealed that β-AR signaling can facilitate tumor growth and cancer progression (12, 13, 20, 35). Most of these studies were performed in immunodeficient mouse models with the goal of identifying a direct action of β-AR signaling on tumor cells or on protumorigenic processes such as angiogenesis. However, it is now clear that the autonomic nervous system may have an important role in regulating immune responses (36). In our study, mice were not submitted to a specific chronic psychological stress stimulus (e.g., restraint stress, social defeat stress, social isolation), but mouse housing conditions, such as subthermoneutral housing (20°C–26°C), have been described as a physiologic model of mild stress generating chronic AR signaling (37). We also chose to treat mice when their tumor reached a palpable size, reflecting a therapeutic approach in which patients are treated once their cancer is diagnosed. Finally, contrary to tumor models (B16, 4T1, and CT26) in which propranolol has already an effect on tumor growth without therapeutic interventions (21, 25), it had no effect on TC1 tumor growth in the absence of the STxBE7 vaccine, thus making it advantageous for analyzing the effects of propranolol blockade on a vaccine-induced antitumor immune response.
The main effect of β-AR signaling blockade by propranolol occurred at the CD8+ T-cell level. Our findings were consistent with studies showing that adrenergic signals impair the antitumor CD8+ T-cell response, a phenomenon that can be reversed by β-blocker treatment (25–27, 38). Interestingly, the inhibition of the CD8+ T-cell response by adrenergic signals has also been shown in the context of infection, such as influenza (39). Here, β-AR signaling blockade by propranolol increased the frequency of specific CD8+ T cells in a model of in vivo antigen vaccination in the absence of tumor. Thus, propranolol could be used as an adjuvant in the treatment of pathologies (e.g., cancer, infectious disease) in which the improvement of the CD8+ T response is required.
Our findings do not imply that β-AR signaling cannot affect non-T cell immune cell populations, in particular myeloid cells. β-AR signaling promotes tumor growth and metastasis formation through the M2 polarization of macrophages (19, 20) and the tumoral infiltration of MDSCs (21, 28). However, we did not observe differences in the frequency of myeloid cells, in accordance with the unaltered chemokine production after propranolol treatment. We also did not observe any apparent effect of propranolol on the activation status of myeloid cells whose MHC II expression was unchanged in propranolol-treated mice. Nevertheless, further experimentation that unravels the effects of β-AR signaling on the suppressive functions of MDSCs and TAMs is needed.
Our study provides insight into the beneficial effect of β-AR blockade by showing that it can potentiate the priming of naïve CD8+ T cells in the TDLN. We found an inhibitory effect of β2-AR signaling on the proliferation of naïve CD8+ T cells, suggesting a direct inhibitory effect on CD8+ T cells during their priming phase in the TDLN. This result could explain the higher frequency of specific CD8+ T cells in the TDLN and tumor of STxBE7-vaccinated mice treated with propranolol. However, we do not exclude the possibility of additional mechanisms as adrenergic signals reduce the antigen presentation, maturation, cytokine production and migration of DCs (40–44), which can affect the T-cell priming. Social disruption induced-chronic stress suppresses antitumor CD8+ T-cell responses following microsphere (PLGA-MS)-based cancer vaccine by impairing DC maturation, migration, and subsequent CD8+ T-cell priming (27). However, in our model, AR agonists did not affect the maturation and the antigen presentation capacities of vaccine-stimulated DCs. This is in accordance with studies showing that DC maturation was not affected by β2-AR signaling (44) and that inhibitory effect of β2-AR signaling on CD8+ T-cell response induced by immune modulating 4-1BB and CD28 antibodies occurred independently of β2-AR expression by DCs (26). Another explanation for the propranolol-induced increase in CD8+ TILs is that propranolol may favor lymphocyte egress from LNs, which is controlled by β2-ARs (45). However, we observed a higher frequency of CD8+ T cells in the TDLN, thereby suggesting that a decisive part of the potentiating effect of propranolol operates through the priming phase of T-cell activation, after TCR engagement.
Our study demonstrated the differential sensitivity of naïve CD8+ T cells versus CD8+ TILs or activated CD8+ T cells to AR agonists. Contrary to naïve CD8+ T cells, CD8+ TILs and activated CD8+ T cells were insensitive to adrenaline/noradrenaline treatment, which was associated with a downregulation of β2-AR expression. Our data corroborate with previous studies reporting differential β2-AR expression in T-cell subsets. Naïve murine CD4+ T cells express more β2-ARs than effector/memory CD4+ T cells (45). In contrast, human memory CD8+ T cells express more β2-ARs than naïve CD8+ T cells (46, 47). Tregs also have lower expression of β2-ARs than naïve CD4+ T cells (48). CD4+ T-cell polarization toward a Th2 profile is accompanied by a loss of β2-AR expression (49, 50), as a result of epigenetic regulation (51). Determining if epigenetic regulation occurs during the activation of naïve CD8+ T cells, leading to the downregulation of β2-ARs in CD8+ TILs and activated CD8+ T cells is a crucial future direction.
To our knowledge, we provide here the demonstration that a CD3-induced calcium response is inhibited by adrenaline or noradrenaline treatment in naïve CD8+ T cells. These data suggest that this inhibition could potentially take place at a very early stage of TCR signaling. Engagement of the β2-AR activates adenylate cyclase, catalyzing the formation of cAMP, a well-known inhibitor of T-cell activation (52). cAMP triggers PKA-mediated activation of Csk, a major regulator of Src kinase activity, for a potential mechanism of cAMP-dependent inhibition of lymphocyte activation (53). However, other PKA-independent signaling pathways, such as the guanine nucleotide exchange protein activated by adenylyl cyclase (EPAC) and β-arrestin, can also be triggered by β2-AR engagement and modulate immune cell functions (17). For instance, downregulation of IL12p70 in DCs mediated by β2-AR activation is related to the inhibition of NF-kB activation and AP-1 signaling, in an PKA-independent manner involving β-arrestin (44). Thus, the exact contribution of PKA-dependent and -independent signaling pathways in the T-cell inhibition induced by β2-AR engagement remains to be investigated.
Immunotherapies have emerged as a great promise in cancer treatment. However, many tumor types remain unresponsive, and the mechanisms hampering cancer immunotherapies remain unclear. Identifying new strategies to improve the efficacy of cancer immunotherapies is thus crucial. In this study, we demonstrated that an induced antitumor CD8+ T-cell response could be greatly improved by β-blockers. These findings are in accordance with studies showing that β2-AR signaling decreases the efficacy of several immunotherapeutic approaches, including therapies using immune-checkpoint inhibitors (25–27) and studies in various cancers (e.g., breast, melanoma, lung, prostate) showing that patients taking β-blockers have better outcomes (11). We demonstrate that β-AR blockade may act as an adjuvant for immunotherapy by enhancing the priming of the antitumor CD8+ T cells. These results strengthen the rationale for using β-blockers in cancer patients to potentiate immunotherapy, specifically cancer vaccines that induce robust antitumor CD8+ T-cell responses. Finally, these results support the idea that psychological stress, which is a powerful provider of adrenergic signals, must be taken into account and managed when treating cancer patients.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: C. Daher, V. Feuillet
Development of methodology: C. Daher, L. Vimeux, E. Donnadieu, V. Feuillet
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C. Daher, L. Vimeux, R. Stoeva, E. Wieduwild, V. Feuillet
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C. Daher, L. Vimeux, R. Stoeva, V. Feuillet
Writing, review, and/or revision of the manuscript: C. Daher, L. Vimeux, G. Bismuth, N. Bercovici, A. Trautmann, V. Feuillet
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): C. Daher, L. Vimeux, B. Lucas, V. Feuillet
Study supervision: V. Feuillet
Other (experimental support and scientific discussion): E. Peranzoni
The authors wish to thank the imaging (IMAG'IC), flow cytometry (CYBIO), and genomic (GENOM'IC) facilities of the Cochin Institute for their helpful advice and their colleagues from the “Dynamic of T cell interactions” team for their helpful discussions. The authors are grateful to Sophie Ugolini for providing β2-AR−/− mice, Agnes Le Bon for providing IFNα, and Eric Tartour and Judger Johannes for providing Shiga vaccines. This study was supported by grants from the French Ligue Nationale contre le Cancer (V. Feuillet), from the Plan Cancer (Tumor Heterogeneity and Ecosystem Program, C16076KS), and from Institut National du Cancer (PLBIO-2016 and R16189KK; E. Donnadieu). C. Daher was supported by a PhD fellowship from the French Ministry of National Education, Research, and Technology.
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.