E-cigarette exposure disrupts antitumor immunity and promotes metastasis

Abstract

Electronic cigarettes (e-cigarettes) are thought to pose low risk of cancer because the components of e-cigarette liquid are not carcinogens. We analyzed the effects of the two major components, PG/VG and nicotine, on tumor development in preclinical models. We found that PG/VG promoted tumor cell migration in migration assays and contributed to more aggressive, metastatic, and immunosuppressive tumors in vivo, aggravated by the presence of nicotine. Whole body exposure of mice to PG/VG and nicotine rendered animals more susceptible to developing tumors with high frequencies of infiltrating proinflammatory macrophages expressing IL-6 and TNFα. Moreover, tumor-infiltrating and circulating T cells in e-cigarette exposed mice showed increased levels of immune checkpoints including CTLA4 and PD-1. Treatment with anti-CTLA4 antibody was able to abrogate metastasis with no detrimental effects on its ability to induce tumor regression in exposed mice. These findings suggest that the major components used in e-cigarette fluid can impact tumor development through induced immunosuppression.

Keywords: electronic cigarettes; immune checkpoint blockade; immunosuppression; metastasis; whole body exposure.

Figure 1

E-cigarette chemicals promote tumor cell migration in vitro. (A) Experimental design for cell scratch assays. Briefly, tumor cells were seeded into inserts containing a cell-free gap (scratch area) between two sides and were cultured overnight in the presence of concentrations of PG/VG from 0 to 20 µM with or without nicotine. Then, inserts were removed and invasion of the scratch area was monitored by time-lapse microscopy. (B) Representative microscope images (40X augmentation) taken at 0, 4, 12 and 24h from 1 of 3 experiments are shown. Vertical white lines show distance from cell fronts in μm, used to calculate percentage of covered scratch areas. (C) Percent of scratch area covered at 24h after insert removal fell by 0.58 (95% CI 0.39, 0.77) %/µM PG/VG (p<.001; Table 1 ) but was not affected by the presence of nicotine (p=.851). We also ran the model including the PG/VG x nicotine interaction and found no significant interaction (p=0.638). (D) Viable cell counts after 7 days increased by 0.060 (0.048, 0.071) x10⁶ cells/µM PG/VG (p<.001; Table 1 ) and dropped by -0.56 (-0.72, -0.39) x 10⁶ in the presence of nicotine (p<0.001). We also ran the model including the PG/VG x nicotine interaction and found a significant interaction (P<0.001). The PG/VG effect was about the same 0.076 (0.064, 0.088) (p<.001), whereas the nicotine main effect dropped to -0.22 ± .11 (-0.42, -0.02) (p=.029) with the interaction term being -.039 (-0.057, -0.021) (p<.001).

Figure 2

Preconditioning with PG/VG increases metastasis in different subcutaneous tumor models. (A) Experimental scheme to test the impact of PG/VG and nicotine on the outcome of subcutaneous tumors. Tumor cells were cultured in the presence of 2.5 μM PG/VG with concentrations of nicotine of 0, 6 and 36 mg/ml for 14 days. Then, preconditioned cells were implanted subcutaneously in the right flanks of wildtype C57/B6 mice (5 mice per experiment, n=3 experiments). Tumors were measured twice weekly and animals were inspected for metastases after euthanization. (B) Viable cells implanted per mouse for each tumor model: melanoma B16, colorectal MC38 or prostate TRAMP-C2. (C) Endpoint tumor volumes for B16, MC38 and TRAMP-C2 models. Error bars represent standard error mean (SEM). Neither PG/VG nor nicotine affected tumor volume (B16: PG/VG p=0.876, nicotine p=0.601; MC38: PG/VG p=0.740, nicotine p=0.914; TRAMP-C2, PG/VG p=0.500, nicotine p=0.130; Table 2 ). (D) Bar plots showing metastasis rates for each subcutaneous tumor model. Logistic regression showed increased odds of metastases associated with exposure to PG/VG (14.9; 95% CI 1.9–116.7; p=0.010) and nicotine (2.9; 1.3–6.9, p=0.012 at 36 mg/mL), controlling for tumor type. Representative pictures with peritoneal metastases are shown for mice exposed to PG/VG with or without nicotine for each tumor model. Control animals (black bars) were not exposed to PG/VG or nicotine. Yellow arrows indicate highly metastasized areas. Metastasis was extremely rare in mice not exposed to PG/VG or nicotine (0/15 in B16 and MC38, 1/15 in TRAMP-C2).

Figure 3

Accelerated tumor progression upon PG/VG preconditioning. (A) Experimental scheme to test the impact of PG/VG and nicotine in a model of disseminated cancer. Luciferase-expressing colorectal cancer MC38 (Luc-MC38) cells were cultured in the presence of PG/VG with concentrations of nicotine of 0, 6 and 36 mg/mL for 14 days. Then, 5 x 10⁵ preconditioned Luc-MC38 cells were injected intravenously in the tail veins of wildtype C57/B6 mice (n=5 per experiment, n=3 experiments). Tumor bioluminescence was monitored twice weekly for 15 days. (B) Representative whole body bioluminescence images on days 4, 8, 11 and 15 post-implantation is shown. (C) Time-course measurement of tumor bioluminescence expressed as Average Radiance for mice involved in experiments described in (A). Error bars represent SEM. (D) Kaplan-Meier survival curves for the 3 experiments. There were significant differences between all four curves (p=0.038, by log rank test). There was no significant difference between the PG/VG, PG/VG plus low nicotine, and PG/VG plus high nicotine curves (p=0.439).

Figure 4

Whole body exposure of mice e-cigarette aerosol leads accelerated tumor growth and more aggressive metastasis. (A) Experimental scheme for whole body exposure experiments. Mice were exposed to e-cigarette aerosol with 2.5 µM PG/VG and 0 or 36 mg/ml nicotine for 1h daily for 4 weeks. Then, mice (n=5 per experiment, n=3 experiments) were challenged intravenously with 5 x 10⁵ Luc-MC38 cells. Tumor bioluminescence was monitored twice weekly. Lungs from three mice per experimental condition were harvested on day 22 post-implantation, weighed, inspected for metastasis burden and immunophenotyped by flow cytometry. (B) Representative whole body bioluminescence images on days 6, 13 and 20 post-implantation is shown. (C) Tumor bioluminescence on day 13 after implantation expressed as average photon radiance was significantly increased by PG/VG (p=0.002; Table 1 ) but not nicotine (p=0.103). (D) Kaplan-Meier survival curves from whole body experiments described in (A). There were significant differences between all three curves (p,0.001, by log rank test). There was no significant difference between the PG/VG and PG/VG plus nicotine curves (p=0.364), suggesting that the survival was reduced by exposure to PG/VG with no additional effect of nicotine effect. (E) Representative images from surgically extracted lungs 22 days after tumor cell injection. Yellow circles indicate areas with metastatic MC38 nodes. (F) Metastatic node count from day 22 lungs increased significantly with exposure to PG/VG (p<0.001; Table 1 ), with a further increase when nicotine was added (p<0.001). (G) Lung weight on day 22 after tumor injection was not affected by PG/VG (p=0.637; Table 1 ), but increased with the addition of nicotine (p=0.013).

Figure 5

Whole body exposure of mice e-cigarette aerosol leads to increased myeloid and lymphoid immunosuppression in the tumor microenvironment. (A) Left, representative flow cytometry pseudocolor dot plot showing gating on lung-infiltrating macrophages from animals exposed to PG/VG with or without 36 mg/ml nicotine as described in Figure 4A . Right, lung-infiltrating macrophage frequency within gated live CD45+ immune cells was not affected by PG/VG (p=0.392; Table 1 ) but increased with the addition of nicotine (p=0.001). (B) Left, representative flow cytometry pseudocolor dot plot showing gating on lung-infiltrating CD8+ T cells. Right, lung-infiltrating CD8+ T cell frequency within gated live CD45+ immune cells fell significantly with exposure to PG/VG (p<0.001; Table 1 ) and about one-third more with the addition of nicotine (p<0.001). (C) Representative flow cytometry density plots showing IL-6 expression in gated lung-infiltrating macrophages. (D) Representative flow cytometry density plots showing TNFα expression in gated lung-infiltrating macrophages. (E) Mean fluorescence intensity of IL-6 in gated lung macrophages was not affected significantly by PG/VG (p=0.075; Table 1 ) but increased significantly with the addition of nicotine (p<0.001). (F) Mean fluorescence intensity of TNFα in gated lung macrophages was not affected significantly by PG/VG (p=0.512; Table 1 ) but increased significantly with the addition of nicotine (p<0.001). (G) Representative flow cytometry density plots showing PD-1 expression in gated lung-infiltrating CD8+ T cells. (H) Representative flow cytometry density plots showing TNFα expression in gated lung- infiltrating CD8+ T cells. (I) Percentage PD-1+ cells in gated lung CD8+ T cells was significantly higher in the presence of PG/VG (p<0.001; Table 1 ) but not affected by addition of nicotine (p=0.335). (J) Mean fluorescence intensity of TNFα in gated lung CD8+ T cells was not significantly different in the presence of PG/VG (p=0.586; Table 1 ) but significantly increased with the addition of nicotine (p<0.001).

Figure 6

Tumors exposed to proinflammatory e-cigarette components are responsive to immune checkpoint blockade. (A) Mean fluorescence intensity of T cell immune checkpoints CTLA4, PD-1, TIM3 and LAG-3 on splenic CD8+ T cells from mice exposed to PG/VG with or without 36 ml/mg nicotine as described in Figure 4A . CTLA4 and TIM3 were significantly elevated only in the presence of nicotine (p=0.011, p=0.015 respectively). PD-1 was significantly elevated in the presence of PG/VG (p=0.024) with no nicotine effect (p=0.431). No significant differences were found for LAG-3 expression. Summarized results can be found in Table 1 . (B) The frequency of immune checkpoint co-expression (none (0), any given one (1), any combination of 2 (2), any combination of 3 (3) or all markers (4)) is shown in splenic CD8+ T cells as assessed by flow cytometry Boolean gating. (C) Percentage of quadruple-positive CTLA4+TIM3+LAG-3+PD-1+ splenic CD8+ T cells. Significantly elevated quadruple-positive cells were found in the presence PG/VG (p<0.001, Table 1 ) with a significant additive nicotine effect (p<0.001). (D) Experimental scheme to test the effect of immune checkpoint inhibition in e-cigarette preconditioned subcutaneous tumors. After preconditioning MC38 cells in the presence of 2.5 μM PG/VG with or without 36 mg/ml nicotine for 14 days, mice (n=5 per experiment) were implanted with tumors subcutaneously. Starting 3 days after implantation., mice were treated with intraperitoneal anti-CTLA4 (aCTLA4) or the relevant isotype control (IgG2k/a) on days 3, 6 and 9 after implantation. (E) Endpoint tumor volumes. Error bars represent standard error mean (SEM). Treatment with anti-CTLA4 showed a significant effect on tumor volume (p<0.001; Table 2 ) but not observed for PG/VG+nicotine (p=0.572). No significant interaction between anti-CTLA4 treatment and presence of PG/VG+nicotine was observed (p=0.934). (F) Metastasis rates: 6/15 (40%) animals with PG/VG-preconditioned MC38 tumors showed peritoneal metastases, which were absent (0/15) in both anti-CTLA4-treated groups and nearly absent (1/15) in control mice. There was no significant difference in metastases among these three groups (p=0.762 by Fisher Exact Test (p=0.762) and a significant increase in the PG/VG+nicotine mice compared to the others (p=0.005 by Fisher Exact Test).