Tobacco smoke exposure is a driver of altered oxidative stress response and immunity in head and neck cancer

Abstract

Background: Exposomes are critical drivers of carcinogenesis. However, how they modulate tumor behavior remains unclear. Extensive clinical data show cigarette smoke to be a key exposome that promotes aggressive tumors, higher rates of metastasis, reduced response to chemoradiotherapy, and suppressed anti-tumor immunity. We sought to determine whether smoke itself can modulate aggressive tumor behavior in head and neck squamous cell carcinoma (HNSCC) through reprogramming of the cellular reductive state.

Methods: Using established human and murine HNSCC cell lines and syngeneic mouse models, we utilized conventional western blotting, steady state and flux metabolomics, RNA sequencing, quantitative proteomics and flow cytometry to analyze the impact of smoke exposure on HNSCC tumor biology and anti-tumor immunity.

Results: Cigarette smoke persistently activated Nrf2 target genes essential for maintenance of the cellular reductive state and survival under conditions of increased oxidative stress in HNSCC regardless of human papillomavirus (HPV) association. In contrast to e-cigarette vapor, conventional cigarette smoke mobilizes cellular metabolism toward oxidative stress adaptation, resulting in development of cross-resistance to cisplatin. In parallel, smoke exposure modulates expression of PDL1 and the secretory phenotype of HNSCC cells resulting in an altered tumor immune microenvironment (TIME) in syngeneic mouse models and downregulated expression of antigen presentation and costimulatory genes in myeloid cells.

Conclusion: The cigarette smoke exposome is a potent activator of the Nrf2 pathway and appears to be the primary trigger for a tripartite phenotype of aggressive HNSCC consisting of: (1) reduced chemotherapy sensitivity, (2) enhanced metastatic potential and (3) suppressed anti-tumor immunity.

Keywords: Glutathione; Keap1; Nrf2; Oxidative stress; PDL1; Tobacco.

Fig. 1

Smoking alters the tumor immune microenvironment and selects for mutations activating the NRF2 pathway. A Two-way hierarchical clustering for the presence of 14 different immune subtypes based on ssGSEA scores identifies a cluster of immunologically cold patient tumors from the OCSCC TCGA cohort that were enriched for patients who smoke or recently quit. The number of patients in the hot and cold clusters according to smoking status are shown in the table directly underneath, along with the numbers expected by chance if there was complete independence and the p value (p < 0.0013) obtained after χ² analysis. B The Nrf2 pathway measured by ssGSEA scores is significantly elevated in OCSCC patients with a smoking history (p < 0.003), with a similar trend observed in LUSC tumors. C Stratification of patient tumors by KEAP1 (red symbols) and NFE2L2(NRF2) (orange symbols) mutations demonstrated they were significantly associated with elevated Nrf2 pathway activation compared to mutated tumors from non-smokers in OCSCC (p = 0.0005) or patients who were wild-type (WT) for both genes (D), regardless of smoking history in LUSC (p < 0.0001). Both smoking and mutation were associated with elevated Nrf2 scores in OCSCC (e.g., p < 0.00001 for both variables) and showed a significant interaction (p = 0.013) with a large effect of smoking on mutation-driven Nrf2 pathway elevation (i.e., Cohen’s d = 1.03). E Pooling of TCGA data from smoking-related cancers (OCSCC, LUSC, LUAD) demonstrates that both NFE2L2(NRF2) and KEAP1 mutations are significantly associated with elevated Nrf2 pathway activation on average compared to WT tumors, but for a minority of mutations (e.g., 10 percentile values indicated with a dashed line) the differences are marginal. F NFE2L2(NRF2) mutations in amino acid motifs involved in Keap1 binding (ETGE and DLG) or their immediate vicinity occurred in 90% of cases when Nrf2 scores were elevated compared to just 30% when Nrf2 activation was marginal (i.e., Nrf2 scores in the lower 10 percentile), which was found to be significant by χ² analysis (p < 0.00001)

Fig. 2

Tobacco exposure activates Nrf2 in both HPV-associated and HPV-independent HNSCC models. A Nrf2 protein levels increased in a dose dependent manner upon acute smoke exposure at the indicated time periods in HPV-associated (upper panel) and HPV-independent (lower panel) HNSCC cell lines. B Nrf2 protein levels increased in UMSCC47 and HN30 cells following chronic exposure to smoke-infused media at the indicated concentrations for 3–6 months. Right-panel lanes in HN30 cells indicate Nrf2 levels following an acute bolus of smoke 4 h before harvesting. C HN30, SCC152 and a murine OCSCC cell line (MOC1) were treated with an additional bolus (6 h) of smoke infused media following chronic exposure at the indicated concentrations. After removing the smoke infused media, cells were allowed to recover for 1 and 2 days prior to measurements of Nrf2 protein levels. β-actin was used as the protein loading control

Fig. 3

Tobacco exposure effects are ROS-mediated. Nrf2 protein levels were measured in nuclear and cytoplasm fractions after acute (6 h) smoke exposure in UDSCC2 cells (A) and in the chronic exposure model of HN30 cells (B). α-tubulin or HSP90 served as loading controls for the cytoplasmic fraction, and histone H3 for served as loading control for the nuclear fraction. C Pretreatment with the ROS scavenger NAC inhibited smoke-induced Nrf2 nuclear translocation following acute smoke exposure of SCC090 and SCC154 cells. D SCC090, SCC152, and SCC154 cells were exposed to 0%, 5%, 10%, and 15% cigarette smoke-infused media in the presence or absence of 3 mM NAC. Relative cell number at the end of a 72 h exposure period was assessed using Hoechst assay. Data are shown as means, normalized to control condition; error bars indicate standard deviation; * denotes p value < 0.05 for the compared values (smoke alone vs smoke + NAC). E Cellular ROS levels were detected by measuring the fluorescence of DCFH-DA (B525-A) in UDSCC2, HN30, HN31 and MOC1 cells after treatment with the indicated concentrations of smoke-infused media for 6 h

Fig. 4

Cross-resistance to smoke and cisplatin is mediated by Nrf2. A Using a 72-h Hoechst assay we compared the cell number following cisplatin (left upper panel) or smoke exposure (right upper panel) in the parental HN30 cell line and its cisplatin-resistant derivative HN30R8. We then compared cell number following smoke exposure (left bottom panel) or cisplatin (right lower panel) in the parental HN30 cell line and its chronically smoke-exposed derivative HN30 3%. B HN30 cells were stably infected with either control shRNA or KEAP1 shRNA resulting in increased Nrf2, Gpx2 and Nqo1 levels when Keap 1 levels were reduced. C HN30 KEAP1 shRNA cells demonstrated increased resistance to both cisplatin and smoke exposure (Hoechst, 72 h). D KEAP1 wild-type (wt) and KEAP1^fl/fl clones A, D, E, and F generated from HNSCC GEMM models were harvested for Nrf-2 and Gpx2 protein measurements. E KEAP1^fl/fl clones E and F demonstrated increased resistance to both cisplatin and smoke compared to the KEAP1 wildtype cells (Hoechst, 72 h). F HN30 cells were stably infected with either human NRF2 overexpression (OE) plasmid or empty vector (EV). Whole cell lysates and cell fractions were analyzed for Nrf2, Keap1, Gpx2, Nqo1 and p65 protein levels. α-Tubulin served as loading control for the cytoplasmic fraction, and histone H3 served as loading control for the nuclear fraction. G HN30 cells with NRF2 OE demonstrated increased resistance to both cisplatin and smoke (Hoechst, 72 h). Hoechst data are shown as means, normalized to control condition; error bars indicate standard deviation; p values are denoted as *p < 0.05, **p < 0.01, and ***p < 0.001

Fig. 5

Smoke exposure activates Nrf2 targets consistently upregulated in cisplatin resistant HNSCC. A NRF2 ssGSEA scores are significantly higher in both smoke-naïve (UMSCC47 and UDSCC2) and chronically exposed cell lines (SCC152 and SCC154) following an 8 h bolus of acute smoke. **** adj p value < 0.0001. B Venn diagram illustrating the number of NRF2 signature genes significantly elevated in 12 clones derived from 3 different cisplatin (CDDP)-resistant HNSCC cell line genetic backgrounds, 4 HPV-associated HNSCC cell lines treated with smoke, and their overlap. A total of 36 common NRF2-target genes were significantly elevated in clones from all 3 different CDDP resistant cell lines (HN30, HN31, and PCI13) and 11 of these genes (listed to the right) overlapped with all 4 HPV-associated HNSCC cell lines exposed to smoke. C Volcano plot of UDSCC2 proteins as a function of fold change and log10 p value at 10% smoke exposure generated using JMP Pro version 17 (SAS Institute Inc., 2023). A cutoff of 1.25 fold change (equivalent to 0.322 log2 fold change) and a false discovery rate (FDR) of 0.05 (equivalent to -log10(p value) of 1.30) were applied. The y-axis represents -log10 p value, while the x-axis displays log2 fold change. Targets associated with metabolism, immunity, and Nrf2 signaling are highlighted. Downregulated targets are indicated in blue, and upregulated targets are shown in red. D Heatmap of proteins significantly elevated (FDR = 0.1) by smoke exposure (10%) from the Molecular Signature Database NRF2 pathway (i.e., NFE2L2.V2), which were identified by mass spectrometry

Fig. 6

Smoke but not e-cigarette vapor shifts metabolism toward an enhanced reductive state. A Metabolomic and transcriptomic data were integrated across 3 distinct cell backgrounds (SCC152, SCC154, UDSCC2) and genes which were consistently upregulated (FDR ≤ 0.05) across all 3 backgrounds, defined within both datasets were identified: Gpx2 (glutathione peroxidase 2), Gsr (glutathione reductase). B SCC152 cells were exposed to smoke (0%, 5%, 15%) for 8 h in the presence of 25 mM ¹³C all labeled glucose (Glc). Following completion of exposure, cell lysates were analyzed for ¹³C incorporation which demonstrated reduced flux toward lactate and TCA with enhanced flux into PPP intermediates. C HN30 cells were exposed to smoke (3%) or e-cigarette vapor (5%) for 8 h. Smoke but not vapor increased Nrf2 total protein levels; β-actin was used as the protein loading control. D In contrast to smoke, vapor generated minimal effects on tumor cell viability (HN31, SCC152) even when adjustments were made for similar nicotine delivery levels (Hoechst, 72 h). Hoechst data are shown as means, normalized to control condition; error bars indicate standard error of the mean; p values are denoted as *p < 0.05. E In parallel (to experiment carried out in panel C, cells were subjected to unbiased steady state metabolomics analysis which demonstrated that smoke, but not vapor shifts tumor metabolism toward an enhanced reductive state. Two-way unsupervised hierarchical clustering of non-exposome metabolites and cell lines illustrating major differences between cells exposed to smoke and minor differences in cells exposed to vapor. Note: Panel E and Figure S3 summarize different aspects of the same experiment

Fig. 7

Nrf2 activation alters PDL1 expression and the secretory phenotype of HNSCC cells. A UDSCC2 and HN30 cells were chronically exposed to cigarette-infused media at specified concentrations for 3–6 months. An additional bolus of 2% smoke was administered for 6 h, followed by a change to fresh media for 24 and 48 h before harvesting for western blot analysis. β-actin was used as the protein loading control. B HN30 cells, chronically exposed to 3% smoke infused media demonstrated higher protein levels of Nrf2 and p65. C These cells secreted higher levels of PGE2 and IL-6 measured via ELISA. D HN30 cells stably overexpressing (OE) NRF2 secreted higher levels of PGE2 (collected over 24 and 48 h) compared to HN30 cells stably overexpressing empty vector (EV). E MOC1 cells stably overexpressing NRF2 secreted higher levels of PGE2 (collected over 24 h) compared to MOC1 cells stably overexpressing empty vector (EV). F KEAP1 fl/fl clone E secreted higher levels of PGE2 (collected over 24 and 48 h) compared to its KEAP1 wild-type (wt) counterpart. ELISA data are shown as means, normalized to control condition; error bars indicate standard deviation; p values are denoted as *p < 0.05, **p < 0.01, and ***p < 0.001

Fig. 8

Smoked exposure modulates anti-tumor immunity. A MOC1 cells were used to generate orthotopic tongue xenografts in immunocompetent mice. MOC1 cells chronically exposed to 4% smoke were used to generate orthotopic tongue xenografts in chronically smoke exposed immunocompetent mice. Tumors were harvested, snap frozen and subjected to RNAseq analysis. HN30 tumor cells exposed to tobacco smoke chronically (4%) followed by an acute bolus (4%) and HN30 cells exposed to control air-infused media were used to generate conditioned media for 24 h. Cell culture media was harvested and added to PBMC derived from three healthy donors which were cultured for an additional 48 h. Cells were then harvested and characterized by flow cytometry. B Myeloid cell gating strategy. C Flow cytometry histograms of antigen presentation marker expression CD1d, CD80, and HLA-DR. D Graphical quantitation of flow histograms shown in C; **p < 0.005, ***p < 0.001 by Student’s paired two-tailed t test. E Mechanistic model of reduced treatment response and immunity in chronic smokers with HNSCC. Clinical studies have definitively linked tobacco smoke exposure to reduced chemo-radiation (CXRT) response and altered immunity in HNSCC. Pre-clinical models of HNSCC have shown that hyper-activation of the Nrf2 pathway can drive resistance to cisplatin (CDDP) and alter the inflammatory profile of HNSCC cells and tumors. The current work proposes that the exposome, specifically smoke exposure activates a self-re-enforcing loop whereby activation of Nrf2 predisposes HNSCC tumors to reduced treatment response and suppressed immunity, generating an aggressive, treatment refractory and highly metastatic phenotype