A single cigarette contains over 7000 hazardous chemicals and 60 carcinogens, including polycyclic hydrocarbon carcinogens such as benzo[a]pyrene (BAP) and nicotine-derived nitrosoaminoketone [13,14]. Polycyclic hydrocarbons are responsible for the formation of bulky DNA adducts and their removal by nucleotide excision repair may result in DNA damage, tumorigenesis, and enhanced mutation burden [15,16,17]. This creates a distinct mutational signature in smokers (smoking signature, SS) characteristic of C>A transversions [18]. Tobacco smoking is also associated with the upregulation of PD-L1, which impairs the inflammatory response, allowing tumor cells to evade the immune system [19,20,21] (Figure 1).

In healthy individuals, the PD-1/PD-L1 axis dampens the immune system to protect the host against autoimmunity. However, tumor cells use this same mechanism to evade the immune system [19]. Tobacco smoking is associated with high tumor PD-L1 expression and has a dose-dependent effect [19,21]. While the exact mechanism responsible for this phenomenon remains unclear, it has been hypothesized that tobacco carcinogens such as BAP may be responsible for upregulating PD-L1 expression [19]. BAP activation is mediated by aryl hydrocarbon receptors (AhR), that detoxify xenobiotics and regulate immune cell function [22]. Wang et al. observed that silencing AhR in mice decreased PD-L1 expression and prolonged survival [19]. Furthermore, treatment with anti-PD-1 agents suppressed tumor formation and significantly enhanced T-cell infiltration. In lung cancer patient samples, they observed higher AhR and PD-L1 expression in smokers compared to never smokers. Patients with high AhR-expressing tumors had greater clinical benefit with pembrolizumab, including in multivariable analysis [19].

Xiao et al. have proposed that mTOR signaling is responsible for stimulating PD-L1 expression through IL-6 [20]. When inhibiting mTOR using rapamycin, a significant reduction in PD-L1 expression was observed in experimental models. The addition of IL-6 recombinant protein to tobacco-cultured cell lines increased PD-L1 expression and partially restored tobacco-stimulated PD-L1 expression in rapamycin-treated cell lines. Together, these findings suggest that tobacco carcinogens play a role in increasing PD-L1 expression, and that these mechanisms may improve response to anti-PD-1 agents.

Chemicals found within tobacco cigarettes can impede the development, effector function and cytokine production of the immune system and alter the tumor immune microenvironment [23,24,25]. Through increasing the production of pro-inflammatory cytokines and reducing anti-inflammatory cytokine production, smoking induces an inflamed tumor microenvironment that promotes tumorigenesis and leads to chronic stimulation of T-cells with subsequent exhaustion and activation-induced cell death [23,26,27,28].

CD8⁺ tumor-infiltrating lymphocytes (TILs) are crucial for eliciting a direct cell-mediated antitumor immune response [29]. A combination of high CD8⁺ and low regulatory T-cell (T_reg) infiltration is associated with prolonged survival in cancer patients [30,31]. High densities of tumor stromal infiltrating CD4⁺ and CD8⁺ T-cells have been significantly associated with smoking status in NSCLC [32] and other lung diseases [33]. However, other studies suggest no association between TIL density and smoking status [34,35]. Hiraoka et al. observed that simultaneously high densities of CD4⁺ and CD8⁺ T-cells have a synergistic effect and are associated with a favorable prognosis in lung cancer [36]. Despite the high abundance of CD8⁺ T-cells in the tumor stroma, they are functionally impaired and poorly responsive to T-cell activating stimuli [26,37]. This may be due to ineffective tumor-antigen presentation by dendritic cells (DC), reduced activity of CD4⁺ T-cells, and regulatory functions by T_reg cells [26].

The highly inflammatory immune response associated with smoking also favors the accumulation and activation of DC [38,39]. However, smoking may also suppress DC function and maturation [26]. In tumor supernatants, Sharma et al. demonstrated that DCs had reduced surface expression of major histocompatibility complex (MHC) class I and II and co-stimulatory molecules (e.g., CD80 and CD86) in vitro, reducing their capacity to present tumor-specific antigens for T-cell activation [40]. Similar findings were observed in mice [27,41] and human DC cell lines [42], where exposure to cigarette smoking extracts hindered DC ability to stimulate and activate antigen-specific T-cells. Thus, cigarette smoking appears to have a profound impact on DC function and subsequent immunosuppression.

CD4⁺ T-cells are important for the initiation and maintenance of CD8⁺ T-cell activity [43]. T_reg cells are a subset of CD4⁺ effector T-cells that maintain immunological homeostasis through immunosuppression [44]. Depletion of T_reg cells has previously been demonstrated to augment anti-tumor T-cell activity [45,46]. Smoking status is associated with a low frequency of CD4⁺ T-cells [47,48] and high frequency of T_reg cells [31]. Kinoshita et al. observed a high FOXP3/CD4 ratio in smokers with lung adenocarcinoma, where FOXP3 regulates T_reg development, and suggested that this was an unfavorable prognostic factor [47]. Likewise, Sato et al. showed significant association between lung adenocarcinomas with a smoking signature (C > A transversions) and FOXP3⁺ T-cells [49]. These studies demonstrate that tumors with smoking signatures were associated low expression of CD4⁺ T-cells and high levels of T_reg cells, leading to reduced tumor-infiltrating CD8⁺ T-cell activity and poor prognosis in NSCLC. T_reg cells can also act on CD8⁺ T-cells indirectly through suppressing DC function [50], preventing proper antigen presentation to T-cells, leading to immunosuppression.

There is mounting evidence that smoking status and smoking signatures in tissues are associated with high somatic mutation rates, neoantigen production, and higher TMB [49,51,52,53,54,55]. There is conflicting evidence as to whether or not this is a dose-dependent effect. In NSCLC patients, Nagahashi et al. showed that current smokers without oncogenic driver mutations had the highest TMB in their samples, (p < 0.01), as compared to former or never smokers [53]. No differences in TMB were observed based on duration or intensity of tobacco exposure or pack years (PY). In contrast, Wang et al. observed that a doubling of smoking PY in NSCLC patients was associated with 1.11-fold increase in TMB (p < 0.001). Doubling of the number of months since quitting smoking was associated with a 0.95-fold decrease in TMB (p < 0.001) [54]. In summary, smoking is strongly associated with TMB, which may be associated with improved treatment outcomes with anti-PD- therapy [55,56].

The NSCLC mutational landscape in smoking and never smoking patients differs considerably. Tobacco smokers often have a tumor genomic profile characteristic of high frequencies of C > A transversions and defective DNA mismatch repair [18,54]. Specific signatures are associated with smoking-induced lung cancer. Nik-Zainal et al. demonstrated that exposure to BAP could induce similar genomic signatures [57]. Never smokers, on the other, have distinct genomic profiles, characteristic of spontaneous deamination of 5-methylcytosine and ultraviolet-induced mutations, respectively [54]. NSCLC patients with a smoking history or tumors with smoking-related signatures had better response rates and survival outcomes with anti-PD-1 treatments than never smokers or those with tumors without smoking-related signatures [51,58].