Cigarette smoking and excessive alcohol consumption have been found to be major risk factors for HPV-negative HNSCC, and their combination has synergistic carcinogenic effects (35). The decrease in smoking rates and the rise of HPV-associated HNSCC in western countries, contribute to the improved outcomes in this patient population (5, 36, 37). However, in certain tumor types such as NSCLC, smokers appear to have better response to ICI compared to never smokers (38–40). One possible explanation for this is that TTF-1 (thyroid transcription factor-1) expression is lower among non-smokers, who are more prone to acquire poorly differentiated lung cancers (41, 42). Despite some discrepancy, there is mounting evidence that links TTF-1 expression to improved ICI efficacy and programmed death-ligand 1 (PD-L1) expression (43–46). Never smokers also have a high incidence of EGFR mutations or anaplastic lymphoma kinase (ALK) rearrangements (47–50), where ICIs are known to be less effective. A further explanation for the superior response to ICI in lung cancer smokers than in HNSCC smokers is that smoking's proinflammatory effect is more pronounced in lung cancer, but its immunosuppressive effect is more prominent in HNSCC (30). In accordance with previous findings, we observed no significant impact of smoking on the survival rate of HNSCC patients treated with ICI.
Overexpression of PD-L1 has been discovered to contribute to T cell exhaustion in persistent infections (51). In similar manner, the expression of PD-L1 by cancers allows them to evade the immune system by downregulating T cells (52). Although PD-1/PD-L1 blockade ICI improved overall survival in a variety of PD-L1-expressing tumors, efficacy remains limited for the vast majority of patients (53). Several studies have demonstrated that nicotine use increases PD-L1 expression (54, 55). Zeleskis et al. showed that smokers had improved immunotherapy efficacy, unless they quit smoking, in which case a rapid rebound effect resulting in decreased PD-L1 expression. Moreover, authors demonstrated that early chemotherapy induces a more consistent high level of PD-L1 expression after cascading chemotherapy (56). However, studies suggest that only increasing the expression of PD-L1 in well-responding tumors would maximize the effectiveness of ICI treatment (57–59). Yang et al. indicated that smoking signature was a better predictor of pathological response in patients with non-small cell lung cancer (NSCLC) than expression of (PD-L1) (60). Preliminary analysis of earlier studies revealed that PD-L1 is expressed in 50–66% of HNSCC and that tumor infiltration by PD-1-positive CD8 lymphocytes, PD-1-positive CD4 lymphocytes, and PD-1-positive Tregs was more prevalent in HPV-positive HNSCCs than in HPV-negative HNSCCs (61–63). In one recent trial, pembrolizumab alone or with chemotherapy significantly increased OS in R/M HNSCC patients with PD-L1 combined positive score (CPS) ≥ 1 (64).
Alcohol consumption appears counterproductive for the ICI since it suppresses the innate immune response by diminishing cell recruitment and disrupting macrophage phagocytosis activity (65–67). It also interferes with T cell stimulation by suppressing the expression of CD80 and CD86 on dendritic cells and reducing their quantity (68, 69). A greater expression of PD-1 on T cells has been reported in individuals with acute alcoholic hepatitis, although we doubt this has a positive effect on ICI due to the aforementioned effects (70). A prior study revealed that the immunological milieu of oral squamous cell carcinoma (OSCC) patients who never smoked or consumed alcohol ,was enhanced with PD-L1 and CD8 T cell infiltration and had a better response to ICI (28). In contrast, another study showed the opposite results for the same type of cancer, with no significant correlation between alcohol consumption and response to ICI (29). In our cohort, we observed no significant impact of alcohol consumption on the OS rate of R/M HNSCC patients treated with ICI. Acute alcohol exposure in mice was reported to stimulate anti-inflammatory cytokine production, but the opposite effect was reported in alcoholic hepatitis patients secondary to a short term of heavy consumption (71, 72). Alcohol consumption also causes gut microbiota shifts and bacterial changes (73–76). There is growing understanding on the impact of gut microbiota as a modulator of efficacy and tolerability of ICI (77). Researchers have found evidence of microbial alteration of anticancer immune responses after transplanting fecal microbiota in vivo (78, 79).
In the past decade, medicinal marijuana has become increasingly prescribed due to its potential benefits as a relaxant, anxiolytic, anti-inflammatory antidepressant, antiemetic, and pain reliever, but it also has several potential side effects (80–86). Some studies have indicated a correlation between cannabinoid and an increase in tumor growth (87, 88). Synthetic agonists of cannabinoids increase the activity of PI3K/AKT and MAP signaling pathway (89). The association between marijuana use and development of head and neck cancer remains controversial. Gillison et al. reported evidence of an association between marijuana use and HPV + HNSCC, while Liang et al. demonstrated that moderate marijuana use significantly reduced the risk of HNSCC in those who began using it later in life (90, 91). Cannabinoid receptors represent a complex pathway and interact with immune system at various levels. Given marijuana use in transplant recipients and auto-immune disorders like ulcerative colitis and rheumatoid arthritis there is concern about potential interaction with ICIs. Hence, with the growing use of marijuana, studies that aim to evaluate the potential association of marijuana use with outcomes in cancer patients improve shared decision making. A survey based study found that about one fourth of cancer patients had high active cannabis use but did not receive information about cannabis use from their oncology providers(92). The growing use of cannabis in palliative oncology and a lack of comfort from oncology providers about discussing cannabis use, highlights the importance of studies like ours that explore potential effect or interaction of cannabis with commonly used oncology treatments.
Active ingredients of cannabis can affect several biological processes and human body also produces endogenous analogs of these ingredients. The complex signaling system composed of these ligands and multiple receptors regulates several physiological processes, including the innate and adaptive immune functions. By influencing immune functions cannabinoids can regulate auto-immunity, inflammation, and antitumor immune response. The endocannabinoid system prevents pathological immune responses and is regarded as the gate keeper. On the level of the bone marrow, cannabis causes the retention of immature B cells, a considerable reduction in CXCR4, and suppression of lymphocyte recovery (93–95). Cannabinoids affect various cellular and cytokine processes leading to immune suppression. There is a four-prong mechanism that leads to these effects: induction of apoptosis (of T cells, macrophages, splenocytes, and thymocytes) (88), inhibition of cell proliferation, inhibition of production of chemokines and cytokines, and induction of T-regs (96, 97). CBD induced CD4 + and CD8 + cell apoptosis is proposed to occur through increased reactive oxygen species generation and increased caspase activity (98). In invitro experiments, human eosinophils T, B, CD-8, and NK cells decrease cytokine production after activation in response to CBD exposure (99). Other proposed mechanisms of reduction of T-cell immunity against cancer by cannabis involve blocking downstream JAK1-STAT signaling (100).
Studies evaluating the interaction of marijuana use on ICI treatment have shown varied results. Biedney et al. and Bar-Sela et al. revealed a substantial connection between cannabis usage during ICI treatment and poorer OS, which they attributed to the anti-inflammatory effects of cannabinoids (101, 102). Bar-Sela et al. also showed a substantial reduction in the time to tumor advancement and a decrease in immune-related side effects (102). Contrarily, Taha et al. found no difference in progression-free survival or OS despite a lower response rate (103). The lack of statistically significant interaction between marijuana and outcomes in patients getting ICI in our study could potentially be due to sample size, under reporting of marijuana use by patients or dose dependent relationship with marijuana use due to the complex ligand and receptor pathway. It is also possible that providers pay less attention to a detailed marijuana history. Within the limitations of a single center retrospective study, we hypothesize that marijuana use in head and neck cancer patients does not modulate the immune system enough to impact outcomes with ICI treatment. Future independent studies should aim to evaluate the interactions between marijuana use, dose intensity and outcomes with various cancer treatments, especially ICI.