Preclinical mouse models are important tools for cancer research. The effect of GM on NSCLC immunotherapy was initially illustrated using a mouse model ( Table 1 ; Figure 1 ). Individual mice were observed to possess varied responses to NSCLC immunotherapy when fed different GM. Additionally, Newsome et al. transplanted the feces of 65 patients with NSCLC undergoing ICI treatment into germ-free wild-type (GF-WT) mice; this transplantation reduced Lewis lung carcinoma (LLC) growth in these mice. Overall, Bacteroides was enriched in the mouse cohort that received transplantation from ICI responders, whereas Ruminococcus was considerably enriched in the non-responding mouse group (24). In the same mouse model, Routy et al. demonstrated that the GM is vital for the response to PD-1 blockade. Specifically, they established that mice receiving fecal microbiota transplantation (FMT) from responders exhibited a greater response to anti-PD-1 treatment than mice receiving FMT from non-responsive individuals. In addition, in mice recolonized with fecal microbiota from non-responsive patients with NSCLC, gavage therapy with Akkermansia muciniphila alone or in combination with Enterococcus hirae alleviated anti-PD1 unresponsiveness (26). Interestingly, they also discovered that the adjuvant impact of A. muciniphila on the anti-PD1 response resulted in a corresponding increase in CD4⁺ T cells expressing the chemokine receptor 9(CCR9), which was initiated via the IL-12-dependent signaling pathway (26). Although both studies employed identical LLC mouse models and anti-PD-1 ICIs, differences in the detected bacteria may be attributed to the different experimental conditions.

Strain-level differences in the GM can also affect the expression and synthesis of relevant immunomodulatory components in NSCLC and can, thereby, modulate immunotherapy efficacy (25). A key study established that Bifidobacterium bifidum is enriched in patients with NSCLC who are responsive to ICI treatment (25). Subsequently, Lee et al. supplemented the GM in syngeneic mouse models with four strains (K57, K18, B06, or R71) of B. bifidum. Although all B. bifidum strains inhibited LLC growth, only two (K57 and K18) exhibited beneficial synergistic effects with anti-PD-1 therapy (25). Therefore, as not all strains exhibit identical effects, elucidating the underlying mechanisms is crucial for creating live biotherapeutic products (27). Importantly, it was determined that certain strains of B. bifidum produce large quantities of peptidoglycan, which activate T cells in NSCLC via peptidoglycan-mediated interferon (IFN)signaling. In mouse models of LLC, this ultimately leads to an enhanced immunotherapy response (25).

Finally, comparable findings have been reported in other preclinical solid tumor models, suggesting that the synergistic effect of GM and ICI therapy may be applicable to ICI-resistant cancer. In preclinical mouse models of melanoma, two studies published in Science revealed that PD-1 and CTLA-4 inhibition only suppressed tumor development in mice carrying Bifidobacterium and Bacteroides fragilis species, respectively (28, 29). Further, Griffin et al. determined that Enterococcus spp. with unique NlpC/p60 peptidoglycan hydrolase activity may produce NOD2-active muropeptides; these muropeptides could then alter the effectiveness of checkpoint blockade immunotherapy in vivo (30). In mouse models of colorectal cancer, Mager et al. revealed that bacterial inosine is associated with the GM-dependent response to ICIs (31). FMT from responsive and non-responsive patients into mice with human epidermal growth factor (HER2)-positive breast cancer mimicked the responses to ICIs observed in animal models of breast cancer (32). In renal cell carcinoma (RCC) tumor-bearing mice, oral gavage with feces from ICI-responding patients with RCC reversed the initial resistance of these mice to ICI treatment (33). In NSCLC, the relevant research is still blank. However, in melanoma, gut dysbiosis attenuated the anti-tumour efficacy of Tim-3 blockade in mice (34). In colon adenocarcinoma tumor-bearing mice, the immunotoxicity of anti-CD137 immunotherapies is dependent on GM (35).

These preclinical studies have successfully demonstrated that GM composition has a clear impact on checkpoint blockade effectiveness, and alteration of the GM can enhance response to ICI treatment. Nevertheless, most preclinical investigations that demonstrated the influence of GM on ICI response have been limited to melanoma and colorectal cancer. Therefore, NSCLC should receive more attention in future pre-clinical experiments into the effect of the GM on ICI response.

The human GM does not necessarily colonize mice in the same manner as it colonizes humans. Furthermore, the mucosal and immunological responses of mice to introduced commensals are not comparable to those in humans (36). Consequently, various research groups have sought to define the function of the GM in NSCLC immunotherapy outcomes to better understand the difference in responders (Rs) and non-responders (NRs) ( Tables 2 , 3 ). Overall, 16S rDNA sequencing and metagenomic shotgun sequencing (MSS) were used to analyze the composition of GM in these studies.

First, Routy et al. established that A. muciniphila is overrepresented in the stool samples of patients with NSCLC who benefit from PD-1 treatment. Specifically, they reported an enrichment of A. muciniphila in R vs. NR, as well as a relative underrepresentation of Bifidobacterium adolescentis and Parabacteroides distasonis in those that responded to PD-1 treatment. Antibiotic (ATB) use has also been observed to have a deleterious influence on the immune checkpoint blockade response (26). Four years later, the same research group used microbiome profiling to prospectively evaluate the benefits of fecal A. muciniphila in a large cohort of patients with advanced NSCLC (n = 338) who were treated with first- or second-line ICIs. In this study, ATB usage and the Clostridium genus were linked to ICI resistance (39). Additionally, Martini et al. provided clinical evidence that two butyrate-producing GM (Agathobacter M104/1 and Blautia SR1/5) may improve immune responses in patients with metastatic and chemorefractory NSCLC (44). Jin et al. also reported that memory T cell and natural killer cell signatures in peripheral blood were higher in patients with advanced NSCLC with high gut bacterial diversity (43). Therefore, patients with NSCLC with high microbiome diversity were determined to exhibit a longer progression-free survival (PFS) than those with a low-diversity microbiome (43). Five other studies reported similar results; the various GM compositions associated with improved ICI responses in these studies are listed in Table 2 . Overall, there are substantial variations between the diversity and composition of GM in the R and NR groups in patients with NSCLC. GM profiling in patients with NSCLC revealed that a higher gut bacterial diversity and an increased abundance of specific bacteria, such as A. muciniphila, Bacteroides vulgatus, P. distasonis, bacterium LF-3, Sutterella wadsworthensis HGA0223, Agathobacter, Blautia, Alistipes, Anaerostipes, Desulfovibrio, Faecalibacterium, Bifidobacterium, Escherichia-Shigella, Olsenella, Ruminococcaceae UCG 13, Methanobrevibacter, Alistipes putredinis, and Prevotella copri, were associated with a stronger T-cell dependent antitumor response, resulting in favorable ICI clinical outcomes in multiple independent trials. Li et al. analyzed the gut microbiota to explore potential causal associations with various subtypes of lung cancer using mendelian randomization. They identified 10 groups associated with lung adenocarcinoma and 9 groups associated with squamous cell lung cancer. Notably, a significant causal relationship between Peptococcaceae and lung adenocarcinoma was observed, as reported in the literature (45). However, the gastrointestinal microbiome does not always play a beneficial role in NSCLC immunotherapy. According to a retrospective investigation of two independent cohorts of patients with NSCLC, Helicobacter pylori has been associated with an evident decrease in the PFS of patients with NSCLC undergoing anti-PD-1 therapy (46) ( Figure 2 ). Studies have shown that gut microbiota plays an important role in the response of NSCLC patients to anti-PD-1 therapy, and that gut microbiota diversity is significantly increased in patients who respond to anti-PD-1 therapy (47). A prospective study of 63 patients with advanced NSCLC taking anti-PD-1 therapy found that β diversity in gut microbiota at baseline was significantly increased in patients with progression-free survival (PFS)≥ 6 months compared with patients with PFS<6 months; The LEfSe analysis suggested that the most significantly associated groups in patients with PFS≥6 months were Paracterium (LDA score=3.8) and Methanobrevibacterium (LDA score=3.4), whereas the PFS<6 months group was rich in Veillonella, Seleniomonas, and Negativicutes. Further analysis showed that Paracobacterium and Methanobrevibacterium species in NSCLC patients enhance their responsiveness to Immune checkpoint inhibitors (42). The results reported in a study showed that in NSCLC receiving first-or second-line immunotherapy to detect its pre-treatment gut microbiota, Akkermansia muciniphila (AKK+) patients were found to have a higher immunotherapy response rate than AKK-patients (28% VS 18%), among which AKK+ patients, the response rate of first-line immunotherapy was as high as 41%. For survival, overall survival(OS) was longer in the AKK+ group than in the AKK-group (18.8 months vs 15.4 months) (39).

In summary, experimental and clinical evidence has demonstrated that GM plays a key role in modifying the effectiveness of immunotherapy in NSCLC. However, these studies have several limitations that should be addressed. For example, most clinical trials have been undertaken on a relatively small number of patients with NSCLC; additionally, current research seems to lack consensus on what GM compositions are advantageous or antagonistic to ICI treatment of NSCLC. Accordingly, several clinical studies are being conducted to address these issues ( Table 3 ). Further, the cause-and-effect relationship between the GM and immunotherapy in NSCLC remains unclear. Therefore, further research is required to determine whether the GM modulates immunotherapy in patients with NSCLC or whether immunotherapy alters the GM of patients with NSCLC. Furthermore, the correlations between B cell responses and the GM are yet to be thoroughly investigated in the context of NSCLC immunotherapy.

Although ICIs are the first-line therapy for NSCLC, several irAEs have been reported. According to a multicenter cohort study, 623 patients with NSCLC treated with PD-(L)1 monotherapy developed multisystem irAEs, the most prevalent of which were pneumonitis thyroiditis, hepatitis thyroiditis, dermatitis thyroiditis, and dermatitis pneumonitis (48). Anti-PD-(L)1-induced checkpoint inhibitor pneumonitis is more common in patients with NSCLC than in those with other malignancies. Disordered T-cell subsets, elevated preexisting and emerging autoantibodies, and imbalanced inflammatory cytokines are three pathways that may explain this PD-(L)1 inhibitor–mediated pulmonary damage (49). Additionally, according to Tang et al., the most frequently observed irAE is checkpoint inhibitor colitis (50).

However, the treatment of irAEs remains a key source of concern in NSCLC immunotherapy. In most cases of irAE, immunotherapy may be withheld or stopped completely, depending on the severity of toxicity (51). Several studies considering the putative link between GM and irAEs in NSCLC have been conducted. Hakozaki et al. obtained pre-ICI fecal samples from 70 Japanese patients with advanced NSCLC who were then treated with ICIs. Ultimately, they determined that Akkermansia, Lactobacillaceae, and Raoultella were associated with milder irAEs, whereas Agathobacter was associated with more severe irAEs (15). Another study in which 37 patients were enrolled indicated that the GM of patients who did not experience irAEs was relatively enriched in Bifidobacterium and Desulfovibrio spp (52). Bifidobacterium, Faecalibacterium, and Agathobacter were also determined to be less abundant in patients with irAEs (53).

Zeng et al. observed notable alterations in butyrate-producing bacteria within a cohort of patients experiencing secondary resistance (SR) to ICIs. There was a downward trend in these bacteria levels upon the onset of secondary resistance. Notably, Bacteroides played a significant role in the variance between the baseline and the occurrence of irAEs. The abundance of Bacteroides declined after the onset of irAEs and subsequently returned to levels comparable to the baseline after remission (54).

Furthermore, a multicenter non-interventional trial (NCT04107168), involving up to 40 UK sites, is currently underway. To examine mechanisms underlying irAEs, longitudinal stool samples were collected prior to NSCLC immunotherapy and following 6 weeks, 3 months, 6 months, and 12 months of treatment. A clear association between GM composition and ICI toxicity was one of the initial results established in this clinical trial (55).

Additionally, the GM has been reported to be associated with irAEs in other cancers. For example, Andrews et al. established that Bacteroides intestinalis mediates immune-related intestinal damage via IL-1β in melanoma (56). The link between irAEs and several Lachnospiraceae species in melanoma has also been verified by McCulloch et al. (57). Further, Mao et al. determined that a higher GM diversity and relative abundance of the Firmicutes phylum may be protective factors against irAEs in patients with hepatobiliary cancer (58). Finally, compared with the post-treatment microbiota of immune-related acute pancreatitis, a low Bacteroidetes/Firmicutes ratio, a low relative abundance of Alistipes and Bacteroides, and a high Lachnospiraceae abundance were observed in the baseline microbiota of individuals with pancreatic cancer (59).

FMT involves the transfer of GM from a known donor to a recipient via the upper or lower gastrointestinal route to restore microbial diversity within the intestine (79). In 1958, Eiseman published the first report on the use of FMT for the treatment of severe pseudomembranous enterocolitis (80). Despite being a well-established therapy for recurrent Clostridium difficile infections, FMT is now being explored in the context of cancer treatment (81).

FMT is a potential therapy for various types of cancer and cancer treatment–associated complications. In preclinical mouse models, Shaikh et al. utilized FMTs from two NSCLC patients who had contrasting responses to ICIs (R and NR); overall, R-FMT mice exhibited superior antitumor responses when paired with ICI treatment compared to that of NR-FMT mice (11). Baruch et al. and Davar et al. reported the first-in-human clinical trials conducted to determine whether FMT can influence the response of patients with metastatic melanoma to anti–PD-1 immunotherapy. Both trials demonstrated signs of therapeutic benefit in some treated patients, including a higher abundance of taxa associated with an anti–PD-1 response and enhanced CD8⁺ T cell activation (82, 83). Several clinical trials of NSCLC are currently being conducted to evaluate whether combining FMT with immunotherapy can enhance the outcomes of patients with NSCLC ( Table 4 ).

Despite the promising results of FMT in patients treated with ICIs, concerns remain regarding its long-term safety. According to a study published in Lancet, 32 (78%) of the 41 individuals receiving FMT treatment experienced adverse events, the majority of which were self-limiting gastrointestinal problems (84). Nonetheless, a randomized, placebo-controlled trial that evaluated the safety of FMT for active peripheral psoriatic arthritis reported no serious adverse events in patients treated with FMT (85). According to the European consensus on FMT in clinical practice, FMT shows an excellent short-term safety profile, with only a few mild adverse events observed. However, long-term safety data are limited. Therefore, theoretically, FMT may transfer potentially dangerous microbial features as well. In conclusion, given that FMT is a highly effective treatment for NSCLC patients, the benefit-to-risk ratio should be considered for each individual (86).

ATBs not only kill pathogenic bacteria but also alter the diversity and composition of the GM, thereby disrupting gut homeostasis (87–89). In the context of NSCLC, ATBs appear to be associated with a reduction in ICI therapeutic impact ( Table 5 ). Specifically, PFS duration was considerably reduced in patients with NSCLC who were treated with ATBs within 2 months of their initial ICI therapy than in those who had not received ATB treatment within this timeframe (96). Furthermore, the use of ATBs before ICI therapy has been demonstrated to reduce GM diversity, and the presence of Ruminococcaceae in patients with advanced NSCLC resulted in unfavorable outcomes (15). The effectiveness of cancer therapies has also been revealed to be impaired in ATB-treated and GF mice (97).

However, considering that many patients with NSCLC require ATB treatment over the course of this disease, the current focus of research is to determine what factors can reduce the effects of ATBs on ICI therapy. First, the use of ATBs during the early immunotherapy period may be associated with poor therapeutic outcomes. Interesting data from Zitvogel et al. demonstrated that patients receiving ATBs immediately before or after initiating ICI therapy exhibited poor outcomes (26). Similarly, in a trial involving 72 patients with NSCLC who were treated with nivolumab, the early use of ATBs had also been associated with a shorter OS (98). A meta-analysis conducted on patients with NSCLC established that exposure to ATBs shortly before or after ICI therapy had a notably harmful impact, whereas ATB use later in the disease progression had no significant effect on survival (99). Second, patients with NSCLC that possessed a higher ATB-ICI exposure ratio (defined as “number of days of ATB administration/number of days of ICI administration” for the entire immunotherapy period) also exhibited poor outcomes (100). Third, the effect of ATBs on ICI effectiveness varies depending on the PD-L1 expression levels in patients with advanced NSCLC. Specifically, a deleterious effect has been observed in patients with ≥50% PD-L1 expression but not in those with <50% PD-L1 expression (101). Finally, with all other clinically relevant parameters constant, patients who received a single course of ATBs were observed to exhibit a shorter median OS and PFS, whereas those who received multiple courses or prolonged ATB therapy exhibited the worst overall outcomes (95). Gut microbiota plays a crucial role in immune response. The use of antibiotics affects the efficacy of ICIs by altering the gut microbiota. Alkan et al. showed that the use of antibiotics at the start of ICIS was associated with reduced OS, PFS and objective response rate (ORR) in NSCLC patients (102). This suggests that gut microbiota diversity may be one of the factors predicting ICI efficacy. Therefore, cumulative ATB use is not recommended.

Similar to ATB, PPIs are negative prognostic markers in patients with NSCLC who are undergoing immunotherapy. In a clinical study involving 169 individuals, PPI use was associated with shorter PFS in patients with NSCLC that were treated with atezolizumab (91). Additionally, PPI use has been observed to negatively influence ICI efficacy in patients with NSCLC who received atezolizumab + carboplatin + paclitaxel treatment (ACP) or atezolizumab + bevacizumab + carboplatin + paclitaxel treatment (ABCP) but not in those who received bevacizumab + carboplatin + paclitaxel treatment (BCP) (91). Moreover, a meta-analysis indicated that PPI treatment is associated with a reduced PFS and OS in patients with advanced cancer who were treated with ICIs (103). Thus, when not clearly warranted, PPI medication should be discontinued; additionally, the use of histamine H2-receptor antagonists should be investigated as an alternative to PPI treatment.

In conclusion, there is a consensus on the negative effects of ATBs and PPIs on patients with NSCLC who are receiving immunotherapy. Given that the prohibition of ATB use is impractical for certain patients, modulating ATB-related GM dysbiosis may be a viable strategy to improve clinical outcomes. Therefore, the use of targeted antibiotics that allow immunopotentiating bacteria to proliferate have been recently proposed (104). However, the dosage, frequency, and delivery methods for targeted ATBs remain to be elucidated.

Recently, probiotics, prebiotics, and postbiotics have gained the attention of clinicians due to their potential to enhance ICI efficacy ( Table 6 ). “Probiotic” was first defined in 1965 (105); nonetheless, the most recent definition of probiotic is as follows: “live microorganisms that, when administered in adequate amounts, confer a health benefit on the host” (106). Prebiotics are fibers that are digestible by individual commensal species; these bacteria metabolize the prebiotics, thereby producing secondary metabolites such as SCFAs (107). Any material released or produced via the metabolic activities of microbes that has a direct or indirect positive impact on the host is regarded as a postbiotic (108). Finally, synbiotics are a mix of prebiotics and probiotics and have been proven advantageous in inflammatory bowel disease; however, their impact on cancer is not yet examined in clinical studies (104).

Probiotics, prebiotics, postbiotics, and synbiotics play important roles in assisting immunotherapy in the treatment of NSCLC. A retrospective study involving 294 patients established that probiotic use is associated with favorable clinical outcomes in patients with advanced or recurrent NSCLC who were subjected to anti-PD-1 monotherapy (109). Alternatively, a retrospective study evaluated 118 patients with advanced NSCLC who were treated with ICIs at Kumamoto University Hospital. Significant improvements in PFS and OS were observed in patients who received probiotic Clostridium butyricum therapy, even in those who had previously undergone ATB therapy before ICI therapy (110). Castalagin, the active ingredient in camu-camu, can concentrate microorganisms (Ruminococcaceae and Alistipes) associated with effective immunotherapeutic responses and can enhance the CD8⁺/FOXP3⁺CD4⁺ ratio in the TME. In preclinical ICI-resistant models, castalagin was determined to re-establish the efficacy of anti–PD-1 in NSCLC (111). Further, in addition to improving survival, 3-IAld (a microbial tryptophan catabolite) has been found to protect mice with ICI-induced colitis from intestinal injury by acting on both the host and microbiota (112). In a direct comparison study, the oral application of a combination of four Clostridiales strains (CC4) was observed to outperform anti-PD-1 therapy in mouse models of NSCLC. Specifically, CC4 therapy has been found to suppress tumor growth by increasing the number of IFN-γ⁺ CD8⁺ T cells and natural killer cells while reducing the number of PD-L1⁺ macrophages (113). A meta-analysis determined that probiotics are favorably associated with the OS and PFS of patients with NSCLC that are being treated with ICIs, but had no effect on ORR (114). Similar results have also been reported in other cancer types. For example, CBM588 appeared to increase the PFS of patients with metastatic RCC who were being treated with nivolumab-ipilimumab (115). Additionally, the adjunctive probiotic Lactobacillus rhamnosus probio-M9 was determined to enhance the effect of anti-PD-1 antitumor therapy by repairing antibiotic-disrupted GM in mouse models of colorectal cancer (116). Chen et al. showed that JK5G postbiotics may attenuate immune-related adverse events irAEs and improve quality of life (QoL) and nutritional levels in advanced NSCLC patients receiving ICIs. JK5G postbiotics can also improve gut microbiota structure, increase Faecalibacterium, Ruminococcaceae relative abundance, fecal butyrate concentration, reduce Escherichia-Shigella relative abundance, and improve tumor microenvironment and inflammation (117). A study reveals that probiotic Lactobacillus reuteri (Lr) uses its released dietary tryptophan catabolite indole-3-aldehyde(I3A), it locally promoted the production of interferon-γ by CD8 T cells, thus bolstering ICI (90).

Overall, opinions have been divided regarding the safety of probiotics. Systemic infections, harmful metabolic activities, excessive immunological activation in sensitive individuals, gene transfer, and gastrointestinal side effects are the potential dangers of this therapeutic strategy. On the other hand, critics argue that the overwhelming body of evidence, including a long history of safe probiotic usage and data from clinical trials, alongside animal and in vitro research, supports the understanding that probiotics are generally safe for most individuals (118). In general, more advantages than disadvantages have been observed with regard to probiotic treatment. Ultimately, probiotics, prebiotics, postbiotics, and synbiotics are potentially viable strategies as they are less complicated and presumably safer than FMT, which has been associated with increased pathogen transmission.

Non-bacterial microorganisms in the GM have received less attention than bacteria, despite their potential importance. Bacteriophages, also known as bacterial viruses or phages, are obligate intracellular parasites that rely on the metabolic apparatus of their bacterial host for multiplication (119). Bacteriophages are the most abundant biological entities on Earth (120). Fluckiger et al. established that E. hirae harbors a bacteriophage that can modify immunological responses. After PD-1 inhibition, mice bearing E. hirae bacteriophages have been observed to exhibit an increase in the CD8⁺ T lymphocyte response. In patients with lung cancer, the presence of bacteriophages in their feces and the expression of a tape measure protein–cross-reactive antigen in tumors has been linked to better survival following PD-1 treatment (121). Dong et al. screened a strain of selective Fusobacterium nucleatum–binding M13 bacteriophages. The introduction of silver nanoparticles onto the surface of the M13 bacteriophages allowed precise scavenging of pro-tumor F. nucleatum; consequently, the host’s immune response was enhanced via APC activation and the tumor-immunosuppressive microenvironment was remodeled. Notably, M13-Ag combined with PD1 inhibitors has been observed to significantly prolong OS in a colorectal cancer mouse model (122). Therefore, the use of bacteriophages to modulate the GM for NSCLC immunotherapy is increasingly gaining importance in both basic and translational research.

Diet plays an essential role in bridging the connection between humans and the microbiota. The GM consumes and utilizes ingested nutrients for basic biological activities; therefore, diet is a pivotal determinant of GM structure and function (123, 124). Recent evidence suggests the use of high-fiber diets for better outcomes in cancer immunotherapy. In a prospective cohort of patients with melanoma that were receiving ICI therapy, intake of a fiber-rich whole-grain diet was strongly correlated with increased PFS. Interestingly, PFS was observed to be the highest in the group that consumed a high-fiber diet but did not consume probiotics. The inhibition of intratumoral IFN-γ T cell responses was assumed to be the cause of the worsened response to immunotherapy observed in those on a low-fiber diet (125). Dietary fermentable fiber content alters the ratio of Firmicutes to Bacteroidetes in the gut and the lungs. The GM digests fiber, which increases the concentration of circulating SCFAs and influencing the lung immune environment (126). In addition to high-fiber diets, compelling evidence has suggested that fasting-mimicking diets, long-term calorie restrictions, and ketogenic diets may improve anticancer immunosurveillance during immunotherapy (12, 127–131). Overall, several studies have focused on the influence of diet on immunotherapy for NSCLC ( Table 7 ).