Lung Microbiome in Lung Cancer: A New Horizon in Cancer Study

Lung cancer is the second most prevalent cancer in men and women (1) and the major cause of cancer-related mortalities. The GLOBOCAN 2020 report shows 2.21 million incidences and 1.79 million deaths annually attributed to lung cancer. Non–small cell lung cancer (NSCLC) and small cell lung cancer (SCLC) are the two histologic subtypes of lung cancer (2). NSCLC cells seem to be larger, whereas SCLC cells seem to be smaller with a round-to-fusiform shape. About 85% of the lung cancer cases are NSCLC, further subclassified into large cell carcinoma, adenocarcinoma, and squamous cell carcinoma (SCC). Late diagnosis due to a lack of regular screening programs, a highly metastatic nature, and quick relapse after therapeutic interventions are the major reasons for the high mortality rate observed in patients with lung cancer. Despite various technological advances and extensive research, the overall survival of patients with lung cancer has not changed significantly in the last four decades (3–5). Modern immune-based therapeutics have improved the survival of patients with lung cancer but only marginally (6). Recent advances in next-generation sequencing methods have allowed us to explore the existence of microbiome residing in the lungs. Commensal microbiota have been shown to regulate various tissue functions, including immune responses (7). In this review, we will focus on the role of lung microbiome in lung cancer progression and in regulating response to therapeutic interventions.

Microbiota refer to a group of many microorganisms, including archaebacteria, bacteria, viruses, and fungi, which colonize in different bodily habitats (8). A microbiome refers to microbiota and associated molecules, including their genes, genomes, secreted factors, and metabolites. The human lungs entail a diverse microbiome critical for human health. An estimated 2.2 × 10³ heterogeneous bacterial genomes per cm² have been detected in a healthy lung (9). The gastrointestinal tract consists of around 100 trillion bacteria in a varied microbial ecosystem. The colon possesses 10¹¹ to 10¹² bacterial cells per milliliter (10), whereas it is estimated that the skin microbiome has 10⁴ to 10⁶ bacteria per cm² (11). Recent studies suggest that bacteria, fungi, or viruses are frequently found in tumors, play a crucial role in cancer therapy, and can be modulated to treat metastasis (12). The recent advancements in next-generation sequencing technologies and analytical methods have provided new information on the lung microbiome, which was earlier thought to be sterile, for its potential in clinical interventions.

This review is organized as follows: The “Microbiome of a Healthy Lung” section briefly highlights the host–microbe interaction in the functioning of healthy lungs. The “Dysbiosis and Lung Cancer” section discusses different possible mechanisms of lung carcinogenesis, establishing the causative association of lung microbiome with lung cancer. It also describes the role of microbiome dysbiosis in lung cancer development and progression. The “Host–Microbiome Interaction in Lung Cancer” section shows the ways in which the lung microbiome facilitates the disease and its interactions with the host in lung cancer. Furthermore, we discuss the ways in which lung microbiota could be used for therapeutic interventions and facilitate early diagnosis and prevention of lung cancer in the “Lung Microbiome: An Emerging Tool in Lung Cancer Diagnostics” section. The “Discussion” section highlights the fact that early screening and therapeutic interventions are essential in improving and conquering lung cancer. Finally, existing gaps in studies and future directions have been discussed in the “Existing Gaps and Future Directions” section.

The healthy lung microbiome community maintains lung homeostasis through microbiome immigration, elimination, and changes in local growth conditions. Studies suggest that microbial immigration occurs primarily by microaspiration and inhalation, whereas microbial elimination is a dynamic process accompanied by cough, mucociliary clearance, and an immune defense system. Local lung growth circumstances like pH, temperature, and nutrient imbalances also assist in maintaining homeostasis (13).

The lung microbiome involvement in influencing inflammatory responses is unclear; however, it is known to maintain immunologic tolerance and modulate immune response (14). The lung microbiome–pulmonary epithelium relationship controls the host pulmonary immune system. Briefly, goblet and club cells of the lungs release mucus to protect epithelial cells and restrict pathogen spread and inflammation, preventing microbiota–host equilibrium. Lung epithelial cells are coated with IgA alveolar surfactant, which aids in lung innate immunity. Epithelial cells form a structural barrier and host protection by secreting antimicrobials, mucins, and pathogen recognition receptors like Toll-like receptor (TLR), nucleotide-binding and oligomerization domain-like (NOD), retinoic acid–inducible gene-like, and C-type lectin receptors that detect specific molecular patterns associated with microorganisms to combat infections via downstream inflammatory signaling pathways. Anti-inflammatory macrophages in alveoli inhibit inflammatory pathways and adaptive immune responses to modify commensal immunologic tolerance (15), and the balance between pathogen defense and commensal immune tolerance determines lung microbiota homeostasis (16, 17). Furthermore, alveolar macrophages and dendritic cell subpopulations maintain the lung microenvironment’s high immunologic tolerance. These cell subpopulations also regulate immunity by generating regulatory T cells and releasing prostaglandin E2, TGFβ, and IL10. Emerging evidence suggests that individuals with a healthy microbiome have more T helper 17 cells and CD4⁺ T cells that release IL1 and IL17 in the airways, indicating that the microbiome impacts T helper 17–mediated immune response (18, 19).

Some epidemiologic studies have investigated the association between gut microbiome and lung cancer by validating the presence of the gut–lung axis and its potential impact on lung cancer (19–22). However, a few investigations have provided preliminary insights into the association of lung microbiomes with lung cancer (23, 24).

Current evidence suggests that the lung microbiome may be associated with carcinogenesis through several possible mechanisms (25–28). These mechanisms include microbiome dysbiosis, genomic instability, microbial metabolism, induction of inflammatory pathways, and immune response in the host (25–28). Le Noci and colleagues demonstrated that microbes and their metabolites trigger TLRs in immunologic and epithelial cells. If prolonged, the inflammatory reactions may cause irreversible damage to healthy cells and impact the formation of a tumor and its progression (29). Another group of researchers found an association between bacteria belonging to the genera Streptococcus, Haemophilus, Prevotella, and Pseudomonas and serum inflammatory biomarkers (30). Certain bacteriotoxins and other proinflammatory factors from microorganisms such as Haemophilus influenzae, Escherichia coli, and Enterobacter spp. and from Moraxella and Legionella genera have been shown to promote carcinogenesis by driving the inflammation in the lungs (31). Similarly, Qin and colleagues (32) indicated that Mycobacterium tuberculosis may cause lung cancer through chronic inflammation–associated carcinogenesis. These sources of evidence have established the causative roles of specific microbes in the development and progression of cancer (33). Some researchers also reviewed the role of the lung microbiome and described its influence on the progression and development of lung cancer and suggested that certain bacterial strains like Veillonella, Prevotella, and Bacteriodetes can “reprogram” local microenvironment cells by releasing immunoregulatory molecules or recruiting peripheral inflammatory immune cells (34, 35). Similarly, it has been reported that cytokines released by Fusobacterium nucleatum lead to the activation of NF-κB signaling, chronic inflammation, and progression of colorectal cancer (35).

Even though there is growing evidence of the association of dysbiosis with lung cancer, studies focused on respiratory microbiota, and their dysbiosis are still in their early stage (36, 37). Dysbiosis in the lung is determined by changes in any of the factors, microbiome immigration, microbiome elimination, and local growth conditions (38). Mclean and colleagues (39) proposed a mechanism by which dysbiosis could cause carcinogenesis. This mechanism includes the following steps: First, dysbiosis causes adaptive immune dysregulation, allowing tumor cells to escape the immune system and promote their growth and progression. Second, microbes and their metabolites can damage host DNA or activate microbe-associated molecular patterns or TLRs to cause downstream inflammatory reactions that promote cancer development. Lastly, microbes stimulate proliferative pathways that might cause host cells to become cancerous. This is summarized in Fig. 1, which displays a comparison between healthy and cancerous lungs.

Interestingly, Tsay and colleagues found enrichment of the taxa Veillonella and Streptococcus in the lower airways of patients with lung cancer. Associated molecular studies showed that these microorganisms upregulate tumor cell proliferation pathways, including PI3K and extracellular signal–regulated kinase signal transduction pathways (40, 41). A few studies also identified a correlation between lung cancer and strains of Staphylococcus (42), Streptococcus (4, 43–45), Veillonella (4, 46, 47), Pseudomonas (48), Megasphaera (46), Bradyrhizobium japonicum (49), and Moraxella (31) and between lung microbiome dysbiosis and lung cancer. A study on lung tissue revealed the enrichment of genus Thermus in advanced stage lung cancer tissues, whereas Legionella was associated with metastases (50). Another study, however, indicated that patients with lung cancer had an abundance of Streptococcus, whereas healthy controls had higher levels of Staphylococcus. It suggests high interindividual variation in the lung microbiome, which highlights the complexity of microbial communities in the respiratory system among different individuals. Further research is needed to unravel the factors contributing to this variation (51, 52). A team of researchers highlighted the peculiar role of commensal microbiota in the tumor microenvironment (TME). They found enrichment of Staphylococcus aureus in adenocarcinoma. It is worth noting that S. aureus can internalize lung cancer–derived exosome-like nanoparticles, which are small, nanosized membrane-bound extracellular vesicles and mediators of intercellular communication. They also found that treating S. aureus with lung adenocarcinoma–derived exosomes shaped like nanoparticles can cause morphologic and functional changes by altering S. aureus shape, adhesion, biofilm formation, and virulence. This study reveals the novel cross-talk between TME and S. aureus and the ways in which lung cancer shapes commensal bacteria using tumor-derived exosome-like nanoparticles (53).

Greathouse and colleagues provided a rationale for a comprehensive examination of the potential role of Acidovorax and other microbes in lung cancer. They found that SCC with the tumor protein 53 variant had higher Acidovorax levels than those in adenocarcinoma (54). Furthermore, a recent study suggested a relationship between specific microbiota obtained from formalin-fixed, paraffin-embedded (FFPE) tissues and lung cancer subtypes. They found that lung adenocarcinoma had significantly more Cyanobacteria than that in SCC. In addition, Cyanobacteria toxin (i.e., microcystin) has been shown to induce procyclic acidic repetitive protein 1 expression in lung cancer and was validated using microcystin-challenged lung adenocarcinoma (A427) cell lines. The inclusion of controls and positive selection based on amplicon presence helped maintain the reliability and accuracy of the microbial DNA data obtained from the FFPE samples. Briefly, commercial host genomic DNA and nucleic acid–free filtered water were used as negative controls in the quality check. The microbial DNA was used as a positive control for bacterial 16S rRNA 500-bp amplicon from FFPE samples. To account for biases caused by uneven sequencing depth, equal numbers of random sequences were selected to ensure the same number of sequences in all samples for downstream analysis. In addition, samples with significantly fewer sequences were removed from the analysis (55). Another study found that lung adenocarcinoma showed enrichment in Brevundimonas, Phenylobacterium, Propionibacterium, Staphylococcus, and Acinetobacter. Meanwhile, SCC had an abundance of Serratia, Kluyvera, Achromobacter, Morganella, Capnocytophaga, Klebsiella, and Enterobacter. Thus, these findings suggest differences in the microbiome in different subtypes of lung cancer (56).

Microbiota affect cancer initiation, progression, host immune system, and treatment response (57). Dong and colleagues (58) revealed the relationship between lung microbiota and tobacco smoking and host gene mutation. They found a high abundance of polycyclic aromatic hydrocarbon–degrading bacteria such as Massilia and Acidovorax in microbiota in the smoker patient’s lung tumor tissue and TP53 mutations. The TP53-mutated tumor microbiome has been shown to dysregulate the p53 signaling pathway. Another interesting study characterized the lung microbiota of patients undergoing immunotherapy treated with immune checkpoint inhibitors PD1 or PDL1. They found that the phyla Proteobacteria, Firmicutes, and Bacteroidetes were enriched in these patients. Furthermore, in the same study, the clinical evidence revealed that increased microbial diversity improved patient survival. They also found that Gammaproteobacteria colonization was associated with lower PDL1 expression. This implies a poor response to checkpoint-based immunotherapy and a decreased overall survival time (59). Zhang and colleagues studied major molecular and cellular mediators of neutrophil–microbiota interactions, which might affect tumor progression. Tumor-associated neutrophils can promote tumorigenesis by enhancing angiogenesis and immune suppression (60). Therefore, an increased understanding of neutrophil–microbiota interactions is key to understanding neutrophil behavior in cancer. These insights into distinct microbiota in various tumor types suggest that intratumoral bacteria may affect neutrophils and other immune cells in the TME (60).

Microorganisms regulate endothelial cells, inflammatory pathways, associated signals, and VEGFs, contributing to angiogenesis, tumor development, and metastasis. Wang and colleagues (27) noted that human papillomavirus, Helicobacter pylori, and H. influenzae could induce chronic inflammation via chemokines, cytokines, and angiogenic mediators that cause inflammation and angiogenesis in lung tumor tissues. The commensal microbiota could be an essential biomarker and regulator of tumorigenesis and cancer therapy response. The bacterial composition and the metabolites of pulmonary microbiota could disrupt various signaling pathways and caused DNA damage, leading to the generation of a procarcinogenic environment. This study provided a unique mechanistic insight into the biology of lung cancer and sheds light on treatment strategies for lung cancer prevention (61). Emerging evidence suggests that microbiota residing in TME can directly impact lung carcinogenesis and treatment responses to chemo- or immunotherapies in cancers originating in the mucosal glands present in the lung, gastrointestinal tract, and skin. The lung microbiota could stimulate tumor-associated myeloid cells to secrete ILs (IL23 and IL1β; ref. 62). This induced T-cell activation, which produced IL17 and other effector molecules in mouse models for tumor cell proliferation. This study demonstrates an association between local microbiota–immune interaction and cancer development (62). Furthermore, the microbiota–autophagy association revealed the influence of microbial community on the process of autophagy to affect physiologic and pathologic reactions involved in cancer progression (63).

Host–microbiota interactions are essential for lung immune homeostasis as lung microbiota dysbiosis and its effect on the local immune system impact inflammation and tolerance (64). The impact of host–microbiota interactions on the airway epithelial cells of the lower respiratory tract in patients with lung cancer has been examined (40). This study showed that the airway microbiota of the lower respiratory tract in patients with lung cancer were enriched with oral taxa Streptococcus and Veillonella and it was correlated with upregulation of lung carcinogenic ERK/PI3K signaling pathway. In vitro, the exposure of airway epithelial cells to Veillonella, Prevotella, and Streptococcus also showed upregulation of these pathways. Understanding interactions between lung tumors and microbes and their metabolites is essential to decipher host–microbiome interactions. The summary of this section is demonstrated in Fig. 2.

Current lung cancer treatments include radiotherapy, immunotherapy, chemotherapy, and surgical procedures. Among these, exploiting the immune system is the latest approach in lung cancer diagnostics. However, the advent of immunotherapy has improved overall survival and progression-free survival marginally by 2 to 3 months compared with chemotherapy. Only 21% of patients with NSCLC and 3.7% of patients with SCLC are eligible for immunotherapy. Unfortunately, patients detected at an advanced stage have a dismal prognosis and few therapy options. Thus, early detection and lung cancer treatment optimization are urgently needed, which can be achieved by a more significant understanding of the commensal lung microbiome (65, 66).

The examination of the microbiome has demonstrated potential as a diagnostic tool in conditions like bronchiectasis. This development was anticipated to significantly impact the diagnostics and management of diseases, such as chronic obstructive pulmonary disease (COPD), by providing clinicians with airway and lung microbiota data. Moreover, the integration of understanding of the lung microbiome with immunomodulatory drugs and immunotherapy has shown potential for enhancing cancer prognosis (67). This suggests that microbiome analysis tests could be used for early prognosis and diagnosis of respiratory diseases like asthma, COPD, bronchiectasis, and lung cancer (68). Several researchers aim to develop lung cancer biomarkers for improved prognosis, diagnosis, risk classification, and therapy options.

In this section, we provide a brief description of the 16S rRNA gene sequencing (16S rRNA-seq) studies in the biological samples collected from different sites of the respiratory tract. We included studies utilizing the 16S rRNA-seq methods to identify microbiome-based diagnostic and prognostic biomarkers for lung cancer.

A recent study revealed that the salivary microbiota of patients with NSCLC have an increased prevalence of the phylum Firmicutes and its two genera, Veillonella and Streptococcus, compared with the healthy controls (69). They also investigated the correlation between systemic inflammation markers and salivary microbiota. They found a strong association between Veillonella and neutrophil–lymphocyte ratio. However, the specific effects of identified salivary microbiota on the molecular mechanisms of NSCLC progression were not evaluated in this study. This limitation was addressed to some extent in another study that assessed the salivary microbiota profiles in females and matched healthy controls, and it showed salivary dysbiosis with decreased microbial variety richness compared with controls (70). This study found genera Sphingomonas and Blastomonas enrichment in the patients, whereas Acinetobacter and Streptococcus were enriched in the controls. Here, the abundance of Streptococcus in healthy controls contradicts most lung cancer studies (43, 51, 71). To explore the molecular mechanism involved in the salivary microbiota–mediated regulation of lung functions, the authors evaluated different molecular markers of lung function using IHC. They found a significant correlation between Blastomonas and Napsin A, which is an IHC marker used in clinical diagnostics of lung cancer. Yan and colleagues used the salivary samples from patients with lung cancer and showed significant enrichment of the genera Capnocytophaga, Selenomonas, and Veillonella. They built a machine learning model to evaluate the combination of Capnocytophaga and Veillonella as potential biomarkers for lung cancer screening. They achieved a significant Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.84 in an independent dataset (47). This study also established the possibility of computational methods to predict the association of salivary microbiota with the occurrence of lung cancer.

The saliva samples from the upper respiratory tract (URT) are used for characterizing the microbiota and analyzing their potential as biomarkers but saliva samples could be confounded by the oral contamination. Therefore, the sputum could be a better sample for assessing the microbial diversity of the lung. Moreover, in another study by Cameron and colleagues, the sputum microbiome was examined for its potential as a noninvasive biomarker for diagnosing patients with lung cancer. High proportions of the genus Streptococcus, followed by genus Granulicatella, were observed in patients with lung cancer. These findings suggest the usage of certain specific bacterial species or phylum or genera as diagnostic biomarkers for lung cancer screening; however, their study is limited by sample size and requires validation of the identified genera in a larger cohort (43). Leng and colleagues also highlighted the potential of the sputum microbiome as a noninvasive biomarker for the early detection and classification of NSCLC. They found that Acidovorax and Veillonella could be used to diagnose SCC with 89% specificity and 80% sensitivity, whereas Capnocytophaga identified adenocarcinoma with 72% sensitivity and 85% specificity (72). Hosgood and colleagues (44) investigated the association of coal burning exposures with lung cancer microbiota in never-smoking Xuanwei women in China using buccal and sputum samples. They found an enrichment of genus Granulicatella, followed by genus Abiotrophia, in sputum microbiota, whereas that of Streptococcus in the salivary microbiota of patients with lung cancer. Additionally, the sputum samples from individuals exposed to smoky coal for heating and cooking showed higher levels of α-diversity of the lung microbiota compared with those who used smokeless coal. This was a pilot study conducted on a particular ethnicity, and its findings must be validated in different ethnicities, involving a larger sample size. None of the studies mentioned above exploring the sputum microbiome–based biomarkers were validated in the lung tissues.

However, a few studies characterized and explored URT samples in combination with samples from other sites. For instance, Lu and colleagues surveyed sputum samples along with fecal microbiota to understand the association of lung and gut microbiota dysbiosis with NSCLC and their roles in distant metastasis. They found enrichment of Pseudomonas aeruginosa in brain metastasis, the deadliest form of metastasis in NSCLC (73). They also validated Pseudomonas potential for differentiating control and NSCLC through a machine learning model and achieved an AUC of 0.83. Moreover, this study has been registered at ClinicalTrial.gov.

A study explored microbial communities from 75 bronchoalveolar lavage fluid (BALF) samples associated with patients with NSCLC (16). Zeng and colleagues (16) revealed high proportions of the phyla Bacteroidetes and Firmicutes along with genera Streptococcus, Veillonella, and Prevotella. Furthermore, this study identified that Prevotella was strongly associated with the NSCLC subtype. In addition, Veillonella was shown to promote lung cancer progression in an NSCLC mouse model (16). Cheng and colleagues (74) profiled the lung microbiota and predicted their metabolic functions using BALF samples of patients with lung cancer. They found that phylum TM7 and genera Sediminibacterium, Gemmiger, Oscillospira, c: TM7-3, Capnocytophaga, and Blautia were enriched in the patients with lung cancer compared with their healthy counterparts. By including the clinical tumor markers (CEA, CYFRA21-1, and NSE) in the machine learning model, they validated the potential of identified microbiota for classifying control versus cancer cases and achieved a significant AUC of 0.84 (74). Another study explored the BALF samples from patients with lung cancer and benign lung diseases. They found the enrichment of phylum Firmicutes in patients with lung cancer, along with Acidobacterium, SARS202_clade. In addition, Bacteroidota was abundance in samples with benign lung disease. Furthermore, they also built a binary classification model using the identified microbiota and achieved an AUC of 0.98 in distinguishing lung cancer from benign lung disease (75). However, they did not through light on the mechanistic role of the identified bacteria in the host.

Furthermore, Gomes and colleagues analyzed lung microbiota using 103 BALF samples. They found the enrichment of Proteobacteria and more diversity in squamous carcinoma (56). They also investigated the clinical relevance of BALF microbiota and found that Enterobacteriaceae was associated with decreased survival. Lee and colleagues (46) explored microbiota in BALF samples of patients with lung cancer, revealing the presence of phyla TM7 and Firmicutes and genera Megasphaera and Veillonella. They showed that smokers had a substantially higher ratio of phylum Firmicutes to Bacteroidetes than nonsmokers. They suggested that species belonging to genera Megasphaera and Veillonella could serve as potential biomarkers for screening lung cancer with an AUC of 0.89. However, this study did not explore the functionality of the BALF microbiota in patients with lung cancer, which could have provided better insights into the mechanistic role of microbiota in carcinogenesis.

Another study evaluated lung microbiome metastases in NSCLC types using sputum and bronchial washing fluid (BWF) samples (71). Huang and colleagues found the enrichment of genus Prevotella in BWF samples whereas Streptococcus in sputum samples. In addition, they also investigated differential microbiota associated with histologic types with distant metastatic stages of lung cancer. They found the enrichment of Actinomyces, Veillonella, Arthrobacter, and Megasphaera in adenocarcinoma without metastasis, whereas Rothia and Capnocytophaga were significantly lower in adenocarcinoma with metastasis.

The sputum, saliva, BALF, protected specimen brushing (PSB), and BWF could be confounded by oral contamination. Lung tumor tissue samples are considered most suitable for biomarker studies as they could be more representative in depicting the microbial ecosystem of lungs. Dong and colleagues (58), analyzed the lung tissue microbiota in tumor and normal tissues from 143 Chinese patients. They reported significant diversity of microbes with higher proportions of Actinobacteria, Proteobacteria, Bacteroidetes, and Firmicutes in these patients. Additionally, more significant numbers of bacteria belonging to the Proteobacteria phylum were also observed in normal tissues. They also revealed significant differences in β-diversity in lung cancer and distant normal samples, whereas no differences were observed in α-diversity. The results reported in Dong and colleagues studies were aligned with other reports published by Koreans, Americans, Italians, and Canadians (76–79), and their study suggested that phylum Proteobacteria dominated lung tissue microbiome regardless of different countries. The possibility of contamination was ruled out by using multiple negative controls. Dong and colleagues also found the prevalence of Pseudoxanthomonas, Massilia, and Phenylobacterium in tumor tissue whereas the higher abundance of Anaerococcus, Cupriavidus, and Brevibacillus in normal samples. Many confounding factors are known to affect lung cancer pathogenesis. Among them, smoking is one of the major confounding factors. Smokers had more enrichment of Massilia and Sphingobacterium in their tumor microbiota and less abundance of Acidovorax. Moreover, Massilia, Acidovorax, and Sphingobacterium can degrade polycyclic aromatic hydrocarbon, which is a carcinogen present in cigarette smoke. Although Massilia and Sphingobacterium were shown to protect the host cells against smoke-induced carcinogens, the comparison of microbiota prevalence between tumor and adjacent normal tissues from the smokers was not analyzed in this study (58). Najafi and colleagues (80) conducted a meta-analysis using lung biopsy and found an abundance of the phylum Actinobacteria and lower levels of Lachnoanaerobaculum, Corynebacterium, and Halomonas in cancerous tissue samples compared with adjacent healthy controls. Their study suggests differences in lung microbiota between tumor and normal samples. However, the authors did not rule out the possibility of Halomonas as a contaminant, and it was also found to be consistent with another study reported by D’Alessandro-Gabazza and colleagues (81) on lung tissue of patients with lung cancer.

Profiling and comparing microbiota in intratumoral and adjacent normal tissues revealed an abundance of pathogenic and proinflammatory bacteria in NSCLC compared with healthy lung samples (82). In this study, a protocol for minimizing contamination, starting from sample collection to sequencing at each step, was used. They found the enrichment of genera Ralstonia, Paracoccus, Micrococcus, Diaphorobacter, and Phascolarctobacterium. However, due to the small sample size, their findings need to be validated in a larger cohort. Nejman and colleagues assessed bacterial function in NSCLC and demonstrated the enrichment of the phyla Actinobacteria, Firmicutes, and Proteobacteria. Their related bacterial species are involved in pathways found to be enriched in smokers. They also found that the α-diversity of taxa is typically greater in nonmalignant lung tissues than in tumor lung tissues. Nonmalignant and tumor tissues have similar β-diversity. Nejman and colleagues (77) took great precautions from sampling to analysis to limit the contamination. Another analysis found the differential abundance of the phyla Actinobacteria, Firmicutes, Proteobacteria, Cyanobacteria, and Bacteroidetes in the microbiome of FFPE lung tissue samples. They also found that Cyanobacteria-derived toxins such as microcystin increase lung cancer development and progression by PARP1 overexpression (55). Another interesting study by Liu and colleagues (51) examined the lung cancer microbiota from paired tumor tissues of 24 patients and 18 healthy individuals. Liu and colleagues reported a significant abundance of Streptococcus in patients with lung cancer, whereas Staphylococcus had higher levels in controls. In addition, decreased microbial diversity was observed in patients with lung cancer compared with controls. They also built a prediction model based on Streptococcus with an AUC of 0.69, showing its potential for classification (51). Yu and colleagues conducted a study on 165 patients and revealed an abundance of phyla Proteobacteria, Firmicutes, and Bacteroidetes. Yu and colleagues (50) found an abundance of the genus Thermus in patients with advanced stages IIIB and IV. In contrast, Legionella was found to be high in metastasizing patients, suggesting its role in lung cancer progression and development via multiple pathways. Moreover, adenocarcinoma displayed a higher abundance of the genus Thermus with increased phylogenetic diversity than SCLC, suggesting that it may be linked to cancer histology.

In order to better understand the association of microbiota identified in different types of samples with the lung tissue samples, we summarize the microbiota identified in different samples collected from URT and lower respiratory tract (LRT), as given in Table 1. It shows the microbiota differences residing in URT and LRT of the lung. The URT microbiota of the lung were collected mainly from saliva and sputum samples. On the other hand, the LRT microbiota were collected from lung tissue either from a resected tumor or through bronchoscopy, BALF, PSB, and airway brushing. Patients with lung cancer exhibit significant heterogeneity in factors such as age, smoking history, genetics, and overall health. These individual variations can influence the lung microbiome (74, 83). Furthermore, the lung microbiome can be influenced by geographic location and environmental conditions (83). The lung microbiome may vary depending on the stage and type of lung cancer. The microenvironment around tumors can be distinct, influencing microbial colonization (66). Prior antibiotic use or ongoing treatments, such as chemotherapy or radiotherapy, can alter the microbial composition in the lungs. These treatments may selectively eliminate certain bacteria, affecting the overall microbiome (84).

Taken together, the increasing 16S rRNA-seq provides a valuable approach to exploring lung microbiome, particularly the bacterial composition, its richness, and diversity in patients with lung cancer and healthy controls. The studies revealed that microbiota dysbiosis was central to lung cancer progression and development, as shown in Table 2. However, the mentioned studies in Table 2 have their limitations. First, 16S rRNA predominantly detects bacteria at the genus level, providing limited information about the specific species. Moreover, identifying different species becomes difficult because of the sequence homology among closely related organisms. This limitation compromises the precision of identifying bacterial species within the lung microbiome. In addition, sequencing hypervariable subregions, such as V3–V4, produces relatively short sequences (100–500 nucleotides) in comparison with the entire sequence of the gene. This can lead to missing variations and failure to capture the complete diversity of the microbial population. Furthermore, 16S rRNA-seq lacks functional information about the identified microorganisms, such as antibiotic resistance or virulence genes. This impedes a thorough comprehension of the possible functions of the detected microorganisms. Therefore, in view of these limitations, additional studies involving techniques like whole-genome sequencing, shotgun metagenomics, metatranscriptomics, and metabolomics are required for better identification and capturing functional information of bacterial species associated with lung cancer (85, 86). Moreover, bulk sequencing methods are unable to capture the rare population of bacteria.

Significant progress has been made in the past several years to elucidate the lung microbiome in lung cancer as a promising target for diagnostics, therapeutics, and prognostication (87). High-throughput sequencing has revolutionized lung microbiota studies. With the advancements in sequencing technologies, the complex and unique compositions of microbial communities in lungs from the LRT and URT have been explored for therapeutic interventions in lung cancer. Compared with microbiota in other mucosal sites, such as the intestine, lung microbiota knowledge remains rather limited. Although current data from research studies provide a shallow understanding of the lung microbiota, several conceptual questions remain unanswered.

There are several areas of gaps existing in the lung microbiome and lung cancer research. First and foremost, understanding the microbiome in the functioning of healthy lungs is crucial to gain insights into host–microbiome interactions in normal individuals. Second, establishing a causative relationship between respiratory microbiota and lung cancer has been most difficult in the past. We also discussed possible microbiome-mediated lung carcinogenesis mechanisms, including microbiome dysbiosis and genomic instability, affecting metabolism, induction of inflammatory pathways, and immune response in the host. Although dysbiosis-induced molecular pathways have been explored, the feedback mechanism between lung microbiome and tumor cells is still ambiguous.

Third, disruption of homeostasis of microbial composition could lead to reduced diversity and expansion of pathobionts, which strongly associates with the risk of lung cancer. A variety of pilot studies have established the importance of unique species of bacteria as biomarkers of lung cancer diagnosis. However, longitudinal and cross-sectional studies with larger sample sizes would be essential to build insightful therapeutic strategies for modulating pulmonary microbiota to diagnose lung cancer. In addition, microbiome-based biomarkers faithfully predict response to modern immune checkpoint inhibitors. Future studies emphasizing microbiome-based biomarkers might help design better immune-based therapies to enhance survival of patients with lung cancer.

Altogether, the microbiome in the lung regulates its development, homeostasis, malignant transformation, and response to therapy. This knowledge can be harnessed to develop better microbiome-based biomarkers for early detection and predictive biomarkers for immunotherapy.

Although various studies have evaluated the microbiome of lung cancer, we still do not have a reliable microbiome-based diagnostic biomarker. The possible reasons for disparities in microbiome composition in individual studies and in combination with others could be attributed to the difference in sampling and sequencing methodology and the choice of the hypervariable region of the 16S rRNA gene. The custom bioinformatics pipeline used could also lead to a difference in the identified microbiota. In addition, databases like SILVA and Greengenes used for identifying the microbiota population also contribute to these differences (88).

Respiratory tract specimen requires special consideration during sample collection, processing, and analysis as contamination is the major concern in research involving lung microbiomes. Therefore, emphasis should be placed on maximizing input biomass through the utilization of as large a sample size as possible, which ensures the reliability of the identified microbiota. The use of alternative methods, such as cultivation of the diagnostic microbiota in mouse models, should be used to validate the findings. Furthermore, certified DNA-free isolation kits can prevent contamination and maintain the integrity of genetic material for analysis. The development and implementation of certified protocols for the analysis of clinical samples can also limit contamination to some extent. Moreover, an extensive relevant literature search can validate the biological plausibility of microbiota identified in clinical samples.

Very few mouse model–based preclinical studies have validated the therapeutic utility of microbiome in lung cancer. Therefore, future translational studies evaluating the clinical utility of microbiome in lung cancer are required. Although tissue biopsies provide accurate local microbiome information, they are invasive in nature and often not available. Therefore, additional studies are required to analyze the microbiome in different sample types from the same patients. However, contamination of LRT samples with oral secretions is a major limitation. Therefore, the development of advanced BALF collection methods with less oral contamination will help us avoid invasive biopsies for future biomarker studies. A large number of lung cancer cases are diagnosed at advanced disease stages, and these patients are often treated with chemotherapy, which is known to compromise the patient's immune system. It is possible that the compromised immune system of these patients favors a different microbiome compared with chemotherapy-naïve patients. Therefore, more studies should also analyze the microbiome data as per the patient’s treatment status. The day-to-day variation of microbiota in saliva or sputum samples has not been studied, but shifts in microbiota could hold diagnostic or therapeutic value. Therefore, more longitudinal studies are required, enabling more detailed analysis of microbiota, which could shed light on day-to-day microbiota dynamics and their potential clinical significance.

No disclosures were reported.

P. Yadav received seed grant (project number I/SEED/PY/20200037) from Indian Institute of Technology Jodhpur; D.K. Ahirwar received grant (SRG/2023/002556) from the Department of Science and Technology, Science and Engineering Research Board, and US Department of Defense (DOD) with grant reference numbers W81XWH-22-1-0001 and W81XWH-22-1-0038.