Lung cancer has the characteristics of high morbidity and high mortality, which seriously endangers human health and quality of life. As the leading cause of cancer death, there would be an estimated 2.2 million new cases and 1.79 million deaths worldwide in 2020 (Sung et al., 2021). Many early-stage lung cancer patients have no symptoms or only have mild symptoms with no specificity, making it difficult to attract attention. However, the diagnosis of lung cancer requires complicated procedures such as pathological biopsy and imaging examinations, and there is a lack of simple bedside detection methods. Therefore, many lung cancer patients are often in the middle or advanced stages when they are diagnosed, so there are always different degrees of cancer metastasis, leading to a poor prognosis (Rudin et al., 2021). The current non-invasive diagnostic method for lung cancer is mainly liquid biopsy. There is a small amount of circulating tumor cells (CTC) shed from the tumor site in the blood of cancer patients, and necrotic cancer cells release a small amount of circulating tumor DNA (ctDNA) into the blood; therefore, it can help to judge the occurrence of cancer by detecting CTC and ctDNA. In addition, non-invasive biomarkers based on protein and microRNA have also been widely studied (I and Cho, 2015; Farooq and Herman, 2020). In recent years, a growing number of studies have discovered the connection between intestinal flora and disease diagnosis, treatment, and prognosis. At present, many studies involve the sequencing of gut microbes in lung cancer patients, and Zheng et al. (2020) have developed an operational taxonomic unit (OTU)-based prediction model for the early diagnosis of lung cancer. However, relying on single omics for prediction does not seem to be sufficient, and so far, the connection and interaction between the gut microbiome and metabolome of lung cancer patients with different stages have not been recorded.

A cohort study found that the diversity of intestinal flora was at a similar level in healthy children and adults, but the composition and function of the microbiome were different (Wen et al., 2019). Herein, we assume that there are differences in the composition or diversity of the intestinal flora of lung cancer patients with different stages. We recruited 30 different-stage lung cancer patients and 15 healthy individuals and performed the corresponding detection and analysis on their stool and serum specimens. The composition of the intestinal flora and serum metabolites was compared by bioinformatics analysis. We are trying to combine microbiology and metabolomics to find out the pathogenesis of lung cancer and potential biomarkers, so as to provide new insights for the diagnosis and treatment of lung cancer in the future.

Our research has proved the changes of intestinal microbiota and serum metabolic spectrum in lung cancer patients. We combined the two omics to find the possible pathogenesis and potential biomarkers of lung cancer.

The control group consisted of individuals undergoing physical examination from the Health Management Center, while lung cancer patients were newly diagnosed without treatment. Medication use is an important consideration when we screen study subjects, which is mainly because any medication, such as antibiotics and anticancer medication, will cause changes in intestinal flora to varying degrees. We have observed differences in the composition and structure of gut microbes between lung cancer patients with different stages and healthy people. Faecalibacterium has a higher abundance in HC, and its relative and absolute abundances in ESLC and NESLC decreased sequentially. Faecalibacterium is considered to be a marker of healthy intestines and is a type of butyrate-producing bacteria (Wang et al., 2007). In previous studies, it was also found that its abundance decreased in lung cancer patients (Gui et al., 2020; Zheng et al., 2020). In addition, ESLC has abundant Klebsiella and Escherichia–Shigella, while NESLC has abundant Prevotella_9 and Escherichia–Shigella. Klebsiella, Escherichia–Shigella, and Prevotella_9 all contain many opportunistic pathogens. The latest research has linked the increase in Prevotella abundance with local and systemic diseases, including rheumatoid arthritis, hypertension, and metabolic disorders (Horta-Baas et al., 2017; Li et al., 2017; Ding et al., 2020).

Previous studies have shown that the changes of intestinal flora in patients with lung cancer are often related to Firmicutes, Bacteroides, and Proteobacteria (Zhang et al., 2018; Liu et al., 2019; Zhuang et al., 2019; Zheng et al., 2020; Georgiou et al., 2021). When using the Kruskal–Wallis algorithm to further identify microbes with different abundances, it was found that the abundance of Firmicutes decreased and Proteobacteria increased in the lung cancer group. Compared with previous studies, the change of Firmicutes was consistent and Proteobacteria was also consistent with most studies, which only decreased in the study of Zhang et al. However, Bacteroides has been found to increase in lung cancer, while Actinomycetes has been found to decrease (Zhuang et al., 2019; Zheng et al., 2020), which was not found in our study, even in the comparison between HC and NESLC. We try to find the reasons for the differences from the included population and experimental design. The study of Zheng et al. included patients with early-stage lung cancer. Although one of its exclusion criteria was antibiotic use in the last 8 weeks, the use of anticancer drugs was not specified. In the study of Zhuang et al., subjects did not use any drugs in the past 3 months and seemed to have not received chemotherapy, but it was uncertain whether there were therapeutic behaviors other than chemotherapy. Our subjects were all newly diagnosed and treated patients, so we consider that the difference may be due to the selection criteria of subjects or the heterogeneity of patients. It is worth mentioning that although no difference was found at the level of Bacteroides and Actinomycota, differential genera mined such as Roseburia, Lachnospira, Olsenella, and Cryptobacterium belonged to Bacteroides or Actinomycota.

When looking for key differential microorganisms, it was found that Anaerostipes, Lachnoclostridium, Roseburia, and Lachnospira were the key differential genera in HC. These genera are closely related to the production of short-chain fatty acids (Liu et al., 2019; Chen et al., 2021; Yi et al., 2021). Short-chain fatty acids play an important role in human health because they regulate the function and differentiation of almost all intestinal immune cells (Surono et al., 2021). It has been found that short-chain fatty acids maintain intestinal homeostasis by promoting the production of IL-10 in Th1 cells (Sun et al., 2018). Xia et al. (2020) made preparations from plant flower buds to induce SCFA-producing bacteria to produce SCFAs to achieve an anticancer effect. Correspondingly, the key differential genus in ESLC was Lactobacillus. Lactobacillus itself is a kind of beneficial bacteria, which acts in preventing pathogenic bacteria from invading and colonizing the intestines, enhances the body’s immunity, and has anticancer effects (Hendler and Zhang, 2018; Eslami et al., 2019; Han et al., 2021). In colorectal cancer, Lactobacillus showed reduced abundance (Loftus et al., 2021). However, in a large cohort study, it was found that the increased abundance of Lactobacillus bacteria in the oral cavity was closely related to the risk of lung cancer (Hosgood et al., 2021). The key differential genus in NESLC was Escherichia_Shigella. The increased abundance of Escherichia/Shigella was considered to be a characteristic of intestinal flora imbalance in Crohn’s disease (Pascal et al., 2017). These indicate that compared with HC, patients with early/advanced lung cancer have different degrees of intestinal flora imbalance. Random forest algorithm, a kind of machine learning algorithm, was used to screen important microorganisms that distinguish differences between groups. The screened genera were mostly the same as the key genera in each group. However, it should be noted that due to the hierarchical structure of the algorithm, there may be no linear relationship between the filtered features and the output. For example, Pseudomonas, although selected as the most important genus, was not a differential genus, and the difference between groups was not statistically significant.

Pathway enrichment analysis was performed using differential metabolites to understand the mechanism of metabolic pathway changes in different samples. Results showed that the aminoacyl-tRNA biosynthesis pathway was enriched in HC vs. ESLC, HC vs. NESLC, and ESLC vs. ESLC, which seemed to be closely related to the progression of lung cancer. In previous studies, the aminoacyl-tRNA biosynthesis pathway was an enrichment pathway for UAP1 expression-related genes in lung adenocarcinoma (Wang et al., 2020). It was closely related to cisplatin resistance in non-small cell lung cancer (Shi et al., 2019). Moreover, it has also been found to be upregulated in gastric cancer, and researchers even proposed a new therapeutic strategy for gastric cancer targeting the aminoacyl-tRNA biosynthesis pathway (Gao et al., 2021). Differential metabolites in this pathway, L-lysine, L-isoleucine, and L-histidine, were significantly reduced in the lung cancer groups, but there was no significant difference between ESLC and NESLC, which meant that these indicators could not be used as reference indicators for staging. L-glutamate was only significantly different between ESLC and NESLC; thus, it was not the best predictor. The serum concentration of L-valine showed differences in the pairwise comparison, so it may be used as a potential marker. In the correlation analysis, we found that the most strongly related genus of L-valine was Lachnospiraceae_UCG-006, which was the differential genus we screened earlier. The level of valine in early-stage non-small cell lung cancer was lower than that in advanced stage (Puchades-Carrasco et al., 2016). In the study of Ni et al. (2019), valine declined in the lung cancer group as a potential biomarker, although this research was based on two existing data sets. In addition, valine has been reported as a potential biomarker for the differential diagnosis of seronegative rheumatoid arthritis and psoriatic arthritis (Souto-Carneiro et al., 2020). Upregulated concentrations of branched-chain amino acids were detected in stool samples from colorectal cancer patients and in gastric tissue fluid from mice with gastric cancer (Yachida et al., 2019; Gu et al., 2020).

It is worth mentioning that although this study describes the changes in the microbiome and metabolome in lung cancer and the relationship between the two, we cannot explain the causal relationship between them. This requires more and larger cohort studies to explore in the future.

This study has some limitations. Prognosis is an important part of disease research. However, due to the inability to track the prognosis of all patients in the short term, we are temporarily unable to study the prognosis of the disease. On the other hand, the strict entry conditions lead to a small number of participants. We hope to expand the sample size and conduct further studies through multicenter cooperation in our future research.

This study describes the characteristics of intestinal flora and serum metabolic profiles of patients with lung cancer in different stages. It reveals that lung cancer may be the result of the mutual regulation of L-valine and Lachnospiraceae_UCG-006 through the aminoacyl-tRNA biosynthesis pathway, and proposes that L-valine may be a potential marker for the diagnosis of lung cancer.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://data.mendeley.com/datasets/8rftx9ybnm/1, genomic sequencing; https://data.mendeley.com/datasets/nj4cz7mmj5/1, metabolomics.

The studies involving human participants were reviewed and approved by the IRB of The Third Xiangya Hospital of Central South University. The patients/participants provided their written informed consent to participate in this study.

Design of the study: SC, RG, and Y-FF. Methodology: SC and X-HZ. Formal analysis: SC and X-HZ. Data curation: SC and X-HZ. Software: SC. Writing—original draft preparation: SC and X-HZ. Writing—review and editing: J-HZ, H-YJ, H-TL, SC, RG, and Y-FF. All authors contributed to the article and approved the submitted version.

This work was supported by the National Natural Science Foundation of China (Grant No. 82002452), and the Fundamental Research Funds for the Central Universities of Central South University (Grant No. 2021zzts1090).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.