A total of 53 LADC samples were included in this study. These patients were admitted to the hospital and underwent surgery between January 2014 and December 2015. None of the patients in this study received antibiotics or chemotherapy before surgery. We collected information on gender, age, Primary tumor dimension, Lymphatic metastasis, clinical stage and followed the patient’s survival as of January 17, 2021. The mean age of the total cohort was 62.2 years, with 50.6% men and 41.5% smokers. The proportion of men in the ever-smokers group was higher than in the never-smokers group (94.1 vs. 31.9) because women rarely smoke. For primary tumor size, 35 patient tumor size is less than 3 cm, 18 patient tumor size is higher than 3 cm. In total, 16 samples belonged to the I–II stages and 6 samples to the III–IV stages in ever-smokers, while 24 samples belonged to the I–II stages and 7 samples to the III–IV stages in never-smokers (Table 1). According to the survival curve of the patients, we found that smoke might be a negative prognostic factor in LADC patients, but this was not statistically significant (Supplementary Figure S1A).

To analyze the composition of the tissue microbial profile, we successfully sequenced the 16S rRNA gene in Formalin-Fixed Paraffin-Embedded (FFPE) tissue samples from ever-smokers (n = 22) or never-smokers (n = 31) LADC. Firstly, we found alpha diversity metrics including Shannon, Chao1, ACE, and Simpson indexes according to amplicon sequence variants (ASVs) were comparable between two groups (Supplementary Figure S1B). We also visualized the Bray-Curtis distance of beta diversity by principal coordinate analysis (PCoA) between ever-smokers and never-smokers (p > 0.05, Supplementary Figure S1C). But we found that Chao1 and ACE index between primary tumor size ≤3 cm and > 3 cm group is significant (Figure 1A).

Then, we conducted the taxonomic assignment to build the microbial abundance profile. The two groups were both enriched by Proteobacteria, as the dominant phylum, followed by Actinobacteriota, Firmicutes and Bacteroidetes (Supplementary Figure S2), which was in consistent with the previous report (Nejman et al., 2020). We observed a significant increase of Acidobacteriota (p = 0.046912) in LADC tissues of never-smokers than ever-smokers. We identified in total 60 bacterial classes, 25 orders, 30 families, and 35 genera with a relative abundance higher than 0.1%. We also performed the analysis at class, order, family, and genera levels (Supplementary Figures S3, S4). At the genus level, the tumor microbiota of LADC in the ever-smokers was enriched with Escherichia_Shigella and Sphingomonas, but did not reach statistical significance (Supplementary Figure S5). While Stenotrophomonas, Ralstonia, and Corynebacterium were identified to be enriched in the never-smokers. Among them, Stenotrophomonas was more than 3.9 fold higher in lung tumors of never-smokers (p < 0.05, Figure 1B). We also found a significant increase of Stenotrophomonas in LADC tissues of primary tumor size >3 cm than ≤3 cm groups (Figure 1C).

Next, we used the LEfSe algorithm to identify the critical bacterial taxa that account for the differences between smoking status or primary tumor size. LEfSe estimates the effect size of significantly different abundant of taxa and ranks them represented by the length of each bar (Segata et al., 2011). We found family Verrucomicrobiaceae and its higher taxonomies, including class Verrucomicrobiae and order Verrucomicrobiales, were significantly enriched in ever-smokers (Figure 1D). There was an increase in the family Burkholderiaceae in never-smokers compared with ever-smokers. Moreover, LADC microbiota of never-smokers was enriched in Acidobacteriota and Stenotrophomonas, in line with phylum and genus level observations (Figure 1D). We also found that the genus Stenotrophomonas was enriched in the group of primary tumor size higher than 3 cm patients through LEfSe analysis (Figure 1E). There was a linear correlation between the relative abundance of Stenotrophomonas and primary tumor size (Supplementary Figure S6). To validate the presence of Stenotrophomonas in LADC tissues, we used FISH with a specific probe against its 16S rRNA. We found the presence of Stenotrophomonas DNA within LADC tissues in never-smokers, but not in that of ever-smokers (Figure 1F). These results indicates that Stenotrophomonas could be identified as biomarker in LADC.

Although various studies have revealed that microbiome contributes to tumor induction and progression (Dapito et al., 2012; Li et al., 2016, 2019; Le Noci et al., 2018; Jin et al., 2019; Tsay et al., 2020; Fu et al., 2022), the causal relationship between Stenotrophomonas and lung cancer remains largely unknown. S. maltophilia, the only species of Stenotrophomonas that infects humans, is considered a “newly emerging pathogen of concern” (An and Berg, 2018). To investigate whether S. maltophilia could colonize in the lung, we harvested lung lobes and cultured the lung homogenate in the nutrient broth after 48 h of bronchial injection of S. maltophilia in mice. We visualized and validated the colonizing ability of S. maltophilia in lung lobes by 16S rRNA gene sequencing (Supplementary Figure S7).

To explore whether S. maltophilia contributes to lung cancer progression, we compared tumor development between mice intratracheal inoculated with S. maltophilia and their control A/J mice in lung cancer model induced by NNK (Akopyan and Bonavida, 2006; Figure 2A). We found that NNK treatment affected the weight gain in mice, but not S. maltophilia treatment (Supplementary Figure S8). Compared with the NNK model group, NNK_ S. maltophilia group displayed heavy tumor burden, with the increased number of lung nodules and area of tumor (Figures 2B,C). At the cellular level, although NNK_ S. maltophilia group did not affect P53 expression, it exhibited increased tumor cell proliferation as demonstrated by immunohistochemical (IHC) analysis of Ki-67 staining (Figure 2C). Together, these results indicate that S. maltophilia plays a profound role in promoting tumor development in NNK induced lung cancer model.

To investigate the underlying mechanisms of S. maltophilia on lung cancer progression, we characterized the transcriptional profile of lung tissue using RNA sequencing (RNA-Seq). Constrained Principal Co-ordinates Analysis (CPCoA) clustering analysis showed that the CTL and NNK groups were clearly separate, while two S. maltophilia treatment groups were relatively close to each other (Figure 3A). These results indicate that the NNK and S. maltophilia treatment groups had distant gene expression signatures.

Gene set enrichment analysis (GSEA) revealed 34 overrepresented gene sets in S. maltophilia group compared to CTL group based on normalized gene set enrichment scores (NES; Supplementary Figure S9). Among them, various inflammatory pathways, including B cell receptor signaling pathway, Th17 cell differentiation, NF-kappa B signaling pathway, as well as regulation of actin cytoskeleton were significantly enriched in S. maltophilia group (Figures 3B–F). Sixteen gene sets were enriched in NNK_ S. maltophilia group compared with NNK group (Supplementary Figure S10). Interestingly, metabolism associated with tumor reprogramming, including propanoate metabolism, citrate cycle (TCA cycle), carbon metabolism, and fatty acid metabolism, were significantly enriched in NNK_ S. maltophilia group (Figures 3G–I). These data suggest that S. maltophilia may trigger inflammatory response, regulate cytoskeleton, and promote cancer metabolism.

Next, differentially expressed genes (DEGs) were identified through pairwise comparisons of groups. Compared with the CTL group, the number of DEGs in the S. maltophilia, NNK, and NNK_S. maltophilia groups was 3,786 (1832/1954; upregulated and downregulated DEGs, respectively), 3,373 (1,335/2038), and 2,283 (1,163/1120) respectively (Figures 3J–L; Supplementary Figure S11; Supplementary Table S1). A total of 1,034 DEGs were identified between NNK and NNK_S. maltophilia groups, among them 564 genes were upregulated in the lung tissues of NNK_S. maltophilia group and 470 genes were downregulated (Figure 3K). NNK_S. maltophilia group and S. maltophilia groups displayed similar gene expression profiles, and only 286 DEGs (the lowest number among all comparisons) were identified between these two groups (Supplementary Table S1). Additionally, we were able to identify 322 common DEGs between the CTL-vs-S. maltophilia, and NNK-vs-NNK_S. maltophilia (Supplementary Figure S12; Supplementary Table S2).

We next calculated the specific KEGG pathway enrichment of upregulated DEGs in S. maltophilia treated groups. Non-small cell lung cancer, Pathways in cancer and Ras signaling pathway enrichment confirmed that NNK induced lung cancer in A/J mice (Supplementary Figure S13). Although S. maltophilia treatment did not trigger lung cancer development in A/J mice, we found that pathway in cancer, MicroRNAs in cancer, Central carbon metabolism in cancer, and Non-small cell lung cancer was enriched in S. maltophilia group compared with CTL group (Figure 4A). KEGG pathway analysis also revealed that mice treated by S. maltophilia turned on many immunity-related signals, including the B cell receptor signaling pathway, Th1 and Th2 cell differentiation, Chemokine signaling pathway, Th17 cell differentiation, T cell receptor signaling pathway, NF-kappa B signaling pathway, and Toll-like Receptor signaling pathway (Figure 4A). Several Toll-like Receptor (TLR) genes were highly upregulated in S. maltophilia groups compared to CTL or NNK group (Figure 4B). Consistent with GSEA, we found propanoate metabolism and fatty acid metabolism was enriched in NNK_ S. maltophilia group compared with NNK group (Figure 4C).

The Apelin signaling pathway was enriched in the S. maltophilia group compared with the CTL group and in NNK_ S. maltophilia group compared with NNK group (Figures 4A,C). We found that 11 genes that belonged to the Apelin signaling pathway were upregulated in NNK_S. maltophilia group (Supplementary Table S3). Interestingly, five of them (Hdac5, Ccn2, Mef2c, Plcb4, Prkag3) were overexpressed in both S. maltophilia and NNK_S. maltophilia group (Figure 4D). Among them, HDAC5 (histone deacetylase 5), as a member of the class IIa family of HDACs, is a well-known oncogene in numerous cancer types, including lung cancer (Zhong et al., 2018; Yang et al., 2021). These results suggest that S. maltophilia may promote lung cancer progression by driving inflammation, regulating tumor cell signaling and metabolism at the transcriptional level.

To explore the possible mechanism of S. maltophilia in promoting LADC in mice, we performed in vitro experiments using cultured A549 cells exposed to the microbial products (Figure 5A). At first, we used three concentrations of bacteria for the in vitro experiment: multiplicity of infection (MOI) of 1, 10, and 100. We found that compared with CTL, 4 h pre-incubation with S. maltophilia_1 and S. maltophilia_2 significantly promoted A549 cell proliferation (Figure 5B). Then to explore the potential factors of S. maltophilia that contribute to tumor cell proliferation, we exposed A549 cells to heat killed or the supernatants from viable bacteria. The result showed that both of them did not promote cell proliferation (Figure 5C). These data suggest that viable S. maltophilia may need to communicate with the cell to stimulate its proliferation.

Stenotrophomonas maltophilia treatment also promoted cell migration in A549 cell lines (Figure 5D). To verify the transcriptome results in mouse lungs, we performed qPCR assays. We found that pre-exposure to S. maltophilia significantly increased the mRNA expression of HDAC5 in A549 cells (Figure 5E). We further confirmed the up-regulated expression of HDAC5 in A549 cells treated with S. maltophilia by western blot (Figure 5F). Considering the tumor-promoting role of HDAC5, we further reasoned if HDAC5 silencing would abrogate the S. maltophilia induced cell proliferation and migration in lung cancer cells. We verified that HDAC5-specific siRNA markedly downregulated its gene expression, as well as inhibited cell proliferation and migration in A549 cells (Figures 6A–C). Further, we found that the cell proliferation and migration stimulated by S. maltophilia were also attenuated by HDAC5 interference (Figures 6A–C). These findings demonstrate that S. maltophilia promotes lung cancer cell proliferation and migration partially through HDAC5.

Formalin-Fixed Paraffin-Embedded (FFPE) tissue samples of 53 patients diagnosed with LADC in the period from Jan 2014 to Dec 2015 were included. All procedures conformed to the Code of Ethics of the World Medical Association (Declaration of Helsinki) and complied with the guidelines of the Institutional Review Board of the Affiliated Hospital of Jiangnan University (IRB: LS2021072).

DNA was extracted from each FFPE sample using the TIANquick FFPE DNA kit (Tiangen, #DP330-02, Beijing, China) according to the manufacturer’s instruction. In addition, we introduced 3 negative controls which are empty tubes that were processed together with the samples for sequencing. All negative controls were processed according to the same protocols.

The DNA was amplified by polymerase chain reaction (PCR) with a bacterial 16S rRNA gene V4 region universe primer pair (341F: 5’-ACT CCT ACG GGA GGC AGC AG-3′ and 806R: 5′-GGA CTA CHV GGG TWT CTA AT-3′; Zhong et al., 2018). Barcoded libraries were generated and the products were sequenced on the HiSeq2500 platform (BGI, Wuhan, China) using the PE300 module. Clean tags were assigned to ASVs using DADA2 (Divisive Amplicon Denoising Algorithm) software in the Quantitative Insights Into Microbial Ecology (QIIME) 2 package, and tags with ≥100% similarity were clustered to the same ASV. Representative ASVs were annotated using Silva_species_assignment_v138 reference database. The ASVs without annotation or annotated to polluted species were removed. Picrust2 (v2.2.0) was used for function predictions.

Common and specific ASVs among groups were compared and displayed in R 3.6.2 with “venn” package (v3.1.1) and Bray-Curtis similarities were calculated in R with “vegan” package (v3.5.1). Alpha diversity was applied to analyze the complexity of species diversity of a sample using Chao 1, ACE, Shannon, and Simpson indices. All indices of our samples were calculated with Mother (v1.31.2). Observed species and Chao 1 were selected to identify community richness, whereas Shannon was used to identify the community diversity. Beta diversity was calculated on both weighted and unweighted UniFrac using QIIME (v1.80). Partial least squares discrimination analysis (PLS-DA) was built using the mixOmics package (v3.2.1) in R. The differential abundance at the phylum, class, order, family, genus, and species levels between groups were performed using LEfSe, with p ≤ 0.01 and LDA ≥ 2.

Stenotrophomonas maltophilia (ATCC #13637 = CGMCC#1.1788 = BCRC#10737 = DSM#50170 = LMG#958 = NBRC#14161 = NCTC#10257) were purchased from the China General Microbiological Culture Collection Center. S. maltophilia were cultured in nutrient broth (Solarbio #n8300, China) at 30°C in an aerobic chamber for 24 to 48 h.

Female A/J mice aged 5 weeks were purchased from GemPharmatech Co., Ltd (Nanjing, China). All mice were housed in a specific-pathogen-free (SPF) environment with access to a standard chow diet and drinking water ad libitum. After a week of acclimation, mice were randomly divided into control group (CTL), bacteria solution (S. maltophilia) group, NNK treatment group (NNK, 100 mg/kg, biweekly for 4 weeks), and NNK + bacteria solution group (NNK_ S. maltophilia; n = 8 per group). The mice were injected intraperitoneally (i.p) or intratracheal inoculation (i.t) at the indicated doses and timepoints (Figure 2A). The intratracheal inoculation of S. maltophilia (1.5 × 10⁶ cfu/mL) were performed as described (DuPage et al., 2009). All mice were treated in accordance to the guidelines of the European Community (Directive 2010/63/EU) and the procedures were approved by the Committee of Ethics in Jiangnan University (#JN. No 20211130a0320401[504]).

The left lobe of the lung from the sacrificed mice was harvested and fixed by 4% paraformaldehyde. After embedding, sections were prepared and stained with hematoxylin–eosin. IHC staining was performed on unstained sections after antigens retrieval with citrate buffer (10 mM Sodium Citrate, 0.05%Tween 20, pH6.0; Jin et al., 2019). The following antibodies were used, including Ki-67 (1: 200 Abcam #ab16667, Cambridge, United Kingdom) and P53 (1: 2000 proteintech #60283-2-Ig, Chicago, United States), secondary antibody (MXB Biotechnologies #KIT-9922, Fuzhou, China). Digitally scanned images of stained slides were created with the Panorama MIDI (3DHISTECH Ltd., Budapest, Hungary).

The A549 cell lines were kindly provided by Stem Cell Bank (Chinese Academy of Sciences). Cell were maintained in RPMI1640 (Gibco #11875093, Waltham, MA, United States) containing 10% FBS (Gibco #10099141C, Waltham, MA, United States), 1% Penicillin–Streptomycin-Amphotericin B Solution (Beyotime #C0224-100 ml, Shanghai, China), cultured in an incubator at 37°C under 95% air and 5%CO₂. Cells were grown and 1 × 10⁶ cells plated in each well (6-well plates with 2 ml of media) and exposed them to different microbial challenges under in vitro experiments. Cells were infected with S. maltophilia (ATCC #13637) at three different multiplicy of infection (MOI) of 1, 10, and 100 for 4 h. Then the cells were washed twice, and incubated with fresh culture medium for 12 h. For heat inactivation, the bacteria were incubated at 60°C for 15 min. Cell viability was detected by a cell counting Kit-8 (Teyebio #TY0312, Shanghai, China) according to the manufacturer’s instruction.

A549 cells were chosen to perform further experiments. A549 cells (2 × 10⁵) were seeded in a 6-well tissue culture plate with 2 mL antibiotic-free RPMI-1640 medium supplemented with 10% FBS. When cells reached 60–80% confluence, the cells were transfected with HDAC5 siRNA, and control vector (Guangzhou RiboBio Co., Ltd., Guangzhou, China) using LipoRNAi™ Transfection Reagent (Beyotime, #C0535, Shanghai, China) according to the manufacturer’s specification. Then, A549 cells were incubated with the compound at 37°C in a CO₂ incubator for 6 h. Following, the transfection mixture was replaced with fresh medium to culture for 48 h. Finally, the A549 cells were assayed using the appropriate protocol.

For Boyden chamber assays, cell suspension (1 × 10⁵cells) was placed into the upper compartment of a 8 μM pore size Transwell chamber (Merck #PTEP24H48, Billerica, MA, United States) in 24-well plate. In each well, serum-free medium was used in the top chamber, while medium containing 10%FBS were used in the bottom chamber. After 24 h culture, cells migrated to the bottom side of the membrane were fixed and stained with the 0.1% crystal violet solution. The migrated cells was counted in five randomly chosen fields per filter from triplicate filters per sample at ×400 magnification. The cell migration index was calculated as the number of cells able to migrate normalized to controls (mean ± SEM).

FISH was conducted with the paraffin-embedded lung tissue according to the protocol described previously (Geller et al., 2017). Briefly, after deparaffinization and rehydration, the slides were incubated with 0.2 M HCl for 12 min at 37°C, and then were incubated with Triton at 37°C for 17 min. Slides were washed twice with PBS and incubated with lysozyme for 15 min at 37°C, and then they were hybridized with the probe specific to ASV4. Before visualization, DAPI was added to the slides. The S. maltophilia probe used for FISH was: GTC GTC CAG TAT CCA CTG C, 5′ modification: Cy3 (Liu et al., 2022). Then the images were captured using Confocal Laser Scanning microscopy (Carl Zeiss Microscopy, United States).

RNAs were extracted from lung tissues and cells using Trizol reagent (Invitrogen, Waltham, United States) and were reversely transcripted into cDNAs. MultiScribe Reverse Transcriptase (Abm #G492, Canada) was used for cDNA synthesis, and then qRT-PCR was performed by using SYBR™ Select Master Mix (Thermo Fisher Scientific, Waltham, United States) according to the manufacturer’s protocols. Gene expression was measured relative to the endogenous reference gene β-actin using the comparative ΔCT method.

Sequences of the specific primer sets are as follows: HDAC5 (NM_001015053.2), forward, 5′-GTG ACA CCG TGT GGA ATG AG-3′; reverse, 5′- AGT CCA CGA TGA GGA CCT TG; β-actin (NM_001101.5), forward, 5’-CTC TTC CAG CCT TCC TTC CT-3′; reverse, 5’-AGC ACT GTG TTG GCG TAC AG-3′.

Cells were lysed in RIPA buffer (Yeasen, Shanghai, China) and followed by 12% SDS-PAGE separation. Separated proteins were transferred onto polyvinylidene difluoride (PVDF) membranes (Merck Millipore, Billerica, MA, United States). The membranes were blocked by 5% bovine albumin in tris-buffered saline plus 0.1% Tween 20 for 1 h at room temperature. The membranes were probed with HDAC5 Antibody (Abcam, #ab55403, Cambridge, United Kingdom) and then probed with anti-rabbit IgG Antibody (Cell Signaling, #7074P2, Danvers, MA, United States). After washing with TBS-T, the membranes were visualized with SuperSignal West Pico PLUS substrate (Thermo Fisher Scientific Inc., Waltham, MA, United States). GAPDH Antibody (Cell Signaling, #5174S, Danvers, MA, United States) was used as a loading control.

The total RNA isolated from each lung tissue sample was applied to RNA-Seq library preparation, by following the protocol described before (Zhu et al., 2018). Briefly, the mRNA was purified by oligo dT-attached magnetic beads and fragmented into small pieces, followed by first-strand and second-strand cDNA synthesis. After PCR amplification, purification, heat denaturation and cyclization, the final cDNA library was sequenced on the BGISEQ-500 platform (BGI-Shenzhen, China) using 50-bp single reads.

The sequencing data was filtered with SOAPnuke (v1.5.2; Li et al., 2008) by (1) Removing reads containing sequencing adapter; (2) Removing reads whose low-quality base ratio (base quality less than or equal to 5) is more than 20%; (3) Removing reads whose unknown base (“N” base) ratio is more than 5%, afterwards clean reads were obtained and stored in FASTQ format. The clean reads were mapped to the reference genome using HISAT2 (v2.0.4; Kim et al., 2015). Bowtie2 (v2.2.5; Langmead and Salzberg, 2012) was applied to align the clean reads to the reference coding gene set, then expression level of gene was calculated by RSEM (v1.2.12; Li and Dewey, 2011). Essentially, differential expression analysis was performed using the DESeq2 (v1.4.5; Love et al., 2014). with |log2 (Fold change)| > 0.4 and Q value ≤0.05. The heatmap was drawn by pheatmap (v1.0.8) according to the gene expression level calculated by RSEM (v1.2.12). GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genome) enrichment analysis of annotated different expressed gene was performed by Phyper based on Hypergeomertric test. The significant levels of terms and pathways were corrected by Q value with a rigorous threshold (Q value ≤0.05) by Bonferroni.

Data are expressed as the mean ± SEM (standard error of the mean). All experiments were repeated at least three times. Student’s two-tailed t-test was used for comparing differences between two groups. One-way or two-way ANOVA followed by Tukey–Kramer’s multiple-comparison test was used for multiple comparisons. Significance was set at p < 0.05. GraphPad Prism software (version 8.0, San Diego, United States) was used for statistical analysis.