High scores!

Lower respiratory tract microflora affects the progress of lung cancer.
Hello everyone.
Today, the editor brings you a high-scoring article published in Cancer Discov.
(JCR Division 1, Chinese Academy of Sciences, IF: 39.
397) this year.
Combine, complement each other, let's enjoy the research results of the masters of learning! Research background and significance Lung cancer is the leading cause of cancer-related deaths worldwide, and its morbidity and mortality rates remain high
.

Targeted therapy can improve the survival rate of patients, but it is only suitable for about 30% of patients with lung adenocarcinoma
.

New immunotherapy treatments have been shown to affect the response of T cells in patients to tumor antigens and improve the survival rate of lung cancer patients.
However, 40% to 60% of patients are still ineffective or resistant to them
.

Existing studies have identified intestinal microbes associated with enhanced anti-tumor immunity and response to immunotherapy in mouse experiments and analysis of cancer patients who responded to immunotherapy
.

While most of these studies focus on intestinal microbes, the correlation between respiratory microbes and lung cancer is rarely mentioned
.

At the same time, studies have shown that the lower respiratory tract of normal individuals usually contains oral bacteria such as Prevotella and Wechella.
It is related to exacerbation of inflammation, and such dysbiosis may trigger changes in host transcriptome signals (PI3K and MAPK)
.

In order to further explore the clinical significance of respiratory microbes in lung cancer, the authors used prospective population cohorts and mouse models to determine the respiratory microbiota that may affect the prognosis of lung cancer
.

Interpretation of results The author’s team recruited 148 subjects with pulmonary nodules from NYU Lung Cancer Biomarker Center between March 2013 and October 2018.
These patients received clinical bronchoscopy.
The lower respiratory tract brush was obtained during the examination for further research, and all samples were obtained before the patient received treatment
.

Among them, 15 patients were non-pulmonary primary tumors (metastasis), 12 patients were benign lung nodules, 38 patients were excluded from the study cohort due to other benign diagnoses, and the remaining 83 patients were ultimately diagnosed as primary lung cancer.
And be included in this study
.

 Among these patients, there are 16S sequencing data, 70 patients have transcriptome data, 40 patients have RECIST (Response Evaluation Criteria In Solid Tumors) evaluation, 75 patients have more than 6 months of clinical follow-up data, 64 patients There are more than 12 months of clinical follow-up data
.

The clinical information of the sample is shown in the table below
.

Table 1: Research cohort clinical information table of the extraction of microbial characteristics related to lung cancer progression (1) In addition to the lower respiratory tract brush, the author also obtained the upper respiratory tract brush (oral area) and the upper respiratory tract brush (oral area) from the patient’s bronchoscopy.
Bronchoscopic examination of background control samples, these samples were included in the 16S sequencing analysis to be compared together
.

The PCoA analysis based on the Bray-Curtis dissimilarity index showed that the β diversity of the bacterial flora of samples from different parts was significantly different (Figure 1a, PERMANOVA, p<0.
001)
.

In the lower respiratory tract samples, the β diversity of small cell lung cancer (SC) and non-small cell lung cancer (NSCLC) is also significantly different (PERMANOVA, p=0.
01)
.

(2) Bacterial differences in samples based on clinical TNM staging classification of NSCLC.
In NSCLC samples, the author divided the samples into I- The two groups IIIA and IIIB-IV
.

PCoA analysis showed that there was a significant difference in the β diversity of the two groups of samples (Figure 1b, PERMANOVA, p=0.
005)
.

Compared with the I-IIIA group, the respiratory tract samples of the IIIB-IV group are closer to that of the oral samples (Figure 1b, Bray Curtis Distance, p<0.
0001)
.

At the same time, the author also compared the flora characteristics of samples from each stage (I-IV), and the results showed that the more advanced patient samples, the more the bacterial flora composition of the respiratory tract samples was closer to that of the oral samples
.

The MiRKAT analysis showed that the difference in microbial community distribution between NSCLC I-IIIA group and IIIB-IV group was not caused by the difference in sample location
.

(3) Differences in the flora of patient samples with different PD-L1 expression The authors studied in some patients (n=39) where the PD-L1 expression value of tumors can be obtained, and found that compared with patients with low PD-L1 expression (0%) , n=16 and 1–79%, n=11), patients with high PD-L1 expression ((≥80%, n=12), the bacterial composition of respiratory samples is closer to that of oral samples (Supplementary Figure 5, Bray Curtis Distance, p<0.
05)
.

(4) The PCoA analysis of the differences in bacterial populations of patient samples with different survival rates determined that patient samples based on 6-month and 12-month survival rates had significant differences in bacterial composition (Figure 1c, PERMANOVA, p<0.
05), among which, and Compared with patients with better prognosis, the bacterial composition of samples from patients with lower survival rates is closer to that of oral samples (Figure 1c, Bray Curtis Distance, p<0.
0001)
.

The multivariate PERMANOVA analysis showed that the difference in the respiratory tract flora of patients with different survival rates has nothing to do with TNM staging
.

(1)(2)(3)(4) The preliminary results show that there are significant differences in the respiratory tract flora of lung cancer patients in different groups.
The higher the degree of malignancy, the more oral symbiotic bacteria in the respiratory tract flora
.

(5) Mining OTUs related to lung cancer progression.
In the previous study, the author has initially explored the differences in the composition of the respiratory tract flora of patients with different progression of lung cancer.
The author uses DESeq analysis to further find specific differences in OTUs (I-IIIA).
vs.
IIIB-IV groups of NSCLC)
.

Based on the mixed-effects model that corrects the influence of the sample source site factors, the author explored the top20 OTUs (ranked by absolute coefficients) of the difference in abundance between the I-IIIA group and the IIIB-IV group in NSCLC.
The results show that the lower respiratory tract of the IIIB-IV group Significantly enriched OTUs considered to be oral symbiotic bacteria in previous studies, such as Haemophilus, Fusobacterium, Gemini, Prevotella, and Streptococcus granulosus (Supplementary Table 3)
.

In the studies in groups I-IIIA and IIIB-IV, the samples with poor survival rates were significantly more significant than those with better survival rates.
Veillonella, Prevotella, and Streptococcus granulosus Increase (Supplementary Figure 9a–d, Supplementary File Table 2–5)
.

In a further study, the authors determined the top20 OTUs related to the overall survival rate of lung cancer based on a mixed-effect model that corrected the smoking status, TNM staging, and treatment types of patients from the sample source.
The enrichment of oral symbiotic bacteria such as Voorella, Streptococcus, Lactobacillus and Gemini are highly correlated in the lower respiratory tract (Supplementary Table 4)
.

Based on the Dirichlet Multinomial Model (DMM), the author determined that the samples can be divided into two categories: the first category consists of all upper respiratory tract samples and 60% of the lower respiratory tract samples, and the second category consists of all bronchoscopy backgrounds.
The control sample and 40% of the lower respiratory tract samples are composed
.

 The lower respiratory tract samples (60%) that are grouped together with the upper respiratory tract samples are mainly enriched in Veillonella, Streptococcus, Prevotella, and Haemophilus.
The author defines these bacterial groups as SPT (Supraglottic predominant taxa)
.

The lower respiratory tract samples (40%) gathered together with the bronchoscopy background control sample are mainly enriched in Flavobacterium and Pseudomonas.
The author defines this type of flora as BPT (background predominant taxa) (Supplementary Figure 12)
.

 Compared with patients with stage I-IIIA NSCLC, lower respiratory tract samples from patients with stage IIIB-IV NSCLC accounted for a higher proportion of samples classified as enriched SPT (Figure 1d, p=0.
006)
.

 The Kaplan-Meier survival analysis showed that in patients with stage I-IIIA NSCLC, the survival time of SPT type was worse than that of BPT type (Figure 1e, p=0.
047).
In patients with stage IIIB-IV NSCLC, SPT type and BPT type There is no statistical difference in survival
.

Furthermore, the authors evaluated the correlation between the flora and clinical RECIST scores in patients with stage IIIB-IV NSCLC.
In this part of patients, the Bray-Curtis dissimilarity index and RECIST scores between upper and lower respiratory tract samples were significantly negative.
Correlation (Spearman r = −0.
48, p=0.
03)
.

Therefore, although the overall survival rate of this part of patients has nothing to do with the classification of SPT and BPT, the result still proves that the RECIST score is positive (tumor progression) and the composition of the lower respiratory tract flora is closer to that of the upper respiratory tract.
In patients with positive RECIST scores, more SPT was enriched in the lower respiratory tract (Supplementary Figure 14)
.

In this part of the results, the author further confirmed the OTUs related to lung cancer progression and defined them as SPT (supraglottic predominant taxa) and BPT (background predominant taxa)
.

Figure 1.
Lung microbiota in lung cancer and cancer survival.
Transcriptome characteristics related to SPT and BPT The author obtained RNA-seq data from the lower respiratory tract samples of 70 patients with NSCLC
.

Different from the microbial characteristics, there is no statistical difference in the Bray-Curtis dissimilarity index between the I-IIIA group and the IIIB-IV group.
The DESeq analysis only found 20 differentially expressed genes between the two groups.
Among patients with survival rate, only a few differentially expressed genes were found (Supplementary File Table 8)
.

Unlike the usual staging grouping, in the sample grouping based on SPT and BPT classification, more differentially expressed genes were mined (Figure 2a, FDR<0.
25)
.

The functional enrichment of these differential genes found that SPT is related to the up-regulation of p53 mutations, PI3K/PTEN, ERK and IL-6/IL-8 and other typical pathways
.

In this part of the results, the author found that SPT and BPT can cause changes in host transcriptome signals.
Multi-omics analysis In order to better explore the host/microbe interaction in lung cancer, the author uses a multi-omics analysis framework (co-occurrence probability network) , In order to evaluate the probability of the co-occurrence of the transcriptome characteristics of the flora and the host, and at the same time use MMvec to calculate the probability of the flora appearing in the I-IIIA advanced stage and the IIIB-IV advanced stage, and add it to the network (Figure 2c).
Finally, the author used the decontam software package to remove noise, and finally confirmed that Veillonella (OTU#585419) is the OTUs most relevant to the progression of lung cancer.
It is also related to cell adhesion molecules, IL-17, cytokines and growth.
Factors, chemokine signaling pathways, TNF, Jak-STAT, PI3K-Akt signaling pathways are highly correlated (Supplementary File Table 10)
.

Based on BLAST, the OTUs were finally compared and annotated as Veillonella parvula
.

In this part of the results, the author finally confirmed that Veillonella parvula is the most relevant bacteria in the respiratory tract flora that is associated with lung cancer progression.
Figure 2.
Airway transcriptome in NSCLC lung cancer based on lung microbiota.
To verify the effect of lower respiratory tract dysbiosis caused by Veillonella parvula on the progression of lung cancer (KP model mice, Figure 3a)
.

The results showed that the lower respiratory tract dysbiosis caused by Veillonella parvula did not affect the survival and weight of wild-type mice, while the lower respiratory tract dysbiosis caused by Veillonella parvula caused the survival time, weight loss, and tumor burden of the KP model lung cancer mice to decrease (Figure 3a, 3b, Supplementary Figure 18a,b)
.

At the same time, the author did a repeated experiment in the third week after the induction of dysbiosis, using transcriptome data, FACS-based T cell profile data and cytokine measurement data to evaluate the model’s immune response to dysbiosis: Ⅰ.
PCoA analysis of transcriptome data The results show that there are significant differences between the four experimental conditions (WT, Dys, LC, LC+Dys), and the transcriptome characteristics of the LC+Dys group have greater changes than the LC group (Supplementary Figure 19a)
.

Ⅱ.
Based on CIBERSORT's tumor immune microenvironment analysis, it is clear that the lower respiratory tract dysbiosis leads to the increase of Th1 cells and the activation of dendritic cells (Supplementary Figure 19b)
.

Ⅲ.
IPA analysis results show that lower respiratory tract dysbiosis also leads to the up-regulation of PI3k/Akt, ERK/MAPK, IL-17A, IL-6/IL-8 and inflammasome signaling pathways (Figure 3c)
.

When comparing the transcriptome signals induced by lower respiratory airway dysbiosis in the mouse model with the transcriptome signals recognized in SPT-type NSCLC patients, the author found that IL-17 signals, chemokines, TOLL-like receptors, and PD-L1 signals The changes in transcriptome characteristic signals related to PI3K-Akt signal are consistent
.

In the end, the authors concluded that the lower respiratory tract dysbiosis caused by Veillonella parvula resulted in an increase in Th17 cells, an increase in IL-17 levels, an increase in PD-1+ T cell expression, and an increase in neutrophils (Figure 3e, Supplementary Figure 22a)
.

The immunohistochemical results of CD4+, CD8+ and neutrophils showed that the increase of these inflammatory cells mainly occurred in lung tissues without tumor invasion (Figure 3e)
.

Figure 3.
Pre-clinical model of lung dysbiosis in lung cancer and cancer survival.
To further evaluate the importance of dysbiosis-induced IL-17 activation in lung cancer progression, anti-IL-17 monoclonal was used after tumorigenesis in the KP mouse model The antibody and isotype control antibody were used to treat the mice for two weeks (Figure 4a)
.

The results showed that compared with the control group, the tumor burden of mice treated with anti-IL-17 monoclonal antibody was significantly reduced (p=0.
0059, Figure 4b)
.

Immune microenvironment analysis showed that the expression of RORγt+CD4+ T cells and neutrophils was significantly reduced in mice treated with anti-IL-17 monoclonal antibody.
The results of immunohistochemistry showed that the reduction of these inflammatory cells occurred in the absence of tumor invasion.
In lung tissue, but not in lung cancer tissue (Figure 4c, 4d)
.

Figure 4.
IL-17 blockade during lung dysbiosis in lung cancer preclinical model.
In this part of the results, the author verifies that the lower respiratory tract dysbiosis caused by Veillonella parvula can promote the inflammatory microenvironment of lung cancer, which is manifested by the increase of Th17 cells and the potential for activation Dendritic cells with antigen-presenting ability and increase in checkpoint inhibitor markers in adjacent tissues
.

The editor concludes that in this study, the author first proposed and confirmed that the imbalance of the lower respiratory tract flora of lung cancer patients will change the host's transcriptome signaling pathway and promote the tumor inflammatory microenvironment, thereby affecting the occurrence and development of lung cancer and patient prognosis
.

After reading this article throughout, the editor can't help feeling the heavy workload of this high-scoring article, the subtle handling of the details, the contextual structure echoes, the combination of dry and wet, and the smooth logic.
I suggest you download the original text to study and learn! If you have questions about Shengxin, please contact: 18501230653 (same number on WeChat) Welcome to follow Shengxin human transcriptome | methylation | resequencing | single cell | m6A | multiomics cytoscape | limma | WGCNA | water bear legend | linux electrophoresis | PCR | A Brief History of Sequencing | Karyotype | NIPT | Basic Experimental Genes | 2019-nCoV | Enrichment Analysis | Joint Analysis | Microenvironmental Plague Hunting | Summary of Ideas | Scholars | Scientific Research | Retraction | PhD Reading | Work