Lung gene expression signatures suggest pathogenic links and molecular markers for pulmonary tuberculosis, adenocarcinoma and sarcoidosis

Previous reports have suggested a link between pulmonary tuberculosis (TB), which is caused by Mycobacterium tuberculosis (Mtb), and the development of lung adenocarcinoma (LUAD) and sarcoidosis. Furthermore, these lung diseases share certain clinical similarities that can challenge differential diagnosis in some cases. Here, through comparison of lung transcriptome-derived molecular signatures of TB, LUAD and sarcoidosis patients, we identify certain shared disease-related expression patterns. We also demonstrate that MKI67, an over-expressed gene shared by TB and LUAD, is a key mediator in Mtb-promoted tumor cell proliferation, migration, and invasion. Moreover, we reveal a distinct ossification-related TB lung signature, which may be associated with the activation of the BMP/SMAD/RUNX2 pathway in Mtb-infected macrophages that can restrain mycobacterial survival and promote osteogenic differentiation of mesenchymal stem cells. Taken together, these findings provide novel pathogenic links and potential molecular markers for better understanding and differential diagnosis of pulmonary TB, LUAD and sarcoidosis.

Overall, the manuscript is interesting and presents mostly original findings which would be of interest to the scientific community.

Reviewer #2 (Remarks to the Author):
This study compared and contrasted TB, sarcoidosis, and adenocarcinoma of the lung by examining the transcriptional signatures and histologic features of tissue samples from patients with these diseases. They found that gene expression related to ECM organization (and particularly collagen matrix remodelling), immune response, cell growth and proliferation were enriched among the 3 disease groups. They further explored matrix remodelling by staining for collagen I and collagen III in tissues samples from each disease group, and found increased expression of collagen types I and III.
My main concerns are that it lacks focus and it doesn't develop a story / explore a concrete hypothesis.
This reads as potentially many papers in 1: Distinguishing gene signatures across pulmonary diseases (this has been done previously -so would have to have a novel bent, such as their use of lung tissue rather than blood as the tissue sample source) Ossification in lung disease: mechanistic and morphological features Pulmonary remodelling in TB, sarcoid and AD TB infection and tumorigenesis The background information could be improved: The authors state that, for AD and sarcoidosis, or for AD and TB, "discriminating these diseases could be challenging". This is incorrect. This has not been reported, and as a respirologist, I have not confused the clinical presentations of AD and sarcoidosis, or AD and TB. The references used to support this claim also do not address diagnostic confusion for AD and these granulomatous diseases.

Method concerns:
The "normal control" tissue is non-affected lung tissue from disease patients. The is not a valid control group, unless they wanted to compare, for each patient, disease and non-diseased lung. Also it wasn't clear how many control samples were derived from each disease category. The suppl. data does not identify this.
Some of the findings are an overstretch: "In conclusion, identification of these potential pathogenic mediators of AD or SA shared by TB indicates the potential causal links between TB infection and lung AD or SA." This is unfounded; shared pathways alone does not imply shared aetiology. In the discussion they use the term "pathogenic link" which is more appropriate.
"These findings indicate that Mtb promotes the proliferation, migration and invasion of A549 cells partially depending on MKI67, which mechanism might also benefit its own dissemination within the host." This is a major statement and could be a major finding if they developed a line of investigation around it.
The finding that "all these diseases drove pulmonary remodeling with various degrees of matrix destruction and fibrous protein deposition" is not surprising but has not before been adequately evaluated or definitely established. However, the interpretation that "collagen matrix remodeling in the lung is a common pathogenic characteristic of pulmonary TB, AD and SA." might (and very likely does!) apply more broadly to pulmonary disease. Other disease groups would have been a good control here.
The figures are generally very good. Some were even stunning. But it wasn't always clear why certain data over others was presented. Scale bars are missing for several histology figures. Small point, for reviewers it would be easier to read the figure legends if they were placed near the figure, and if the figures were labeled. The data for the "Transwell migration and invasion assays" in figure 3d is not clearly demonstrated as currently presented.
Minor points: CXR stages in sarcoidosis are not based on the 1999 statement bur rather are Scadding stages No need to abbreviate sarcoidosis (to SA) -this is not standard in the literature. AD is not typically described as or considered to be a "widespread chronic disease". It wasn't clear who read the slides of tissue sections. A pathologist expert in these diseases? With regards to the H37Rv MTb strains used -it would be good practice to acknowledge the source of the strains. The validity of using freshly thawed frozen stock to infect cells is controversial (bacteria are not going to be actively replicating and at a great level of fitness).
Generally, infection occurs in exponential phase of growth.

Reviewer #3 (Remarks to the Author):
Overall this is a substantial and well-presented article that provides important insights into the molecular mechanisms underlying the link between lung calcification and ossification in TB, as well as the identification of potential markers to better distinguish between TB, AD and SA.
However, I do not find the initial data (Figs 1 and 2) particularly novel. It is not surprising to me that three diseases well known to cause significant damage to the lung would have similar molecular signatures related to immune response, metabolic processes and pulmonary remodeling, or that signature genes between TB and AD involved in cell growth and proliferation were most directly correlated with tumor development. Similarly the involvement of MKI67 is not surprising in this context. However, the idea of Mtb directly promoting the proliferation, migration and invasion of human lung adenoma A549 cells is interesting (Fig 3), but the authors use a CCK-8 assay to measure cell proliferation and this only determines cell viability, a proper cell proliferation assay should be performed. This part of the study could also be improved, especially considering the rest of the article is better developed. To add strength to this section it would be beneficial if the authors could show that Mtb-secreted PtpA regulates MKI67 in the A549 cells.
The discussion should include a section "limitations of the study": potential points to discuss: 1) the caveat/strength of normal controls; they are actually derived from the patients. 2) the authors argue consistently that it is very difficult to differentiate between the three diseases, but they by themselves obviously did not have problems doing so; how high is the likelihood of misdiagnosing?

Reviewers' comments: Reviewer #1
Remarks to the author: This is an interesting study that presents a comprehensive comparison of pulmonary TB lung molecular signature with that of lung adenocarcinoma and sarcoidosis, revealing the potential pathogenic links and differences among these diseases. The main aspects of the study involve the transcriptome and comparison between shared and unique molecular signatures of lung adenocarcinoma, sarcoidosis and pulmonary tuberculosis. They revealed that infection with Mtb induces upregulation of MKI67 mRNA in lung adenocarcinoma cell line and that MKI67 is involved in migration and invasion of A549 cells. They showed that MMP8 was specifically increased in TB lungs as compared to the other groups and verified this finding by qPCR and IHC. They presented that BRCA1 and PCNA was specifically increased in lung adenocarcinoma as compared to the other groups, which was also verified by qPCR and IHC. Moreover, they presented that AA metabolism-related genes (including PLA2G6, PLA2G7, AKR1C1, AKR1C3, LTA4H and PTGER4) were positively correlated to sarcoidosis. Furthermore, they show that Mtb activates BMP/SMAD/RUNX2 signaling pathway and induces pulmonary ossification in TB patients. R: We thank the reviewer for the encouraging comments on our manuscript.
Specific comments about manuscript: 1. The study refers to Mtb infection for TB, while for the whole genome analysis only 2 out of 5 had confirmed culture-Mtb assay by 2. Two out of 5 TB analyzed by WGS received treatment before surgery, although the treatment was unsuccessful this is something that should be taken into account when analyzing the microenvironment, which could definitely have an impact on it and should be added to the discussion. R: We thank the reviewer for raising this concern. Actually, in most situations, TB patients would receive anti-TB drug therapy rather than directly adopting surgical treatment, which makes it difficult for us to collect a sufficient number of untreated TB lung samples within a reasonable length of time, thus we included some TB patients who had failed anti-TB drug therapy before tissue sampling, and tissue samples obtained from those patients could not exclude the potential effects of TB drug treatment on TB lung microenvironment. We have pointed out this limitation in the revised "Discussion" section (in the second to the last paragraph).
3. Age was not matched in the study and should be added to the discussion since it could have an impact as well. R: We thank the reviewer for pointing out this issue. Because the peak incidence ages of pulmonary TB and sarcoidosis are relatively younger than that of lung cancers 4. The number of samples used for the validation of expression by qPCR, IHC were not mention for MMP8, BRCA1, PCNA and the AA metabolism related genes.

R: We thank the reviewer for pointing out this issue. We have added the missing number of samples in the legends of revised Figs. 2 and 5 and Supplementary Figs. 4 and 7.
5. They mention lung adenocarcinoma as a chronic disease, which should be changed to terminal one. R: We thank the reviewer for raising this issue, and we have revised the manuscript carefully to avoid describing lung adenocarcinoma as a chronic lung disease.
6. There are discrepancies between supplementary Table 1 and Table 2 regarding the NC; numbers in Signs of Calcification; Signs of fibrosis; The tables should be checked. R: We thank the reviewer for pointing this out. We have carefully revised the content of Supplementary Tables 1 and 2 to make sure that they are correct.
Overall, the manuscript is interesting and presents mostly original findings which would be of interest to the scientific community. R: We thank the reviewer for this encouraging comment.

Reviewer #2
Remarks to the author: This study compared and contrasted TB, sarcoidosis, and adenocarcinoma of the lung by examining the transcriptional signatures and histologic features of tissue samples from patients with these diseases. They found that gene expression related to ECM organization (and particularly collagen matrix remodelling), immune response, cell growth and proliferation were enriched among the 3 disease groups. They further explored matrix remodelling by staining for collagen I and collagen III in tissues samples from each disease group, and found increased expression of collagen types I and III. My main concerns are that it lacks focus and it doesn't develop a story / explore a concrete hypothesis. This reads as potentially many papers in 1: 1) Distinguishing gene signatures across pulmonary diseases (this has been done previously -so would have to have a novel bent, such as their use of lung tissue rather than blood as the tissue sample source); 2) Ossification in lung disease: mechanistic and morphological features Pulmonary remodelling in TB, sarcoid and AD; 3) TB infection and tumorigenesis. R: We thank the reviewer for raising this issue. As described in the first two paragraphs of the manuscript, TB infection has long been linked to lung cancer (such as AD) and sarcoidosis, and the underlying mechanism remains not fully understood. in some cases, which may challenge the diagnosis and delay the treatment. Therefore, in this study, we performed comprehensive analysis of both the lung molecular similarities and differences among TB, AD and sarcoidosis, with a primary aim to unravel both the potential pathogenic links among these diseases and the distinct molecular signatures of them (as is briefly described in line 59-66): For investigation of the pathogenic links, we revealed the molecular similarities among TB, AD and sarcoidosis based on lung transcriptomes (Fig. 1). We then performed molecular experiments using additional samples to further confirm those omics-derived results (Fig. 2). Thereafter, we noticed and focused on an important pathogenic mediator (MKI67) shared by TB and AD to exemplify how Mtb infection could influence tumor cells via this molecule (Fig.  3).
In parallel, for investigation of the pathogenic differences, we revealed the distinct modular signatures of TB, AD and sarcoidosis based on lung transcriptomic data (Fig. 4). We then performed experiments using additional samples to further verify the specific molecular markers derived from modular analysis (Fig. 5). Thereafter, we focused on an important TB-specific lung signature-ossification, to investigate the underlying mechanisms (Fig. 6).
To summarize, this is an interdisciplinary and comprehensive study combining bioinformatic analysis, experimental data and clinical findings to reveal both pathogenic links and differences among three important lung diseases (Supplementary Fig. 12). We have revised the manuscript to make it more logically understandable.
The background information could be improved: The authors state that, for AD and sarcoidosis, or for AD and TB, "discriminating these diseases could be challenging". This is incorrect. This has not been reported, and as a respirologist, I have not confused the clinical presentations of AD and sarcoidosis, or AD and TB. The references used to support this claim also do not address diagnostic confusion for AD and these granulomatous diseases. R: We thank the reviewer for pointing out this issue and we are sorry for making this misunderstanding. Actually, it is not that difficult to differentiate between the three diseases in most cases, but it is difficult to differentiate between these diseases in some cases, especially those share similar clinical characteristics. We have revised our manuscript accordingly. In this study, we selected patients with typical characteristics and definite diagnosis of each disease. Specifically, the included TB, AD or sarcoidosis patients were definitively diagnosed by the specialists in Beijing Chest Hospital according to the diagnostic criteria, and all sample types were verified by a certified pathologist, as described in the "Methods" section (line 440-455). We also added information regarding the likelihood of misdiagnosing of these three lung diseases (line 367-378).
Method concerns: 1. The "normal control" tissue is non-affected lung tissue from disease patients. The is not a valid control group, unless they wanted to compare, for each patient, disease and non-diseased lung. Also it wasn't clear how many control samples were derived from each disease category. The suppl. data does not identify this. R: We thank the reviewer for raising this issue. We included the required information in the "Methods" section (in the first paragraph of this section). As to the control group, since lung biopsy tissues from healthy people were not available, we thus used disease-uninvolved lung tissues from patients as controls. We have pointed out this limitation in the revised "Discussion" section (in the second to the last paragraph).
2. Some of the findings are an overstretch: "In conclusion, identification of these potential pathogenic mediators of AD or SA shared by TB indicates the potential causal links between TB infection and lung AD or SA." This is unfounded; shared pathways alone does not imply shared aetiology. In the discussion they use the term "pathogenic link" which is more appropriate. R: We thank the reviewer for pointing this out. We have revised the manuscript accordingly to the reviewer's suggestions (line 135-138).
"These findings indicate that Mtb promotes the proliferation, migration and invasion of A549 cells partially depending on MKI67, which mechanism might also benefit its own dissemination within the host." This is a major statement and could be a major finding if they developed a line of investigation around it. R: We thank the reviewer for pointing this out. We have performed more experiments (revised Fig. 3 and Supplementary Fig. 1) to improve this part of the study (line 142-168) and we also revised our "Discussion" part accordingly (line 335-348).
3. The finding that "all these diseases drove pulmonary remodeling with various degrees of matrix destruction and fibrous protein deposition" is not surprising but has not before been adequately evaluated or definitely established. However, the interpretation that "collagen matrix remodeling in the lung is a common pathogenic characteristic of pulmonary TB, AD and SA." might (and very likely does!) apply more broadly to pulmonary disease. Other disease groups would have been a good control here. R: As pointed out by the reviewer, collagen matrix remodeling in the lung might apply more broadly to pulmonary disease. We thus revised our "Discussion" according to the reviewer's suggestions as follows: "It should be noted that apart from TB, AD and sarcoidosis, some other lung diseases such as chronic obstructive pulmonary disease, idiopathic pulmonary fibrosis and asthma may also exhibit disordered lung matrix organization (Gu et al., Matrix Biol., 2018), and the underlying mechanistic links among those diseases warrant further investigation." We also agree with the reviewer that other disease groups would have been a good control here. However, in this study, we focused on pulmonary TB as well as the other two potentially related lung diseases, AD and sarcoidosis, to reveal their pathogenic links and differences.
4. The figures are generally very good. Some were even stunning. But it wasn't always clear why certain data over others was presented. Scale bars are missing for several histology figures. Small point, for reviewers it would be easier to read the figure legends if they were placed near the figure, and if the figures were labeled. The data for the "Transwell migration and invasion assays" in figure 3d is not clearly demonstrated as currently presented.