Identification of a Two-Gene Signature and Establishment of a Prognostic Nomogram Predicting Overall Survival in Diffuse-Type Gastric Cancer
, Jiannan Xu 2,†, Songcheng Yin 1, Huabin Wang 3, Guangyao Liu 1, Xinghan Jin 1, Junchang Zhang 1, Huijin Wang 1, Han Wang 1, Huan Li 1, Jianming Liang 1, Yulong He 1,4,* and Changhua Zhang 1,*
Round 1
Reviewer 1 Report
Title: Identification of a Two-gene Signature and Establishment of a Prognostic Nomogram Predicting Overall Survival in Diffuse-3 type Gastric Cancer
Summary: The authors used four published GEO datasets of gastric cancer to identify predictor genes in diffuse vs. intestinal-type gastric cancer. They performed bioinformatic analyses, identifying two differentially upregulated and two downregulated genes using three datasets. Further narrowed using the fourth dataset to one upregulated gene (MEF2C) and one downregulated gene (TRIM15). The author used Bioinformatic analysis to define a prognostic prediction model which compared TNM classification of the tumor and gene signature of MEF2C and TRIM15. Authors identified that high MEF2C and low TRIM15 gene signatures could serve as a predictor of overall survival in diffuse-type gastric cancer.
Although the authors identified DEG by comparing the intestinal-type vs. diffuse-type gastric cancer, the data analysis was performed only on diffuse-type gastric cancer, which I believe the analysis should be performed for both types to conclude the importance of specific genes as specific survival predictors for only diffuse-type gastric cancer. Also, the lack of verification using qPCR or protein analysis is a limiting factor for the findings in the manuscript.
Comments:
Method Section
The method section lacks the source or method for KM plotter survival analysis presented in the manuscript in figure 3, line 160.
Results section
1. In Table 1, line 135:
GSE47007 dataset, Case no; reported as 30 in column three, but the sample size column stated 13 intestinal-18 diffuse, the sum of 13+18 would be 31; please check the numbers. And the same is for GSE62254 Case no column is 300, while the sample size column reported as 146-intestinal, 134-diffuse, 17-mixed, and 2-indeterminate, the sum of the case number will be 299; please check the numbers.
2. In Line 138.
the author stated, “Two upregulated and one downregulated gene were identified (Figure 2A).”, while in figure 2A the Venn diagram shows two upregulated and two downregulated genes. Also, the author stated the validation in dataset GSE62254, while the referred figure (figure 2A) has no such dataset. I believe the authors intend to refer to Figure 2B; please correct this figure reference.
3. In section 3.3, lines 144-151.
the manuscript started by identifying the DEG between intestinal vs. diffused type gastric cancer. I understand the interest of the authors is in diffuse gastric cancer, and he identified the 4 DEG between diffuse and intestinal-type gastric cancer, but I think the analysis in table-2 should also be performed in intestinal-type gastric cancers cases from the same sued dataset (GSE62254), to confirm that MEF2C and TRIM15 genes are specific to diffuse-type gastric cancer as the author concludes. Additionally, COL4A3 did not correlate with diffuse-type gastric cancer in the authors data, but COL4A3 might correlate with the TNM status of intestinal-type gastric cancer, which will add further strength to the author’s work.
4. In section 3.4, lines 153-162 and the corresponding figure 3:
The author needs to cite the source of the KM plotter he used. Also, is there a reason the author used (not the best Jetset prob ”yellow coded”) for his KM plotter for COL4A3 and MEF2C Instead of the (recommended best Jetset probe “green coded)? as the reported probes in the manuscript are: 214641_at, and 207968_s_at?”, The author only used the recommended best Jetset probe for the TRIM15 gene.
5. In section 3.4, lines 153-162 and the corresponding figure 3
In KM-survival analysis, authors should state either in the results section or figure legend that they used the best cutoff value of the expression, not the median expression, and should report the FDR value (as in the author analysis, FDR might be high for certain genes due to using best cutoff value instead of median expression for COL4A3, MEF2C, TRIM15) which affects the p-value interpretation.
6. In figure 3, line 160:
I recommend that the KM plotter for tested genes (COL4A3, TRIM15, MEF2C) should also be reported/displayed for intestinal-type gastric cancer, not only diffuse-type gastric cancer, if the author wants to conclude that the identified genes correlation with survival is specific for diffuse-type gastric cancer only.
7. In section 3.8, line 225:
there is no reference to any figure (I assume there should be a reference to figure 7), Also, it would be helpful if the GSEA for the mentioned pathways is added to the figure (or as a supplementary figure).
8. In line 232
the author ends the paragraph with the sentence (and so on), a better explanation is needed.
9. In section 4, llines286-289:
Authors stated that “Our GSEA results showed that biological process, including glycosaminoglycan biosynthesis chondroitin sulfate signaling pathway, melanogenesis and neurotrophin signal pathway were enriched in diffuse-type gastric cancer” The melanogenesis and neurotrophin signal pathways are not mentioned in the GSEA results in section 3.8. Also, what is the significance of the mentioned pathways to your results? Please discuss.
I recommend if it is feasible to perform MEF2C and TRIM15 labeling on diffuse-type gastric cancer and intestinal-type gastric cancer FFPE to verify the expression of these genes at the protein levels.
Author Response
Comments:
Method Section
The method section lacks the source or method for KM plotter survival analysis presented in the manuscript in figure 3, line 160.
Response: Thanks for your constructive suggestion, which is highly appreciated. We added “2.3. Kaplan-Meier Plotter analysis” in Method section. In addition, we added the source and citation of Kaplan-Meier Plotter in revised version.
Results section
1. In Table 1, line 135:
GSE47007 dataset, Case no; reported as 30 in column three, but the sample size column stated 13 intestinal-18 diffuse, the sum of 13+18 would be 31; please check the numbers. And the same is for GSE62254 Case no column is 300, while the sample size column reported as 146-intestinal, 134-diffuse, 17-mixed, and 2-indeterminate, the sum of the case number will be 299; please check the numbers.
Response: Thanks for your suggestions. We are sorry to make your confusion, because we made a mistake. We check carefully again the GEO dataset (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi). For GSE47007, sample size of diffuse type is 12 (GSM No.: GSM1142850, GSM1142851, GSM1142852, GSM1142853, GSM1142854, GSM1142855, GSM1142856, GSM1142857, GSM1142858, GSM1142859, GSM1142860, GSM1142861), and sample size of intestinal type is 18 (GSM No.: GSM1142862, GSM1142863, GSM1142864, GSM1142865, GSM1142866, GSM1142867, GSM1142868, GSM1142869, GSM1142870, GSM1142871, GSM1142872, GSM1142873, GSM1142874, GSM1142875, GSM1142876, GSM1142877, GSM1142878, GSM1142879). For GSE62254, sample size of indeterminate type is 3 (sample ID: T534, T374 and T376). For GSE22377 and GSE38749, the manuscript described the correct sample size.
2. In Line 138.
the author stated, “Two upregulated and one downregulated gene were identified (Figure 2A).”, while in figure 2A the Venn diagram shows two upregulated and two downregulated genes. Also, the author stated the validation in dataset GSE62254, while the referred figure (figure 2A) has no such dataset. I believe the authors intend to refer to Figure 2B; please correct this figure reference.
Response: We apologize for our mistake. We intended to refer to Figure 2B in line 138. We have made the amendments in revised version.
3. In section 3.3, lines 144-151.
the manuscript started by identifying the DEG between intestinal vs. diffused type gastric cancer. I understand the interest of the authors is in diffuse gastric cancer, and he identified the 4 DEG between diffuse and intestinal-type gastric cancer, but I think the analysis in table-2 should also be performed in intestinal-type gastric cancers cases from the same sued dataset (GSE62254), to confirm that MEF2C and TRIM15 genes are specific to diffuse-type gastric cancer as the author concludes. Additionally, COL4A3 did not correlate with diffuse-type gastric cancer in the authors data, but COL4A3 might correlate with the TNM status of intestinal-type gastric cancer, which will add further strength to the author’s work.
Response: Thanks for your suggestion. In our study, by compared with intestinal type gastric cancer, we identified MEF2C high expression and TRIM15 low expression in diffuse type gastric cancer. Furthermore, these two genes expression were identified as risk factors of overall survival in diffuse type gastric cancer. We constructed a prognostic nomogram and validated the performance of the nomogram in TCGA dataset. Thus, we drew our research conclusion. We found that several previous studies only performed analysis in specific type of tumor. For example, Zhang F, et al. only explored gene expression, survival difference in intestinal type of gastric cancer[1]. Zhang, J, et al. also performed further analysis in intestinal type of gastric cancer[2]. Besides, Lee, I. S, et al. only performed bioinformatic analysis in diffuse type of gastric cancer[3]. Based on our results, MEF2C and TRIM15 may be predictors of overall survival in diffuse type gastric cancer. More future work on this field will be done in our laboratory to adequate funding.
Reference:
[1] Zhang F.; Maswikiti E.P.; Wei Y.; Wu W.; Li Y. Construction and Validation of a Novel Prognostic Signature for Intestinal Type of Gastric Cancer. Dis Markers. 2021, 2021, 5567392.
[2] Zhang J.; Liu X.; Yu G.; Liu L.; Wang J.; Chen X.; Bian Y.; Ji Y.; Zhou X.; Chen Y.; et al. UBE2C Is a Potential Biomarker of Intestinal-Type Gastric Cancer With Chromosomal Instability. Front Pharmacol. 2018, 9, 847.
[3] Lee I.S.; Sahu D.; Hur H.; Yook J.H.; Kim B.S.; Goel A. Discovery and validation of an expression signature for recurrence prediction in high-risk diffuse-type gastric cancer. Gastric Cancer. 2021, 24, 655-665.
4. In section 3.4, lines 153-162 and the corresponding figure 3:
The author needs to cite the source of the KM plotter he used. Also, is there a reason the author used (not the best Jetset prob ”yellow coded”) for his KM plotter for COL4A3 and MEF2C Instead of the (recommended best Jetset probe “green coded)? as the reported probes in the manuscript are: 214641_at, and 207968_s_at?”, The author only used the recommended best Jetset probe for the TRIM15 gene.
Response: Thanks for your constructive suggestion. We added the source of Kaplan-Meier Plotter and citation of a reference in Method section 2.3. of revised manuscript. Also, we selected the recommended best Jetset probe “green coded” to analyze the survival difference of COL4A3 and MEF2C in revised manuscript.
5. In section 3.4, lines 153-162 and the corresponding figure 3
In KM-survival analysis, authors should state either in the results section or figure legend that they used the best cutoff value of the expression, not the median expression, and should report the FDR value (as in the author analysis, FDR might be high for certain genes due to using best cutoff value instead of median expression for COL4A3, MEF2C, TRIM15) which affects the p-value interpretation.
Response: Thanks for your kind suggestion. We make change in result in paragraph 1 of Result section 3.4.
6. In figure 3, line 160:
I recommend that the KM plotter for tested genes (COL4A3, TRIM15, MEF2C) should also be reported/displayed for intestinal-type gastric cancer, not only diffuse-type gastric cancer, if the author wants to conclude that the identified genes correlation with survival is specific for diffuse-type gastric cancer only.
Response: Thanks for your suggestion. We established a prognostic nomogram and validated the performance of the nomogram in TCGA dataset. Thus, we drew our research conclusion. We found that several previous studies only performed analysis in specific type of tumor. For example, Zhang F, et al. only explored gene expression, survival difference in intestinal type of gastric cancer[1]. Zhang, J, et al. also performed further analysis in intestinal type of gastric cancer[2]. Besides, Lee, I. S, et al. only performed bioinformatic analysis in diffuse type of gastric cancer[3]. Based on our results, MEF2C and TRIM15 may be predictors of overall survival in diffuse type gastric cancer. More future work on this field will be done in our laboratory to adequate funding.
Reference:
[1] Zhang F.; Maswikiti E.P.; Wei Y.; Wu W.; Li Y. Construction and Validation of a Novel Prognostic Signature for Intestinal Type of Gastric Cancer. Dis Markers. 2021, 2021, 5567392.
[2] Zhang J.; Liu X.; Yu G.; Liu L.; Wang J.; Chen X.; Bian Y.; Ji Y.; Zhou X.; Chen Y.; et al. UBE2C Is a Potential Biomarker of Intestinal-Type Gastric Cancer With Chromosomal Instability. Front Pharmacol. 2018, 9, 847.
[3] Lee I.S.; Sahu D.; Hur H.; Yook J.H.; Kim B.S.; Goel A. Discovery and validation of an expression signature for recurrence prediction in high-risk diffuse-type gastric cancer. Gastric Cancer. 2021, 24, 655-665.
7. In section 3.8, line 225:
there is no reference to any figure (I assume there should be a reference to figure 7), Also, it would be helpful if the GSEA for the mentioned pathways is added to the figure (or as a supplementary figure).
Response: Thanks for your kind suggestion. We added the mentioned pathways in supplementary figure 1. According to the level of NES, the top 13 pathways was showed in Supplement Figure 1.
8. In line 232
the author ends the paragraph with the sentence (and so on), a better explanation is needed.
Response: Thanks for your suggestion. We apologized to make your confusion. We ranked the results of GSEA according to the level of NES. Then we showed the top 13 pathways in Supplementary figure 1. We have made change in revised version.
9. In section 4, llines286-289:
Authors stated that “Our GSEA results showed that biological process, including glycosaminoglycan biosynthesis chondroitin sulfate signaling pathway, melanogenesis and neurotrophin signal pathway were enriched in diffuse-type gastric cancer” The melanogenesis and neurotrophin signal pathways are not mentioned in the GSEA results in section 3.8. Also, what is the significance of the mentioned pathways to your results? Please discuss.
Response: Thanks for your suggestion. In the result of GSEA, many biological pathways were enriched in diffuse type gastric cancer, including glycosaminoglycan biosynthesis chondroitin sulfate signaling pathway, melanogenesis and neurotrophin signal pathway. By reviewing the literatures, we found that glycosaminoglycan biosynthesis chondroitin sulfate signaling pathway correlated closely with cancer[4-6]. The melanogenesis and neurotrophin signal pathway were not relevant to cancer. We made change in revised version.
Reference:
4. Clausen T.M.; Pereira M.A.; Al Nakouzi N.; Oo H.Z.; Agerbaek M.O.; Lee S.; Orum-Madsen M.S.; Kristensen A.R.; El-Naggar A.; Grandgenett P.M.; et al. Oncofetal Chondroitin Sulfate Glycosaminoglycans Are Key Players in Integrin Signaling and Tumor Cell Motility. Mol Cancer Res. 2016, 14, 1288-1299.
5. Iida J.; Dorchak J.; Clancy R.; Slavik J.; Ellsworth R.; Katagiri Y.; Pugacheva E.N.; van Kuppevelt T.H.; Mural R.J.; Cutler M.L.; et al. Role for chondroitin sulfate glycosaminoglycan in NEDD9-mediated breast cancer cell growth. Exp Cell Res. 2015, 330, 358-370.
6. Pantazaka E.; Papadimitriou E. Chondroitin sulfate-cell membrane effectors as regulators of growth factor-mediated vascular and cancer cell migration. Biochim Biophys Acta. 2014, 1840, 2643-2650.
I recommend if it is feasible to perform MEF2C and TRIM15 labeling on diffuse-type gastric cancer and intestinal-type gastric cancer FFPE to verify the expression of these genes at the protein levels.
Response:
Thank you very much for your common. We understand that MEF2C and TRIM15 should be validated in RNA and protein level. Currently, we are collecting the specimens of intestinal type and diffuse type gastric cancer. In the future, we will further explore the specific mechanism of MEF2C and TRIM15 in vivo and vitro subject to adequate funding. In this study, we identified MEF2C and TRIM15 were correlated with overall survival in diffuse type gastric cancer. And the prognostic nomogram integrated these two genes showed good capacity of predicting overall survival of diffuse type gastric cancer. Therefore, we drew our research conclusion. Recently, some similar high-scoring articles published were analyzed totally based on GEO dataset and TCGA database. For instance, Molecular Characteristics, Clinical Significance, and Cancer Immune Interactions of Angiogenesis-Associated Genes in Gastric Cancer (Frontiers in Immunology, IF: 8.786), Development and Validation of a Prognostic Nomogram for Gastric Cancer Based on DNA Methylation-Driven Differentially Expressed Genes (International Journal of Biological Science, IF: 10.75), Development and validation of a hypoxia-immune-based microenvironment gene signature for risk stratification in gastric cancer (Journal of Translational Medicine, IF: 8.44). These studies have certain reference significance for future clinical research.
Reviewer 2 Report
The aim to develop a novel prognostic model for patients with gastric cancer is an important and ambitious initiative. But, as the authors said, it may be hard to promote the utilization of multi-genome sequencing during clinical practice due to its high price and practicability.
In this idea, there would be more to discuss regarding the possibility of developing a mixed TNM and genetic prognostic model.
I also suggest adding in the discussion chapter other prognostic factors from the literature, citing the relevant references.
Author Response
Response to Reviewer 2
Comments and Suggestions for Authors
The aim to develop a novel prognostic model for patients with gastric cancer is an important and ambitious initiative. But, as the authors said, it may be hard to promote the utilization of multi-genome sequencing during clinical practice due to its high price and practicability.
In this idea, there would be more to discuss regarding the possibility of developing a mixed TNM and genetic prognostic model.
I also suggest adding in the discussion chapter other prognostic factors from the literature, citing the relevant references.
Response: Thanks for your constructive suggestion. TNM staging system only consider tumor invasion depth, lymph node metastasis and distant metastasis. The biological characteristics of tumor such as immune infiltration status, drug response and intracellular signal pathways were not reflected on TNM staging system. However, the genomic sequence of tumor was an effective tool to uncover heterogeneous malignance. Several tumor biomarkers can help guide treatment decisions, including Human Epidermal Growth Factor Re-ceptor-2 (HER2), Programmed Cell Death-Ligand 1 (PDL1) and Vascular Endothelial Growth Factor Receptor (VEGFR). Our prognostic nomogram based on MEF2C, TRIM15 expression and TNM stage information showed a good capacity of predicting overall survival of diffuse type of gastric cancer. Similarly, several previous studies integrated clinical features and risk score based on expression level of risk genes into a novel prognostic nomogram. The predictive value of their integrated nomograms also better than using the risk factor alone. These studies and our present study have certain reference significance for future clinical research. (We have added the above description in the section of discussion).
