Using single cell sequencing to explore the difference in tumor- associated macrophages between left and right colon cancer and evaluate the prognosis of patient

doi:10.21203/rs.3.rs-4154339/v1

Download PDF

Research Article

Using single cell sequencing to explore the difference in tumor- associated macrophages between left and right colon cancer and evaluate the prognosis of patient

https://doi.org/10.21203/rs.3.rs-4154339/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

This research investigates the role of TAMs in left and right colon cancer progression, utilizing scRNA-seq to identify variances in TAMs and their microenvironment. We gathered macrophages from 12 left and 6 right colon cancer tissue samples, identifying 325 marker genes. Patients were stratified into three subgroups based on the expression of marker gene, which facilitated the examination of tumor microenvironment differences, immune cell infiltration, and immune checkpoint expression differences. WGCNA further identified 27 TAM characteristic genes. Two genes, CXCR4 and RGS2, were selected as prognostic genes through multivariate Cox regression analysis. Using these genes, we have created a TAM-related prognosis evaluation model, which was validated using external datasets. Our findings suggest that TAM-related genes have prognostic significance in colon cancer, offering potential therapeutic targets for left and right colon cancer treatment.

left colon cancer

right colon cancer

prognosis

tumor-associated macrophage

single-cell RNA sequencing

Colon cancer (CC) is a malignant tumor located in the digestive tract, with the second highest mortality rate among cancers. In the past few years, there has been a notable rise in the prevalence of colon cancer. ^[1]. The estimated number of newly diagnosed colon cancer cases is approaching 600,000 in China in 2022, accounting for about one eighth of the same year's cancer cases. The mortality rate exceeded 300000, accounting for one tenth of the same year's cancer mortality rate^[2]. However, according to research data, the timely identification and intervention of colon cancer have the potential to significantly improve patients' chances of survival^[3]. According to the embryology and physiology of the colon, we can divide colon cancer into left colon cancer (LCCs) and right colon cancer (RCCs) Research data shows that these two types of colon cancer have different pathways and clinical manifestations, and this difference holds immense significance in the context of colon cancer treatment^[4–6].Therefore, we need to enhance our comprehension of the disparities between left and right colon cancer, as well as the potential targets of their onset, in order to advance research on the therapy and prognosis of colon cancer. Cancer is a process of tumor cell evolution, involving the constant, dynamic and interaction between tumor microenvironment (TME) and cancer cells ^[7–8]. TME includes non cancer cells and components present in tumors. Among them, macrophages exist in all tissues and have functions such as phagocytosis and tissue repair. During tissue infection or cancer, monocytes recruit to the site of infection or disease through chemotactic signals and differentiate into macrophages. TAMs derived from monocytes, are attracted to tumors and influenced by the tumor microenvironment. Consequently, they have a crucial impact on promoting tumor malignancy and progression. ^[9]. TAM has multiple cell types in TME and exert significant influence on tumor development and progression ^[8]. In early tumor development, TAM recruitment and polarization have a mediating anti-tumor effect ^[10]. However, in cancer, TME can affect the metabolism of TAM by directly exchanging metabolites or cytokines, thereby promoting the development of cancer ^[11]. The tumor promoting function of TAM makes it a powerful target for cancer treatment.

In this research, we performed an extensive analysis of the subpopulations of macrophages related to left and right colon cancer using scRNA-seq method, using 1500 samples from 6 datasets included in the GEO database. Macrophage marker genes were obtained and grouped. After obtaining the differences between each group, immune related analysis was performed. Then, weighted gene co expression network analysis (WGCNA) was used to screen the module genes with the highest immune correlation, build a model of genes related to prognosis. Finally, external validation was conducted based on GSE data. In addition, according to ROC analysis, our constructed prognostic model is more accurate than other prognostic indicators. The findings of this research are conducive to a preliminary understanding of pivotal genes implicated in the progression of colon cancer and provide ideas for targeted clinical treatment.

Data collection and processing

We downloaded selected left and right half macrophage single-cell transcript sequencing data sets (GSM3067368, GSM3067369, GSM3067370, GSM3067373, GSM3067374, GSM3067375) from Gene Expression Omnibus (GEO) database(www.ncbi.nlm.nih.gov/). and patient clinical information. To process the scrap-seq data, we first utilized the Seurat package to control data quality, which requires mitochondrial genes to account for less than 10% of the total genes. Normalization was then performed to obtain mutable genes, scatter plots were drawn, and the first ten mutable genes were annotated to determine the principal components exhibiting high heterogeneity in bowel cancer patients, and principal component analysis (PCA) was conducted on the genes with high variation. To verify if our PCA can map cells to a two-dimensional plane, the cell distribution should exhibit sufficient dispersion under the two principal components.

Infiltration and survival analysis of macrophage subsets associated with collector cancer

The default values of some parameters are used and the result of dimension reduction is given. The genes of different LCC and RCC cells were compared with those of normal people, and the average expression values of p value, q value and the exponential function of gene difference multiple were calculated to categorize cells and plot the path network of different cluster genes. UMAP was used again on myeloma cells to classify and map the pathway network of cluster genes. The differential genes were normalized by PCA analysis. The t-CINE algorithm was used to separate cells into distinct clusters, and the "Fingermarks" function was used to identify deferentially expressed genes in different clusters. Use heat maps to show highly expressed gene clusters. Next, epithelial cell types were labeled with Singer packages to further detect marker genes in epithelial cell clusters. Consensus clustering was carried out and patients were classified according to marker gene expression. Observe the relation of k value cumulative density function (CDF). Kaplan-Meier survival analysis was conducted for the classified patients. The clinical correlation of different patients' survival time was observed.

Single cell sequencing and acquisition of tumor macrophage-related genes

We conducted single cell RNA sequencing on colon cells extracted from colorectal patients, and filtered the gene from the cell matrix(< 200 transcripts/cells, > 10% mitochondrial genes) and genes (< 10 cells/genes). The matrix was then imported into R package Seurat (v3.1.2) for subsequent analysis ^[12]. The level of gene expression was standardized by ensuring that the count of unique molecular identifiers in each cell (UMI/ cell) matched the UMI median, then logarithmic conversion was performed ^[13]. Finally, visual single-cell cluster analysis was carried out on the extracted cells unsupervised by t-SNE, and annotation was made on all kinds of cells. Then we obtained the differentially expressed genes (DEGs) of each cluster through applying “FindAllMarkers” function. Meanwhile, marker genes of various cells were evaluated and their expression in different cells was shown using heat map. After quality control, we utilized Monocle2v2.5.4 (R package) ^[14] and placed the gene into Monocle's Reversed Graph Embedding algorithm to shape the trajectory. Then we obtained the cell trajectory and the pseudotime value of each cell. In "pseudotime" on a single cell sorting, the trajectory curve displayed track and analyzed the differences between each branch. Following that, we used Gene Ontology (GO) database (http://geneontology.org/) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases (www.genome.jp/kegg/) to expand the function and signaling pathway of marker genes ^[15]. Kaplan-Meier analysis was employed to explore the effect of macrophage genotyping characteristics on the prognosis of colorectal cancer.

Tumor microenvironment, immune cell infiltration, and immune cell checkpoint Analysis in patient population

To assess variations in tumor microenvironment across the three clusters, we utilized the limma software to compute the ESTIMATE score, immunological score, stroma score, and tumor purity, and the violin diagram to show and compare the variances between the groups. Next, CIBERSORT algorithm was employed to evaluate the infiltration of 22 types of immune cells, and the differences across clusters were displayed using a boxplot. The patients were categorized into two groups, high-risk and low-risk groups based on the 6 kinds of immune cells with obvious differential expression, and corresponding survival curves were drawn for survival analysis. Finally, limma packets were utilized to compare the infiltration difference of different immune checkpoints in the three groups of samples. The immune checkpoints with significant differences were screened and survival analysis was conducted to explore the correlation between the difference of immune checkpoint infiltration and patient survival.

prognosis of obtaining and analysis of related genes

Monocle is adopted to improve the time series analysis, and it will be checked after marker gene immunotherapy is evaluated using WGCNA package, followed by variance analysis to filter and prognosis related genes. The obtained marker genes were grouped, the missing genes were checked and the outliers were deleted. After no significant outliers were found, we the overall correlation of all samples in the dataset was determined. The soft threshold power is determined as 19 according to the minimum power when the scale-free topological fitting index reaches a higher value. After that, the modules were correlated with prognostic data and important genes were identified. The minimal gene module was established, resulting in generation of 2 gene modules. Then, correlation analysis between modules and traits was conducted to determine the modules most associated with prognostic expression to obtain related genes in this module. Volcanic maps were generated to illustrate differentially expressed genes in colorectal cancer samples, while heat maps were employed to demonstrate differential gene expression in colorectal cancer compared to normal.

Construction and verification of TAM related prognostic model

To identify genes associated with TAM, we conducted univariate COX regression analysis. The risk score formula was obtained from the multivariate COX regression analysis, and the risk score for each colon cancer patient in the training set and test set were calculated. Subsequently, patients were categorized into low risk group and high risk group according to the optimal threshold of the risk score. The survival rates of the two groups were assessed and compared through the utilization of Kaplan-Meier survival curve and log-rank test. 1-year, 3-year, and 5-year receiver operating characteristic (ROC) curves were plotted employed the "survival ROC" package to assess the effectiveness of the prognostic model. We then validated the prognostic model in the test dataset. According to the prognostic characteristics and clinical features of the samples, an independent prognostic analysis was conducted to ascertain their prognostic significance. The "rms" package was utilized to draw a column plot, and the performance was evaluated by calibration curve.

Statistical analysis

Kaplan-Meier method and logarithmic rank test were utilized for conducting survival analysis. The Wilcoxon test was utilized to assess the disparities between the two groups and the Kruskal-Wallis test was utilized to evaluate the variances among the three groups. p < 0.05 was considered statistically significant.

Analysis of cell difference in patients with LCC VS RCC

By t-SNE analysis, we divided the cells into 30 clusters based on gene expression standardization (Fig. 1A), categorized the clusters (Fig. 1B) and found that Cluster12 and Cluster14 were mainly from RCC patients, and Cluster3 and Cluster16 were mainly from LCC patients. We performed annotation and categorization of the clusters (Fig. 1C) and identified distinct cellular subpopulations including tumor cells, normal epithelial cells, immune cells, and stromal cells, with Cluster1, Cluster17, and Cluster20 being myeloid cells, and their numbers differed significantly in LCC and RCC patients. Next, we performed further clustering analysis on myeloid cells (Fig. 1D) and grouped the cells into 13 clusters. After classifying and annotating the cell sources (Fig. 1E-F), we found that Cluster1 and Cluster3 mainly originated from RCC patients, while Cluster4, Cluster8 and Cluster10 mostly stemmed from LCC patients. Meanwhile, we discovered that these cells could be classified into SPP1 + TAMs, C1QC + TAMs and otherMyeloids. TAMs of Cluster1 and Cluster5 were dramatically different in LCC and RCC patients.

Data processing and quality control

We selected six datasets, GSM3067368, GSM3067369, GSM3067370, GSM3067373, GSM3067374 and GSM3067375, analyzed and quality controlled their cell sequencing (Fig. 1G). The number of RNAs and mitochondrial genes expressed by cells in all six datasets was within a certain period, and we were unable to locate any counts of mitochondrial sequencing in the dataset. Vlnplots correlation analysis suggested that nCount was positively correlated with nFeature (Fig. 1H) with a correlation coefficient of 0.38. This indicates that cell viablility and sequencing quality are high, and nCount correlates with nFeature, which can be further analyzed. The 1500 variable genes were plotted as a scatter plot (Fig. 1I) and the top 10 variable genes were annotated. 15 PCs (Fig. 1J) showed high heterogeneity in intestinal cancer patients, then the PCA clustering results of cells were displayed at the PC (Fig. 1K), with the cellular distribution of the two main components PC1 and PC2 showing clear dispersion, revealing our PCA can map cells to a two-dimensional plane.

Highly variable gene analysis

PCA algorithm was applied for analyzing highly variable genes, and four different PCs were obtained, while heat maps displayed the number of highly variable genes in each group. (Fig. 2A-B)

Cluster Analysis and marker Gene Analysis of cluster

UMAP clustering was performed by Seurat 3.1.1 through its PCA and t-SNE analysis, and the gene expression standardization was utilized to partition the cells into 6 distinct clusters. We annotated these 6 clusters and found only C5 were macrophages, while the other groups were epithelial cells (Fig. 3A-B). After clustering to acquire 325 marker genes, we displayed marker genes of each cluster using heat maps (Fig. 3C), in which CEACAM6 and MET highly expressed in C0, OLFM4 and LCN2 highly expressed in C1, GKN2, GKN1 and PSCA highly expressed in C2, PGC and LIPF highly expressed in C3, PRCI highly expressed in C4, LAPTM and TYROPB highly expressed in C5 and MZB1 show low expression in all clusters. We chose the C0 group which accounted for the largest proportion of the six cluster groups for further analysis, selected the KRT20 and CAPZA2 genes with the greatest difference in expression in this group, and analyzed distribution of the two genes in other groups (Fig. 3D). The expression levels of the two genes in the six cluster groups were different, and both were mostly expressed in C0 group, while KPT20 was not expressed in the C1, C3, C4 and C5 groups. The Bubble map showed the expression level of 10 marker genes in 6 clusters, of which 10 marker genes both low expressed in C0, C2 and C5, OLFM4 highly expressed in C1, and PDIA2 highly expressed in C3 (Fig. 3E). The detailed expression of 10 genes in each group is shown in Fig. 3F.

Pseudo-time series analysis and enrichment analysis

To study the differences among the cells, we ranked the cells along the pseudo-time trajectory according to the gene expression above, and three states were obtained, in which macrophages only existed in state 2, the epithelial cells mainly existed in state 1 and 3 and there may be significant differences between cells of Cluster0 and Cluster5 (Fig. 4A-C). Meanwhile, the pseudo-time values of Cluster0 and Cluster5 were smaller, while those of other clusters were larger (Fig. 4D). In a bid to prove the effectiveness of cell annotation to Cluster5 and the function of candidate marker genes in Cluster5, we performed GO and KEGG analysis on the candidate marker genes in Cluster5 (Fig. 4E-F). The majority of macrophage-associated gene activities, such as humoral immune response and collagen, have been discovered to be intimately connected with macrophages. These findings prove the correctness of annotating the cells in Cluster5. The findings from GO and KEGG analysis further substantiate that these candidate genes have the potential to exert significant influence on the proliferation and differentiation of macrophages.

Cluster screening of related genes

Consensus clustering patients according to the expression of marker gene, we can see that when k ≤ 3, the area under the CDF curve increases remarkably, but when k > 3, there is no substantial increment observed in the area under the CDF curve. Therefore, when patients are divided into 3 clusters, the expression is obvious (Fig. 5A-L). Survival analysis showed that it was meaningful only when the marker genes were divided into three clusters (p < 0.050). The prognosis of patients in three clusters was different with Cluster 1 demonstrating the best prognosis and cluster 2 displaying the worst(Fig. 5M).

Expression of clustering-related genes

The clinical correlation of typing only correlated with N (p = 0.012) and T (p = 0.024) (p < 0.050), but not with age and sex (Fig. 6A-D). The Down-regulated gene in Branch 2 of C3 was higher than that of C1, and The Down-regulated gene in Branch 2 of C2 was the lowest. The Up-regulated gene in Branch 2 of C2 was higher than that of C3, and The Up-regulated gene in Branch 2 of C1 was the lowest (Fig. 6E-F). The ESTIMATE Score and Immune Score of C2 is the highest and C1 is the lowest. The Stromal Score of C2 is the highest and C3 is the lowest. The Tumor Purity of C2 is the lowest while C1 is the highest (Fig. 6G-J).

Immune examination and analysis

The differences in the infiltration of 22 kinds of immune cells in 3 different clusters were compared (Fig. 7A). In the borplot, we can see there are dramatic differences in the expression of Macrophages M0, T cells CD4 memory resting, T cells CD4 memory activated and other cells in 3 clusters (Fig. 7B). In survival analysis of patients expressing different immune cells, those with low expression of B cells memory, Mast cells resting and T cells CD4 memory resting, high expression of plasma cells, T cells CD4 memory activated and T cells follicular helper had higher survival possibility (Fig. 7C-H).

Analysis of gene expression associated immune checkpoint

In terms of gene expression related to immune checkpoints, there were many differentially expressed genes among the 3 groups, such as YTHDF1, PVR and LDHA, whose expression level significantly increased in C3 (Fig. 7I). This discovery made it possible to find new targets of immunotherapy. Survival analysis suggested that high expression of CD80, CTLA4, ICOS, IFNG, JAK2, LAG3 and TNFRF9 in patients with different checkpoint genes could lead to better survival(Fig. 7J-P).

WGCNA analysis screens for colon cancer-associated macrophage genes

WGCNA was employed to pinpoint genes linked to macrophages in colon cancer. A dataset comprising 378 samples from the TCGA was used to create a matrix, where 325 genes were pre-selected, and subsequent clustering analysis led to the exclusion of two outlier samples (Fig. 8A). The optimal soft threshold power was determined to be 19 (Fig. 8B-C). The clustering from the WGCNA analysis resulted in the identification of two gene modules (Fig. 8D). Module correlation with clinical parameters highlighted the turquoise module's significant association with tumor grading (with a correlation coefficient of 0.36 and a p-value less than 0.001). Gene expression differences across various groups were analyzed and visualized using a heatmap for the 27 filtered genes (Fig. 8E). Additionally, volcano plots identified substantial expression level changes, with four genes being down-regulated and 23 being up-regulated (Fig. 8F).

Screening for TAM-related prognostic genes

27 genes associated with prognosis were identified through single-cell RNA sequencing, immune system correlation analysis, and WGCNA. Univariate COX regression analysis pinpointed four prognostic genes related to the tumor microenvironment, specifically FCGR2A, CXCR4, DUSP1, and RGS2. The forest plot indicated that these four genes are potential risk factors that affect survival outcomes (Fig. 8G). Further multivariate COX regression analysis identified CXCR4 and RGS2 as the most significant genes in this context.

Construction of TAM-related prognostic model

The risk scores for each sample in the training set, which included 371 TCGA samples, were computed based on the expression levels of prognostic genes. Patients were subsequently classified into either a low-risk or high-risk group using the risk score's optimal threshold value. The Kaplan-Meier survival analysis revealed that the high-risk group experienced a higher mortality rate compared to the low-risk group, signifying a significantly poorer prognosis for the former (Fig. 8H). To assess the effectiveness of the risk model, ROC curves were generated, with the area under the curve (AUC) values being 0.641, 0.613, and 0.619 for predictions at 1, 3, and 5 years, respectively.

Validation of prognostic models

Aimed at testing the independence of the model, we analyzed age, sex, grade, stage and risk score by univariate (Fig. 9A) and multivariate analysis (Fig. 9B). The results revealed that the risk score has the potential to function as an independent prognostic factor for survival and was superior to other prognostic indicators. Based on risk score of the prognostic characteristics of patients and other clinical indicators, the construction of a nomogram was used to predict the survival rate of patients more roundly (Fig. 9C), and used the calibration curve to evaluate its performance (Fig. 9D). The results show that the nomogram is reliable for the prediction of 3-year overall survival.

Colorectal cancer (CRC) is globally recognized as the third most prevalent form of cancer and has the second highest fatality rate among all types of cancer^[16]. This research demonstrates that the early detection of CRC can lead to a reduction in mortality rates. Colonoscopy, recognized for its high sensitivity and specificity, is the preferred approach for screening CRC. However, the accessibility of this procedure is influenced by the patient's financial means and the proficiency of the performing physician^[17]. Additionally, due to variations in anatomy and histology, left-sided CRC (arising from the splenic flexure, descending, and sigmoid colon) and right-sided CRC (originating in the cecum, ascending colon, and hepatic flexure) are distinct clinical entities with notable disparities in treatment efficacy and prognosis^[18–19]. Numerous studies have corroborated that right-sided CRC tends to have a more unfavorable prognosis compared to left-sided CRC^[20].

Therefore, accurate and cost-effective early screening for both types of colorectal cancer constitutes a primary objective in ongoing research endeavors. Guo W and his team utilized single-cell sequencing to map the differences between the left and right halves of colorectal cancer for the first time. Digging deep into cell types, transcriptomics, and the tumor microenvironment, the team analyzed the cellular and molecular features of the two diseases and discovered a set of genes that are specifically expressed in left-sided CRC. We learned from the research of his team that the differences in cancer cell signaling pathways, tumor microenvironment and immune cell status play a crucial role in the prognosis of left and right colorectal cancer.^[20].

In previous retrospective systematic studies, researchers have generally found that the TP53 mutation rate of left colorectal cancer is pretty high, while the BRAF, PIK3CA, CTNNB1, SMAD4 and KRAS mutation rates of right colorectal cancer are pretty high, and the incidence of mucinous histology, CIMP and MSI are all very high ^[21–27].

DongXF et al. used immunohistochemistry to explore the expression of WTAP(up-regulated in malignant tumors and considered to be a cancer gene) in the tissues of CC patients, and results showed that WTAP was the highest expression in samples of left colorectal cancer, showing a positive correlation, and a negative correlation with poorly differentiated tumors, providing new ideas for early screening and treatment of left colorectal cancer ^[27]. Because of the different molecular mechanisms of the two diseases, there are differences in prognosis and treatment. Currently, surgical resection is still a common choice for non-metastatic colorectal cancer, but it has been reported that the incidence of postoperative complications and surgical site infection in the right half of colorectal cancer is relatively low, and there is no significant difference in other postoperative symptoms ^[28–29]. Overall, right colorectal cancer has a poor prognosis and a higher mortality rate than left colorectal cancer, which may be due to the higher likelihood of positive MSI in right colorectal cancer ^[30] as well as the fact that left colorectal cancer can be screened relatively early by endoscopy ^[30].

Single-cell sequencing technology is a highly effective and innovative tool to decipher molecular maps based on single cells, which has an increasingly extensive application and promising prospect in cancer research ^[31]. In our study, we analyzed cellular and genetic differences between LCC and RCC using single- cell sequencing. In addition, we constructed a macrophage-based prognostic model for CC patients based on single-cell sequencing data. WGCNA was used to further screen out four TAM-associated prognostic genes related to CRC prognosis, namely FCGR2A, CXCR4, DUSP1 and RGS2. These genes were utilized to build our prognostic model. Our study also analyzed the infiltration of immune cells, and we found that some differentially expressed genes in patient macrophages were significantly associated with survival. ROC curve showed that the area under AUC at 1, 3 and 5 years were 0.641, 0.613 and 0.619 respectively.

Analyses of age, sex, grade, stage, and risk score were conducted using both univariate and multivariate methods to assess the model's prognostic independence. The findings indicated that the risk score stands out as a significant prognostic element. The calibration curves were utilized to appraise the model's performance, further corroborating its dependability. Prior research, including single-cell RNA sequencing, has illustrated that macrophage diversity increases in inflammatory contexts like cancer^[32]. The advancement of CRC is frequently linked to a systemic inflammatory reaction, which involves the release of inflammatory mediators into the bloodstream. These mediators contribute to tumor growth and dissemination via various pathways, such as enhancing tumor cell proliferation and conditioning premetastatic sites to support further metastasis^[33–34].

A survival analysis focusing on immune checkpoints was conducted, revealing that certain checkpoints, including YTHDF1, PVR, and LDHA, are overexpressed in the C3 group, suggesting potential novel targets for immunotherapeutic interventions. The expression levels of immune checkpoint genes like CD80, CTLA4, ICOS, IFNG, JAK2, LAG3, and TNFRF9 vary among patients, with increased expression being associated with improved survival outcomes. Notably, CD80, a costimulatory molecule, is recognized as a highly effective factor for enhancing immune cell activity and curbing the advancement of cancer^[35].

While our research has yielded novel single-cell insights, it is not without its constraints. The absence of immunohistochemical assays and additional verification for model genes is a notable gap in this study. It is imperative that future studies are multi-centric to substantiate our findings. Moreover, due to restrictions in threshold determination and methodological constraints, a comprehensive analysis of all sequenced genes was not feasible, potentially impacting the effective screening and identification of macrophage-related genes. The sample size used in this study is relatively limited, which may affect the model's accuracy. The AUC for 1, 3, and 5 years was 0.641, 0.613, and 0.619, respectively, suggesting that there is a need to either increase the sample size or utilize multiple testing sets to further validate the model.

In summary, we differentiated LCC and RCC from the single-cell sequencing level, obtained two prognostic related genes (CXCR4 and RGS2) by WGCNA analysis and constructed and verified a prognostic model based on CXCR4 and RGS2. We observed a correlation between TAM-related genes and the prognostic of colon cancer and a possibly promising therapeutic target for both left and right colon cancer.

List of Abbreviation
single-cell RNA sequencing	scRNA-seq
Colon cancer	CC
left colon cancer	LCC
right colon cancer	RCC
Colorectal cancer	CRC
tumor microenvironment	TME
weighted gene coexpression network analysis	WGCNA
principal component analysis	PCA
cumulative density function	CDF
differentially expressed genes	DEGs
Gene Ontology database	GO
Kyoto Encyclopedia of Genes and Genomes database	KEGG
t-distributed stochastic neighbor embedding analysis	t-SNE
receiver operating characteristic curves	ROC
area under the curve	AUC

Ethics approval and consent to participate

This study was conducted in accordance with the recommendations of the Ethics Committee of the First Second Affiliated Hospital of Nanchang University. This study was approved by the Ethics Committee of the First Affiliated Hospital of Nanchang University. All the authors consent to participate the study.

Consent for publication

Not applicable.

Availability of data and material

All the data is readily available. Please contact us, if it is needed.

Competing interests

The authors declare that they have no competing interests.

Funding

The study was supported by Applied Research Incubation Program of Jiangxi Province (2021BAG70046), Nature Science Foundation of Jiangxi Province (20202BABL206019), and Science and Technology Plan Project of Jiangxi Provincial Health Care Commission (202210346).

Authors' contributions

Kaili Liao: research design and draft the manuscript.

Yujie Hu, Hanqing Zhao, Bing Sun: draft and revise the manuscript.

Yuxin Fu, Jingyan Zhang, Xiaomeng Sun: literature search and collate references

Zhenfang Xiong, Xiaozhong Wang: review and modification of the manuscript and writing guidance

Acknowledgements:

Not Applicable.

Hossain MS, Karuniawati H, Jairoun AA, Urbi Z, Ooi J, John Aet al. Colorectal cancer: a review of carcinogenesis, global epidemiology, current challenges, risk factors, preventive and treatment strategies. Cancers. 2022; 14(7)doi: 10.3390/cancers14071732.
Xia C, Dong X, Li H, Cao M, Sun D, He Set al. Cancer statistics in china and united states, 2022: profiles, trends, and determinants. Chinese Med J-Peking. 2022; 135(5): 584 − 90.doi: 10.1097/CM9.0000000000002108.
Qaderi SM, Dickman PW, de Wilt J, Verhoeven R. Conditional survival and cure of patients with colon or rectal cancer: a population-based study. J Natl Compr Canc Ne. 2020; 18(9): 1230-7.doi: 10.6004/jnccn.2020.7568.
Gervaz P, Bucher P, Morel P. Two colons-two cancers: paradigm shift and clinical implications. J Surg Oncol. 2004; 88(4): 261-6.doi: 10.1002/jso.20156.
Iacopetta B. Are there two sides to colorectal cancer? Int J Cancer. 2002; 101(5): 403-8.doi: 10.1002/ijc.10635.
Lee MS, Menter DG, Kopetz S. Right versus left colon cancer biology: integrating the consensus molecular subtypes. J Natl Compr Canc Ne. 2017; 15(3): 411-9.doi: 10.6004/jnccn.2017.0038.
Liggett LA, DeGregori J. Changing mutational and adaptive landscapes and the genesis of cancer. Bba-Rev Cancer. 2017; 1867(2): 84–94.doi: 10.1016/j.bbcan.2017.01.005.
Xiao Y, Yu D. Tumor microenvironment as a therapeutic target in cancer. Pharmacol Therapeut. 2021; 221: 107753.doi: 10.1016/j.pharmthera.2020.107753.
Cassetta L, Pollard JW. Tumor-associated macrophages. Curr Biol. 2020; 30(6): R246-8.doi: 10.1016/j.cub.2020.01.031.
Vitale I, Manic G, Coussens LM, Kroemer G, Galluzzi L. Macrophages and metabolism in the tumor microenvironment. Cell Metab. 2019; 30(1): 36–50.doi: 10.1016/j.cmet.2019.06.001.
Mehla K, Singh PK. Metabolic regulation of macrophage polarization in cancer. Trends Cancer. 2019; 5(12): 822 − 34.doi: 10.1016/j.trecan.2019.10.007.
Satija R, Farrell JA, Gennert D, Schier AF, Regev A. Spatial reconstruction of single-cell gene expression data. Nat Biotechnol. 2015; 33(5): 495–502.doi: 10.1038/nbt.3192.
Chen K, Wang Q, Li M, Guo H, Liu W, Wang Fet al. Single-cell rna-seq reveals dynamic change in tumor microenvironment during pancreatic ductal adenocarcinoma malignant progression. Ebiomedicine. 2021; 66: 103315.doi: 10.1016/j.ebiom.2021.103315.
Qiu X, Hill A, Packer J, Lin D, Ma YA, Trapnell C. Single-cell mrna quantification and differential analysis with census. Nat Methods. 2017; 14(3): 309 − 15.doi: 10.1038/nmeth.4150.
Qu X, Zhao X, Lin K, Wang N, Li X, Li Set al. M2-like tumor-associated macrophage-related biomarkers to construct a novel prognostic signature, reveal the immune landscape, and screen drugs in hepatocellular carcinoma. Front Immunol. 2022; 13: 994019.doi: 10.3389/fimmu.2022.994019.
Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal Aet al. Global cancer statistics 2020: globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries. Ca-Cancer J Clin. 2021; 71(3): 209 − 49.doi: 10.3322/caac.21660.
Zygulska AL, Pierzchalski P. Novel diagnostic biomarkers in colorectal cancer. Int J Mol Sci. 2022; 23(2)doi: 10.3390/ijms23020852.
Hong TS, Clark JW, Haigis KM. Cancers of the colon and rectum: identical or fraternal twins? Cancer Discov. 2012; 2(2): 117 − 21.doi: 10.1158/2159-8290.CD-11-0315.
Guo W, Zhang C, Wang X, Dou D, Chen D, Li J. Resolving the difference between left-sided and right-sided colorectal cancer by single-cell sequencing. JCI Insight. 2022; 7(1)doi: 10.1172/jci.insight.152616.
Mehrvarz SA, Advani S, Overman MJ, Manyam G, Kee BK, Fogelman DRet al. Association of smad4 mutation with patient demographics, tumor characteristics, and clinical outcomes in colorectal cancer. Plos One. 2017; 12(3): e173345.doi: 10.1371/journal.pone.0173345.
Rosty C, Young JP, Walsh MD, Clendenning M, Sanderson K, Walters RJet al. Pik3ca activating mutation in colorectal carcinoma: associations with molecular features and survival. Plos One. 2013; 8(6): e65479.doi: 10.1371/journal.pone.0065479.
Albuquerque C, Baltazar C, Filipe B, Penha F, Pereira T, Smits Ret al. Colorectal cancers show distinct mutation spectra in members of the canonical wnt signaling pathway according to their anatomical location and type of genetic instability. Gene Chromosome Canc. 2010; 49(8): 746 − 59.doi: 10.1002/gcc.20786.
Takahashi Y, Sugai T, Habano W, Ishida K, Eizuka M, Otsuka Ket al. Molecular differences in the microsatellite stable phenotype between left-sided and right-sided colorectal cancer. Int J Cancer. 2016; 139(11): 2493 − 501.doi: 10.1002/ijc.30377.
Yamauchi M, Morikawa T, Kuchiba A, Imamura Y, Qian ZR, Nishihara Ret al. Assessment of colorectal cancer molecular features along bowel subsites challenges the conception of distinct dichotomy of proximal versus distal colorectum. Gut. 2012; 61(6): 847 − 54.doi: 10.1136/gutjnl-2011-300865.
Zhang J, Zheng J, Yang Y, Lu J, Gao J, Lu Tet al. Molecular spectrum of kras, nras, braf and pik3ca mutations in chinese colorectal cancer patients: analysis of 1,110 cases. Sci Rep-Uk. 2015; 5: 18678.doi: 10.1038/srep18678.
Loree JM, Pereira A, Lam M, Willauer AN, Raghav K, Dasari Aet al. Classifying colorectal cancer by tumor location rather than sidedness highlights a continuum in mutation profiles and consensus molecular subtypes. Clin Cancer Res. 2018; 24(5): 1062-72.doi: 10.1158/1078-0432.CCR-17-2484.
Dong XF, Wang Y, Tang CH, Huang BF, Du Z, Wang Qet al. Upregulated wtap expression in colorectal cancer correlates with tumor site and differentiation. Plos One. 2022; 17(2): e263749.doi: 10.1371/journal.pone.0263749.
Hinojosa MW, Konyalian VR, Murrell ZA, Varela JE, Stamos MJ, Nguyen NT. Outcomes of right and left colectomy at academic centers. Am Surgeon. 2007; 73(10): 945-8.
Kwaan MR, Al-Refaie WB, Parsons HM, Chow CJ, Rothenberger DA, Habermann EB. Are right-sided colectomy outcomes different from left-sided colectomy outcomes?: Study of patients with colon cancer in the acs nsqip database. Jama Surg. 2013; 148(6): 504 − 10.doi: 10.1001/jamasurg.2013.1205.
Dahlin AM, Palmqvist R, Henriksson ML, Jacobsson M, Eklof V, Rutegard Jet al. The role of the cpg island methylator phenotype in colorectal cancer prognosis depends on microsatellite instability screening status. Clin Cancer Res. 2010; 16(6): 1845-55.doi: 10.1158/1078-0432.CCR-09-2594.
Lei Y, Tang R, Xu J, Wang W, Zhang B, Liu Jet al. Applications of single-cell sequencing in cancer research: progress and perspectives. J Hematol Oncol. 2021; 14(1): 91.doi: 10.1186/s13045-021-01105-2.
Bassler K, Schulte-Schrepping J, Warnat-Herresthal S, Aschenbrenner AC, Schultze JL. The myeloid cell compartment-cell by cell. Annu Rev Immunol. 2019; 37: 269 − 93.doi: 10.1146/annurev-immunol-042718-041728.
Tuomisto AE, Makinen MJ, Vayrynen JP. Systemic inflammation in colorectal cancer: underlying factors, effects, and prognostic significance. World J Gastroentero. 2019; 25(31): 4383 − 404.doi: 10.3748/wjg.v25.i31.4383.
Dolan RD, Lim J, McSorley ST, Horgan PG, McMillan DC. The role of the systemic inflammatory response in predicting outcomes in patients with operable cancer: systematic review and meta-analysis. Sci Rep-Uk. 2017; 7(1): 16717.doi: 10.1038/s41598-017-16955-5.
Scarpa M, Marchiori C, Scarpa M, Castagliuolo I. Cd80 expression is upregulated by tp53 activation in human cancer epithelial cells. Oncoimmunology. 2021; 10(1): 1907912.doi: 10.1080/2162402X.2021.1907912.

Hossain MS, Karuniawati H, Jairoun AA, Urbi Z, Ooi J, John Aet al. Colorectal cancer: a review of carcinogenesis, global epidemiology, current challenges, risk factors, preventive and treatment strategies. Cancers. 2022; 14(7)doi: 10.3390/cancers14071732.
Xia C, Dong X, Li H, Cao M, Sun D, He Set al. Cancer statistics in china and united states, 2022: profiles, trends, and determinants. Chinese Med J-Peking. 2022; 135(5): 584 − 90.doi: 10.1097/CM9.0000000000002108.
Qaderi SM, Dickman PW, de Wilt J, Verhoeven R. Conditional survival and cure of patients with colon or rectal cancer: a population-based study. J Natl Compr Canc Ne. 2020; 18(9): 1230-7.doi: 10.6004/jnccn.2020.7568.
Gervaz P, Bucher P, Morel P. Two colons-two cancers: paradigm shift and clinical implications. J Surg Oncol. 2004; 88(4): 261-6.doi: 10.1002/jso.20156.
Iacopetta B. Are there two sides to colorectal cancer? Int J Cancer. 2002; 101(5): 403-8.doi: 10.1002/ijc.10635.
Lee MS, Menter DG, Kopetz S. Right versus left colon cancer biology: integrating the consensus molecular subtypes. J Natl Compr Canc Ne. 2017; 15(3): 411-9.doi: 10.6004/jnccn.2017.0038.
Liggett LA, DeGregori J. Changing mutational and adaptive landscapes and the genesis of cancer. Bba-Rev Cancer. 2017; 1867(2): 84–94.doi: 10.1016/j.bbcan.2017.01.005.
Xiao Y, Yu D. Tumor microenvironment as a therapeutic target in cancer. Pharmacol Therapeut. 2021; 221: 107753.doi: 10.1016/j.pharmthera.2020.107753.
Cassetta L, Pollard JW. Tumor-associated macrophages. Curr Biol. 2020; 30(6): R246-8.doi: 10.1016/j.cub.2020.01.031.
Vitale I, Manic G, Coussens LM, Kroemer G, Galluzzi L. Macrophages and metabolism in the tumor microenvironment. Cell Metab. 2019; 30(1): 36–50.doi: 10.1016/j.cmet.2019.06.001.
Mehla K, Singh PK. Metabolic regulation of macrophage polarization in cancer. Trends Cancer. 2019; 5(12): 822 − 34.doi: 10.1016/j.trecan.2019.10.007.
Satija R, Farrell JA, Gennert D, Schier AF, Regev A. Spatial reconstruction of single-cell gene expression data. Nat Biotechnol. 2015; 33(5): 495–502.doi: 10.1038/nbt.3192.
Chen K, Wang Q, Li M, Guo H, Liu W, Wang Fet al. Single-cell rna-seq reveals dynamic change in tumor microenvironment during pancreatic ductal adenocarcinoma malignant progression. Ebiomedicine. 2021; 66: 103315.doi: 10.1016/j.ebiom.2021.103315.
Qiu X, Hill A, Packer J, Lin D, Ma YA, Trapnell C. Single-cell mrna quantification and differential analysis with census. Nat Methods. 2017; 14(3): 309 − 15.doi: 10.1038/nmeth.4150.
Qu X, Zhao X, Lin K, Wang N, Li X, Li Set al. M2-like tumor-associated macrophage-related biomarkers to construct a novel prognostic signature, reveal the immune landscape, and screen drugs in hepatocellular carcinoma. Front Immunol. 2022; 13: 994019.doi: 10.3389/fimmu.2022.994019.
Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal Aet al. Global cancer statistics 2020: globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries. Ca-Cancer J Clin. 2021; 71(3): 209 − 49.doi: 10.3322/caac.21660.
Zygulska AL, Pierzchalski P. Novel diagnostic biomarkers in colorectal cancer. Int J Mol Sci. 2022; 23(2)doi: 10.3390/ijms23020852.
Hong TS, Clark JW, Haigis KM. Cancers of the colon and rectum: identical or fraternal twins? Cancer Discov. 2012; 2(2): 117 − 21.doi: 10.1158/2159-8290.CD-11-0315.
Guo W, Zhang C, Wang X, Dou D, Chen D, Li J. Resolving the difference between left-sided and right-sided colorectal cancer by single-cell sequencing. JCI Insight. 2022; 7(1)doi: 10.1172/jci.insight.152616.
Mehrvarz SA, Advani S, Overman MJ, Manyam G, Kee BK, Fogelman DRet al. Association of smad4 mutation with patient demographics, tumor characteristics, and clinical outcomes in colorectal cancer. Plos One. 2017; 12(3): e173345.doi: 10.1371/journal.pone.0173345.
Rosty C, Young JP, Walsh MD, Clendenning M, Sanderson K, Walters RJet al. Pik3ca activating mutation in colorectal carcinoma: associations with molecular features and survival. Plos One. 2013; 8(6): e65479.doi: 10.1371/journal.pone.0065479.
Albuquerque C, Baltazar C, Filipe B, Penha F, Pereira T, Smits Ret al. Colorectal cancers show distinct mutation spectra in members of the canonical wnt signaling pathway according to their anatomical location and type of genetic instability. Gene Chromosome Canc. 2010; 49(8): 746 − 59.doi: 10.1002/gcc.20786.
Takahashi Y, Sugai T, Habano W, Ishida K, Eizuka M, Otsuka Ket al. Molecular differences in the microsatellite stable phenotype between left-sided and right-sided colorectal cancer. Int J Cancer. 2016; 139(11): 2493 − 501.doi: 10.1002/ijc.30377.
Yamauchi M, Morikawa T, Kuchiba A, Imamura Y, Qian ZR, Nishihara Ret al. Assessment of colorectal cancer molecular features along bowel subsites challenges the conception of distinct dichotomy of proximal versus distal colorectum. Gut. 2012; 61(6): 847 − 54.doi: 10.1136/gutjnl-2011-300865.
Zhang J, Zheng J, Yang Y, Lu J, Gao J, Lu Tet al. Molecular spectrum of kras, nras, braf and pik3ca mutations in chinese colorectal cancer patients: analysis of 1,110 cases. Sci Rep-Uk. 2015; 5: 18678.doi: 10.1038/srep18678.
Loree JM, Pereira A, Lam M, Willauer AN, Raghav K, Dasari Aet al. Classifying colorectal cancer by tumor location rather than sidedness highlights a continuum in mutation profiles and consensus molecular subtypes. Clin Cancer Res. 2018; 24(5): 1062-72.doi: 10.1158/1078-0432.CCR-17-2484.
Dong XF, Wang Y, Tang CH, Huang BF, Du Z, Wang Qet al. Upregulated wtap expression in colorectal cancer correlates with tumor site and differentiation. Plos One. 2022; 17(2): e263749.doi: 10.1371/journal.pone.0263749.
Hinojosa MW, Konyalian VR, Murrell ZA, Varela JE, Stamos MJ, Nguyen NT. Outcomes of right and left colectomy at academic centers. Am Surgeon. 2007; 73(10): 945-8.
Kwaan MR, Al-Refaie WB, Parsons HM, Chow CJ, Rothenberger DA, Habermann EB. Are right-sided colectomy outcomes different from left-sided colectomy outcomes?: Study of patients with colon cancer in the acs nsqip database. Jama Surg. 2013; 148(6): 504 − 10.doi: 10.1001/jamasurg.2013.1205.
Dahlin AM, Palmqvist R, Henriksson ML, Jacobsson M, Eklof V, Rutegard Jet al. The role of the cpg island methylator phenotype in colorectal cancer prognosis depends on microsatellite instability screening status. Clin Cancer Res. 2010; 16(6): 1845-55.doi: 10.1158/1078-0432.CCR-09-2594.
Lei Y, Tang R, Xu J, Wang W, Zhang B, Liu Jet al. Applications of single-cell sequencing in cancer research: progress and perspectives. J Hematol Oncol. 2021; 14(1): 91.doi: 10.1186/s13045-021-01105-2.
Bassler K, Schulte-Schrepping J, Warnat-Herresthal S, Aschenbrenner AC, Schultze JL. The myeloid cell compartment-cell by cell. Annu Rev Immunol. 2019; 37: 269 − 93.doi: 10.1146/annurev-immunol-042718-041728.
Tuomisto AE, Makinen MJ, Vayrynen JP. Systemic inflammation in colorectal cancer: underlying factors, effects, and prognostic significance. World J Gastroentero. 2019; 25(31): 4383 − 404.doi: 10.3748/wjg.v25.i31.4383.
Dolan RD, Lim J, McSorley ST, Horgan PG, McMillan DC. The role of the systemic inflammatory response in predicting outcomes in patients with operable cancer: systematic review and meta-analysis. Sci Rep-Uk. 2017; 7(1): 16717.doi: 10.1038/s41598-017-16955-5.
Scarpa M, Marchiori C, Scarpa M, Castagliuolo I. Cd80 expression is upregulated by tp53 activation in human cancer epithelial cells. Oncoimmunology. 2021; 10(1): 1907912.doi: 10.1080/2162402X.2021.1907912.

Table 1

Results of the multivariate cox regression analysis of colon cancer model genes
ID	coef	HR	HR.95L	HR.95H	P value
CXCR4	0.150354094	1.162245714	0.992836444	1.360561559	0.061411228
RGS2	0.194712706	1.214961885	1.033477275	1.428316247	0.018329631

Table 2

Results of multivariate cox analysis of prognostic factors
ID	HR	HR.95L	HR.95H	P value
Age	1.037776343	1.018943226	1.056957552	7.23941317662298e-05
Gender	1.190813852	0.822449376	1.724164028	0.355057598878788
Grade	1.214430089	0.849862144	1.735387851	0.286098699741763
Stage	1.585645819	1.265126174	1.987369098	6.30496480067888e-05
Risk Score	2.153646871	1.264162965	3.668984915	0.00476738565691238

No competing interests reported.

GA.jpg
Graphical Abstract Image

Download PDF

Version 1

posted

You are reading this latest preprint version

Using single cell sequencing to explore the difference in tumor- associated macrophages between left and right colon cancer and evaluate the prognosis of patient

Status:

Version 1

Abstract

Introduction

Methods

Data collection and processing

Infiltration and survival analysis of macrophage subsets associated with collector cancer

Single cell sequencing and acquisition of tumor macrophage-related genes

Tumor microenvironment, immune cell infiltration, and immune cell checkpoint Analysis in patient population

Construction and verification of TAM related prognostic model

Statistical analysis

Results

Data processing and quality control

Highly variable gene analysis

Cluster Analysis and marker Gene Analysis of cluster

Pseudo-time series analysis and enrichment analysis

Cluster screening of related genes

Expression of clustering-related genes

Immune examination and analysis

Analysis of gene expression associated immune checkpoint

WGCNA analysis screens for colon cancer-associated macrophage genes

Screening for TAM-related prognostic genes

Construction of TAM-related prognostic model

Validation of prognostic models

Discussion

Conclusion

Abbreviations

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1