SPP1+ macrophages: A malignant macrophage subset in the colorectal cancer microenvironment revealed by single-cell sequencing

SPP1+ macrophages: A malignant macrophage subset in the colorectal cancer microenvironment revealed by single-cell sequencing | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article SPP1+ macrophages: A malignant macrophage subset in the colorectal cancer microenvironment revealed by single-cell sequencing Jianyong Zheng, Zhenyu Xie, Liaoran Niu, Gaozan Zheng, Ruikai Li, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-1984416/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Accumulating single-cell studies suggest that SPP1 + macrophages are key players in the tumor microenvironment. However, a systematic investigation of SPP1 + macrophages in colorectal cancer (CRC) has not been conducted. A total of eight single-cell RNA-seq datasets and 16 bulk RNA-seq datasets were included in this study. On the basis of existing research, we propose the SPP1 + macrophage model paradigm, which can explain the clinical features and functional changes of macrophages in CRC better than the M1/M2 polarization theory. We identified four macrophage subsets from CRC myeloid cells: FCN1 + macrophages, C1QC + macrophages, SPP1 + macrophages, and MKI67 + macrophages. Inflammation, phagocytosis, malignancy, and proliferation were identified as the most prominent features related to each of the four macrophage subsets. Our results show that SPP1 + macrophages can serve as markers of CRC occurrence, progression, metastasis and a poor prognosis and exhibit enhanced transcription of genes associated with angiogenesis, epithelial-mesenchymal transition, glycolysis, hypoxia, and immunosuppressive signatures. CellPhoneDB analysis further indicated that SPP1 may mediate crosstalk between SPP1 + macrophages and other cells via the SPP1-CD44, SPP1-PTGER4 and SPP1-a4b1 complex axes. Additionally, our study suggests that SPP1 + macrophages are associated with the benefit of immune checkpoint blockade (ICB) therapy and that immunotherapy targeting SPP1 + macrophages is expected to improve the prognosis of CRC patients. Furthermore, anti-CSF1R treatment is more likely to preferentially deplete CSF1R-enriched C1QC + macrophages versus SPP1 + macrophages, which may account for the minimal effect of this monotherapy. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Introduction Colorectal cancer (CRC) is the third most common malignancy and the second leading cause of cancer-related death 1 . Following the great success of immune checkpoint blockade (ICB) in various advanced solid tumors, interest in immunotherapy for CRC is increasing 2 . However, immunotherapy targeting PD-1 is only effective in patients with mismatch repair deficiency (dMMR)/high levels of microsatellite instability (MSI-H), which is present in 5% of metastatic CRC cases 3 – 5 . Therefore, it is necessary to understand the complex mechanisms of cellular and molecular interactions in the CRC tumor microenvironment (TME) and to search for potential targets to develop new immunotherapies. Macrophages are key mediators in the TME and are involved in multiple aspects of tumor immunity 6 . M1 (inflammatory; antitumoral) and M2 (anti-inflammatory; protumoral) macrophage polarization systems have been used to characterize the activation state of macrophages in vitro 7 . However, macrophages exhibit a more complex phenotype in vivo, which contradicts this simple in vitro classification 8 . Macrophages in the TME are often referred to as tumor-associated macrophages (TAMs). TAM is a heterogeneous cell type 9 that contributes to angiogenesis 10 , extracellular matrix (ECM) remodeling 6 , promotes epithelial-mesenchymal transition (EMT) of tumor cells 11 , and activates immunosuppression 12 , promoting the progression and metastasis of CRC 13 . Immunotherapies targeting TAMs, such as those reducing the number of TAMs in the TME by blocking the CSF1-CSF1R axis 14 – 16 , have been applied in the clinic with minimal efficacy as monotherapy in malignancies 16 , 17 . Accumulating single-cell studies have reported that SPP1 + macrophages are important players in the TME 9,18−20 , are characterized as malignancy driven and are associated with a poor prognosis in CRC patients 9 . However, the function of SPP1 + macrophages remains unclear. Our single-cell analysis indicated that SPP1 is not only a specific biomarker for SPP1 + macrophages in CRC but also a key functional molecule. Secreted phosphoprotein 1 ( SPP1 ), also called osteopontin ( OPN ), is overexpressed in most types of cancer 21 – 23 and is associated with a poor prognosis in breast, cervical, bladder, colorectal, head and neck, liver, lung, and esophageal cancers 21 , 23 . SPP1 can promote CRC progression and metastasis through EMT 24 and hypoxia pathways 25 . It has been reported that SPP1 functions by binding to its receptors CD44 and integrin to regulate several signaling pathways (ERK, JNK1, and PI3K/Akt) 26–31 . However, research on SPP1 + macrophages is still in its infancy, and systematic research is lacking. The recent rapid development of single-cell RNA-sequencing (scRNA-seq) has revealed the complexity of myeloid cells, especially the predominant macrophages, in a variety of tumors 9 , 19 , 32 , 33 . However, unlike the major cell types, which have distinctive features and classification markers, there is no generally accepted classification criterion for myeloid cells. Based on the CRC single-cell public dataset, we constructed the largest CRC myeloid cell atlas to date and performed a systematic study of SPP1 + macrophages. Accumulating single-cell evidence suggests that SPP1 + macrophages may be the key to uncovering the heterogeneity and functional changes of macrophages during CRC liver metastasis (CRLM) 9 , 18 . On the basis of existing data, we further proposed the SPP1 + macrophage model, which can explain the clinical characteristics and functional changes of macrophages in CRC better than the M1/M2 polarization theory; this model is also expected to help aid in the development of macrophage-targeted immunotherapeutic strategies. Result ScRNA-seq revealed the myeloid cell landscape of CRC The assessed tissue types of the CRLM dataset included colorectal cancer (CRC), normal colorectum (NC), liver metastases (LM), normal liver (NL), and peripheral blood mononuclear cells (PBMCs) (Fig. 1 a). NC and NL samples included both adjacent and healthy tissue. The CRC single-cell data used in this study were obtained from seven public datasets (GSE132465, GSE144735, GSE146771, GSE164522, GSE178318, GSE178341, and Wu et al .)(Fig. 1 b). The GSE164522 and Wu et al . datasets were also further defined as CRLM datasets as they included CRC, NC, LM, NL and PBMC samples. After data processing and annotation, we identified nine main cell types in the seven CRC datasets: T cells, NK cells, B cells, plasma cells, myeloid cells, mast cells, neutrophils, epithelial cells and stromal cells (Fig. S1a). The proportion of myeloid cells among all immune cells was significantly higher (6/6) in CRC samples than in NC samples (Fig. 1 c and Table S1), suggesting that myeloid cells may play an important role in the CRC tumor immune microenvironment. Myeloid cells were then screened, reclustered and annotated (Fig. S1b). We defined a total of 14 well-characterized myeloid cell subsets, including one plasmacytoid DC subset (pDC-LILRA4), four cDC subsets (cDC1-CLEC9A, cDC2-CD1C, cDC3-LAMP3, and cDC4-LTB), four macrophage subsets (Macro-C1QC, Macro-FCN1, Macro-MKI67, and Macro-SPP1), three doublet subsets (Doublets-B, Doublets-Epi and Doublets-T), and monocyte and Kupffer cell subsets (Fig. S1b and Table S1). We then analyzed the tissue distribution of these myeloid cell subsets in each of the seven cohorts using odds ratios (ORs) (Fig. 1 d). The data showed that cDC1-CLEC9A and cDC2-CD1C were relatively enriched in NC samples, while cDC3-LAMP3 and Macro-SPP1 were relatively enriched in CRC samples; the remaining myeloid subsets were not found to have stable and significant enrichment in NC or CRC samples. In addition, monocytes and Kupffer cells were significantly enriched in PBMC and NL samples, respectively. To construct a more robust and accurate model of the CRC myeloid cell lineage, CRC myeloid cells in the GSE132465, GSE144735, GSE146771, GSE178318, and GSE178341 datasets were integrated into the large cohort CRC-Mix (n = 48,815) (Fig. S1c). The CRC-Mix data were then reclustered and annotated after harmony correction for batch effects. Finally, we identified 12 myeloid cell subsets in CRC tissue, including pDC-LILRA4, cDC1-CLEC9A, cDC2-CD1C, cDC3-LAMP3, cDC4-LTB, Macro-C1QC, Macro-FCN1, Macro-MKI67, Macro-SPP1, Doublets-B, Doublets-Epi and Doublets-T subsets (Fig. 1 e and S1d). In addition, a summary table including representative genes was made to facilitate the identification of these 12 myeloid cell subsets by other researchers, and details of differentially expressed genes (DEGs) between subsets are in the Supplementary Material (Fig. 1 f and Table S2). Changes in monocytes/macrophages during CRLM Single-cell sequencing provides the opportunity to study genetic and functional differences at the cellular level, allowing us to explore changes in the transcriptional profile and functional pathways of monocytes/macrophages in the TME during CRLM. To observe changes in the macrophage transcriptome during CRC occurrence, we identified DEGs of monocytes/macrophages between NC and CRC samples (GSE164522) and specifically annotated the marker genes of the four major macrophage subtypes in the heatmap (Fig. 2 a and Table S3). The results showed that the expression of SPP1 and MKI67 was significantly elevated in CRC samples, while C1QC was highly expressed in NC samples. The same method was used to identify DEGs of monocytes/macrophages in six CRC single-cell datasets (NC vs. CRC) (Fig. 2 b and Table S3). The common DEGs (n = 216) in these six datasets were then analyzed by Metascape to further explore the functional and pathway changes in monocytes/macrophages during CRC occurrence (Fig. 2 c and Fig. S2a). Similarly, to observe changes in the macrophage transcriptome during liver metastasis of CRC, we identified DEGs of monocytes/macrophages between CRC and LM samples (GSE164522). The results showed that the expression of SPP1 and FCN1 was significantly elevated in LM samples, while the expression of C1QC was higher in CRC samples (Fig. 2 d). The same method was used to identify DEGs of monocytes/macrophages in three single-cell datasets (CRC vs. LM) (Fig. 2 e and Table S4). Common DEGs (n = 1,383) in the three datasets were then analyzed by Metascape to further explore the functional and pathway changes in monocytes/macrophages during liver metastasis of CRC (Figs. 2 f and S2b). To characterize the origin and differentiation fate of monocytes/macrophages during CRC occurrence, development, and liver metastasis, we observed trajectories of monocyte/macrophage subsets, marker genes and tissue types based on pseudo-temporal analysis of Monocle2 (Wu et al. ) (Fig. 2 g-i). The results show that monocytes from PBMCs first differentiate into FCN1 + macrophages after entering colon or liver tissue, and then FCN1 + macrophages polarize toward C1QC + macrophages and SPP1 + macrophages, while Kupffer cells in NL samples are relatively isolated (Fig. 2 g and 2 h). The pseudo-chronological results of monocyte/macrophage marker genes also showed that the expression of CD14 and FCN1 decreased significantly when monocytes/macrophages moved from the PBMC-enriched area to the solid tissue-enriched area (Fig. 2 i), suggesting the rapid differentiation of monocytes into macrophages after entering tissues. The expression of SPP1 was specifically elevated in CRC and LM-concentrated regions, suggesting that SPP1 + macrophages play an important role in the occurrence and metastasis of CRC. In addition, the expression of MARCO was concentrated in the NL-rich region. To explore the functional heterogeneity of monocytes/macrophages during CRLM, gene set variation analysis (GSVA) was used to analyze hallmark pathways of different monocyte/macrophage subsets (Wu et al. ) (Fig. 2 j). The results showed a significant enrichment of angiogenesis-, epithelial mesenchymal transition-, glycolysis-, and hypoxia-related genes in SPP1 + macrophages, and the results were also reliably validated in multiple datasets (Fig. S2c-i). Notably, these hallmark pathways were also significantly altered in monocytes/macrophages during CRC occurrence and LM (Fig. S2a and S2b), suggesting that SPP1 + macrophages play a dominant role in monocytes/macrophages changes during CRLM. Tissue distribution and potential mechanisms of SPP1 + macrophages Next, our study focused on the tissue distribution characteristics and biological functions of SPP1 + macrophages. Five scRNA-seq datasets (GSE132465, GSE144735, GSE146771, GSE164522, and GSE178341) showed a significant increase (5/6) in the ratio of SPP1 + macrophages to all macrophages in CRC (NC vs. CRC) (Fig. 3 a and Table S1), and data from three scRNA-seq datasets (GSE164522, GSE178318, and Wu et al .) further showed a significant increase (3/3) in the SPP1 + macrophage proportion in LM samples (CRC vs. LM) (Fig. 3 b and Table S1). The increasing proportion of SPP1 + macrophages during CRC development and liver metastases suggests its critical role in CRLM. To explore whether the increased proportion of SPP1 + macrophages is cancer specific, we integrated CRLM macrophage (GSE164522) and hepatocellular carcinoma (HCC) macrophage data (GSE156625) and performed reclustering. The results showed that the proportion of SPP1 + macrophages in HCC was not significantly different from that in adjacent tissues but was significantly lower than that in CRLM samples (Fig. 3 c and Table S1). This suggests that the elevation of SPP1 + macrophages is only observed in specific cancer tissues (e.g., primary foci and liver metastases of CRC) but is not significant in HCC. Then, we explored the biological function of SPP1 + macrophages in CRC based on the large CRC-Mix dataset. Macrophages in CRC-Mix included C1QC + macrophages (35.5%, n = 13,640), FCN1 + macrophages (30.6%, n = 11,784), SPP1 + macrophages (27.7%, n = 10,647), and MKI67 + macrophages (6.2%, n = 2,403) (Fig. 3 d). MKI67 + macrophages have transcriptional profile highly similar to C1QC + macrophages but specifically express proliferation-related genes such as MKI67 (Fig. 3 h). GSVA and signature scores also showed enrichment of the proliferation signature in MKI67 + macrophages (Figs. 2 j and 3 a). Considering the low proportion and well-defined proliferation function of MKI67 + macrophages, our follow-up study focused on the three major macrophages subsets (C1QC + macrophages, FCN1 + macrophages and SPP1 + macrophages) in CRC. To validate the function of each macrophage subset in the literature and our GSVA results, we assessed angiogenesis, EMT, glycolysis, hypoxia, phagocytosis, and MHC-II signatures in different macrophage subsets in CRC-Mix (Figs. 3 d-I and S3b-d). The results showed that C1QC + macrophages had the highest phagocytosis, MHC-II score and classical MHC class I genes ( HLA-A , HLA-B , HLA-C ) (Fig. S3c and S3d), which indicated that C1QC + macrophages were the dominant macrophages for antigen presentation and phagocytosis of pathogens. These features also support a protective role of C1QC + macrophages in the CRC TME. A recent CRC single-cell study reported that SPP1-TAMs showed the highest angiogenesis score and the lowest phagocytosis score compared with C1QC-TAMs and MKI67-TAMs 9 . Our study included FCN1 + macrophages and further showed that FCN1 + macrophages had higher angiogenesis scores and lower phagocytosis scores than SPP1 + macrophages (Fig. S3b and S3c). This suggests that although SPP1 + macrophages are characterized by high angiogenesis and low phagocytosis, these are not unique features that constitute the specificity of SPP1 + macrophages. In addition, our results showed specific elevation of EMT, glycolysis, and hypoxia scores in SPP1 + macrophages (Fig. 3 e-g), which is consistent with the GSVA results. The classical M1/M2 polarization theory is based on inflammatory signatures to describe the heterogeneity of macrophages, and we also assessed the inflammatory, anti-inflammatory, M1, and M2 signatures of different macrophage subsets in Mix-CRC (Fig. 3 h and 3 i). FCN1 + macrophages showed the highest inflammatory score and the lowest anti-inflammatory score (Fig. 3 h); similarly, FCN1 + macrophages also showed higher M1 scores and lower M2 scores than C1QC + macrophages and SPP1 + macrophages (Fig. 3 i). In addition, we performed enrichment analysis of elevated DEGs in different macrophage subsets separately using Metascape and observed enrichment of the inflammatory response pathway in FCN1 + macrophages (Fig. 3 h). The above evidence suggests an inflammatory profile of FCN1 + macrophages. In contrast to FCN1 + macrophages, C1QC + macrophages and SPP1 + macrophages exhibited more anti-inflammatory features. Enrichment analysis also showed enrichment of the negative regulation of immune system process pathway in C1QC + macrophages. Studies in glioma 35 and HCC 36 have reported the correlation of SPP1 expression with the M2 phenotype, and our data further showed that SPP1 + macrophages have a higher M2 signature than FCN1 + macrophages and C1QC + macrophages (Fig. 3 i). In addition, SPP1 + macrophages exhibited specific expression of CD274 ( PD-L1 ) and HLA-G (a major immune checkpoint) 37 (Fig. S3d), suggesting an immunosuppressive characteristic of SPP1 + macrophages. Notably, consistent with the M1/M2 signature found in multiple tumor scRNA-seq studies 38 , we also observed a significant positive correlation between the M1 signature and the M2 signature in CRC macrophages (Fig. 3 i). To explore the specific molecular mechanism by which SPP1 + macrophages exert their biological functions, we used CellPhoneDB to identify the expression of potential crosstalk molecules based on ligand‒receptor interactions between different macrophage subsets and major cell types (GSE178341) (Fig. 3 j and k). The most abundant receptor‒ligand pair interactions were concentrated between different macrophage subsets, and SPP1 + macrophages also showed relatively abundant interactions with stromal cells, epithelial cells, and T cells (Fig. 3 j). Interestingly, we found three pairs of receptor‒ligand interactions specific for SPP1 + macrophages: SPP1 - CD44 , SPP1 - PTGER4 , and the SPP1 - a4b1 complex (Fig. 3 k). SPP1 + macrophages express high levels of SPP1 , and the receptor CD44 is widely present in different cells, while PTGER4 expression is enriched in epithelial cells and T cells, and integrin a4b1 expression is enriched in T cells and plasma cells. These findings reveal the potential molecular mechanisms of SPP1 + macrophages and indicate that SPP1 is not only a key marker gene of SPP1 + macrophages but also a core molecule for the biological functions of SPP1 + macrophages. In summary, SPP1 + macrophages are malignant cells that are specifically enriched in CRC, with characteristics of angiogenesis, EMT, glycolysis, hypoxia, and immunosuppression. SPP1, a novel macrophage marker in CRC Single-cell data have revealed the important role of SPP1 + macrophages in CRC. The existing bulk RNA-seq dataset with a large number of samples and rich clinical information is also undoubtedly valuable for studying SPP1 + macrophages, and SPP1 can be the key biomarker to unlocking this treasure trove. Next, we first need to answer two fundamental and important questions: whether SPP1 is a specific marker of SPP1 + macrophages and how SPP1 differs from classical macrophage markers. To describe changes in the proportion of macrophages among immune cells during CRC occurrence, we compared the proportions of 22 immune cell types in 10 CRC bulk RNA-seq datasets based on the CIBERSORT algorithm (NC vs. CRC) (Fig. 4 a). The results, which showed the same significant trend in the comparison of half or more datasets (n ≥ 5), are considered to have high credibility. The data showed that the proportions of M0 macrophages and M1 macrophages were significantly increased in CRC, while the proportion of M2 macrophages was significantly decreased. Next, to describe changes in the numbers of macrophages during CRC occurrence, we compared the expression levels of five classical macrophage markers in 10 CRC bulk datasets (NC vs. CRC) (Fig. 4 a). The results did not show significant changes in the pan-macrophage marker CD68 or the M1 macrophage markers CD68 and iNOS ( NOS2 ), while the M2 macrophage markers CD163 and CD206 ( MRC1 ) were significantly downregulated in CRC. The above results of changes in the proportion and number of macrophages during CRC occurrence seem to contradict our single-cell results that the proportion of myeloid cells is significantly elevated in CRC. We speculate that this is related to the overall decline in immune cells in the tumor immune microenvironment during CRC occurrence. Therefore, we used the immune score based on the ESTIMATE algorithm and the immune cell marker CD45 ( PTPRC ) to evaluate the difference in the level of immune infiltration between NC and CRC samples (Fig. 4 a), and the results showed a significant decrease in immune infiltration in CRC tissues. We further assessed the ratio of macrophages to immune cells using CD68 / CD45 , and the results showed that CD68 / CD45 was significantly elevated in CRC. These findings indicated that the number of macrophages in tissues did not change significantly during the occurrence of CRC, while the decrease in the number of overall immune cells resulted in a significant increase in the proportion of macrophages. Unlike classical macrophage markers whose levels were unchanged or decreased, SPP1 (9/9) and SPP1 / CD68 (6/6) were significantly upregulated in CRC (Fig. 4 a). Correlation analysis between SPP1 expression and immune cell proportions (CIBERSORT) showed that SPP1 expression in TCGA-CRC was positively correlated with four myeloid cell types, including neutrophils, M0 macrophages, M2 macrophages, and mast cells (Fig. 4 b). The correlation analysis between SPP1 and six types of immune cells in TCGA-CRC based on TIMER showed that SPP1 had the highest correlation with macrophages (Fig. S4a). In addition, correlation analysis showed that SPP1 was highly positively correlated with CD45 , CD68 , CD86 , CD163 , and CD206 but negatively correlated with iNOS (Fig. S4b). This evidence indicates that SPP1 expression is highly correlated with the infiltration of macrophages, especially M2 macrophages, in CRC. Single-cell sequencing data further revealed that SPP1 expression was only enriched in specific myeloid cells in CRC, and the enrichment pattern was almost consistent with our definition of SPP1 + macrophage subsets (Figs. 4 c, d, S4c and d). In other myeloid cells, other immune cells, and nonimmune cells, SPP1 is expressed at very low levels. These results demonstrate that SPP1 is indeed a specific marker for SPP1 + macrophages in CRC. The IHC staining from The Human Protein Atlas (THPA, https://www.proteinatlas.org/ ) database 39 also verified that SPP1 protein was specifically expressed in CRC samples, while no staining was detected in NC, NL and HCC samples (Fig. 4 e). In addition, to assess the function of the SPP1 gene in CRC, we used the Gene Set Enrichment Analysis (GSEA) function of LinkedOmics to analyze the coexpressed genes of SPP1 in TCGA-CRC (Fig. 4 f). The top five significantly enriched GO BP terms were extracellular structure organization, cell adhesion mediated by integrin, cellular response to lipoprotein particle stimulus, cellular response to VEGF stimulus and macrophage activation. In summary, SPP1 is a specific biomarker of SPP1 + macrophages and is significantly elevated in CRC. Clinical value of SPP1 + macrophages in CRC (bulk RNA-seq) Five CRC bulk datasets (TCGA, GSE14333, GSE17536, GSE41258, and GSE161158) were used to compare SPP1 expression in different stages (stage I/II vs. stage III/IV) (Fig. 5 a). The results showed elevated SPP1 expression in patients with stage III/IV (advanced) CRC (5/5). We also used three CRC bulk datasets (TCGA, GSE39582, and GSE41258) to compare the expression of SPP1 in samples with different TNM classifications (T1/2 vs. T3/4, N0 vs. N1-4, and M0 vs. M1) (Fig. 5 b-d). The results showed higher SPP1 expression in T3/4 (3/3) and N1-3 (2/3) samples than in T1/2 and N0 samples. The close relationship between SPP1 expression level and the degree of tumor infiltration and lymph node metastasis suggests an important role of SPP1 + macrophages in CRC progression. Furthermore, we compared the expression of SPP1 between CRC and LM samples in six bulk datasets (GSE6988, GSE41258, GSE41568, GSE68468, GSE131418 Consortium, and GSE131418 MCC) (Fig. 5 e). SPP1 expression was significantly higher in LM samples than in CRC samples (4/6), suggesting an important role for SPP1 + macrophages in CRC metastasis. We also observed a further decrease in immune infiltration in LM samples compared to CRC samples (5/6). Macrophage enrichment indicates a poor prognosis in most tumors, but macrophage enrichment in CRC tends to be associated with a favorable prognosis 13 , 40 , 41 . Univariate Cox analysis of four datasets (TCGA, GSE17536, GSE39582, and GSE41258) indicated that SPP1 expression was significantly associated with poor overall survival (OS) (3/4) (Fig. 5 f). We performed the same analysis for traditional macrophage markers: iNOS showed a protective effect, while CD68 , CD86 , CD163 and CD206 were not significantly related to CRC progression, metastasis or prognosis. Considering the colinearity between SPP1 + macrophages ( SPP1 ) and macrophages ( CD68 ), we used SPP1 / CD68 coexpression status to study the relationship between SPP1 + macrophages and the clinical features of CRC (Table 1 ). The TCGA-CRC samples were divided into high (H- SPP1 / CD68 ) and low (L- SPP1 / CD68 ) groups according to the median value of SPP1 / CD68 . Patients in the H- SPP1 / CD68 group had a higher mean age; a higher proportion of right-side tumors, advanced T classification, lymph node metastasis, MSI-H status and high TMB; and shorter survival time (OS and PFI). In addition, GSEA verified the enrichment of hallmark pathways such as EMT, angiogenesis and hypoxia in the H- SPP1 / CD68 group (Fig. 5 g). Table 1 Comparison of clinical parameters between the L- SPP1 / CD68 and H- SPP1 / CD68 groups in CRC. Clinical parameters L- SPP1 / CD68 ( N = 309) H- SPP1 / CD68 ( N = 308) p value Age (years) Mean ± SD 64.7 ± 12.7 67.8 ± 12.5 0.002 a NA 3 0 Gender Female 140 (45.8%) 146 (47.4%) 0.682 b Male 166 (54.2%) 162 (52.6%) NA 3 0 Location Left 180 (61.2%) 160 (53.2%) 0.047 b Right 114 (38.8%) 141 (46.8%) NA 15 7 Pathologic stage I 62 (21.3%) 41 (13.5%) 0.078 b II 109 (37.5%) 117 (38.6%) III 82 (28.2%) 96 (31.7%) IV 38 (13.1%) 49 (16.2%) NA 18 5 T classification T1 14 (4.6%) 5 (1.6%) 0.005 b T2 63 (20.7%) 41 (13.4%) T3 199 (65.5%) 219 (71.3%) T4 28 (9.2%) 42 (13.7%) NA 5 1 N classification N0 182 (60.1%) 165 (53.7%) 0.025 b N1 77 (25.4%) 71 (23.1%) N2 44 (14.5%) 71 (23.1%) NA 6 1 Metastasis M0 216 (85.4%) 236 (82.8%) 0.417 b M1 37 (14.6%) 49 (17.2%) NA 56 23 MSI-H Yes 32 (10.5%) 50 (16.8%) 0.024 b No 272 (89.5%) 247 (83.2%) NA 5 11 TMB High 32 (11.7%) 50 (20.5%) 0.006 b Low 242 (88.3%) 194 (79.5%) NA 35 64 OS (years) Mean (95%CI) 7.97 (6.94–9.01) 6.49 (5.43–7.54) 0.046 c Outcome event 55/299 68/294 PFI (years) Mean (95%CI) 8.09 (7.08–9.09) 6.03 (4.99–7.06) 0.030 c Outcome event 67/298 86/293 a Mann-Whitney test, b Chi-square test, c Log-rank test. TMB, tumor mutation burden; high TMB (> 10mut/MB); CI, confidence interval; OS, overall survival; PFI, progression-free interval; Location left includes the cecum to the transverse colon; Location left includes the splenic flexure to the rectum. In summary, SPP1 + macrophages can be used as a malignant biomarker to evaluate CRC progression, metastasis and prognosis. Significance of SPP1 + macrophages for CRC treatment Kaplan‒Meier survival analysis showed that the H- SPP1 / CD68 group had shorter OS than the L- SPP1 / CD68 group (log-rank, P < 0.001) (Fig. 6 a). CRC single-cell data revealed the relationship between preoperative chemotherapy and SPP1 expression in macrophages (Fig. 6 b and c). Patients with partial response (PR) or stable disease (SD) were defined as responders, while patients with progressive disease (PD) were defined as nonresponders. The results showed that SPP1 expression in macrophages was significantly decreased in patients treated with preoperative chemotherapy, and SPP1 expression in macrophages was further decreased in responders (those with PR/SD) compared with nonresponders (those with PD). This suggests that preoperative chemotherapy may exert a therapeutic effect by specifically reducing the number of SPP1 + macrophages in the tumor immune microenvironment. We also speculate that immunotherapy targeting SPP1 + macrophages may improve the prognosis of CRC patients with a poor prognosis. Interestingly, both scRNA-seq data and bulk RNA-seq data showed a strong association between SPP1 + macrophages and genome instability and mutation (dMMR, MSI-H and high TMB) of CRC (Fig. 6 d-I and Table 1 ), suggesting that CRC patients with high SPP1 + macrophages had the characteristics of immunotherapy benefit. We also used SubMap algorithms to predict immunotherapy responsiveness in the L- SPP1 / CD68 and H- SPP1 / CD68 groups (Fig. 6 j). The results indicated that the H- SPP1 / CD68 group was more likely to respond to anti-PD-1 therapy (Bonferroni corrected P = 0.021) and anti-CTLA-4 therapy (Bonferroni corrected P = 0.023). In addition, the results from the GSE39582 dataset also validate that patients with high SPP1 expression are more likely to respond to anti-PD-1 therapy and anti-CTLA-4 therapy (Fig. 6 k). The development and maintenance of macrophages requires a continuous supply of the nutritional factor CSF1 , and CSF1R is also essential for the development, survival, recruitment, and proliferation of most macrophages 42 , 43 . CRC-Mix data showed that CSF1R expression was significantly enriched in C1QC + macrophages but not SPP1 + macrophages (Fig. 6 l), which was also validated in two other scRNA-seq cohorts (GSE164522 and Wu et al .) (Fig. S5a and b). This suggests that immunotherapy that blocks CSF1R preferentially depletes C1QC + macrophages over SPP1 + macrophages. Discussion We investigated SPP1 + macrophages in CRC from various aspects; we assessed their cell origin and tissue distribution, clinical value as a marker of CRC malignancy, potential functional pathways and cell‒cell interactions, and implications for treatment (Fig. 7 a-e). Based on the literature and the results of this study, we propose the SPP1 + macrophage model theory. Compared with the M1/M2 polarization theory, this theory can better explain the changes in macrophages in the whole process of CRLM and guide clinical diagnosis, condition assessment and treatment. Importantly, the SPP1 + macrophage model provides a unique single-cell perspective to help us understand the heterogeneity of macrophages in the CRC tumor immune microenvironment. We identified four macrophage subsets from CRC tissue, including FCN1 + macrophages, C1QC + macrophages, SPP1 + macrophages and MKI67 + macrophages (Fig. 7 a). Enrichment analysis and phenotypic signatures indicated that inflammation, phagocytosis, malignancy and proliferation were the most prominent features related to these four macrophage subsets. SPP1 + macrophages were specifically enriched in primary foci and liver metastases of CRC but were almost absent in normal colorectal and liver tissues and HCC tissues. Evidence from trajectory analysis and tissue distribution analysis of macrophage subsets supports the following hypothesis regarding the origin and differentiation of SPP1 + macrophages: monocytes from peripheral blood are recruited into colonic tissues and rapidly differentiate into macrophages, which differentiate into monocyte-like FCN1 + macrophages that polarize toward C1QC + macrophages and SPP1 + macrophages. Subsequently, SPP1 + macrophages that are enriched in CRC primary foci are transferred to the liver along with tumor tissue and further enriched (Fig. 7 b). The expansion of myeloid cells in CRC tissue suggests their active role in shaping the TME 44 . Our data also confirmed the significant increase in the proportion of myeloid cells and macrophages in CRC. However, our further study found that the number of macrophages in tissues did not change significantly during CRC tumorigenesis, but the decrease in the number of overall immune cells resulted in a significant increase in the proportion of macrophages (Fig. 7 c). The decrease in immune infiltration during the occurrence and metastasis of CRC suggests that antitumor immunity weakens as the disease progresses, while the importance of macrophages in the TME continues to increase. Accumulating single-cell studies have reported the association of SPP1 + TAMs with malignancies, especially CRC 19 , 20 , 32 , 33 . A recent study also found that SPP1 + TAMs in CRC were associated with LM 9 . Numerous studies have reported that SPP1 is a diagnostic and prognostic marker for CRC 22 , 28 , 45 , 46 . Our results integrate the content of the above studies. We found that SPP1 is a specific marker of SPP1 + macrophages, and SPP1 + macrophages can be used as an indicator to assess the malignancy of CRC. The number and proportion of SPP1 + macrophages constantly increase during the occurrence, progression and metastasis of CRC and indicate a poor prognosis (Fig. 7 c). These results support that SPP1 + macrophages in the TME may play an important driving role in the occurrence, progression and metastasis of CRC. SPP1 + macrophages are also expected to be a comprehensive marker for CRC diagnosis, disease severity assessment and prognosis evaluation. The "classically activated" M1 and "alternately activated" M2 macrophage polarization systems have been widely used to describe the activation state of macrophages in vitro 7 . However, macrophages in vivo exhibit a more complex phenotype, which contradicts this simple in vitro classification 8 . Our studies in CRC also demonstrate the limitations of this in vitro polarization model. First, the M1/M2 polarization theory considers M1 and M2 activation states as mutually exclusive discrete states, while multiple recent single-cell studies have shown that M1 and M2 gene signatures are coexpressed in macrophage subsets in nearly all cancer types 19 , 32 , 38 , 47 . Our data also found that M1 and M2 signatures coexisted in CRC macrophages and were significantly positively correlated. Second, according to the M1/M2 polarization theory, proinflammatory M1 macrophages are thought to prevent tumor progression, while M2 macrophages promote tumor growth and metastasis 34 , 47 . This theory can well explain the changes in macrophages in the TME in some tumors 34 , 48 . However, our results based on CIBERSORT showed that the proportion of M1 macrophages was significantly increased in CRC, while the proportion of M2 macrophages was significantly decreased, which contradicted the M1/M2 polarization theory. Furthermore, in our study of CRC progression, metastasis and prognosis, no significant differences were observed in M1/M2 markers except for the M1 marker iNOS , which presented a protective effect. The above evidence indicates that the M1/M2 polarization theory is not applicable in the CRC TME, which also suggests that the malignant features of SPP1 + macrophages are not derived from their M2 signatures. The biological function of SPP1 + macrophages is still unclear. A recent study suggested that SPP1 + TAMs promote CRC through enhancing angiogenesis and attenuating phagocytosis 9 . However, our study found that FCN1 + macrophages exhibited higher angiogenesis signatures and lower phagocytosis signatures than SPP1 + macrophages, suggesting that SPP1 + macrophages also promote tumors through other important pathways. Various analyses in this study confirmed the enrichment of EMT, hypoxia, glycolysis, and immunosuppressive signatures in SPP1 + macrophages, suggesting that SPP1 + macrophages may promote CRC through related pathways (Fig. 7 d). Furthermore, our study found that SPP1 is also a key functional molecule of SPP1 + macrophages. CellPhoneDB predicted that SPP1 + macrophages interact with other cells through the SPP1 - CD44 , SPP1 - PTGER4 , and SPP1 - a4b1 complex axes. The CD44 receptor plays an important role in tumor EMT, invasion and metastasis 49 , PTGER4 50 , and integrin a4b1 51 have also been reported to be involved in cancer progression and metastasis. This suggests that SPP1 + macrophages may promote CRC progression and metastasis through the interaction of secreted SPP1 with receptors CD44 , PTGER4 , and integrin a4b1 (Fig. 7 d). The understanding of SPP1 + macrophages is expected to provide new options for the treatment of CRC. Chemotherapy-induced macrophage phenotype switching has been reported in breast cancer 52 . We also found that preoperative chemotherapy significantly reduced SPP1 expression in CRC macrophages, particularly in responders. This suggests that effective chemotherapy may specifically inhibit SPP1 + macrophages. Moreover, it suggests the possibility of immunotherapy targeting SPP1 + macrophages. Interestingly, we found that although SPP1 + macrophages predict a poor prognosis, patients with high SPP1 + macrophage levels show higher levels of genome instability and mutation characteristics (MSI-H, dMMR, and high TMB) (Fig. 7 d). dMMR/MSI-H status is a recognized immunotherapy biomarker in colorectal cancer and other solid tumors 53 . In addition, a high TMB also predicts better therapeutic efficacy of ICBs in several solid tumors, including CRC 54 – 56 . This suggests that CRC patients with high SPP1 + macrophages are more likely to benefit from ICB immunotherapy. This notion is further supported by the prediction results of SubMap algorithm and the enrichment of PD-L1 and HLA-G in SPP1 + macrophages. In addition, a large number of studies have reported that targeted blockade of SPP1 can inhibit tumor progression and metastasis 21,25,57−59 . Substantial evidence from this study supports the development of drugs targeting SPP1 + macrophages as a viable option for CRC immunotherapy. Notably, our SPP1 + macrophage theory also reveals a possible problem in current macrophage-targeted immunotherapy research. Inhibition of the CSF1-CSF1R axis to reduce TAM levels is a major research direction of macrophage-targeted immunotherapy 6,14−16 . However, we found that CSF1R was significantly enriched in C1QC + macrophages compared to other macrophages, suggesting that CSF1R blockade may preferentially deplete the C1QC + macrophage subset while sparing SPP1 + macrophages. Anti-CSF1R therapy may not be sufficient to deplete all tumor-promoting SPP1 + macrophages, which may contribute to the poor efficacy of anti-CSF1R monotherapy in the Renca mouse tumor model and in human cancer patients 16 , 17 , 19 . However, this study has some limitations. First, the lack of a consensus classification system for myeloid cell subsets identified by single-cell sequencing has become a key obstacle limiting further research on subsets. Therefore, we mainly referred to the study of Zhang et al . 19 for the classification of macrophage subsets and only identified four macrophage subsets that are ubiquitous and well characterized, while some rare macrophage subsets (such as CXCL10- and ISG-enriched macrophage subsets 18 , 32 , 60 ) were generally incorporated into the macrophage subsets with similar transcriptomes. Second, we observed specific enrichment of SPP1 + macrophages in LM samples. These SPP1 + macrophages may be derived from peripheral blood monocytes acclimated by the TME of the LM or from the abnormal activation and proliferation of transferred SPP1 + macrophages. However, our current techniques remain unable to determine the origin of SPP1 + macrophages in metastases. Third, extensive data in this study suggest that CRC patients with high SPP1 + macrophage levels are more likely to benefit from immunotherapy, but this evidence is all based on correlational inferences. We have not yet found a large-scale CRC immunotherapy dataset that can be used to validate the hypothesis. Conversely, a recently published study suggested that SPP1 + macrophages reduced the efficacy of PD-L1 blockade treatment 24 . Therefore, additional in-house clinical data are needed to validate the relationship between SPP1 + macrophages and ICB immunotherapy. In conclusion, based on public scRNA-seq and bulk RNA-seq datasets, we propose the SPP1 + macrophage model paradigm, which can explain the characteristics and changes of macrophages in CRC better than the M1/M2 polarization theory. We found that SPP1 + macrophages can serve as a malignant marker of CRC occurrence, progression, metastasis and poor prognosis. We also explored the underlying mechanisms of SPP1 + macrophages and suggest that immunotherapy targeting SPP1 + macrophages (in terms of abundance and function) is a promising CRC therapeutic strategy. Materials And Methods Material The public datasets used in this study are from the GEO and TCGA and the website that stores the scRNA-seq data of Wu et al . These datasets include eight scRNA-seq datasets and 16 bulk RNA-seq datasets (Table 2 ). To ensure the stability and comparability of single-cell data, only scRNA-seq datasets based on the 10x Genomics platform were included in this study; seven CRC datasets (GSE132465, GSE144735, GSE146771, GSE164522, GSE178318, GSE178341, and Wu et al .) and one HCC dataset (GSE156625) met this criterion. GSE164522 and Wu et al . were further defined as CRLM datasets due to their tissue sample types, including CRC, NC, LM, NC, and PBMC. The bulk RNA-seq dataset includes high-throughput sequencing data from TCGA-CRC and 15 GEO microarray datasets (GSE6988, GSE14333, GSE17536, GSE20842, GSE20916, GSE39582, GSE41258, GSE44076, GSE44861, GSE68468, GSE83889, GSE87211, GSE106582, GSE131418 and GSE161158). Transcriptome data and clinical information of TCGA-CRC were downloaded from UCSC Xena ( http://xena.ucsc.edu/ ), and single nucleotide variation data (VarScan) were downloaded from the Genomic Data Commons (GDC) portal ( https://portal.gdc.cancer.gov/ ). The transcriptome and clinical information of the GEO dataset were downloaded from the GEO database ( https://www.ncbi.nlm.nih.gov/geo/ ). Table 2 Source of scRNA-seq and bulk RNA-seq datasets. Deposited Data Platform Source Identifier Human CRC scRNA-seq dataset 10x Genomics Lee et al ., 2020 44 GEO: GSE132465 Human CRC scRNA-seq dataset 10x Genomics Lee et al ., 2020 44 GEO: GSE144735 Human CRC scRNA-seq dataset 10x Genomics Zhang et al ., 2020 19 GEO: GSE146771 Human CRC scRNA-seq dataset 10x Genomics Che et al ., 2021 61 GEO: GSE178318 Human CRC scRNA-seq dataset 10x Genomics Pelka et al ., 2021 60 GEO: GSE178341 Human CRLM scRNA-seq dataset 10x Genomics Liu et al ., 2022 9 GEO: GSE164522 Human CRLM scRNA-seq dataset 10x Genomics Wu et al ., 2022 18 http://www.cancerdiversity.asia/scCRLM Human HCC scRNA-seq dataset 10x Genomics Sharma et al ., 2020 62 GEO: GSE156625 Human CRC bulk RNA-seq dataset Illumina Hiseq - TCGA Human CRC bulk RNA-seq dataset GPL4811 Ki et al ., 2007 63 GEO: GSE6988 Human CRC bulk RNA-seq dataset GPL570 Jorissen et al ., 2009 64 GEO: GSE14333 Human CRC bulk RNA-seq dataset GPL570 Smith et al ., 2010 65 GEO: GSE17536 Human CRC bulk RNA-seq dataset GPL4133 Gaedcke et al ., 2010 66 GEO: GSE20842 Human CRC bulk RNA-seq dataset GPL570 Skrzypczak et al ., 2010 67 GEO: GSE20916 Human CRC bulk RNA-seq dataset GPL570 Marisa et al ., 2013 68 GEO: GSE39582 Human CRC bulk RNA-seq dataset GPL96 Sheffer et al ., 2009 69 GEO: GSE41258 Human CRC bulk RNA-seq dataset GPL13667 Solé et al ., 2014 70 GEO: GSE44076 Human CRC bulk RNA-seq dataset GPL3921 Ryan et al ., 2014 71 GEO: GSE44861 Human CRC bulk RNA-seq dataset GPL96 Sheffer et al ., 2009 69 GEO: GSE68468 Human CRC bulk RNA-seq dataset GPL10558 Kwon et al ., 2017 72 GEO: GSE83889 Human CRC bulk RNA-seq dataset GPL13497 Hu et al ., 2018 73 GEO: GSE87211 Human CRC bulk RNA-seq dataset GPL10558 - GEO: GSE106582 Human CRC bulk RNA-seq dataset GPL15048 Kamal et al ., 2019 74 GEO: GSE131418 Human CRC bulk RNA-seq dataset GPL570 Szeglin et al ., 2022 75 GEO: GSE161158 Basic analysis workflow for scRNA-seq data The R package Seurat (version 4.1.0) converts raw unique molecular identifier (UMI) count matrices to Seurat objects and then filters the data according to the following criteria: (1) filter genes that are expressed in at least 10 cells; (2) filter cells that have unique feature counts over 6,000 or less than 500; and (3) filter cells that have > 20% mitochondrial counts. After quality control, "LogNormalize" was used to normalize the expression of all cells with a scale factor of 10,000. The top 2000 highly variable genes were identified based on the mean and dispersion. During scaling of the highly variable genes, the "var.to.regress" option was used to regress percent mitochondrial content. The results are obtained by principal component analysis (PCA) for linear dimensionality reduction. We used the FindClusters function on 50 principal components with a resolution of 0.8 for preliminary clustering and annotation. Afterward, the visualization of cell clustering is completed by the nonlinear dimensionality reduction UMAP method 76 . Finally, the main cell types were identified and annotated based on canonical marker genes. Subsequent annotation of myeloid cell subsets should follow the same workflow described above for major cell types. Annotation of major cell types and myeloid cell subsets The main basis for classifying clusters into specific major cell types and myeloid cell subsets was as follows: (1) expression level of key marker genes in clusters; (2) the relative distance between clusters in the UMAP graph; and (3) the distance between clusters evaluated by Cluster tree algorithm. The following key marker genes of major cell types were used in this study: T cells ( CD3D ), NK cells ( KLRF1 ), B cells ( MS4A1 ), plasma cells ( MZB1 ), myeloid cells ( LYZ ), mast cells ( TPSAB1 ), neutrophils ( FCGR3B ), epithelial cells ( EPCAM ), and stromal cells ( DCN ). Myeloid cells were defined in this study based on the similarity of cell clusters rather than strictly following biological definitions. Therefore, in the cluster annotation of major cell types, real myeloid cells such as neutrophils and mast cells were independently named, while pDCs of lymphoid origin 77 , Kupffer cells derived from embryonic yolk sac 78 , and even doublets were all classified as myeloid cells. The myeloid cell subsets were identified based on markers as follows: pDC-LILRA4 ( LILRA4 ), cDC1-CLEC9A ( CLEC9A ), cDC2-CD1C ( CD1C ), cDC3-LAMP3 ( LAMP3 ), cDC4-LTB ( LTB ), Macro-C1QC ( C1QC ), Macro-FCN1 ( FCN1 ), Macro-MKI67 ( MKI67 ), Macro-SPP1 ( SPP1 ), Doublets-B ( CD79A, MS4A1 and MZB1 ), Doublets-Epi ( EPCAM ), Doublets-T ( CD3D ), Monocytes ( CD14 ), and Kupffer cells ( MARCO ). Since CRC macrophages were the focus of this study, the major monocyte/macrophage subpopulation in PBMCs of the CRLM dataset was only broadly defined as monocytes. In the datasets without PBMC samples, we did not define the monocyte subset, and because of the high similarity in the transcriptome characteristics of monocytes and Macro-FCN1, the potential monocytes in these datasets are usually classified into the Macro-FCN1 subset. Similarly, in datasets without NL samples, we do not define a subset of Kupffer cells. Macro-C1QC, Macro-FCN1, Macro-MKI67, Macro-SPP1, monocytes, and Kupffer cells in this study were collectively referred to as monocytes/macrophages; macrophages included Macro-C1QC, Macro-FCN1, Macro-MKI67, and Macro-SPP1 subsets. Data integration and batch correction To construct the large sample data of CRC myeloid cells, CRC myeloid cell data from five independent datasets (GSE132465, GSE144735, GSE146771, GSE178318, and GSE178341) were combined into CRC-Mix data. After correcting for batch effects using the harmony algorithm 79 , the CRC-Mix data were reclustered into groups and annotated. In addition, to compare the differences in SPP1 + macrophage proportions in CRLM and HCC, myeloid cells of GSE156625 and GSE164522 were integrated in the same way, corrected for batch effects by the harmony algorithm, reclustered into groups and annotated. Expression difference analysis and representative genes The Findallmarkers function in Seurat was used to identify marker genes for each cluster or subset. In addition, to explore the functional changes in monocytes/macrophages during CRLM, we also used the Findmarkers function to compare the monocyte/macrophage transcriptomes of NC and CRC, as well as CRC and LM. The parameters assert="RNA" and slot="counts" allowed for the inclusion of genes with low variability in the comparison. Genes with significant expression differences (p < 0.05) in the comparison between groups were defined as DEGs. DEGs with the same trend in multiple datasets were defined as common DEGs, and these genes were used for subsequent functional enrichment analysis. In brief, for myeloid cell subsets in the CRC-Mix dataset, genes that met these criteria were considered representative genes: 1) adjusted P value 0.5; 3) pct.1 (percentage of gene expression detected in the first group) > 0.25; 4) The above three-step filtering process was also performed on the myeloid cell subsets defined before harmony correction, and the intersection genes were selected. These genes were further checked manually to ensure their expression specificity on the corresponding subsets. Tissue distribution of myeloid cell subsets To understand the distribution preference of each myeloid cell subtype in different tissues, we evaluated the enrichment of each myeloid cell subtype using the OR in seven CRC single-cell datasets. OR > 1 indicates the relative enrichment of the cell subtype in a specific tissue. Monocyte/macrophage developmental trajectory in CRLM To construct the monocyte/macrophage developmental trajectory of colorectal cancer liver metastases, we used the "Monocle" package (version 2.22.0) 80 to align monocytes/macrophages in pseudotime order. The DDRTree method implemented with the reduceDimension function of Monocle 2 was used for dimensionality reduction and construction of pseudo-temporal order. Cell‒cell interaction analysis We used Python-based (version 3.7) CellphoneDB 81 to evaluate the interactions of macrophages with other cells. The putative ligands and receptors were determined based on whether they were expressed on each cell. To efficiently assess the closeness of cell interactions based on the number of receptor‒ligand pairs, we randomly sampled 1,000 cells per population from the Macro-C1QC, Macro-FCN1, Macro-MKI67, Macro-SPP1 and major cell types in the GSE178318 dataset. Defining phenotype scores To understand the characteristics between different macrophage subsets, scores for different phenotypes were obtained by the AddModuleScore function in the "Seurat" package. These scores were defined by the mean expression of phenotype-related genes. The related genes of M1 macrophages and M2 macrophages were defined as M1 score and M2 score, respectively. The M1 score considers the expression of CCL5, CCR7, CD40, CD86, CXCL9, CXCL10, CXCL11, IDO1, IL1A, IL1B, IL6, IRF1, and IRF5 , while M2 score considers the expression of CCL4, CCL13, CCL18, CCL20, CCL22, CD276, CLEC7A, CTSA, CTSB, CTSC, CTSD, FN1, IL4R, IRF4, LYVE1, MMP9, MMP14, MMP19, MSR1, TGFB1, TGFB2, TGFB3, TNFSF8, TNFSF12, VEGFA, VEGFB, and VEGFC 47 . Inflammatory score-related genes included IL1B, IL6, S100A8, and S100A9 , while APOE, CD163, MAF, SELENOP, and SEPP1 were used to define the anti-inflammatory score 44 . Glycolysis score-related genes included ALDOA, ALDOB, ALDOC, ENO1, ENO2, ENO3, GAPDH, GPI, HK1, HK2, HK3, LDHA, PFKL, PFKM, PFKP, PGAM1, PGAM4, PGK1, PKLR, PKM, SLC2A1, and TPI1 82 . Hypoxia score-related genes included ACOT7, ADM, ALDOA, CDKN3, ENO1, LDHA, MIF, NDRG1, P4HA1, PGAM1, SLC2A1, TPI1, TUBB6, and VEGFA 82 . EMT score-related genes included ADAM12, ADAMTS12, ADAMTS2, AEBP1, ANGPTL2, ANTXR1, AP1G1, ATP8B1, AXL, BNC2, CALD1, CDH1, CDH2, CDS1, CGN, CLDN4, CMTM3, CNOT1, CNRIP1, COL10A1, COL1A1, COL1A2, COL3A1, COL5A1, COL5A2, COL6A1, COL6A2, COL6A3, COL8A1, CTNND1, DACT1, DYNC1LI2, EMP3, ERBB3, ESRP1, ESRP2, F11R, FAP, FBN1, FN1, FSTL1, GALNT3, GPC6, GPR56, GRHL2, GYPC, HOOK1, HTRA1, INHBA, IRF6, ITGA11, LOXL2, LRRC15, MAP7, MARVELD2, MARVELD3, MMP2, MSRB3, MYO5B, NAP1L3, NID2, OCLN, OLFML2B, PCOLCE, PDGFRB, PMP22, POSTN, PRSS8, SPARC, SPINT1, SPOCK1, SULF1, SYT11, THBS2, VCAN, VIM and ZEB2 83 . Proliferation score-related genes included AURKA, BUB1, CCNB1, CCND1, CCNE1, DEK, E2F1, FEN1, FOXM1, H2AFZ, HMGB2, MCM2, MCM3, MCM4, MCM5, MCM6, MKI67, MYBL2, PCNA, PLK1, TOP2A, TYMS , and ZWINT 84 . Angiogenesis score-related genes included ANGPTL4, CXCL8, VCAN, and VEGFA 19 . Phagocytosis score-related genes included C1QB, CD163, MERTK, and MRC1 19 . MHC-II score-related genes included HLA-DMA, HLA-DMB, HLA-DOA, HLA-DOB, HLA-DPA1, HLA-DPB1, HLA-DQA1, HLA-DQA2, HLA-DQB1, HLA-DQB2, HLA-DRA, HLA-DRB1, and HLA-DRB5. Immune infiltration and macrophage markers The immune score calculated by the ESTIMATE algorithm was used to assess the overall immune infiltration of each sample 85 . The immune cell marker CD45 was also used to assess the level of immune cell infiltration in tissues. CIBERSORT, a deconvolution algorithm based on transcriptional profiles, was used to calculate the proportions of 22 immune cells, including M0, M1, and M2 macrophages, in each sample 86 . To understand the clinical features of SPP1 as a specific macrophage marker in CRC, some classic macrophage markers were selected for comparison; these included the pan-macrophage marker CD68, M1 macrophage markers CD86 and iNOS , and M2 macrophage markers CD163 and CD206 34,87 . In addition, the web server TIMER ( http://timer.cistrome.org/ ) was used to evaluate the correlation of SPP1 with the six immune cell and macrophage markers mentioned above. 88 Tumor mutation burden Simple nucleotide variation data from TCGA-CRC were used to calculate TMB, which is defined as the number of mutations per megabase. The visualization of the detailed gene mutation status of the high and low SPP1/CD68 groups was implemented in the R package "maftools" 89 . Prediction of immunotherapy response We used the SubMap algorithm to predict the clinical response to anti-PD1 and anti-CTLA4 immunotherapy. SubMap was used to compare the similarity of different expression profiles; this feature can reflect the response to treatment 90 . The expression profiles and associated annotation data of 47 melanoma patients used to define high- and low-risk groups were obtained from the Supplementary Materials of Lu et al . 91 In addition, samples in GSE39582 were divided into high and low SPP1 groups based on the median SPP1 expression. Survival analysis OS and PFI were regarded as the outcome events in the datasets (OS in GSE17536, GSE39582, and GSE17536; OS/PFI in TCGA). Survival analysis was performed based on the Kaplan‒Meier algorithm. The log-rank test was used to calculate p values between groups. Univariate Cox models were constructed by Cox proportional hazards regression. These analyses were implemented with "survival" and "survminer" in R. Functional and pathway enrichment analysis In this study, four different gene sets were used for functional and pathway enrichment analysis, including Gene Ontology (GO) Biological Process (BP), Kyoto Encyclopedia of Genes and Genomes (KEGG), WikiPathways, and Hallmark gene sets 92 . Metascape ( http://metascape.org/ ), a portal for gene function annotation analysis, was used for enrichment analysis of DEGs in this study (GO BP, KEGG, WikiPathways, and Hallmark gene sets) 93 . Gene set variation analysis (GSVA) implemented in the "GSVA" package (version 1.44.2) was used for gene set enrichment analysis of different macrophage subsets (Hallmark gene sets) 94 . LinkedOmics ( http://www.linkedomics.org/ ) is an online source portal for analyzing multiomics data of 32 cancers in TCGA 95 . In this study, LinkedOmics was used to perform GSEA of SPP1 expression profiles in TCGA-CRC (GO BP). In addition, GSEA software (version 4.2.3) ( https://www.gsea-msigdb.org/gsea/ ) was used to explore differences in functional enrichment between the L- SPP1/CD68 and H- SPP1/CD68 groups (Hallmark gene sets, permutations = 1000) 96 . | Normalized enrichment score (NES) | ≥1.0, false discovery rate (FDR) q-val ≤ 0.25, and p value < 0.05 were considered statistically significant. Statistical analysis R (version 4.1.3), GraphPad Prism (version 9), Excel, SPSS (version 26), and Python (version 3.7) were used for statistical analysis. The Mann‒Whitney U test was used to compare the differences between two groups. The Pearson or Spearman method was used for the correlation test. The log-rank method was used to calculate the P value in survival between groups. A chi-square test was used to compare clinical parameters between the high and low SPP1/CD68 groups. Data in bar plots of cell proportions for single-cell datasets are presented as the mean ± SEM (small sample size), while the mean ± SD is used for the remaining data. P value < 0.05 was considered statistically significant. Declarations Code availability The codes used in this study are available on Github (https://github.com/LPC19970117/SPP1-macrophages). Data availability The data used in this study are all from public datasets. Details can be found in the Methods section of the manuscript. Acknowledgments We would like to thank the researchers contributing to public databases and those providing data related to human medicine. We especially thank Dr. Wu for providing the CRLM data. This work was supported by the National Natural Science Foundation of China (82072655). Conflict of interest The authors declare no conflict of interest. Author contributions ZYX designed the study. ZYX and LRN collected the data. ZYX, FF and GZZ analyzed the data. ZYX, RKL, HJD, LLD, HZW, GMR and XYD visualized the data. ZYX and LRN drafted the manuscript. JZ and JYZ revised the manuscript. All authors have read and approved the final manuscript. All authors contributed to the article and approved the submitted version. Supplementary information accompanies the manuscript on the Cell Discovery website http://www.nature.com/celldisc’ References Siegel, R. L., Miller, K. D., Fuchs, H. E. & Jemal, A. Cancer statistics, 2022. CA Cancer J. Clin. 72 , 7–33 (2022). Fan, A. et al. Immunotherapy in colorectal cancer: current achievements and future perspective. Int. J. Biol. Sci. 17 , 3837–3849 (2021). Zaanan, A. et al. Role of deficient DNA mismatch repair status in patients with stage III colon cancer treated with FOLFOX adjuvant chemotherapy: a pooled analysis from 2 randomized clinical trials. JAMA Oncol. 4 , 379–383 (2018). Nordholm-Carstensen, A., Krarup, P. M., Morton, D., Harling, H. & Danish Colorectal Cancer Group. Mismatch repair status and synchronous metastases in colorectal cancer: a nationwide cohort study. Int. J. Cancer 137 , 2139–2148 (2015). Le, D. T. et al. PD-1 blockade in tumors with mismatch-repair deficiency. N. Engl. J. Med. 372 , 2509–2520 (2015). DeNardo, D. G. & Ruffell, B. Macrophages as regulators of tumour immunity and immunotherapy. Nat. Rev. Immunol. 19 , 369–382 (2019). Vogel, D. Y. et al. Human macrophage polarization in vitro: maturation and activation methods compared. Immunobiology 219 , 695–703 (2014). Ginhoux, F., Schultze, J. L., Murray, P. J., Ochando, J. & Biswas, S. K. New insights into the multidimensional concept of macrophage ontogeny, activation and function. Nat. Immunol. 17 , 34–40 (2016). Liu, Y. et al. Immune phenotypic linkage between colorectal cancer and liver metastasis. Cancer Cell 40 , 424–437.e5 (2022). Khantakova, D., Brioschi, S. & Molgora, M. Exploring the impact of TREM2 in tumor-associated macrophages. Vaccines (Basel) 10 , 943 (2022). Ding, J. et al. M2 macrophage-derived G-CSF promotes trophoblasts EMT, invasion and migration via activating PI3K/Akt/Erk1/2 pathway to mediate normal pregnancy. J. Cell. Mol. Med. 25 , 2136–2147 (2021). Wang, Z. et al. Polarization of intestinal tumour-associated macrophages regulates the development of schistosomal colorectal cancer. J. Cancer 12 , 1033–1041 (2021). Cortese, N. et al. Macrophages in colorectal cancer liver metastases. Cancers (Basel) 11 , 633 (2019). Pyonteck, S. M. et al. CSF-1R inhibition alters macrophage polarization and blocks glioma progression. Nat. Med. 19 , 1264–1272 (2013). Lee, J. H. et al. A phase I study of pexidartinib, a colony-stimulating factor 1 receptor inhibitor, in Asian patients with advanced solid tumors. Invest. New Drugs 38 , 99–110 (2020). Ries, C. H. et al. Targeting tumor-associated macrophages with anti-CSF-1R antibody reveals a strategy for cancer therapy. Cancer Cell 25 , 846–859 (2014). Papadopoulos, K. P. et al. First-in-human study of AMG 820, a monoclonal anti-colony-stimulating factor 1 receptor antibody, in patients with advanced solid tumors. Clin. Cancer Res. 23 , 5703–5710 (2017). Wu, Y. et al. Spatiotemporal immune landscape of colorectal cancer liver metastasis at single-cell level. Cancer Discov. 12 , 134–153 (2022). Zhang, L. et al. Single-cell analyses inform mechanisms of myeloid-targeted therapies in colon cancer. Cell 181 , 442–459.e29 (2020). Qi, J. et al. Single-cell and spatial analysis reveal interaction of FAP(+) fibroblasts and SPP1(+) macrophages in colorectal cancer. Nat. Commun. 13 , 1742 (2022). Feng, S. et al. SPP1 as a key gene in the lymph node metastasis and a potential predictor of poor prognosis in head and neck carcinoma. J. Oral Pathol. Med. http://dx.doi.org/10.1111/jop.13333 (2022). Liu, Y. et al. A pan-cancer analysis of the oncogenic role of secreted phosphoprotein 1 (SPP1) in human cancers. Ann. Transl. Med. 10 , 279 (2022). Zhao, K., Ma, Z. & Zhang, W. Comprehensive analysis to identify SPP1 as a prognostic biomarker in cervical cancer. Front. Genet. 12 , 732822 (2021). Xu, C. et al. SPP1, analyzed by bioinformatics methods, promotes the metastasis in colorectal cancer by activating EMT pathway. Biomed. Pharmacother. 91 , 1167–1177 (2017). Wei, J. et al. Characterizing intercellular communication of pan-cancer reveals SPP1 + tumor-associated macrophage expanded in hypoxia and promoting cancer malignancy through single-cell RNA-seq data. Front. Cell Dev. Biol. 9 , 749210 (2021). Rao, G. et al. Reciprocal interactions between tumor-associated macrophages and CD44-positive cancer cells via osteopontin/CD44 promote tumorigenicity in colorectal cancer. Clin. Cancer Res. 19 , 785–797 (2013). Senbanjo, L. T. & Chellaiah, M. A. CD44: a multifunctional cell surface adhesion receptor is a regulator of progression and metastasis of cancer cells. Front. Cell Dev. Biol. 5 , 18 (2017). Assidi, M. et al. Prognostic value of Osteopontin (SPP1) in colorectal carcinoma requires a personalized molecular approach. Tumour Biol. 41 , 1010428319863627 (2019). Pietras, A. et al. Osteopontin-CD44 signaling in the glioma perivascular niche enhances cancer stem cell phenotypes and promotes aggressive tumor growth. Cell Stem Cell 14 , 357–369 (2014). Pang, X. et al. Osteopontin as a multifaceted driver of bone metastasis and drug resistance. Pharmacol. Res. 144 , 235–244 (2019). Clay, R. & Siddiqi, S. A. in Theranostic Approach for Pancreatic Cancer (eds Nagaraju, G. P. & Ahmad, S.) (Academic Press, 2019). Cheng, S. et al. A pan-cancer single-cell transcriptional atlas of tumor infiltrating myeloid cells. Cell 184 , 792–809.e23 (2021). Ren, X. et al. Insights gained from single-cell analysis of immune cells in the tumor microenvironment. Annu. Rev. Immunol. 39 , 583–609 (2021). Wu, K. et al. Redefining tumor-associated macrophage subpopulations and functions in the tumor microenvironment. Front. Immunol. 11 , 1731 (2020). Szulzewsky, F. et al. Glioma-associated microglia/macrophages display an expression profile different from M1 and M2 polarization and highly express Gpnmb and Spp1. PLoS One 10 , e0116644 (2015). Liu, L. et al. Construction of TME and Identification of crosstalk between malignant cells and macrophages by SPP1 in hepatocellular carcinoma. Cancer Immunol. Immunother. 71 , 121–136 (2022). Loustau, M. et al. HLA-G neo-expression on tumors. Front. Immunol. 11 , 1685 (2020). Zhang, Y. & Zhang, Z. The history and advances in cancer immunotherapy: understanding the characteristics of tumor-infiltrating immune cells and their therapeutic implications. Cell. Mol. Immunol. 17 , 807–821 (2020). Ponten, F., Jirstrom, K. & Uhlen, M. The human protein Atlas–a tool for pathology. J. Pathol. 216 , 387–393 (2008). Li, J. et al. Tumor-associated macrophage infiltration and prognosis in colorectal cancer: systematic review and meta-analysis. Int. J. Colorectal Dis. 35 , 1203–1210 (2020). Bruni, D., Angell, H. K. & Galon, J. The immune contexture and immunoscore in cancer prognosis and therapeutic efficacy. Nat. Rev. Cancer 20 , 662–680 (2020). Pridans, C. et al. Pleiotropic impacts of macrophage and microglial deficiency on development in rats with targeted mutation of the Csf1r locus. J. Immunol. 201 , 2683–2699 (2018). Guilliams, M., Thierry, G. R., Bonnardel, J. & Bajenoff, M. Establishment and maintenance of the macrophage niche. Immunity 52 , 434–451 (2020). Lee, H. O. et al. Lineage-dependent gene expression programs influence the immune landscape of colorectal cancer. Nat. Genet. 52 , 594–603 (2020). Amilca-Seba, K. et al. Osteopontin (OPN/SPP1), a mediator of tumor progression, is regulated by the mesenchymal transcription factor Slug/SNAI2 in colorectal cancer (CRC). Cells 11 , 1808 (2022). Tu, Y., Chen, C. & Fan, G. Association between the expression of secreted phosphoprotein - related genes and prognosis of human cancer. BMC Cancer 19 , 1230 (2019). Azizi, E. et al. Single-cell map of diverse immune phenotypes in the breast tumor microenvironment. Cell 174 , 1293–1308.e36 (2018). Xie, Z. et al. Immune cell confrontation in the papillary thyroid carcinoma microenvironment. Front. Endocrinol. (Lausanne) 11 , 570604 (2020). Chen, C., Zhao, S., Karnad, A. & Freeman, J. W. The biology and role of CD44 in cancer progression: therapeutic implications. J. Hematol. Oncol. 11 , 64 (2018). He, X. et al. Tumor-initiating stem cell shapes its microenvironment into an immunosuppressive barrier and pro-tumorigenic niche. Cell Rep. 36 , 109674 (2021). Schlesinger, M. & Bendas, G. Contribution of very late antigen-4 (VLA-4) integrin to cancer progression and metastasis. Cancer Metastasis Rev. 34 , 575–591 (2015). Liu, J. et al. The IRENA lncRNA converts chemotherapy-polarized tumor-suppressing macrophages to tumor-promoting phenotypes in breast cancer. Nat. Cancer 2 , 457–473 (2021). Sidaway, P. MSI-H: a truly agnostic biomarker? Nat. Rev. Clin. Oncol. 17 , 68 (2020). Yarchoan, M., Hopkins, A. & Jaffee, E. M. Tumor mutational burden and response rate to PD-1 inhibition. N. Engl. J. Med. 377 , 2500–2501 (2017). Cristescu, R. et al. Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy. Science 362 , eaar3593 (2018). Panda, A. et al. Identifying a clinically applicable mutational burden threshold as a potential biomarker of response to immune checkpoint therapy in solid tumors. JCO Precis. Oncol. 2017 , PO.17.00146 (2017). Klement, J. D. et al. Osteopontin blockade immunotherapy increases cytotoxic T lymphocyte lytic activity and suppresses colon tumor progression. Cancers (Basel) 13 , 1006 (2021). Deng, G. et al. BET inhibitor suppresses melanoma progression via the noncanonical NF-kappaB/SPP1 pathway. Theranostics 10 , 11428–11443 (2020). Farrokhi, V., Chabot, J. R., Neubert, H. & Yang, Z. Assessing the feasibility of neutralizing osteopontin with various therapeutic antibody modalities. Sci. Rep. 8 , 7781 (2018). Pelka, K. et al. Spatially organized multicellular immune hubs in human colorectal cancer. Cell 184 , 4734–4752.e20 (2021). Che, L. H. et al. A single-cell atlas of liver metastases of colorectal cancer reveals reprogramming of the tumor microenvironment in response to preoperative chemotherapy. Cell Discov. 7 , 80 (2021). Sharma, A. et al. Onco-fetal reprogramming of endothelial cells drives immunosuppressive macrophages in hepatocellular carcinoma. Cell 183 , 377–394.e21 (2020). Ki, D. H. et al. Whole genome analysis for liver metastasis gene signatures in colorectal cancer. Int. J. Cancer 121 , 2005–2012 (2007). Jorissen, R. N. et al. Metastasis-associated gene expression changes predict poor outcomes in patients with dukes stage B and C colorectal cancer. Clin. Cancer Res. 15 , 7642–7651 (2009). Smith, J. J. et al. Experimentally derived metastasis gene expression profile predicts recurrence and death in patients with colon cancer. Gastroenterology 138 , 958–968 (2010). Gaedcke, J. et al. Mutated KRAS results in overexpression of DUSP4, a MAP-kinase phosphatase, and SMYD3, a histone methyltransferase, in rectal carcinomas. Genes Chromosomes Cancer 49 , 1024–1034 (2010). Skrzypczak, M. et al. Modeling oncogenic signaling in colon tumors by multidirectional analyses of microarray data directed for maximization of analytical reliability. PLoS One 5 , e13091 (2010). Marisa, L. et al. Gene expression classification of colon cancer into molecular subtypes: characterization, validation, and prognostic value. PLoS Med. 10 , e1001453 (2013). Sheffer, M. et al. Association of survival and disease progression with chromosomal instability: a genomic exploration of colorectal cancer. Proc. Natl. Acad. Sci. U. S. A. 106 , 7131–7136 (2009). Sole, X. et al. Discovery and validation of new potential biomarkers for early detection of colon cancer. PLoS One 9 , e106748 (2014). Ryan, B. M. et al. Germline variation in NCF4, an innate immunity gene, is associated with an increased risk of colorectal cancer. Int. J. Cancer 134 , 1399–1407 (2014). Kwon, Y. et al. Prognosis of stage III colorectal carcinomas with FOLFOX adjuvant chemotherapy can be predicted by molecular subtype. Oncotarget 8 , 39367–39381 (2017). Hu, Y. et al. Colorectal cancer susceptibility loci as predictive markers of rectal cancer prognosis after surgery. Genes Chromosomes Cancer 57 , 140–149 (2018). Kamal, Y., Schmit, S. L., Hoehn, H. J., Amos, C. I. & Frost, H. R. Transcriptomic differences between primary colorectal adenocarcinomas and distant metastases reveal metastatic colorectal cancer subtypes. Cancer Res. 79 , 4227–4241 (2019). Szeglin, B. C. et al. A SMAD4-modulated gene profile predicts disease-free survival in stage II and III colorectal cancer. Cancer Rep. (Hoboken) 5 , e1423 (2022). McInnes, L. & Healy, J. UMAP: uniform manifold approximation and projection for dimension reduction. arXiv:1802.03426 (2018). Mulder, K. et al. Cross-tissue single-cell landscape of human monocytes and macrophages in health and disease. Immunity 54 , 1883–1900.e5 (2021). Guilliams, M. & Scott, C. L. Does niche competition determine the origin of tissue-resident macrophages? Nat. Rev. Immunol. 17 , 451–460 (2017). Korsunsky, I. et al. Fast, sensitive and accurate integration of single-cell data with Harmony. Nat. Methods 16 , 1289–1296 (2019). Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32 , 381–386 (2014). Efremova, M., Vento-Tormo, M., Teichmann, S. A. & Vento-Tormo, R. CellPhoneDB: inferring cell-cell communication from combined expression of multi-subunit ligand-receptor complexes. Nat. Protoc. 15 , 1484–1506 (2020). Wei, J. et al. Characterization of glycolysis-associated molecules in the tumor microenvironment revealed by pan-cancer tissues and lung cancer single cell data. Cancers (Basel) 12 , 1788 (2020). Mak, M. P. et al. A patient-derived, Pan-cancer EMT signature identifies global molecular alterations and immune target enrichment following epithelial-to-mesenchymal transition. Clin. Cancer Res. 22 , 609–620 (2016). Whitfield, M. L., George, L. K., Grant, G. D. & Perou, C. M. Common markers of proliferation. Nat. Rev. Cancer 6 , 99–106 (2006). Yoshihara, K. et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat. Commun. 4 , 2612 (2013). Newman, A. M. et al. Robust enumeration of cell subsets from tissue expression profiles. Nat. Methods 12 , 453–457 (2015). Zhao, Y. et al. Prognostic value and clinicopathological roles of phenotypes of tumour-associated macrophages in colorectal cancer. J. Cancer Res. Clin. Oncol. 145 , 3005–3019 (2019). Li, T. et al. TIMER: a web server for comprehensive analysis of tumor-infiltrating immune cells. Cancer Res. 77 , e108–e110 (2017). Mayakonda, A., Lin, D. C., Assenov, Y., Plass, C. & Koeffler, H. P. Maftools: efficient and comprehensive analysis of somatic variants in cancer. Genome Res. 28 , 1747–1756 (2018). Hoshida, Y., Brunet, J. P., Tamayo, P., Golub, T. R. & Mesirov, J. P. Subclass mapping: identifying common subtypes in independent disease data sets. PLoS One 2 , e1195 (2007). Lu, X. et al. Immune signature-based subtypes of cervical squamous cell carcinoma tightly associated with human papillomavirus type 16 expression, molecular features, and clinical outcome. Neoplasia 21 , 591–601 (2019). Liberzon, A. et al. The molecular signatures database (MSigDB) hallmark gene set collection. Cell Syst. 1 , 417–425 (2015). Zhou, Y. et al. Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat. Commun. 10 , 1523 (2019). Hanzelmann, S., Castelo, R. & Guinney, J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics 14 , 7 (2013). Vasaikar, S. V., Straub, P., Wang, J. & Zhang, B. LinkedOmics: analyzing multi-omics data within and across 32 cancer types. Nucleic Acids Res. 46 , D956–D963 (2018). Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. U. S. A. 102 , 15545–15550 (2005). Additional Declarations (Not answered) Supplementary Files figS1.tif Fig. S1. Construction of the CRC-Mix cohort and annotation of myeloid cells. figS2.tif Fig. S2. Hallmark gene set enrichment analysis of monocytes/macrophages. figS3.tif Fig. S3. Phenotypes of different macrophage subsets. figS4.tif Fig. S4. Enrichment of SPP1 expression in macrophages. figS5.tif Fig. S5. Enrichment of CSF1R expression in C1QC+ macrophages. TableS1.xlsx Table S1. Major cell type composition, myeloid cell subset composition, and clinical information for each sample.Table S2. Differentially expressed genes in myeloid subgroups in CRC-Mix TableS2.xlsx Table S2. Differentially expressed genes in myeloid subgroups in CRC-Mix TableS3.xlsx Table S3. Differentially expressed genes in monocyte/macrophage between NC and CRC. TableS4.xlsx Table S4. Differentially expressed genes in monocyte/macrophage between CRC and LM. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-1984416","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":133579089,"identity":"d7b3e82e-5dd9-4409-b5f8-310d8dc473f7","order_by":0,"name":"Jianyong Zheng","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAuklEQVRIie3PsQqDMBCA4YSApXCSNWLBV3ArgrSvkmJxcujoGBB8hkj7HJ0jrqFdC+6d4xvU4gucW6H5pxvugztCfL5fTBBJXJ0D52oFodqWu0gbPCFs2w55qiRSJFf1jol9QkoMdVOFIPRmyvhSj7BnikXdHUGYkGWs7QiZMgELMST4krB9QGokksBCzAoihCwybQuIdN/gfkl0dXq5+nDkvOndhCHzaWexDFSh9uc2g8Ou+nw+35/2AbgmNi8n9eZ6AAAAAElFTkSuQmCC","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":true,"prefix":"","firstName":"Jianyong","middleName":"","lastName":"Zheng","suffix":""},{"id":133579090,"identity":"fc54314b-b098-4956-94a6-9913b13dc4e9","order_by":1,"name":"Zhenyu Xie","email":"","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zhenyu","middleName":"","lastName":"Xie","suffix":""},{"id":133579091,"identity":"61cef70b-c611-47d3-9c57-923e85818a20","order_by":2,"name":"Liaoran Niu","email":"","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Liaoran","middleName":"","lastName":"Niu","suffix":""},{"id":133579092,"identity":"f8a61bd5-3279-496b-8b35-5a20eeb6277a","order_by":3,"name":"Gaozan Zheng","email":"","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Gaozan","middleName":"","lastName":"Zheng","suffix":""},{"id":133579093,"identity":"c35d6ca6-5123-4afb-b520-22c597ea97ee","order_by":4,"name":"Ruikai Li","email":"","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ruikai","middleName":"","lastName":"Li","suffix":""},{"id":133579094,"identity":"0e1ee5ca-a3a9-4adb-a2ca-0ced816bd6fd","order_by":5,"name":"Hanjun Dan","email":"","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hanjun","middleName":"","lastName":"Dan","suffix":""},{"id":133579095,"identity":"ea61eb95-f92c-4bbc-bf49-ccbc8d5c6a80","order_by":6,"name":"Lili Duan","email":"","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Lili","middleName":"","lastName":"Duan","suffix":""},{"id":133579096,"identity":"919ce5d5-fdde-41d2-918e-d3574098d986","order_by":7,"name":"Hongze Wu","email":"","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hongze","middleName":"","lastName":"Wu","suffix":""},{"id":133579097,"identity":"f27f96e2-b063-44d0-8ccc-e7acf8f40bb3","order_by":8,"name":"Guangming Ren","email":"","orcid":"","institution":"Xi'an Medical University","correspondingAuthor":false,"prefix":"","firstName":"Guangming","middleName":"","lastName":"Ren","suffix":""},{"id":133579098,"identity":"08680cb7-bf2e-4364-afbd-98d7a49ec1f5","order_by":9,"name":"Xinyu Dou","email":"","orcid":"","institution":"Xi'an Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xinyu","middleName":"","lastName":"Dou","suffix":""},{"id":133579099,"identity":"6da1ac2b-67fd-4680-ba9a-714fccc6996b","order_by":10,"name":"Fan Feng","email":"","orcid":"","institution":"Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Fan","middleName":"","lastName":"Feng","suffix":""},{"id":133579100,"identity":"e536d093-e535-4484-9f4f-6236588b90e6","order_by":11,"name":"Jian Zhang","email":"","orcid":"https://orcid.org/0000-0001-9129-9092","institution":"School of Basic Medicine, Fourth Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jian","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2022-08-22 02:40:48","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-1984416/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-1984416/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":26187036,"identity":"cb5b76fa-ed19-4a70-8abb-c33d6c1a210b","added_by":"auto","created_at":"2022-09-07 16:30:30","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":259431,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eScRNA-seq revealed the myeloid cell landscape of CRC.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e. Tissue types of colorectal cancer liver metastases, including CRC, NC, LM, and NL samples and PBMCs. CRC, colorectal cancer; NC, normal colorectum (adjacent colorectum); LM, liver metastases; NL, normal liver (adjacent); PBMC, peripheral blood mononuclear cell. NC and NL samples include both adjacent and healthy tissue.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb\u003c/strong\u003e. Summary of the seven CRC single-cell datasets in this study. The green box indicates the sample origin of the CRC-Mix dataset. Numbers indicate sample size. Other cells, nonimmune cells including malignant cells and stromal cells; MMR, mismatch repair; MSI, microsatellite instability.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e. Barplots showing comparison of the proportion of myeloid cells to immune cells between the NC and CRC groups in six CRC scRNA-seq datasets. *, p \u0026lt; 0.05; ****, p \u0026lt; 0.0001. Sample size above bar; mean ± SEM; Mann‒Whitney U test.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed\u003c/strong\u003e. Tissue\u003cem\u003e\u003cstrong\u003e \u003c/strong\u003e\u003c/em\u003edistribution preference of each myeloid subset estimated by OR.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ee\u003c/strong\u003e. UMAP plot of 12 myeloid subsets from the CRC-Mix dataset.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ef\u003c/strong\u003e. Representative genes and functional properties of myeloid subsets of the CRC-Mix dataset.\u003c/p\u003e","description":"","filename":"Onlinefig1.png","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/ab324423c1da3ea321ddb304.png"},{"id":26187153,"identity":"20097d02-d147-452f-b2fc-dfda915054cb","added_by":"auto","created_at":"2022-09-07 16:35:30","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":311631,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eChanges in monocytes/macrophages during CRLM.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e. Volcano plot showing DEGs in monocytes/macrophages between NC and CRC samples in the GSE164552 dataset.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb\u003c/strong\u003e. Venn diagrams showing the number of common DEGs that were either up- or downregulated across the six scRNA-seq datasets (NC vs. CRC). Representative common DEGs are shown in the right box.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e. Metascape bar graph showing functional enrichment analysis results for common DEGs (NC vs. CRC). The top 20 enriched terms are displayed. The red, blue, and green backgrounds indicate that the enriched terms are based on GO BP, KEGG, and WikiPathways gene sets, respectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed\u003c/strong\u003e. Volcano plot showing DEGs in monocytes/macrophages between CRC and LM samples in the GSE164552 dataset.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ee\u003c/strong\u003e. Venn diagrams showing the number of common DEGs that were either up- or downregulated across the three scRNA-seq datasets (CRC vs. LM). Representative common DEGs are shown in the right box.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ef\u003c/strong\u003e. Metascape bar graph showing functional enrichment analysis of common DEGs (CRC vs. LM). The top 20 enriched terms are displayed.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eg-i\u003c/strong\u003e shows trajectory analysis of monocytes/macrophages (Wu\u003cem\u003e et al\u003c/em\u003e.)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eg\u003c/strong\u003e. Cells were color-coded by monocyte/macrophage subset.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eh\u003c/strong\u003e. Cells were color-coded by tissue type. Arrows indicate the pseudotime direction.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ei\u003c/strong\u003e. Expression levels of monocyte/macrophage marker genes in pseudotime order. The color of the dots shows the tissue types of the cells.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ej\u003c/strong\u003e. Heatmap showing the GSVA score of 50 hallmark pathwaysof different monocyte/macrophage subsets (Wu\u003cem\u003e et al\u003c/em\u003e.)\u003c/p\u003e","description":"","filename":"Onlinefig2.png","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/f2ee09df789fc36e4d8e106c.png"},{"id":26187037,"identity":"6675f92a-fb46-4da4-9bfd-219fc31607a6","added_by":"auto","created_at":"2022-09-07 16:30:30","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":298825,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTissue distribution and potential mechanisms of SPP1+ macrophages.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e. Bar graphs showing the ratio of SPP1+ macrophages to all macrophages in the six scRNA-seq datasets (NC vs. CRC). *, p \u0026lt; 0.05; **, p \u0026lt; 0.01; ***, p \u0026lt; 0.001; ****, p \u0026lt; 0.0001; ns, not significant. Sample size above bar; mean ± SEM; Mann‒Whitney U test.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb\u003c/strong\u003e. Bar graphs showing the ratio of SPP1+ macrophages to all macrophages in the three scRNA-seq datasets (CRC vs. LM).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e. Bar graphs showing the ratio of SPP1+ macrophages to all macrophages in different tissue types (CRC, NC, LM, NL and HCC) in the GSE164522 and GSE156625 datasets.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed\u003c/strong\u003e. Pie charts show the proportions of four macrophage subsets in the CRC-Mix dataset.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ee\u003c/strong\u003e-\u003cstrong\u003eg\u003c/strong\u003e. The expression of (\u003cstrong\u003ee\u003c/strong\u003e) EMT, (\u003cstrong\u003ef\u003c/strong\u003e) glycolysis, and (\u003cstrong\u003eg\u003c/strong\u003e) hypoxia signatures in the three major macrophage subsets in the CRC-Mix dataset.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eh\u003c/strong\u003e. Heatmap showing the mean expression levels of the top 10 significant DEGs in the four macrophage subsets (CRC-Mix) (left). Exp, z score normalized mean expression. Metascape performs enrichment analysis of DEGs in each subset of macrophages, and the top five enriched terms based on GO BP, KEGG, and WikiPathways gene sets are displayed (right). Lollipop plot showing the expression of inflammatory and anti-inflammatory signatures in the four macrophage subsets (top).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ei\u003c/strong\u003e. The upper and right box plots show the expression of M1 and M2 signatures in the three major macrophage subsets, respectively (CRC-Mix). Scatter plot showing the correlation between the M1 and M2 signatures in macrophages. Pearson correlation and linear regression.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ej\u003c/strong\u003e. Heatmap showing cellular interactions among the four macrophage subsets and six major cell types (GSE178341). The redder the color is, the higher the predicted number of intercellular receptor‒ligand pairs.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ek\u003c/strong\u003e. Receptor‒ligand interactions among four macrophage subsets and six major cell types (GSE178341). P values are indicated by circle size; colors indicate the average expression levels of ligand and receptor or molecules in the corresponding cell types.\u003c/p\u003e","description":"","filename":"Onlinefig3.png","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/cb4bb31a08a60c28c058f793.png"},{"id":26187405,"identity":"47ca59b3-8ac1-4bc8-aa88-bb2dd2b116d9","added_by":"auto","created_at":"2022-09-07 16:40:31","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":337059,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSPP1, a novel macrophage marker in CRC.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e. Heatmap showing the proportion of 22 immune cells (CIBERSORT), immune score (ESTIMATE), expression of markers (\u003cem\u003eCD45\u003c/em\u003e, \u003cem\u003eCD68\u003c/em\u003e, \u003cem\u003eCD86\u003c/em\u003e, \u003cem\u003eiNOS\u003c/em\u003e, \u003cem\u003eCD163\u003c/em\u003e, \u003cem\u003eCD206\u003c/em\u003e, and \u003cem\u003eSPP1\u003c/em\u003e), \u003cem\u003eCD68\u003c/em\u003e/\u003cem\u003eCD45\u003c/em\u003e, and \u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003ein 10 bulk RNA-seq datasets (NC vs. CRC). The color of the bricks represents significance; the redder the color is, the higher the value in the CRC compared to NC, and the opposite is true for blue. The expression of \u003cem\u003eSPP1\u003c/em\u003e in GSE106582 shows an example of brick color meaning. Method (right), dataset source (top), sample size (bottom).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb\u003c/strong\u003e. Correlation analysis between SPP1 expression and the proportion of immune cells in TCGA-CRC (COAD/READ) (CIBERSORT, n=647). Significantly positively and negatively correlated immune cells are shown in red and blue, respectively. Spearman correlation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e. Violin plots showing the expression of \u003cem\u003eSPP1\u003c/em\u003e in CD45- cells, other immune cells, other myeloid cells, and SPP1+ macrophages (GSE178341).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed\u003c/strong\u003e. UMAP plot showing the distribution of CD45- cells, other immune cells, other myeloid cells, and SPP1+ macrophages (left). UMAP of iterative subsets of cells from the global level to immune cells to myeloid cells showing the enrichment of SPP1 expression levels in different cells (right).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ee\u003c/strong\u003e. Images of immunohistochemical staining showing SPP1 expression in NC, CRC, NL and HCC samples. Data from THPA. Scale bar, 200 μm.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ef\u003c/strong\u003e. Significantly enriched GO biological process terms in genes coexpressed with SPP1 in the TCGA-CRC cohort. FDR, false discovery rate.\u003c/p\u003e","description":"","filename":"Onlinefig4.png","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/148fc176309e823ec3d66690.png"},{"id":26187040,"identity":"f233ca56-a37f-4d2a-8b72-aa0986d84fe9","added_by":"auto","created_at":"2022-09-07 16:30:30","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":186664,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eClinical value of SPP1+ macrophages in CRC (bulk RNA-seq).\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e. Heatmap showing the immune score (ESTIMATE) and expression of markers (\u003cem\u003eCD45\u003c/em\u003e, \u003cem\u003eCD68\u003c/em\u003e, \u003cem\u003eCD86\u003c/em\u003e, \u003cem\u003eiNOS\u003c/em\u003e, \u003cem\u003eCD163\u003c/em\u003e, \u003cem\u003eCD206\u003c/em\u003e, and \u003cem\u003eSPP1\u003c/em\u003e) in five bulk RNA-seq CRC datasets (stage I/II vs. stage III/IV). The color of the bricks represents significance; the redder the color is, the higher the value in the CRC compared to NC, and the bluer the color is, the lower the value. dataset source (top), sample size (bottom). *, GSE14333 uses the Duke stage.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb\u003c/strong\u003e. Heatmap showing the immune score and expression of markers (\u003cem\u003eCD45\u003c/em\u003e, \u003cem\u003eCD68\u003c/em\u003e, \u003cem\u003eCD86\u003c/em\u003e, \u003cem\u003eiNOS\u003c/em\u003e, \u003cem\u003eCD163\u003c/em\u003e, \u003cem\u003eCD206\u003c/em\u003e, and \u003cem\u003eSPP1\u003c/em\u003e) in three bulk RNA-seq CRC datasets (stage I/II vs. stage III/IV).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e. Heatmap showing the immune score and expression of markers (\u003cem\u003eCD45\u003c/em\u003e, \u003cem\u003eCD68\u003c/em\u003e, \u003cem\u003eCD86\u003c/em\u003e, \u003cem\u003eiNOS\u003c/em\u003e, \u003cem\u003eCD163\u003c/em\u003e, \u003cem\u003eCD206\u003c/em\u003e, and \u003cem\u003eSPP1\u003c/em\u003e) in three bulk RNA-seq CRC datasets (N0 vs. stage N1-3). Six N3 samples were present in GSE39582.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed\u003c/strong\u003e. Heatmap showing the immune score and expression of markers (\u003cem\u003eCD45\u003c/em\u003e, \u003cem\u003eCD68\u003c/em\u003e, \u003cem\u003eCD86\u003c/em\u003e, \u003cem\u003eiNOS\u003c/em\u003e, \u003cem\u003eCD163\u003c/em\u003e, \u003cem\u003eCD206\u003c/em\u003e, and \u003cem\u003eSPP1\u003c/em\u003e) in three bulk RNA-seq CRC datasets (M0 vs. M1).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ee\u003c/strong\u003e. Heatmap showing the immune score and expression of markers (\u003cem\u003eCD45\u003c/em\u003e, \u003cem\u003eCD68\u003c/em\u003e, \u003cem\u003eCD86\u003c/em\u003e, \u003cem\u003eiNOS\u003c/em\u003e, \u003cem\u003eCD163\u003c/em\u003e, \u003cem\u003eCD206\u003c/em\u003e, and \u003cem\u003eSPP1\u003c/em\u003e) in six bulk RNA-seq CRC datasets (CRC vs. LM).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ef\u003c/strong\u003e.\u003cem\u003e \u003c/em\u003eUnivariate Cox analysis of macrophage markers in four CRC datasets (TCGA, GSE17536, GSE39582, and GSE41258).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eg\u003c/strong\u003e. GSEA was performed between the high and low \u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e groups, and the top 10 enriched hallmark terms in the H-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e group are displayed. CI, confidence interval; GSEA, gene set enrichment analysis; HR, hazard ratio.\u003c/p\u003e","description":"","filename":"Onlinefig5.png","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/6d53af2c492e1f444dc8f31e.png"},{"id":26187048,"identity":"9bff4426-60b8-4072-aacf-f7367434f060","added_by":"auto","created_at":"2022-09-07 16:30:31","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":219125,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSignificance of SPP1+ macrophages for CRC treatment.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e. Kaplan‒Meier survival curve showing that \u003cem\u003eSPP1/CD6\u003c/em\u003e8 were associated with worse OS in TCGA-CRC cohorts (n = 593). Log-rank test.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb\u003c/strong\u003e-\u003cstrong\u003ec\u003c/strong\u003e. Box plots showing the effect of preoperative chemotherapy on SPP1 expression in macrophages in the scRNA-seq cohorts of (B) GSE178318 and (C) Wu \u003cem\u003eet al\u003c/em\u003e. R, responders; NR, nonresponders. Mann‒Whitney U test.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed\u003c/strong\u003e-\u003cstrong\u003ei\u003c/strong\u003e. The association between SPP1+ macrophages and genome instability and mutations.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed\u003c/strong\u003e. Comparison of the ratio of SPP1+ macrophages to all macrophages in MSI-H versus MSS CRC (CRC-Mix scRNA-seq cohort).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ee\u003c/strong\u003e. Comparison of the ratio of SPP1+ macrophages to all macrophages in dMMR versus pMMR CRC (CRC-Mix scRNA-seq cohort).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ef\u003c/strong\u003e. Comparison of \u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e in MSI-H versus MSS CRC (TCGA-CRC bulk RNA-seq cohort).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eg\u003c/strong\u003e. Comparison of \u003cem\u003eSPP1\u003c/em\u003e expression in MSI-H versus MSS CRC (GSE39582 bulk RNA-seq cohort).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eh\u003c/strong\u003e. The waterfall plot shows the mutation profiles of the top 30 mutated genes in the L-\u003cem\u003eSPP1/CD68\u003c/em\u003e group. Colors of annotations indicate different mutation types (bottom).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ei\u003c/strong\u003e. The waterfall plot shows the mutation profiles of the top 30 mutated genes in the H-\u003cem\u003eSPP1/CD68\u003c/em\u003e group.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ej\u003c/strong\u003e. SubMap analysis in TCGA-CRC showed that the H-\u003cem\u003eSPP1/CD68\u003c/em\u003e group was more sensitive to CTLA4 inhibitors (Bonferroni-corrected P = 0.023) and PD1 inhibitors (Bonferroni-corrected P = 0.021) than the L-\u003cem\u003eSPP1/CD68\u003c/em\u003e group.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ek\u003c/strong\u003e. SubMap analysis in GSE39582 showed that the H-\u003cem\u003eSPP1\u003c/em\u003e group was more sensitive to the CTLA4 inhibitor (Bonferroni-corrected P = 0.023) and PD1 inhibitor (Bonferroni-corrected P = 0.024).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003el\u003c/strong\u003e. The expression of \u003cem\u003eCSF1R\u003c/em\u003e in the three major macrophage subsets (CRC-Mix scRNA-seq cohort).\u003c/p\u003e","description":"","filename":"Onlinefig6.png","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/3032f3c01338a784186ed6e1.png"},{"id":26187044,"identity":"e5f9ec14-9d03-471e-b0cf-e6e27767ef3c","added_by":"auto","created_at":"2022-09-07 16:30:31","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":126001,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSchematic diagramof the SPP1+ macrophage model in CRC.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea\u003c/strong\u003e. Classification of macrophage subsets in CRC. Macrophages in the scRNA-seq CRC cohort were divided into FCN1+ macrophages, C1QC+ macrophages, SPP1+ macrophages and MKI67+ macrophages. Inflammation, phagocytosis, malignancy, and proliferation were defined as core features of these four macrophage subsets.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb\u003c/strong\u003e. Cell origin and tissue distribution of SPP1+ macrophage. Model of the developmental trajectory of monocyte/macrophage lineages in different sample types (PBMCs and NC, CRC, NL, LM and HCC tissues) from CRLM and HCC patients.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec\u003c/strong\u003e. SPP1+ macrophages as a malignant marker of CRC. The number and proportion of SPP1+ macrophages continued to increase during the occurrence, progression and metastasis of CRC. The sizes of the blue, yellow and red dots represent the numbers of all immune cells, all macrophages, and SPP1+ macrophages in the sample, respectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed\u003c/strong\u003e. Potential mechanisms of SPP1+ macrophages. SPP1+ macrophages exert biological functions through potentially related pathways (left). Red indicates that the signature is specifically elevated in SPP1+ macrophages compared to other macrophages, yellow indicates that the signature is elevated in SPP1+ macrophages compared to C1QC+ macrophages, and blue indicates that the signature is decreased in SPP1+ macrophages compared to C1QC+ macrophages. \u003cem\u003ePD-L1\u003c/em\u003eexpression was elevated in SPP1+ macrophages compared to other macrophages. Crosstalk between SPP1+ macrophages and other cell subsets (right).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ee\u003c/strong\u003e. Implications of SPP1+ macrophages for CRC immunotherapy. Patients with a high proportion of SPP1+ macrophages had a worse prognosis, along with signatures related to genome instability and mutations (dMMR, MSI-H, and H-TMB). Immunotherapy has potential to improve outcomes in patients with a high proportion of SPP1+ macrophages, whereas targeting \u003cem\u003eCSF1R\u003c/em\u003e is less effective.\u003c/p\u003e","description":"","filename":"Onlinefig7.png","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/b1777684f275706aaec09521.png"},{"id":26187637,"identity":"47061fa5-f5cd-4f25-9fb5-30558cc1ae9b","added_by":"auto","created_at":"2022-09-07 16:45:40","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1010996,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/7734d060-bb9a-4d6c-bedc-b73db5ec17b3.pdf"},{"id":26187038,"identity":"0beacf1f-e770-42da-85d5-4f87118f271d","added_by":"auto","created_at":"2022-09-07 16:30:30","extension":"tif","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":4438528,"visible":true,"origin":"","legend":"\u003cp\u003eFig. S1. Construction of the CRC-Mix cohort and annotation of myeloid cells.\u003c/p\u003e","description":"","filename":"figS1.tif","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/02cd21461aef0adb4ae9d340.tif"},{"id":26187154,"identity":"005e4a9b-9804-4d96-a286-6d6a772593e7","added_by":"auto","created_at":"2022-09-07 16:35:30","extension":"tif","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":4664468,"visible":true,"origin":"","legend":"\u003cp\u003eFig. S2. Hallmark gene set enrichment analysis of monocytes/macrophages.\u003c/p\u003e","description":"","filename":"figS2.tif","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/a9678caaffdb2f81c1ca743a.tif"},{"id":26187404,"identity":"03563099-c647-4770-b423-345b21887c25","added_by":"auto","created_at":"2022-09-07 16:40:30","extension":"tif","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":1097756,"visible":true,"origin":"","legend":"\u003cp\u003eFig. S3. Phenotypes of different macrophage subsets.\u003c/p\u003e","description":"","filename":"figS3.tif","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/db38116f5a906218e4f7f5b6.tif"},{"id":26187407,"identity":"fa4f6132-6a86-463e-8d50-b2f9729804ce","added_by":"auto","created_at":"2022-09-07 16:40:31","extension":"tif","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":3718252,"visible":true,"origin":"","legend":"\u003cp\u003eFig. S4. Enrichment of SPP1 expression in macrophages.\u003c/p\u003e","description":"","filename":"figS4.tif","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/bfd687c16450e4d59a0910e3.tif"},{"id":26187158,"identity":"b69e419f-fc0e-4582-a3f9-25a112f20264","added_by":"auto","created_at":"2022-09-07 16:35:31","extension":"tif","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":673788,"visible":true,"origin":"","legend":"\u003cp\u003eFig. S5. Enrichment of CSF1R expression in C1QC+ macrophages.\u003c/p\u003e","description":"","filename":"figS5.tif","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/18f8eaced764e9a06d7805e3.tif"},{"id":26187631,"identity":"19d6a351-3d57-4e4d-8fbc-d08ca5025de2","added_by":"auto","created_at":"2022-09-07 16:45:31","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":733677,"visible":true,"origin":"","legend":"\u003cp\u003eTable S1. Major cell type composition, myeloid cell subset composition, and clinical information for each sample.Table S2. Differentially expressed genes in myeloid subgroups in CRC-Mix\u003c/p\u003e","description":"","filename":"TableS1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/768c0ec8b833595842a78ee6.xlsx"},{"id":26187051,"identity":"c13d6f19-bbd7-4d5d-aa95-8e50c75d0581","added_by":"auto","created_at":"2022-09-07 16:30:31","extension":"xlsx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":10898916,"visible":true,"origin":"","legend":"\u003cp\u003eTable S2. Differentially expressed genes in myeloid subgroups in CRC-Mix\u003c/p\u003e","description":"","filename":"TableS2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/7cb0bc648442b2e0668c51fb.xlsx"},{"id":26187160,"identity":"ac66acd7-4962-4f5b-94c0-9c5a7f39b445","added_by":"auto","created_at":"2022-09-07 16:35:31","extension":"xlsx","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":3897261,"visible":true,"origin":"","legend":"\u003cp\u003eTable S3. Differentially expressed genes in monocyte/macrophage between NC and CRC.\u003c/p\u003e","description":"","filename":"TableS3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/4c963f9f75a33c5e56080ca1.xlsx"},{"id":26187049,"identity":"d9b05381-ed2e-46ad-808e-5201fdf45717","added_by":"auto","created_at":"2022-09-07 16:30:31","extension":"xlsx","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":3048646,"visible":true,"origin":"","legend":"\u003cp\u003eTable S4. Differentially expressed genes in monocyte/macrophage between CRC and LM.\u003c/p\u003e","description":"","filename":"TableS4.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-1984416/v1/0dc8ee8b38bfacd0e835e73e.xlsx"}],"financialInterests":"(Not answered)","formattedTitle":"SPP1+ macrophages: A malignant macrophage subset in the colorectal cancer microenvironment revealed by single-cell sequencing","fulltext":[{"header":"Introduction","content":"\u003cp\u003eColorectal cancer (CRC) is the third most common malignancy and the second leading cause of cancer-related death\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. Following the great success of immune checkpoint blockade (ICB) in various advanced solid tumors, interest in immunotherapy for CRC is increasing\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. However, immunotherapy targeting PD-1 is only effective in patients with mismatch repair deficiency (dMMR)/high levels of microsatellite instability (MSI-H), which is present in 5% of metastatic CRC cases\u003csup\u003e\u003cspan additionalcitationids=\"CR4\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. Therefore, it is necessary to understand the complex mechanisms of cellular and molecular interactions in the CRC tumor microenvironment (TME) and to search for potential targets to develop new immunotherapies.\u003c/p\u003e \u003cp\u003eMacrophages are key mediators in the TME and are involved in multiple aspects of tumor immunity\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. M1 (inflammatory; antitumoral) and M2 (anti-inflammatory; protumoral) macrophage polarization systems have been used to characterize the activation state of macrophages in vitro\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. However, macrophages exhibit a more complex phenotype in vivo, which contradicts this simple in vitro classification\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. Macrophages in the TME are often referred to as tumor-associated macrophages (TAMs). TAM is a heterogeneous cell type\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e that contributes to angiogenesis\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e, extracellular matrix (ECM) remodeling\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e, promotes epithelial-mesenchymal transition (EMT) of tumor cells\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e, and activates immunosuppression\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e, promoting the progression and metastasis of CRC\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. Immunotherapies targeting TAMs, such as those reducing the number of TAMs in the TME by blocking the \u003cem\u003eCSF1-CSF1R\u003c/em\u003e axis\u003csup\u003e\u003cspan additionalcitationids=\"CR15\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e, have been applied in the clinic with minimal efficacy as monotherapy in malignancies\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eAccumulating single-cell studies have reported that SPP1\u0026thinsp;+\u0026thinsp;macrophages are important players in the TME\u003csup\u003e9,18\u0026minus;20\u003c/sup\u003e, are characterized as malignancy driven and are associated with a poor prognosis in CRC patients\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. However, the function of SPP1\u0026thinsp;+\u0026thinsp;macrophages remains unclear. Our single-cell analysis indicated that \u003cem\u003eSPP1\u003c/em\u003e is not only a specific biomarker for SPP1\u0026thinsp;+\u0026thinsp;macrophages in CRC but also a key functional molecule. Secreted phosphoprotein 1 (\u003cem\u003eSPP1\u003c/em\u003e), also called osteopontin (\u003cem\u003eOPN\u003c/em\u003e), is overexpressed in most types of cancer\u003csup\u003e\u003cspan additionalcitationids=\"CR22\" citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e and is associated with a poor prognosis in breast, cervical, bladder, colorectal, head and neck, liver, lung, and esophageal cancers\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. \u003cem\u003eSPP1\u003c/em\u003e can promote CRC progression and metastasis through EMT\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e and hypoxia pathways\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. It has been reported that \u003cem\u003eSPP1\u003c/em\u003e functions by binding to its receptors \u003cem\u003eCD44\u003c/em\u003e and integrin to regulate several signaling pathways (ERK, JNK1, and PI3K/Akt)\u003csup\u003e26\u0026ndash;31\u003c/sup\u003e. However, research on SPP1\u0026thinsp;+\u0026thinsp;macrophages is still in its infancy, and systematic research is lacking.\u003c/p\u003e \u003cp\u003eThe recent rapid development of single-cell RNA-sequencing (scRNA-seq) has revealed the complexity of myeloid cells, especially the predominant macrophages, in a variety of tumors\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e,\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. However, unlike the major cell types, which have distinctive features and classification markers, there is no generally accepted classification criterion for myeloid cells. Based on the CRC single-cell public dataset, we constructed the largest CRC myeloid cell atlas to date and performed a systematic study of SPP1\u0026thinsp;+\u0026thinsp;macrophages. Accumulating single-cell evidence suggests that SPP1\u0026thinsp;+\u0026thinsp;macrophages may be the key to uncovering the heterogeneity and functional changes of macrophages during CRC liver metastasis (CRLM)\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. On the basis of existing data, we further proposed the SPP1\u0026thinsp;+\u0026thinsp;macrophage model, which can explain the clinical characteristics and functional changes of macrophages in CRC better than the M1/M2 polarization theory; this model is also expected to help aid in the development of macrophage-targeted immunotherapeutic strategies.\u003c/p\u003e"},{"header":"Result","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eScRNA-seq revealed the myeloid cell landscape of CRC\u003c/h2\u003e \u003cp\u003eThe assessed tissue types of the CRLM dataset included colorectal cancer (CRC), normal colorectum (NC), liver metastases (LM), normal liver (NL), and peripheral blood mononuclear cells (PBMCs) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). NC and NL samples included both adjacent and healthy tissue. The CRC single-cell data used in this study were obtained from seven public datasets (GSE132465, GSE144735, GSE146771, GSE164522, GSE178318, GSE178341, and Wu \u003cem\u003eet al\u003c/em\u003e.)(Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb). The GSE164522 and Wu \u003cem\u003eet al\u003c/em\u003e. datasets were also further defined as CRLM datasets as they included CRC, NC, LM, NL and PBMC samples.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eAfter data processing and annotation, we identified nine main cell types in the seven CRC datasets: T cells, NK cells, B cells, plasma cells, myeloid cells, mast cells, neutrophils, epithelial cells and stromal cells (Fig. S1a). The proportion of myeloid cells among all immune cells was significantly higher (6/6) in CRC samples than in NC samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec and Table S1), suggesting that myeloid cells may play an important role in the CRC tumor immune microenvironment.\u003c/p\u003e \u003cp\u003eMyeloid cells were then screened, reclustered and annotated (Fig. S1b). We defined a total of 14 well-characterized myeloid cell subsets, including one plasmacytoid DC subset (pDC-LILRA4), four cDC subsets (cDC1-CLEC9A, cDC2-CD1C, cDC3-LAMP3, and cDC4-LTB), four macrophage subsets (Macro-C1QC, Macro-FCN1, Macro-MKI67, and Macro-SPP1), three doublet subsets (Doublets-B, Doublets-Epi and Doublets-T), and monocyte and Kupffer cell subsets (Fig. S1b and Table S1).\u003c/p\u003e \u003cp\u003eWe then analyzed the tissue distribution of these myeloid cell subsets in each of the seven cohorts using odds ratios (ORs) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ed). The data showed that cDC1-CLEC9A and cDC2-CD1C were relatively enriched in NC samples, while cDC3-LAMP3 and Macro-SPP1 were relatively enriched in CRC samples; the remaining myeloid subsets were not found to have stable and significant enrichment in NC or CRC samples. In addition, monocytes and Kupffer cells were significantly enriched in PBMC and NL samples, respectively.\u003c/p\u003e \u003cp\u003eTo construct a more robust and accurate model of the CRC myeloid cell lineage, CRC myeloid cells in the GSE132465, GSE144735, GSE146771, GSE178318, and GSE178341 datasets were integrated into the large cohort CRC-Mix (n\u0026thinsp;=\u0026thinsp;48,815) (Fig. S1c). The CRC-Mix data were then reclustered and annotated after harmony correction for batch effects. Finally, we identified 12 myeloid cell subsets in CRC tissue, including pDC-LILRA4, cDC1-CLEC9A, cDC2-CD1C, cDC3-LAMP3, cDC4-LTB, Macro-C1QC, Macro-FCN1, Macro-MKI67, Macro-SPP1, Doublets-B, Doublets-Epi and Doublets-T subsets (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee and S1d). In addition, a summary table including representative genes was made to facilitate the identification of these 12 myeloid cell subsets by other researchers, and details of differentially expressed genes (DEGs) between subsets are in the Supplementary Material (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ef and Table S2).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eChanges in monocytes/macrophages during CRLM\u003c/h2\u003e \u003cp\u003eSingle-cell sequencing provides the opportunity to study genetic and functional differences at the cellular level, allowing us to explore changes in the transcriptional profile and functional pathways of monocytes/macrophages in the TME during CRLM.\u003c/p\u003e \u003cp\u003eTo observe changes in the macrophage transcriptome during CRC occurrence, we identified DEGs of monocytes/macrophages between NC and CRC samples (GSE164522) and specifically annotated the marker genes of the four major macrophage subtypes in the heatmap (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea and Table S3). The results showed that the expression of \u003cem\u003eSPP1\u003c/em\u003e and \u003cem\u003eMKI67\u003c/em\u003e was significantly elevated in CRC samples, while \u003cem\u003eC1QC\u003c/em\u003e was highly expressed in NC samples. The same method was used to identify DEGs of monocytes/macrophages in six CRC single-cell datasets (NC vs. CRC) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb and Table S3). The common DEGs (n\u0026thinsp;=\u0026thinsp;216) in these six datasets were then analyzed by Metascape to further explore the functional and pathway changes in monocytes/macrophages during CRC occurrence (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec and Fig. S2a).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSimilarly, to observe changes in the macrophage transcriptome during liver metastasis of CRC, we identified DEGs of monocytes/macrophages between CRC and LM samples (GSE164522). The results showed that the expression of \u003cem\u003eSPP1\u003c/em\u003e and \u003cem\u003eFCN1\u003c/em\u003e was significantly elevated in LM samples, while the expression of \u003cem\u003eC1QC\u003c/em\u003e was higher in CRC samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed). The same method was used to identify DEGs of monocytes/macrophages in three single-cell datasets (CRC vs. LM) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ee and Table S4). Common DEGs (n\u0026thinsp;=\u0026thinsp;1,383) in the three datasets were then analyzed by Metascape to further explore the functional and pathway changes in monocytes/macrophages during liver metastasis of CRC (Figs.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ef and S2b).\u003c/p\u003e \u003cp\u003eTo characterize the origin and differentiation fate of monocytes/macrophages during CRC occurrence, development, and liver metastasis, we observed trajectories of monocyte/macrophage subsets, marker genes and tissue types based on pseudo-temporal analysis of Monocle2 (Wu \u003cem\u003eet al.\u003c/em\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eg-i). The results show that monocytes from PBMCs first differentiate into FCN1\u0026thinsp;+\u0026thinsp;macrophages after entering colon or liver tissue, and then FCN1\u0026thinsp;+\u0026thinsp;macrophages polarize toward C1QC\u0026thinsp;+\u0026thinsp;macrophages and SPP1\u0026thinsp;+\u0026thinsp;macrophages, while Kupffer cells in NL samples are relatively isolated (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eg and \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eh). The pseudo-chronological results of monocyte/macrophage marker genes also showed that the expression of \u003cem\u003eCD14\u003c/em\u003e and \u003cem\u003eFCN1\u003c/em\u003e decreased significantly when monocytes/macrophages moved from the PBMC-enriched area to the solid tissue-enriched area (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ei), suggesting the rapid differentiation of monocytes into macrophages after entering tissues. The expression of \u003cem\u003eSPP1\u003c/em\u003e was specifically elevated in CRC and LM-concentrated regions, suggesting that SPP1\u0026thinsp;+\u0026thinsp;macrophages play an important role in the occurrence and metastasis of CRC. In addition, the expression of \u003cem\u003eMARCO\u003c/em\u003e was concentrated in the NL-rich region.\u003c/p\u003e \u003cp\u003eTo explore the functional heterogeneity of monocytes/macrophages during CRLM, gene set variation analysis (GSVA) was used to analyze hallmark pathways of different monocyte/macrophage subsets (Wu \u003cem\u003eet al.\u003c/em\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ej). The results showed a significant enrichment of angiogenesis-, epithelial mesenchymal transition-, glycolysis-, and hypoxia-related genes in SPP1\u0026thinsp;+\u0026thinsp;macrophages, and the results were also reliably validated in multiple datasets (Fig. S2c-i). Notably, these hallmark pathways were also significantly altered in monocytes/macrophages during CRC occurrence and LM (Fig. S2a and S2b), suggesting that SPP1\u0026thinsp;+\u0026thinsp;macrophages play a dominant role in monocytes/macrophages changes during CRLM.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eTissue distribution and potential mechanisms of SPP1\u003csup\u003e+\u003c/sup\u003e macrophages\u003c/h2\u003e \u003cp\u003eNext, our study focused on the tissue distribution characteristics and biological functions of SPP1\u0026thinsp;+\u0026thinsp;macrophages. Five scRNA-seq datasets (GSE132465, GSE144735, GSE146771, GSE164522, and GSE178341) showed a significant increase (5/6) in the ratio of SPP1\u0026thinsp;+\u0026thinsp;macrophages to all macrophages in CRC (NC vs. CRC) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea and Table S1), and data from three scRNA-seq datasets (GSE164522, GSE178318, and Wu \u003cem\u003eet al\u003c/em\u003e.) further showed a significant increase (3/3) in the SPP1\u0026thinsp;+\u0026thinsp;macrophage proportion in LM samples (CRC vs. LM) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb and Table S1). The increasing proportion of SPP1\u0026thinsp;+\u0026thinsp;macrophages during CRC development and liver metastases suggests its critical role in CRLM. To explore whether the increased proportion of SPP1\u0026thinsp;+\u0026thinsp;macrophages is cancer specific, we integrated CRLM macrophage (GSE164522) and hepatocellular carcinoma (HCC) macrophage data (GSE156625) and performed reclustering. The results showed that the proportion of SPP1\u0026thinsp;+\u0026thinsp;macrophages in HCC was not significantly different from that in adjacent tissues but was significantly lower than that in CRLM samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec and Table S1). This suggests that the elevation of SPP1\u0026thinsp;+\u0026thinsp;macrophages is only observed in specific cancer tissues (e.g., primary foci and liver metastases of CRC) but is not significant in HCC.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThen, we explored the biological function of SPP1\u0026thinsp;+\u0026thinsp;macrophages in CRC based on the large CRC-Mix dataset. Macrophages in CRC-Mix included C1QC\u0026thinsp;+\u0026thinsp;macrophages (35.5%, n\u0026thinsp;=\u0026thinsp;13,640), FCN1\u0026thinsp;+\u0026thinsp;macrophages (30.6%, n\u0026thinsp;=\u0026thinsp;11,784), SPP1\u0026thinsp;+\u0026thinsp;macrophages (27.7%, n\u0026thinsp;=\u0026thinsp;10,647), and MKI67\u0026thinsp;+\u0026thinsp;macrophages (6.2%, n\u0026thinsp;=\u0026thinsp;2,403) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed). MKI67\u0026thinsp;+\u0026thinsp;macrophages have transcriptional profile highly similar to C1QC\u0026thinsp;+\u0026thinsp;macrophages but specifically express proliferation-related genes such as \u003cem\u003eMKI67\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eh). GSVA and signature scores also showed enrichment of the proliferation signature in MKI67\u0026thinsp;+\u0026thinsp;macrophages (Figs.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ej and \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea). Considering the low proportion and well-defined proliferation function of MKI67\u0026thinsp;+\u0026thinsp;macrophages, our follow-up study focused on the three major macrophages subsets (C1QC\u0026thinsp;+\u0026thinsp;macrophages, FCN1\u0026thinsp;+\u0026thinsp;macrophages and SPP1\u0026thinsp;+\u0026thinsp;macrophages) in CRC.\u003c/p\u003e \u003cp\u003eTo validate the function of each macrophage subset in the literature and our GSVA results, we assessed angiogenesis, EMT, glycolysis, hypoxia, phagocytosis, and MHC-II signatures in different macrophage subsets in CRC-Mix (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed-I and S3b-d). The results showed that C1QC\u0026thinsp;+\u0026thinsp;macrophages had the highest phagocytosis, MHC-II score and classical MHC class I genes (\u003cem\u003eHLA-A\u003c/em\u003e, \u003cem\u003eHLA-B\u003c/em\u003e, \u003cem\u003eHLA-C\u003c/em\u003e) (Fig. S3c and S3d), which indicated that C1QC\u0026thinsp;+\u0026thinsp;macrophages were the dominant macrophages for antigen presentation and phagocytosis of pathogens. These features also support a protective role of C1QC\u0026thinsp;+\u0026thinsp;macrophages in the CRC TME. A recent CRC single-cell study reported that SPP1-TAMs showed the highest angiogenesis score and the lowest phagocytosis score compared with C1QC-TAMs and MKI67-TAMs\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. Our study included FCN1\u0026thinsp;+\u0026thinsp;macrophages and further showed that FCN1\u0026thinsp;+\u0026thinsp;macrophages had higher angiogenesis scores and lower phagocytosis scores than SPP1\u0026thinsp;+\u0026thinsp;macrophages (Fig. S3b and S3c). This suggests that although SPP1\u0026thinsp;+\u0026thinsp;macrophages are characterized by high angiogenesis and low phagocytosis, these are not unique features that constitute the specificity of SPP1\u0026thinsp;+\u0026thinsp;macrophages. In addition, our results showed specific elevation of EMT, glycolysis, and hypoxia scores in SPP1\u0026thinsp;+\u0026thinsp;macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ee-g), which is consistent with the GSVA results.\u003c/p\u003e \u003cp\u003eThe classical M1/M2 polarization theory is based on inflammatory signatures to describe the heterogeneity of macrophages, and we also assessed the inflammatory, anti-inflammatory, M1, and M2 signatures of different macrophage subsets in Mix-CRC (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eh and \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ei). FCN1\u0026thinsp;+\u0026thinsp;macrophages showed the highest inflammatory score and the lowest anti-inflammatory score (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eh); similarly, FCN1\u0026thinsp;+\u0026thinsp;macrophages also showed higher M1 scores and lower M2 scores than C1QC\u0026thinsp;+\u0026thinsp;macrophages and SPP1\u0026thinsp;+\u0026thinsp;macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ei). In addition, we performed enrichment analysis of elevated DEGs in different macrophage subsets separately using Metascape and observed enrichment of the inflammatory response pathway in FCN1\u0026thinsp;+\u0026thinsp;macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eh). The above evidence suggests an inflammatory profile of FCN1\u0026thinsp;+\u0026thinsp;macrophages. In contrast to FCN1\u0026thinsp;+\u0026thinsp;macrophages, C1QC\u0026thinsp;+\u0026thinsp;macrophages and SPP1\u0026thinsp;+\u0026thinsp;macrophages exhibited more anti-inflammatory features. Enrichment analysis also showed enrichment of the negative regulation of immune system process pathway in C1QC\u0026thinsp;+\u0026thinsp;macrophages. Studies in glioma\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e and HCC\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e have reported the correlation of SPP1 expression with the M2 phenotype, and our data further showed that SPP1\u0026thinsp;+\u0026thinsp;macrophages have a higher M2 signature than FCN1\u0026thinsp;+\u0026thinsp;macrophages and C1QC\u0026thinsp;+\u0026thinsp;macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ei). In addition, SPP1\u0026thinsp;+\u0026thinsp;macrophages exhibited specific expression of \u003cem\u003eCD274\u003c/em\u003e (\u003cem\u003ePD-L1\u003c/em\u003e) and \u003cem\u003eHLA-G\u003c/em\u003e (a major immune checkpoint)\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e (Fig. S3d), suggesting an immunosuppressive characteristic of SPP1\u0026thinsp;+\u0026thinsp;macrophages. Notably, consistent with the M1/M2 signature found in multiple tumor scRNA-seq studies\u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e, we also observed a significant positive correlation between the M1 signature and the M2 signature in CRC macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ei).\u003c/p\u003e \u003cp\u003eTo explore the specific molecular mechanism by which SPP1\u0026thinsp;+\u0026thinsp;macrophages exert their biological functions, we used CellPhoneDB to identify the expression of potential crosstalk molecules based on ligand‒receptor interactions between different macrophage subsets and major cell types (GSE178341) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ej and k). The most abundant receptor‒ligand pair interactions were concentrated between different macrophage subsets, and SPP1\u0026thinsp;+\u0026thinsp;macrophages also showed relatively abundant interactions with stromal cells, epithelial cells, and T cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ej). Interestingly, we found three pairs of receptor‒ligand interactions specific for SPP1\u0026thinsp;+\u0026thinsp;macrophages: \u003cem\u003eSPP1\u003c/em\u003e-\u003cem\u003eCD44\u003c/em\u003e, \u003cem\u003eSPP1\u003c/em\u003e-\u003cem\u003ePTGER4\u003c/em\u003e, and the \u003cem\u003eSPP1\u003c/em\u003e-\u003cem\u003ea4b1 complex\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ek). SPP1\u0026thinsp;+\u0026thinsp;macrophages express high levels of \u003cem\u003eSPP1\u003c/em\u003e, and the receptor \u003cem\u003eCD44\u003c/em\u003e is widely present in different cells, while \u003cem\u003ePTGER4\u003c/em\u003e expression is enriched in epithelial cells and T cells, and integrin a4b1 expression is enriched in T cells and plasma cells. These findings reveal the potential molecular mechanisms of SPP1\u0026thinsp;+\u0026thinsp;macrophages and indicate that SPP1 is not only a key marker gene of SPP1\u0026thinsp;+\u0026thinsp;macrophages but also a core molecule for the biological functions of SPP1\u0026thinsp;+\u0026thinsp;macrophages.\u003c/p\u003e \u003cp\u003eIn summary, SPP1\u0026thinsp;+\u0026thinsp;macrophages are malignant cells that are specifically enriched in CRC, with characteristics of angiogenesis, EMT, glycolysis, hypoxia, and immunosuppression.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eSPP1, a novel macrophage marker in CRC\u003c/h2\u003e \u003cp\u003eSingle-cell data have revealed the important role of SPP1\u0026thinsp;+\u0026thinsp;macrophages in CRC. The existing bulk RNA-seq dataset with a large number of samples and rich clinical information is also undoubtedly valuable for studying SPP1\u0026thinsp;+\u0026thinsp;macrophages, and \u003cem\u003eSPP1\u003c/em\u003e can be the key biomarker to unlocking this treasure trove. Next, we first need to answer two fundamental and important questions: whether \u003cem\u003eSPP1\u003c/em\u003e is a specific marker of SPP1\u0026thinsp;+\u0026thinsp;macrophages and how \u003cem\u003eSPP1\u003c/em\u003e differs from classical macrophage markers.\u003c/p\u003e \u003cp\u003eTo describe changes in the proportion of macrophages among immune cells during CRC occurrence, we compared the proportions of 22 immune cell types in 10 CRC bulk RNA-seq datasets based on the CIBERSORT algorithm (NC vs. CRC) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). The results, which showed the same significant trend in the comparison of half or more datasets (n\u0026thinsp;\u0026ge;\u0026thinsp;5), are considered to have high credibility. The data showed that the proportions of M0 macrophages and M1 macrophages were significantly increased in CRC, while the proportion of M2 macrophages was significantly decreased. Next, to describe changes in the numbers of macrophages during CRC occurrence, we compared the expression levels of five classical macrophage markers in 10 CRC bulk datasets (NC vs. CRC) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). The results did not show significant changes in the pan-macrophage marker \u003cem\u003eCD68\u003c/em\u003e or the M1 macrophage markers \u003cem\u003eCD68\u003c/em\u003e and \u003cem\u003eiNOS\u003c/em\u003e (\u003cem\u003eNOS2\u003c/em\u003e), while the M2 macrophage markers \u003cem\u003eCD163\u003c/em\u003e and \u003cem\u003eCD206\u003c/em\u003e (\u003cem\u003eMRC1\u003c/em\u003e) were significantly downregulated in CRC.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe above results of changes in the proportion and number of macrophages during CRC occurrence seem to contradict our single-cell results that the proportion of myeloid cells is significantly elevated in CRC. We speculate that this is related to the overall decline in immune cells in the tumor immune microenvironment during CRC occurrence. Therefore, we used the immune score based on the ESTIMATE algorithm and the immune cell marker \u003cem\u003eCD45\u003c/em\u003e (\u003cem\u003ePTPRC\u003c/em\u003e) to evaluate the difference in the level of immune infiltration between NC and CRC samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea), and the results showed a significant decrease in immune infiltration in CRC tissues. We further assessed the ratio of macrophages to immune cells using \u003cem\u003eCD68\u003c/em\u003e/\u003cem\u003eCD45\u003c/em\u003e, and the results showed that \u003cem\u003eCD68\u003c/em\u003e/\u003cem\u003eCD45\u003c/em\u003e was significantly elevated in CRC. These findings indicated that the number of macrophages in tissues did not change significantly during the occurrence of CRC, while the decrease in the number of overall immune cells resulted in a significant increase in the proportion of macrophages.\u003c/p\u003e \u003cp\u003eUnlike classical macrophage markers whose levels were unchanged or decreased, SPP1 (9/9) and \u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e (6/6) were significantly upregulated in CRC (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). Correlation analysis between SPP1 expression and immune cell proportions (CIBERSORT) showed that SPP1 expression in TCGA-CRC was positively correlated with four myeloid cell types, including neutrophils, M0 macrophages, M2 macrophages, and mast cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb). The correlation analysis between \u003cem\u003eSPP1\u003c/em\u003e and six types of immune cells in TCGA-CRC based on TIMER showed that SPP1 had the highest correlation with macrophages (Fig. S4a). In addition, correlation analysis showed that SPP1 was highly positively correlated with \u003cem\u003eCD45\u003c/em\u003e, \u003cem\u003eCD68\u003c/em\u003e, \u003cem\u003eCD86\u003c/em\u003e, \u003cem\u003eCD163\u003c/em\u003e, and \u003cem\u003eCD206\u003c/em\u003e but negatively correlated with \u003cem\u003eiNOS\u003c/em\u003e (Fig. S4b). This evidence indicates that SPP1 expression is highly correlated with the infiltration of macrophages, especially M2 macrophages, in CRC. Single-cell sequencing data further revealed that SPP1 expression was only enriched in specific myeloid cells in CRC, and the enrichment pattern was almost consistent with our definition of SPP1\u0026thinsp;+\u0026thinsp;macrophage subsets (Figs.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec, d, S4c and d). In other myeloid cells, other immune cells, and nonimmune cells, SPP1 is expressed at very low levels. These results demonstrate that SPP1 is indeed a specific marker for SPP1\u0026thinsp;+\u0026thinsp;macrophages in CRC.\u003c/p\u003e \u003cp\u003eThe IHC staining from The Human Protein Atlas (THPA, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.proteinatlas.org/\u003c/span\u003e\u003cspan address=\"https://www.proteinatlas.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) database\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e also verified that SPP1 protein was specifically expressed in CRC samples, while no staining was detected in NC, NL and HCC samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ee). In addition, to assess the function of the SPP1 gene in CRC, we used the Gene Set Enrichment Analysis (GSEA) function of LinkedOmics to analyze the coexpressed genes of SPP1 in TCGA-CRC (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ef). The top five significantly enriched GO BP terms were extracellular structure organization, cell adhesion mediated by integrin, cellular response to lipoprotein particle stimulus, cellular response to VEGF stimulus and macrophage activation.\u003c/p\u003e \u003cp\u003eIn summary, \u003cem\u003eSPP1\u003c/em\u003e is a specific biomarker of SPP1\u0026thinsp;+\u0026thinsp;macrophages and is significantly elevated in CRC.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eClinical value of SPP1\u0026thinsp;+\u0026thinsp;macrophages in CRC (bulk RNA-seq)\u003c/h2\u003e \u003cp\u003eFive CRC bulk datasets (TCGA, GSE14333, GSE17536, GSE41258, and GSE161158) were used to compare \u003cem\u003eSPP1\u003c/em\u003e expression in different stages (stage I/II vs. stage III/IV) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea). The results showed elevated \u003cem\u003eSPP1\u003c/em\u003e expression in patients with stage III/IV (advanced) CRC (5/5). We also used three CRC bulk datasets (TCGA, GSE39582, and GSE41258) to compare the expression of \u003cem\u003eSPP1\u003c/em\u003e in samples with different TNM classifications (T1/2 vs. T3/4, N0 vs. N1-4, and M0 vs. M1) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eb-d). The results showed higher \u003cem\u003eSPP1\u003c/em\u003e expression in T3/4 (3/3) and N1-3 (2/3) samples than in T1/2 and N0 samples. The close relationship between \u003cem\u003eSPP1\u003c/em\u003e expression level and the degree of tumor infiltration and lymph node metastasis suggests an important role of SPP1\u0026thinsp;+\u0026thinsp;macrophages in CRC progression.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFurthermore, we compared the expression of SPP1 between CRC and LM samples in six bulk datasets (GSE6988, GSE41258, GSE41568, GSE68468, GSE131418 Consortium, and GSE131418 MCC) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ee). SPP1 expression was significantly higher in LM samples than in CRC samples (4/6), suggesting an important role for SPP1\u0026thinsp;+\u0026thinsp;macrophages in CRC metastasis. We also observed a further decrease in immune infiltration in LM samples compared to CRC samples (5/6).\u003c/p\u003e \u003cp\u003eMacrophage enrichment indicates a poor prognosis in most tumors, but macrophage enrichment in CRC tends to be associated with a favorable prognosis\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e,\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. Univariate Cox analysis of four datasets (TCGA, GSE17536, GSE39582, and GSE41258) indicated that \u003cem\u003eSPP1\u003c/em\u003e expression was significantly associated with poor overall survival (OS) (3/4) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ef). We performed the same analysis for traditional macrophage markers: \u003cem\u003eiNOS\u003c/em\u003e showed a protective effect, while \u003cem\u003eCD68\u003c/em\u003e, \u003cem\u003eCD86\u003c/em\u003e, \u003cem\u003eCD163\u003c/em\u003e and \u003cem\u003eCD206\u003c/em\u003e were not significantly related to CRC progression, metastasis or prognosis.\u003c/p\u003e \u003cp\u003eConsidering the colinearity between SPP1\u0026thinsp;+\u0026thinsp;macrophages (\u003cem\u003eSPP1\u003c/em\u003e) and macrophages (\u003cem\u003eCD68\u003c/em\u003e), we used \u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e coexpression status to study the relationship between SPP1\u0026thinsp;+\u0026thinsp;macrophages and the clinical features of CRC (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The TCGA-CRC samples were divided into high (H-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e) and low (L-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e) groups according to the median value of \u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e. Patients in the H-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e group had a higher mean age; a higher proportion of right-side tumors, advanced T classification, lymph node metastasis, MSI-H status and high TMB; and shorter survival time (OS and PFI). In addition, GSEA verified the enrichment of hallmark pathways such as EMT, angiogenesis and hypoxia in the H-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e group (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eg).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparison of clinical parameters between the L-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e and H-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e groups in CRC.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinical parameters\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eL-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e\u003c/p\u003e \u003cp\u003e(\u003cem\u003eN\u003c/em\u003e\u0026thinsp;=\u0026thinsp;309)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eH-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e\u003c/p\u003e \u003cp\u003e(\u003cem\u003eN\u003c/em\u003e\u0026thinsp;=\u0026thinsp;308)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003ep\u003c/em\u003e value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAge (years)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e64.7\u0026thinsp;\u0026plusmn;\u0026thinsp;12.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e67.8\u0026thinsp;\u0026plusmn;\u0026thinsp;12.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.002\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eGender\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e140 (45.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e146 (47.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.682\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e166 (54.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e162 (52.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLocation\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLeft\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e180 (61.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e160 (53.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.047\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRight\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e114 (38.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e141 (46.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePathologic stage\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e62 (21.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e41 (13.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.078\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eII\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e109 (37.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e117 (38.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIII\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e82 (28.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e96 (31.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIV\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e38 (13.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e49 (16.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eT classification\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e14 (4.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5 (1.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.005\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e63 (20.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e41 (13.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e199 (65.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e219 (71.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e28 (9.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e42 (13.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eN classification\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e182 (60.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e165 (53.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.025\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e77 (25.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e71 (23.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e44 (14.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e71 (23.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMetastasis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eM0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e216 (85.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e236 (82.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.417\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eM1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e37 (14.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e49 (17.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMSI-H\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e32 (10.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50 (16.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.024\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e272 (89.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e247 (83.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTMB\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e32 (11.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50 (20.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.006\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLow\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e242 (88.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e194 (79.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eOS (years)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean (95%CI)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7.97 (6.94\u0026ndash;9.01)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6.49 (5.43\u0026ndash;7.54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.046\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOutcome event\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e55/299\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e68/294\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePFI (years)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean (95%CI)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8.09 (7.08\u0026ndash;9.09)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6.03 (4.99\u0026ndash;7.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.030\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOutcome event\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e67/298\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e86/293\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003e\u003csup\u003ea\u003c/sup\u003e Mann-Whitney test, \u003csup\u003eb\u003c/sup\u003e Chi-square test, \u003csup\u003ec\u003c/sup\u003e Log-rank test. TMB, tumor mutation burden; high TMB (\u0026gt;\u0026thinsp;10mut/MB); CI, confidence interval; OS, overall survival; PFI, progression-free interval; Location left includes the cecum to the transverse colon; Location left includes the splenic flexure to the rectum.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eIn summary, SPP1\u0026thinsp;+\u0026thinsp;macrophages can be used as a malignant biomarker to evaluate CRC progression, metastasis and prognosis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eSignificance of SPP1\u0026thinsp;+\u0026thinsp;macrophages for CRC treatment\u003c/h2\u003e \u003cp\u003eKaplan‒Meier survival analysis showed that the H-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e group had shorter OS than the L-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e group (log-rank, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea). CRC single-cell data revealed the relationship between preoperative chemotherapy and SPP1 expression in macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eb and c). Patients with partial response (PR) or stable disease (SD) were defined as responders, while patients with progressive disease (PD) were defined as nonresponders. The results showed that SPP1 expression in macrophages was significantly decreased in patients treated with preoperative chemotherapy, and SPP1 expression in macrophages was further decreased in responders (those with PR/SD) compared with nonresponders (those with PD). This suggests that preoperative chemotherapy may exert a therapeutic effect by specifically reducing the number of SPP1\u0026thinsp;+\u0026thinsp;macrophages in the tumor immune microenvironment. We also speculate that immunotherapy targeting SPP1\u0026thinsp;+\u0026thinsp;macrophages may improve the prognosis of CRC patients with a poor prognosis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eInterestingly, both scRNA-seq data and bulk RNA-seq data showed a strong association between SPP1\u0026thinsp;+\u0026thinsp;macrophages and genome instability and mutation (dMMR, MSI-H and high TMB) of CRC (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ed-I and Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), suggesting that CRC patients with high SPP1\u0026thinsp;+\u0026thinsp;macrophages had the characteristics of immunotherapy benefit. We also used SubMap algorithms to predict immunotherapy responsiveness in the L-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e and H-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ej). The results indicated that the H-\u003cem\u003eSPP1\u003c/em\u003e/\u003cem\u003eCD68\u003c/em\u003e group was more likely to respond to anti-PD-1 therapy (Bonferroni corrected \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.021) and anti-CTLA-4 therapy (Bonferroni corrected P\u0026thinsp;=\u0026thinsp;0.023). In addition, the results from the GSE39582 dataset also validate that patients with high \u003cem\u003eSPP1\u003c/em\u003e expression are more likely to respond to anti-PD-1 therapy and anti-CTLA-4 therapy (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ek).\u003c/p\u003e \u003cp\u003eThe development and maintenance of macrophages requires a continuous supply of the nutritional factor \u003cem\u003eCSF1\u003c/em\u003e, and \u003cem\u003eCSF1R\u003c/em\u003e is also essential for the development, survival, recruitment, and proliferation of most macrophages\u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e,\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e. CRC-Mix data showed that \u003cem\u003eCSF1R\u003c/em\u003e expression was significantly enriched in C1QC\u0026thinsp;+\u0026thinsp;macrophages but not SPP1\u0026thinsp;+\u0026thinsp;macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003el), which was also validated in two other scRNA-seq cohorts (GSE164522 and Wu \u003cem\u003eet al\u003c/em\u003e.) (Fig. S5a and b). This suggests that immunotherapy that blocks \u003cem\u003eCSF1R\u003c/em\u003e preferentially depletes C1QC\u0026thinsp;+\u0026thinsp;macrophages over SPP1\u0026thinsp;+\u0026thinsp;macrophages.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eWe investigated SPP1\u0026thinsp;+\u0026thinsp;macrophages in CRC from various aspects; we assessed their cell origin and tissue distribution, clinical value as a marker of CRC malignancy, potential functional pathways and cell‒cell interactions, and implications for treatment (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ea-e). Based on the literature and the results of this study, we propose the SPP1\u0026thinsp;+\u0026thinsp;macrophage model theory. Compared with the M1/M2 polarization theory, this theory can better explain the changes in macrophages in the whole process of CRLM and guide clinical diagnosis, condition assessment and treatment. Importantly, the SPP1\u0026thinsp;+\u0026thinsp;macrophage model provides a unique single-cell perspective to help us understand the heterogeneity of macrophages in the CRC tumor immune microenvironment.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe identified four macrophage subsets from CRC tissue, including FCN1\u0026thinsp;+\u0026thinsp;macrophages, C1QC\u0026thinsp;+\u0026thinsp;macrophages, SPP1\u0026thinsp;+\u0026thinsp;macrophages and MKI67\u0026thinsp;+\u0026thinsp;macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ea). Enrichment analysis and phenotypic signatures indicated that inflammation, phagocytosis, malignancy and proliferation were the most prominent features related to these four macrophage subsets.\u003c/p\u003e \u003cp\u003eSPP1\u0026thinsp;+\u0026thinsp;macrophages were specifically enriched in primary foci and liver metastases of CRC but were almost absent in normal colorectal and liver tissues and HCC tissues. Evidence from trajectory analysis and tissue distribution analysis of macrophage subsets supports the following hypothesis regarding the origin and differentiation of SPP1\u0026thinsp;+\u0026thinsp;macrophages: monocytes from peripheral blood are recruited into colonic tissues and rapidly differentiate into macrophages, which differentiate into monocyte-like FCN1\u0026thinsp;+\u0026thinsp;macrophages that polarize toward C1QC\u0026thinsp;+\u0026thinsp;macrophages and SPP1\u0026thinsp;+\u0026thinsp;macrophages. Subsequently, SPP1\u0026thinsp;+\u0026thinsp;macrophages that are enriched in CRC primary foci are transferred to the liver along with tumor tissue and further enriched (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eb).\u003c/p\u003e \u003cp\u003eThe expansion of myeloid cells in CRC tissue suggests their active role in shaping the TME\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e. Our data also confirmed the significant increase in the proportion of myeloid cells and macrophages in CRC. However, our further study found that the number of macrophages in tissues did not change significantly during CRC tumorigenesis, but the decrease in the number of overall immune cells resulted in a significant increase in the proportion of macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ec). The decrease in immune infiltration during the occurrence and metastasis of CRC suggests that antitumor immunity weakens as the disease progresses, while the importance of macrophages in the TME continues to increase.\u003c/p\u003e \u003cp\u003eAccumulating single-cell studies have reported the association of SPP1\u0026thinsp;+\u0026thinsp;TAMs with malignancies, especially CRC\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e,\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. A recent study also found that SPP1\u0026thinsp;+\u0026thinsp;TAMs in CRC were associated with LM\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. Numerous studies have reported that \u003cem\u003eSPP1\u003c/em\u003e is a diagnostic and prognostic marker for CRC\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e,\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e,\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e. Our results integrate the content of the above studies. We found that SPP1 is a specific marker of SPP1\u0026thinsp;+\u0026thinsp;macrophages, and SPP1\u0026thinsp;+\u0026thinsp;macrophages can be used as an indicator to assess the malignancy of CRC. The number and proportion of SPP1\u0026thinsp;+\u0026thinsp;macrophages constantly increase during the occurrence, progression and metastasis of CRC and indicate a poor prognosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ec). These results support that SPP1\u0026thinsp;+\u0026thinsp;macrophages in the TME may play an important driving role in the occurrence, progression and metastasis of CRC. SPP1\u0026thinsp;+\u0026thinsp;macrophages are also expected to be a comprehensive marker for CRC diagnosis, disease severity assessment and prognosis evaluation.\u003c/p\u003e \u003cp\u003eThe \"classically activated\" M1 and \"alternately activated\" M2 macrophage polarization systems have been widely used to describe the activation state of macrophages in vitro\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. However, macrophages in vivo exhibit a more complex phenotype, which contradicts this simple in vitro classification\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. Our studies in CRC also demonstrate the limitations of this in vitro polarization model. First, the M1/M2 polarization theory considers M1 and M2 activation states as mutually exclusive discrete states, while multiple recent single-cell studies have shown that M1 and M2 gene signatures are coexpressed in macrophage subsets in nearly all cancer types\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e,\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e,\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e. Our data also found that M1 and M2 signatures coexisted in CRC macrophages and were significantly positively correlated. Second, according to the M1/M2 polarization theory, proinflammatory M1 macrophages are thought to prevent tumor progression, while M2 macrophages promote tumor growth and metastasis\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e,\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e. This theory can well explain the changes in macrophages in the TME in some tumors\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e,\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e. However, our results based on CIBERSORT showed that the proportion of M1 macrophages was significantly increased in CRC, while the proportion of M2 macrophages was significantly decreased, which contradicted the M1/M2 polarization theory. Furthermore, in our study of CRC progression, metastasis and prognosis, no significant differences were observed in M1/M2 markers except for the M1 marker \u003cem\u003eiNOS\u003c/em\u003e, which presented a protective effect. The above evidence indicates that the M1/M2 polarization theory is not applicable in the CRC TME, which also suggests that the malignant features of SPP1\u0026thinsp;+\u0026thinsp;macrophages are not derived from their M2 signatures.\u003c/p\u003e \u003cp\u003eThe biological function of SPP1\u0026thinsp;+\u0026thinsp;macrophages is still unclear. A recent study suggested that SPP1\u0026thinsp;+\u0026thinsp;TAMs promote CRC through enhancing angiogenesis and attenuating phagocytosis\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. However, our study found that FCN1\u0026thinsp;+\u0026thinsp;macrophages exhibited higher angiogenesis signatures and lower phagocytosis signatures than SPP1\u0026thinsp;+\u0026thinsp;macrophages, suggesting that SPP1\u0026thinsp;+\u0026thinsp;macrophages also promote tumors through other important pathways. Various analyses in this study confirmed the enrichment of EMT, hypoxia, glycolysis, and immunosuppressive signatures in SPP1\u0026thinsp;+\u0026thinsp;macrophages, suggesting that SPP1\u0026thinsp;+\u0026thinsp;macrophages may promote CRC through related pathways (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ed). Furthermore, our study found that SPP1 is also a key functional molecule of SPP1\u0026thinsp;+\u0026thinsp;macrophages. CellPhoneDB predicted that SPP1\u0026thinsp;+\u0026thinsp;macrophages interact with other cells through the \u003cem\u003eSPP1\u003c/em\u003e-\u003cem\u003eCD44\u003c/em\u003e, \u003cem\u003eSPP1\u003c/em\u003e-\u003cem\u003ePTGER4\u003c/em\u003e, and \u003cem\u003eSPP1\u003c/em\u003e-\u003cem\u003ea4b1 complex\u003c/em\u003e axes. The \u003cem\u003eCD44\u003c/em\u003e receptor plays an important role in tumor EMT, invasion and metastasis\u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e, \u003cem\u003ePTGER4\u003c/em\u003e\u003csup\u003e50\u003c/sup\u003e, and integrin a4b1\u003csup\u003e51\u003c/sup\u003e have also been reported to be involved in cancer progression and metastasis. This suggests that SPP1\u0026thinsp;+\u0026thinsp;macrophages may promote CRC progression and metastasis through the interaction of secreted \u003cem\u003eSPP1\u003c/em\u003e with receptors \u003cem\u003eCD44\u003c/em\u003e, \u003cem\u003ePTGER4\u003c/em\u003e, and integrin a4b1 (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ed).\u003c/p\u003e \u003cp\u003eThe understanding of SPP1\u0026thinsp;+\u0026thinsp;macrophages is expected to provide new options for the treatment of CRC. Chemotherapy-induced macrophage phenotype switching has been reported in breast cancer\u003csup\u003e\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e. We also found that preoperative chemotherapy significantly reduced \u003cem\u003eSPP1\u003c/em\u003e expression in CRC macrophages, particularly in responders. This suggests that effective chemotherapy may specifically inhibit SPP1\u0026thinsp;+\u0026thinsp;macrophages. Moreover, it suggests the possibility of immunotherapy targeting SPP1\u0026thinsp;+\u0026thinsp;macrophages. Interestingly, we found that although SPP1\u0026thinsp;+\u0026thinsp;macrophages predict a poor prognosis, patients with high SPP1\u0026thinsp;+\u0026thinsp;macrophage levels show higher levels of genome instability and mutation characteristics (MSI-H, dMMR, and high TMB) (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ed). dMMR/MSI-H status is a recognized immunotherapy biomarker in colorectal cancer and other solid tumors\u003csup\u003e\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e. In addition, a high TMB also predicts better therapeutic efficacy of ICBs in several solid tumors, including CRC\u003csup\u003e\u003cspan additionalcitationids=\"CR55\" citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e\u003c/sup\u003e. This suggests that CRC patients with high SPP1\u0026thinsp;+\u0026thinsp;macrophages are more likely to benefit from ICB immunotherapy. This notion is further supported by the prediction results of SubMap algorithm and the enrichment of \u003cem\u003ePD-L1\u003c/em\u003e and \u003cem\u003eHLA-G\u003c/em\u003e in SPP1\u0026thinsp;+\u0026thinsp;macrophages. In addition, a large number of studies have reported that targeted blockade of SPP1 can inhibit tumor progression and metastasis\u003csup\u003e21,25,57\u0026minus;59\u003c/sup\u003e. Substantial evidence from this study supports the development of drugs targeting SPP1\u0026thinsp;+\u0026thinsp;macrophages as a viable option for CRC immunotherapy.\u003c/p\u003e \u003cp\u003eNotably, our SPP1\u0026thinsp;+\u0026thinsp;macrophage theory also reveals a possible problem in current macrophage-targeted immunotherapy research. Inhibition of the \u003cem\u003eCSF1-CSF1R\u003c/em\u003e axis to reduce TAM levels is a major research direction of macrophage-targeted immunotherapy\u003csup\u003e6,14\u0026minus;16\u003c/sup\u003e. However, we found that \u003cem\u003eCSF1R\u003c/em\u003e was significantly enriched in C1QC\u0026thinsp;+\u0026thinsp;macrophages compared to other macrophages, suggesting that \u003cem\u003eCSF1R\u003c/em\u003e blockade may preferentially deplete the C1QC\u0026thinsp;+\u0026thinsp;macrophage subset while sparing SPP1\u0026thinsp;+\u0026thinsp;macrophages. Anti-CSF1R therapy may not be sufficient to deplete all tumor-promoting SPP1\u0026thinsp;+\u0026thinsp;macrophages, which may contribute to the poor efficacy of anti-CSF1R monotherapy in the Renca mouse tumor model and in human cancer patients\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eHowever, this study has some limitations. First, the lack of a consensus classification system for myeloid cell subsets identified by single-cell sequencing has become a key obstacle limiting further research on subsets. Therefore, we mainly referred to the study of Zhang \u003cem\u003eet al\u003c/em\u003e.\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e for the classification of macrophage subsets and only identified four macrophage subsets that are ubiquitous and well characterized, while some rare macrophage subsets (such as \u003cem\u003eCXCL10-\u003c/em\u003e and ISG-enriched macrophage subsets\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e,\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e\u003c/sup\u003e) were generally incorporated into the macrophage subsets with similar transcriptomes. Second, we observed specific enrichment of SPP1\u0026thinsp;+\u0026thinsp;macrophages in LM samples. These SPP1\u0026thinsp;+\u0026thinsp;macrophages may be derived from peripheral blood monocytes acclimated by the TME of the LM or from the abnormal activation and proliferation of transferred SPP1\u0026thinsp;+\u0026thinsp;macrophages. However, our current techniques remain unable to determine the origin of SPP1\u0026thinsp;+\u0026thinsp;macrophages in metastases. Third, extensive data in this study suggest that CRC patients with high SPP1\u0026thinsp;+\u0026thinsp;macrophage levels are more likely to benefit from immunotherapy, but this evidence is all based on correlational inferences. We have not yet found a large-scale CRC immunotherapy dataset that can be used to validate the hypothesis. Conversely, a recently published study suggested that SPP1\u0026thinsp;+\u0026thinsp;macrophages reduced the efficacy of PD-L1 blockade treatment\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. Therefore, additional in-house clinical data are needed to validate the relationship between SPP1\u0026thinsp;+\u0026thinsp;macrophages and ICB immunotherapy.\u003c/p\u003e \u003cp\u003eIn conclusion, based on public scRNA-seq and bulk RNA-seq datasets, we propose the SPP1\u0026thinsp;+\u0026thinsp;macrophage model paradigm, which can explain the characteristics and changes of macrophages in CRC better than the M1/M2 polarization theory. We found that SPP1\u0026thinsp;+\u0026thinsp;macrophages can serve as a malignant marker of CRC occurrence, progression, metastasis and poor prognosis. We also explored the underlying mechanisms of SPP1\u0026thinsp;+\u0026thinsp;macrophages and suggest that immunotherapy targeting SPP1\u0026thinsp;+\u0026thinsp;macrophages (in terms of abundance and function) is a promising CRC therapeutic strategy.\u003c/p\u003e"},{"header":"Materials And Methods","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eMaterial\u003c/h2\u003e \u003cp\u003eThe public datasets used in this study are from the GEO and TCGA and the website that stores the scRNA-seq data of Wu \u003cem\u003eet al\u003c/em\u003e. These datasets include eight scRNA-seq datasets and 16 bulk RNA-seq datasets (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). To ensure the stability and comparability of single-cell data, only scRNA-seq datasets based on the 10x Genomics platform were included in this study; seven CRC datasets (GSE132465, GSE144735, GSE146771, GSE164522, GSE178318, GSE178341, and Wu \u003cem\u003eet al\u003c/em\u003e.) and one HCC dataset (GSE156625) met this criterion. GSE164522 and Wu \u003cem\u003eet al\u003c/em\u003e. were further defined as CRLM datasets due to their tissue sample types, including CRC, NC, LM, NC, and PBMC. The bulk RNA-seq dataset includes high-throughput sequencing data from TCGA-CRC and 15 GEO microarray datasets (GSE6988, GSE14333, GSE17536, GSE20842, GSE20916, GSE39582, GSE41258, GSE44076, GSE44861, GSE68468, GSE83889, GSE87211, GSE106582, GSE131418 and GSE161158). Transcriptome data and clinical information of TCGA-CRC were downloaded from UCSC Xena (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://xena.ucsc.edu/\u003c/span\u003e\u003cspan address=\"http://xena.ucsc.edu/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), and single nucleotide variation data (VarScan) were downloaded from the Genomic Data Commons (GDC) portal (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://portal.gdc.cancer.gov/\u003c/span\u003e\u003cspan address=\"https://portal.gdc.cancer.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The transcriptome and clinical information of the GEO dataset were downloaded from the GEO database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSource of scRNA-seq and bulk RNA-seq datasets.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeposited Data\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePlatform\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSource\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIdentifier\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC scRNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10x Genomics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLee \u003cem\u003eet al\u003c/em\u003e., 2020\u003csup\u003e44\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE132465\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC scRNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10x Genomics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLee \u003cem\u003eet al\u003c/em\u003e., 2020\u003csup\u003e44\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE144735\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC scRNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10x Genomics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eZhang \u003cem\u003eet al\u003c/em\u003e., 2020\u003csup\u003e19\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE146771\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC scRNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10x Genomics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eChe \u003cem\u003eet al\u003c/em\u003e., 2021\u003csup\u003e61\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE178318\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC scRNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10x Genomics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePelka \u003cem\u003eet al\u003c/em\u003e., 2021\u003csup\u003e60\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE178341\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRLM scRNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10x Genomics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLiu \u003cem\u003eet al\u003c/em\u003e., 2022\u003csup\u003e9\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE164522\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRLM scRNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10x Genomics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWu \u003cem\u003eet al\u003c/em\u003e., 2022\u003csup\u003e18\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.cancerdiversity.asia/scCRLM\u003c/span\u003e\u003cspan address=\"http://www.cancerdiversity.asia/scCRLM\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman HCC scRNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10x Genomics\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSharma \u003cem\u003eet al\u003c/em\u003e., 2020\u003csup\u003e62\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE156625\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIllumina Hiseq\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTCGA\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL4811\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eKi \u003cem\u003eet al\u003c/em\u003e., 2007\u003csup\u003e63\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE6988\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL570\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eJorissen \u003cem\u003eet al\u003c/em\u003e., 2009\u003csup\u003e64\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE14333\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL570\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSmith \u003cem\u003eet al\u003c/em\u003e., 2010\u003csup\u003e65\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE17536\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL4133\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGaedcke \u003cem\u003eet al\u003c/em\u003e., 2010\u003csup\u003e66\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE20842\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL570\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSkrzypczak \u003cem\u003eet al\u003c/em\u003e., 2010\u003csup\u003e67\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE20916\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL570\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMarisa \u003cem\u003eet al\u003c/em\u003e., 2013\u003csup\u003e68\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE39582\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSheffer \u003cem\u003eet al\u003c/em\u003e., 2009\u003csup\u003e69\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE41258\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL13667\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSol\u0026eacute; \u003cem\u003eet al\u003c/em\u003e., 2014\u003csup\u003e70\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE44076\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL3921\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRyan \u003cem\u003eet al\u003c/em\u003e., 2014\u003csup\u003e71\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE44861\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSheffer \u003cem\u003eet al\u003c/em\u003e., 2009\u003csup\u003e69\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE68468\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL10558\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eKwon \u003cem\u003eet al\u003c/em\u003e., 2017\u003csup\u003e72\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE83889\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL13497\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHu \u003cem\u003eet al\u003c/em\u003e., 2018\u003csup\u003e73\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE87211\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL10558\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE106582\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL15048\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eKamal \u003cem\u003eet al\u003c/em\u003e., 2019\u003csup\u003e74\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE131418\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHuman CRC bulk RNA-seq dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL570\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSzeglin \u003cem\u003eet al\u003c/em\u003e., 2022\u003csup\u003e75\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGEO: GSE161158\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eBasic analysis workflow for scRNA-seq data\u003c/h2\u003e \u003cp\u003eThe R package Seurat (version 4.1.0) converts raw unique molecular identifier (UMI) count matrices to Seurat objects and then filters the data according to the following criteria: (1) filter genes that are expressed in at least 10 cells; (2) filter cells that have unique feature counts over 6,000 or less than 500; and (3) filter cells that have \u0026gt;\u0026thinsp;20% mitochondrial counts. After quality control, \"LogNormalize\" was used to normalize the expression of all cells with a scale factor of 10,000. The top 2000 highly variable genes were identified based on the mean and dispersion. During scaling of the highly variable genes, the \"var.to.regress\" option was used to regress percent mitochondrial content. The results are obtained by principal component analysis (PCA) for linear dimensionality reduction. We used the \u003cem\u003eFindClusters\u003c/em\u003e function on 50 principal components with a resolution of 0.8 for preliminary clustering and annotation. Afterward, the visualization of cell clustering is completed by the nonlinear dimensionality reduction UMAP method\u003csup\u003e\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e\u003c/sup\u003e. Finally, the main cell types were identified and annotated based on canonical marker genes. Subsequent annotation of myeloid cell subsets should follow the same workflow described above for major cell types.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eAnnotation of major cell types and myeloid cell subsets\u003c/h2\u003e \u003cp\u003eThe main basis for classifying clusters into specific major cell types and myeloid cell subsets was as follows: (1) expression level of key marker genes in clusters; (2) the relative distance between clusters in the UMAP graph; and (3) the distance between clusters evaluated by Cluster tree algorithm.\u003c/p\u003e \u003cp\u003eThe following key marker genes of major cell types were used in this study: T cells (\u003cem\u003eCD3D\u003c/em\u003e), NK cells (\u003cem\u003eKLRF1\u003c/em\u003e), B cells (\u003cem\u003eMS4A1\u003c/em\u003e), plasma cells (\u003cem\u003eMZB1\u003c/em\u003e), myeloid cells (\u003cem\u003eLYZ\u003c/em\u003e), mast cells (\u003cem\u003eTPSAB1\u003c/em\u003e), neutrophils (\u003cem\u003eFCGR3B\u003c/em\u003e), epithelial cells (\u003cem\u003eEPCAM\u003c/em\u003e), and stromal cells (\u003cem\u003eDCN\u003c/em\u003e).\u003c/p\u003e \u003cp\u003eMyeloid cells were defined in this study based on the similarity of cell clusters rather than strictly following biological definitions. Therefore, in the cluster annotation of major cell types, real myeloid cells such as neutrophils and mast cells were independently named, while pDCs of lymphoid origin\u003csup\u003e\u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e77\u003c/span\u003e\u003c/sup\u003e, Kupffer cells derived from embryonic yolk sac\u003csup\u003e\u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e\u003c/sup\u003e, and even doublets were all classified as myeloid cells.\u003c/p\u003e \u003cp\u003eThe myeloid cell subsets were identified based on markers as follows: pDC-LILRA4 (\u003cem\u003eLILRA4\u003c/em\u003e), cDC1-CLEC9A (\u003cem\u003eCLEC9A\u003c/em\u003e), cDC2-CD1C (\u003cem\u003eCD1C\u003c/em\u003e), cDC3-LAMP3 (\u003cem\u003eLAMP3\u003c/em\u003e), cDC4-LTB (\u003cem\u003eLTB\u003c/em\u003e), Macro-C1QC (\u003cem\u003eC1QC\u003c/em\u003e), Macro-FCN1 (\u003cem\u003eFCN1\u003c/em\u003e), Macro-MKI67 (\u003cem\u003eMKI67\u003c/em\u003e), Macro-SPP1 (\u003cem\u003eSPP1\u003c/em\u003e), Doublets-B (\u003cem\u003eCD79A, MS4A1\u003c/em\u003e and \u003cem\u003eMZB1\u003c/em\u003e), Doublets-Epi (\u003cem\u003eEPCAM\u003c/em\u003e), Doublets-T (\u003cem\u003eCD3D\u003c/em\u003e), Monocytes (\u003cem\u003eCD14\u003c/em\u003e), and Kupffer cells (\u003cem\u003eMARCO\u003c/em\u003e).\u003c/p\u003e \u003cp\u003eSince CRC macrophages were the focus of this study, the major monocyte/macrophage subpopulation in PBMCs of the CRLM dataset was only broadly defined as monocytes. In the datasets without PBMC samples, we did not define the monocyte subset, and because of the high similarity in the transcriptome characteristics of monocytes and Macro-FCN1, the potential monocytes in these datasets are usually classified into the Macro-FCN1 subset. Similarly, in datasets without NL samples, we do not define a subset of Kupffer cells.\u003c/p\u003e \u003cp\u003eMacro-C1QC, Macro-FCN1, Macro-MKI67, Macro-SPP1, monocytes, and Kupffer cells in this study were collectively referred to as monocytes/macrophages; macrophages included Macro-C1QC, Macro-FCN1, Macro-MKI67, and Macro-SPP1 subsets.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eData integration and batch correction\u003c/h2\u003e \u003cp\u003eTo construct the large sample data of CRC myeloid cells, CRC myeloid cell data from five independent datasets (GSE132465, GSE144735, GSE146771, GSE178318, and GSE178341) were combined into CRC-Mix data. After correcting for batch effects using the harmony algorithm\u003csup\u003e\u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e79\u003c/span\u003e\u003c/sup\u003e, the CRC-Mix data were reclustered into groups and annotated.\u003c/p\u003e \u003cp\u003eIn addition, to compare the differences in SPP1\u0026thinsp;+\u0026thinsp;macrophage proportions in CRLM and HCC, myeloid cells of GSE156625 and GSE164522 were integrated in the same way, corrected for batch effects by the harmony algorithm, reclustered into groups and annotated.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eExpression difference analysis and representative genes\u003c/h2\u003e \u003cp\u003eThe \u003cem\u003eFindallmarkers\u003c/em\u003e function in Seurat was used to identify marker genes for each cluster or subset. In addition, to explore the functional changes in monocytes/macrophages during CRLM, we also used the \u003cem\u003eFindmarkers\u003c/em\u003e function to compare the monocyte/macrophage transcriptomes of NC and CRC, as well as CRC and LM. The parameters assert=\"RNA\" and slot=\"counts\" allowed for the inclusion of genes with low variability in the comparison. Genes with significant expression differences (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) in the comparison between groups were defined as DEGs. DEGs with the same trend in multiple datasets were defined as common DEGs, and these genes were used for subsequent functional enrichment analysis.\u003c/p\u003e \u003cp\u003eIn brief, for myeloid cell subsets in the CRC-Mix dataset, genes that met these criteria were considered representative genes: 1) adjusted P value\u0026thinsp;\u0026lt;\u0026thinsp;0.01 with all features; 2) log fold-change of the average expression\u0026thinsp;\u0026gt;\u0026thinsp;0.5; 3) pct.1 (percentage of gene expression detected in the first group)\u0026thinsp;\u0026gt;\u0026thinsp;0.25; 4) The above three-step filtering process was also performed on the myeloid cell subsets defined before harmony correction, and the intersection genes were selected. These genes were further checked manually to ensure their expression specificity on the corresponding subsets.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eTissue distribution of myeloid cell subsets\u003c/h2\u003e \u003cp\u003eTo understand the distribution preference of each myeloid cell subtype in different tissues, we evaluated the enrichment of each myeloid cell subtype using the OR in seven CRC single-cell datasets. OR\u0026thinsp;\u0026gt;\u0026thinsp;1 indicates the relative enrichment of the cell subtype in a specific tissue.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eMonocyte/macrophage developmental trajectory in CRLM\u003c/h2\u003e \u003cp\u003eTo construct the monocyte/macrophage developmental trajectory of colorectal cancer liver metastases, we used the \"Monocle\" package (version 2.22.0)\u003csup\u003e\u003cspan citationid=\"CR80\" class=\"CitationRef\"\u003e80\u003c/span\u003e\u003c/sup\u003e to align monocytes/macrophages in pseudotime order. The \u003cem\u003eDDRTree\u003c/em\u003e method implemented with the \u003cem\u003ereduceDimension\u003c/em\u003e function of Monocle 2 was used for dimensionality reduction and construction of pseudo-temporal order.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eCell‒cell interaction analysis\u003c/h2\u003e \u003cp\u003eWe used Python-based (version 3.7) CellphoneDB\u003csup\u003e\u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e81\u003c/span\u003e\u003c/sup\u003e to evaluate the interactions of macrophages with other cells. The putative ligands and receptors were determined based on whether they were expressed on each cell. To efficiently assess the closeness of cell interactions based on the number of receptor‒ligand pairs, we randomly sampled 1,000 cells per population from the Macro-C1QC, Macro-FCN1, Macro-MKI67, Macro-SPP1 and major cell types in the GSE178318 dataset.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eDefining phenotype scores\u003c/h2\u003e \u003cp\u003eTo understand the characteristics between different macrophage subsets, scores for different phenotypes were obtained by the \u003cem\u003eAddModuleScore\u003c/em\u003e function in the \"Seurat\" package. These scores were defined by the mean expression of phenotype-related genes.\u003c/p\u003e \u003cp\u003eThe related genes of M1 macrophages and M2 macrophages were defined as M1 score and M2 score, respectively. The M1 score considers the expression of \u003cem\u003eCCL5, CCR7, CD40, CD86, CXCL9, CXCL10, CXCL11, IDO1, IL1A, IL1B, IL6, IRF1, and IRF5\u003c/em\u003e, while M2 score considers the expression of \u003cem\u003eCCL4, CCL13, CCL18, CCL20, CCL22, CD276, CLEC7A, CTSA, CTSB, CTSC, CTSD, FN1, IL4R, IRF4, LYVE1, MMP9, MMP14, MMP19, MSR1, TGFB1, TGFB2, TGFB3, TNFSF8, TNFSF12, VEGFA, VEGFB, and VEGFC\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eInflammatory score-related genes included \u003cem\u003eIL1B, IL6, S100A8, and S100A9\u003c/em\u003e, while \u003cem\u003eAPOE, CD163, MAF, SELENOP, and SEPP1\u003c/em\u003e were used to define the anti-inflammatory score\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eGlycolysis score-related genes included \u003cem\u003eALDOA, ALDOB, ALDOC, ENO1, ENO2, ENO3, GAPDH, GPI, HK1, HK2, HK3, LDHA, PFKL, PFKM, PFKP, PGAM1, PGAM4, PGK1, PKLR, PKM, SLC2A1, and TPI1\u003c/em\u003e\u003csup\u003e82\u003c/sup\u003e. Hypoxia score-related genes included \u003cem\u003eACOT7, ADM, ALDOA, CDKN3, ENO1, LDHA, MIF, NDRG1, P4HA1, PGAM1, SLC2A1, TPI1, TUBB6, and VEGFA\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e82\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eEMT score-related genes included \u003cem\u003eADAM12, ADAMTS12, ADAMTS2, AEBP1, ANGPTL2, ANTXR1, AP1G1, ATP8B1, AXL, BNC2, CALD1, CDH1, CDH2, CDS1, CGN, CLDN4, CMTM3, CNOT1, CNRIP1, COL10A1, COL1A1, COL1A2, COL3A1, COL5A1, COL5A2, COL6A1, COL6A2, COL6A3, COL8A1, CTNND1, DACT1, DYNC1LI2, EMP3, ERBB3, ESRP1, ESRP2, F11R, FAP, FBN1, FN1, FSTL1, GALNT3, GPC6, GPR56, GRHL2, GYPC, HOOK1, HTRA1, INHBA, IRF6, ITGA11, LOXL2, LRRC15, MAP7, MARVELD2, MARVELD3, MMP2, MSRB3, MYO5B, NAP1L3, NID2, OCLN, OLFML2B, PCOLCE, PDGFRB, PMP22, POSTN, PRSS8, SPARC, SPINT1, SPOCK1, SULF1, SYT11, THBS2, VCAN, VIM\u003c/em\u003e and \u003cem\u003eZEB2\u003c/em\u003e\u003csup\u003e83\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eProliferation score-related genes included \u003cem\u003eAURKA, BUB1, CCNB1, CCND1, CCNE1, DEK, E2F1, FEN1, FOXM1, H2AFZ, HMGB2, MCM2, MCM3, MCM4, MCM5, MCM6, MKI67, MYBL2, PCNA, PLK1, TOP2A, TYMS\u003c/em\u003e, and \u003cem\u003eZWINT\u003c/em\u003e \u003csup\u003e\u003cem\u003e\u003cspan citationid=\"CR84\" class=\"CitationRef\"\u003e84\u003c/span\u003e\u003c/em\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eAngiogenesis score-related genes included \u003cem\u003eANGPTL4, CXCL8, VCAN, and VEGFA\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. Phagocytosis score-related genes included \u003cem\u003eC1QB, CD163, MERTK, and MRC1\u003c/em\u003e\u003csup\u003e19\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eMHC-II score-related genes included \u003cem\u003eHLA-DMA, HLA-DMB, HLA-DOA, HLA-DOB, HLA-DPA1, HLA-DPB1, HLA-DQA1, HLA-DQA2, HLA-DQB1, HLA-DQB2, HLA-DRA, HLA-DRB1, and HLA-DRB5.\u003c/em\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eImmune infiltration and macrophage markers\u003c/h2\u003e \u003cp\u003eThe immune score calculated by the ESTIMATE algorithm was used to assess the overall immune infiltration of each sample\u003csup\u003e\u003cspan citationid=\"CR85\" class=\"CitationRef\"\u003e85\u003c/span\u003e\u003c/sup\u003e. The immune cell marker \u003cem\u003eCD45\u003c/em\u003e was also used to assess the level of immune cell infiltration in tissues. CIBERSORT, a deconvolution algorithm based on transcriptional profiles, was used to calculate the proportions of 22 immune cells, including M0, M1, and M2 macrophages, in each sample\u003csup\u003e\u003cspan citationid=\"CR86\" class=\"CitationRef\"\u003e86\u003c/span\u003e\u003c/sup\u003e. To understand the clinical features of SPP1 as a specific macrophage marker in CRC, some classic macrophage markers were selected for comparison; these included the pan-macrophage marker CD68, M1 macrophage markers \u003cem\u003eCD86\u003c/em\u003e and \u003cem\u003eiNOS\u003c/em\u003e, and M2 macrophage markers \u003cem\u003eCD163\u003c/em\u003e and \u003cem\u003eCD206\u003c/em\u003e\u003csup\u003e34,87\u003c/sup\u003e. In addition, the web server TIMER (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://timer.cistrome.org/\u003c/span\u003e\u003cspan address=\"http://timer.cistrome.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was used to evaluate the correlation of \u003cem\u003eSPP1\u003c/em\u003e with the six immune cell and macrophage markers mentioned above.\u003csup\u003e\u003cspan citationid=\"CR88\" class=\"CitationRef\"\u003e88\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eTumor mutation burden\u003c/h2\u003e \u003cp\u003eSimple nucleotide variation data from TCGA-CRC were used to calculate TMB, which is defined as the number of mutations per megabase. The visualization of the detailed gene mutation status of the high and low \u003cem\u003eSPP1/CD68\u003c/em\u003e groups was implemented in the R package \"maftools\"\u003csup\u003e\u003cspan citationid=\"CR89\" class=\"CitationRef\"\u003e89\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003ePrediction of immunotherapy response\u003c/h2\u003e \u003cp\u003eWe used the SubMap algorithm to predict the clinical response to anti-PD1 and anti-CTLA4 immunotherapy. SubMap was used to compare the similarity of different expression profiles; this feature can reflect the response to treatment\u003csup\u003e\u003cspan citationid=\"CR90\" class=\"CitationRef\"\u003e90\u003c/span\u003e\u003c/sup\u003e. The expression profiles and associated annotation data of 47 melanoma patients used to define high- and low-risk groups were obtained from the Supplementary Materials of Lu \u003cem\u003eet al\u003c/em\u003e.\u003csup\u003e\u003cspan citationid=\"CR91\" class=\"CitationRef\"\u003e91\u003c/span\u003e\u003c/sup\u003e In addition, samples in GSE39582 were divided into high and low SPP1 groups based on the median SPP1 expression.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003eSurvival analysis\u003c/h2\u003e \u003cp\u003eOS and PFI were regarded as the outcome events in the datasets (OS in GSE17536, GSE39582, and GSE17536; OS/PFI in TCGA). Survival analysis was performed based on the Kaplan‒Meier algorithm. The log-rank test was used to calculate p values between groups. Univariate Cox models were constructed by Cox proportional hazards regression. These analyses were implemented with \"survival\" and \"survminer\" in R.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eFunctional and pathway enrichment analysis\u003c/h2\u003e \u003cp\u003eIn this study, four different gene sets were used for functional and pathway enrichment analysis, including Gene Ontology (GO) Biological Process (BP), Kyoto Encyclopedia of Genes and Genomes (KEGG), WikiPathways, and Hallmark gene sets\u003csup\u003e\u003cspan citationid=\"CR92\" class=\"CitationRef\"\u003e92\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eMetascape (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://metascape.org/\u003c/span\u003e\u003cspan address=\"http://metascape.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), a portal for gene function annotation analysis, was used for enrichment analysis of DEGs in this study (GO BP, KEGG, WikiPathways, and Hallmark gene sets)\u003csup\u003e\u003cspan citationid=\"CR93\" class=\"CitationRef\"\u003e93\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eGene set variation analysis (GSVA) implemented in the \"GSVA\" package (version 1.44.2) was used for gene set enrichment analysis of different macrophage subsets (Hallmark gene sets)\u003csup\u003e\u003cspan citationid=\"CR94\" class=\"CitationRef\"\u003e94\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eLinkedOmics (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.linkedomics.org/\u003c/span\u003e\u003cspan address=\"http://www.linkedomics.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) is an online source portal for analyzing multiomics data of 32 cancers in TCGA\u003csup\u003e\u003cspan citationid=\"CR95\" class=\"CitationRef\"\u003e95\u003c/span\u003e\u003c/sup\u003e. In this study, LinkedOmics was used to perform GSEA of \u003cem\u003eSPP1\u003c/em\u003e expression profiles in TCGA-CRC (GO BP).\u003c/p\u003e \u003cp\u003eIn addition, GSEA software (version 4.2.3) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gsea-msigdb.org/gsea/\u003c/span\u003e\u003cspan address=\"https://www.gsea-msigdb.org/gsea/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was used to explore differences in functional enrichment between the L-\u003cem\u003eSPP1/CD68\u003c/em\u003e and H-\u003cem\u003eSPP1/CD68\u003c/em\u003e groups (Hallmark gene sets, permutations\u0026thinsp;=\u0026thinsp;1000)\u003csup\u003e96\u003c/sup\u003e. | Normalized enrichment score (NES) | \u0026ge;1.0, false discovery rate (FDR) q-val\u0026thinsp;\u0026le;\u0026thinsp;0.25, and p value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 were considered statistically significant.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec25\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eR (version 4.1.3), GraphPad Prism (version 9), Excel, SPSS (version 26), and Python (version 3.7) were used for statistical analysis. The Mann‒Whitney U test was used to compare the differences between two groups. The Pearson or Spearman method was used for the correlation test. The log-rank method was used to calculate the P value in survival between groups. A chi-square test was used to compare clinical parameters between the high and low \u003cem\u003eSPP1/CD68\u003c/em\u003e groups. Data in bar plots of cell proportions for single-cell datasets are presented as the mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SEM (small sample size), while the mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD is used for the remaining data. \u003cem\u003eP\u003c/em\u003e value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eCode availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe codes used in this study are available on Github (https://github.com/LPC19970117/SPP1-macrophages).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data used in this study are all from public datasets. Details can be found in the Methods section of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to thank the researchers contributing to public databases and those providing data related to human medicine. We especially thank Dr. Wu for providing the CRLM data. This work was supported by the National Natural Science Foundation of China (82072655).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no\u0026nbsp;conflict of interest.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eZYX designed the study. ZYX and LRN collected the data. ZYX, FF and GZZ analyzed the data. ZYX, RKL, HJD, LLD, HZW, GMR and XYD visualized the data. ZYX and LRN drafted the manuscript. JZ and JYZ revised the manuscript. All authors have read and approved the final manuscript. All authors contributed to the article and approved the submitted version.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary information accompanies the manuscript on the Cell Discovery website http://www.nature.com/celldisc\u0026rsquo;\u003c/strong\u003e\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eSiegel, R. L., Miller, K. D., Fuchs, H. E. \u0026amp; Jemal, A. Cancer statistics, 2022. CA Cancer J. Clin. \u003cb\u003e72\u003c/b\u003e, 7\u0026ndash;33 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFan, A. \u003cem\u003eet al.\u003c/em\u003e Immunotherapy in colorectal cancer: current achievements and future perspective. Int. J. Biol. Sci. \u003cb\u003e17\u003c/b\u003e, 3837\u0026ndash;3849 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZaanan, A. \u003cem\u003eet al.\u003c/em\u003e Role of deficient DNA mismatch repair status in patients with stage III colon cancer treated with FOLFOX adjuvant chemotherapy: a pooled analysis from 2 randomized clinical trials. JAMA Oncol. \u003cb\u003e4\u003c/b\u003e, 379\u0026ndash;383 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNordholm-Carstensen, A., Krarup, P. M., Morton, D., Harling, H. \u0026amp; Danish Colorectal Cancer Group. Mismatch repair status and synchronous metastases in colorectal cancer: a nationwide cohort study. Int. J. Cancer \u003cb\u003e137\u003c/b\u003e, 2139\u0026ndash;2148 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLe, D. T. \u003cem\u003eet al.\u003c/em\u003e PD-1 blockade in tumors with mismatch-repair deficiency. N. Engl. J. Med. \u003cb\u003e372\u003c/b\u003e, 2509\u0026ndash;2520 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDeNardo, D. G. \u0026amp; Ruffell, B. Macrophages as regulators of tumour immunity and immunotherapy. Nat. Rev. Immunol. \u003cb\u003e19\u003c/b\u003e, 369\u0026ndash;382 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVogel, D. Y. \u003cem\u003eet al.\u003c/em\u003e Human macrophage polarization in vitro: maturation and activation methods compared. Immunobiology \u003cb\u003e219\u003c/b\u003e, 695\u0026ndash;703 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGinhoux, F., Schultze, J. L., Murray, P. J., Ochando, J. \u0026amp; Biswas, S. K. New insights into the multidimensional concept of macrophage ontogeny, activation and function. Nat. Immunol. \u003cb\u003e17\u003c/b\u003e, 34\u0026ndash;40 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, Y. \u003cem\u003eet al.\u003c/em\u003e Immune phenotypic linkage between colorectal cancer and liver metastasis. Cancer Cell \u003cb\u003e40\u003c/b\u003e, 424\u0026ndash;437.e5 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhantakova, D., Brioschi, S. \u0026amp; Molgora, M. Exploring the impact of TREM2 in tumor-associated macrophages. Vaccines (Basel) \u003cb\u003e10\u003c/b\u003e, 943 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDing, J. \u003cem\u003eet al.\u003c/em\u003e M2 macrophage-derived G-CSF promotes trophoblasts EMT, invasion and migration via activating PI3K/Akt/Erk1/2 pathway to mediate normal pregnancy. J. Cell. Mol. Med. \u003cb\u003e25\u003c/b\u003e, 2136\u0026ndash;2147 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, Z. \u003cem\u003eet al.\u003c/em\u003e Polarization of intestinal tumour-associated macrophages regulates the development of schistosomal colorectal cancer. J. Cancer \u003cb\u003e12\u003c/b\u003e, 1033\u0026ndash;1041 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCortese, N. \u003cem\u003eet al.\u003c/em\u003e Macrophages in colorectal cancer liver metastases. Cancers (Basel) \u003cb\u003e11\u003c/b\u003e, 633 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePyonteck, S. M. \u003cem\u003eet al.\u003c/em\u003e CSF-1R inhibition alters macrophage polarization and blocks glioma progression. Nat. Med. \u003cb\u003e19\u003c/b\u003e, 1264\u0026ndash;1272 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee, J. H. \u003cem\u003eet al.\u003c/em\u003e A phase I study of pexidartinib, a colony-stimulating factor 1 receptor inhibitor, in Asian patients with advanced solid tumors. Invest. New Drugs \u003cb\u003e38\u003c/b\u003e, 99\u0026ndash;110 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRies, C. H. \u003cem\u003eet al.\u003c/em\u003e Targeting tumor-associated macrophages with anti-CSF-1R antibody reveals a strategy for cancer therapy. Cancer Cell \u003cb\u003e25\u003c/b\u003e, 846\u0026ndash;859 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePapadopoulos, K. P. \u003cem\u003eet al.\u003c/em\u003e First-in-human study of AMG 820, a monoclonal anti-colony-stimulating factor 1 receptor antibody, in patients with advanced solid tumors. Clin. Cancer Res. \u003cb\u003e23\u003c/b\u003e, 5703\u0026ndash;5710 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu, Y. \u003cem\u003eet al.\u003c/em\u003e Spatiotemporal immune landscape of colorectal cancer liver metastasis at single-cell level. Cancer Discov. \u003cb\u003e12\u003c/b\u003e, 134\u0026ndash;153 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, L. \u003cem\u003eet al.\u003c/em\u003e Single-cell analyses inform mechanisms of myeloid-targeted therapies in colon cancer. Cell \u003cb\u003e181\u003c/b\u003e, 442\u0026ndash;459.e29 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQi, J. \u003cem\u003eet al.\u003c/em\u003e Single-cell and spatial analysis reveal interaction of FAP(+) fibroblasts and SPP1(+) macrophages in colorectal cancer. Nat. Commun. \u003cb\u003e13\u003c/b\u003e, 1742 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFeng, S. \u003cem\u003eet al.\u003c/em\u003e SPP1 as a key gene in the lymph node metastasis and a potential predictor of poor prognosis in head and neck carcinoma. J. Oral Pathol. Med. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1111/jop.13333\u003c/span\u003e\u003cspan address=\"10.1111/jop.13333\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, Y. \u003cem\u003eet al.\u003c/em\u003e A pan-cancer analysis of the oncogenic role of secreted phosphoprotein 1 (SPP1) in human cancers. Ann. Transl. Med. \u003cb\u003e10\u003c/b\u003e, 279 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, K., Ma, Z. \u0026amp; Zhang, W. Comprehensive analysis to identify SPP1 as a prognostic biomarker in cervical cancer. Front. Genet. \u003cb\u003e12\u003c/b\u003e, 732822 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu, C. \u003cem\u003eet al.\u003c/em\u003e SPP1, analyzed by bioinformatics methods, promotes the metastasis in colorectal cancer by activating EMT pathway. Biomed. Pharmacother. \u003cb\u003e91\u003c/b\u003e, 1167\u0026ndash;1177 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWei, J. \u003cem\u003eet al.\u003c/em\u003e Characterizing intercellular communication of pan-cancer reveals SPP1 + tumor-associated macrophage expanded in hypoxia and promoting cancer malignancy through single-cell RNA-seq data. Front. Cell Dev. Biol. \u003cb\u003e9\u003c/b\u003e, 749210 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRao, G. \u003cem\u003eet al.\u003c/em\u003e Reciprocal interactions between tumor-associated macrophages and CD44-positive cancer cells via osteopontin/CD44 promote tumorigenicity in colorectal cancer. Clin. Cancer Res. \u003cb\u003e19\u003c/b\u003e, 785\u0026ndash;797 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSenbanjo, L. T. \u0026amp; Chellaiah, M. A. CD44: a multifunctional cell surface adhesion receptor is a regulator of progression and metastasis of cancer cells. Front. Cell Dev. Biol. \u003cb\u003e5\u003c/b\u003e, 18 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAssidi, M. \u003cem\u003eet al.\u003c/em\u003e Prognostic value of Osteopontin (SPP1) in colorectal carcinoma requires a personalized molecular approach. Tumour Biol. \u003cb\u003e41\u003c/b\u003e, 1010428319863627 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePietras, A. \u003cem\u003eet al.\u003c/em\u003e Osteopontin-CD44 signaling in the glioma perivascular niche enhances cancer stem cell phenotypes and promotes aggressive tumor growth. Cell Stem Cell \u003cb\u003e14\u003c/b\u003e, 357\u0026ndash;369 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePang, X. \u003cem\u003eet al.\u003c/em\u003e Osteopontin as a multifaceted driver of bone metastasis and drug resistance. Pharmacol. Res. \u003cb\u003e144\u003c/b\u003e, 235\u0026ndash;244 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eClay, R. \u0026amp; Siddiqi, S. A. in Theranostic Approach for Pancreatic Cancer (eds Nagaraju, G. P. \u0026amp; Ahmad, S.) (Academic Press, 2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCheng, S. \u003cem\u003eet al.\u003c/em\u003e A pan-cancer single-cell transcriptional atlas of tumor infiltrating myeloid cells. Cell \u003cb\u003e184\u003c/b\u003e, 792\u0026ndash;809.e23 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRen, X. \u003cem\u003eet al.\u003c/em\u003e Insights gained from single-cell analysis of immune cells in the tumor microenvironment. Annu. Rev. Immunol. \u003cb\u003e39\u003c/b\u003e, 583\u0026ndash;609 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu, K. \u003cem\u003eet al.\u003c/em\u003e Redefining tumor-associated macrophage subpopulations and functions in the tumor microenvironment. Front. Immunol. \u003cb\u003e11\u003c/b\u003e, 1731 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSzulzewsky, F. \u003cem\u003eet al.\u003c/em\u003e Glioma-associated microglia/macrophages display an expression profile different from M1 and M2 polarization and highly express Gpnmb and Spp1. PLoS One \u003cb\u003e10\u003c/b\u003e, e0116644 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, L. \u003cem\u003eet al.\u003c/em\u003e Construction of TME and Identification of crosstalk between malignant cells and macrophages by SPP1 in hepatocellular carcinoma. Cancer Immunol. Immunother. \u003cb\u003e71\u003c/b\u003e, 121\u0026ndash;136 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLoustau, M. \u003cem\u003eet al.\u003c/em\u003e HLA-G neo-expression on tumors. Front. Immunol. \u003cb\u003e11\u003c/b\u003e, 1685 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, Y. \u0026amp; Zhang, Z. The history and advances in cancer immunotherapy: understanding the characteristics of tumor-infiltrating immune cells and their therapeutic implications. Cell. Mol. Immunol. \u003cb\u003e17\u003c/b\u003e, 807\u0026ndash;821 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePonten, F., Jirstrom, K. \u0026amp; Uhlen, M. The human protein Atlas\u0026ndash;a tool for pathology. J. Pathol. \u003cb\u003e216\u003c/b\u003e, 387\u0026ndash;393 (2008).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, J. \u003cem\u003eet al.\u003c/em\u003e Tumor-associated macrophage infiltration and prognosis in colorectal cancer: systematic review and meta-analysis. Int. J. Colorectal Dis. \u003cb\u003e35\u003c/b\u003e, 1203\u0026ndash;1210 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBruni, D., Angell, H. K. \u0026amp; Galon, J. The immune contexture and immunoscore in cancer prognosis and therapeutic efficacy. Nat. Rev. Cancer \u003cb\u003e20\u003c/b\u003e, 662\u0026ndash;680 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePridans, C. \u003cem\u003eet al.\u003c/em\u003e Pleiotropic impacts of macrophage and microglial deficiency on development in rats with targeted mutation of the Csf1r locus. J. Immunol. \u003cb\u003e201\u003c/b\u003e, 2683\u0026ndash;2699 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuilliams, M., Thierry, G. R., Bonnardel, J. \u0026amp; Bajenoff, M. Establishment and maintenance of the macrophage niche. Immunity \u003cb\u003e52\u003c/b\u003e, 434\u0026ndash;451 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee, H. O. \u003cem\u003eet al.\u003c/em\u003e Lineage-dependent gene expression programs influence the immune landscape of colorectal cancer. Nat. Genet. \u003cb\u003e52\u003c/b\u003e, 594\u0026ndash;603 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmilca-Seba, K. \u003cem\u003eet al.\u003c/em\u003e Osteopontin (OPN/SPP1), a mediator of tumor progression, is regulated by the mesenchymal transcription factor Slug/SNAI2 in colorectal cancer (CRC). Cells \u003cb\u003e11\u003c/b\u003e, 1808 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTu, Y., Chen, C. \u0026amp; Fan, G. Association between the expression of secreted phosphoprotein - related genes and prognosis of human cancer. BMC Cancer \u003cb\u003e19\u003c/b\u003e, 1230 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAzizi, E. \u003cem\u003eet al.\u003c/em\u003e Single-cell map of diverse immune phenotypes in the breast tumor microenvironment. Cell \u003cb\u003e174\u003c/b\u003e, 1293\u0026ndash;1308.e36 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXie, Z. \u003cem\u003eet al.\u003c/em\u003e Immune cell confrontation in the papillary thyroid carcinoma microenvironment. Front. Endocrinol. (Lausanne) \u003cb\u003e11\u003c/b\u003e, 570604 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, C., Zhao, S., Karnad, A. \u0026amp; Freeman, J. W. The biology and role of CD44 in cancer progression: therapeutic implications. J. Hematol. Oncol. \u003cb\u003e11\u003c/b\u003e, 64 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe, X. \u003cem\u003eet al.\u003c/em\u003e Tumor-initiating stem cell shapes its microenvironment into an immunosuppressive barrier and pro-tumorigenic niche. Cell Rep. \u003cb\u003e36\u003c/b\u003e, 109674 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchlesinger, M. \u0026amp; Bendas, G. Contribution of very late antigen-4 (VLA-4) integrin to cancer progression and metastasis. Cancer Metastasis Rev. \u003cb\u003e34\u003c/b\u003e, 575\u0026ndash;591 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, J. \u003cem\u003eet al.\u003c/em\u003e The IRENA lncRNA converts chemotherapy-polarized tumor-suppressing macrophages to tumor-promoting phenotypes in breast cancer. Nat. Cancer \u003cb\u003e2\u003c/b\u003e, 457\u0026ndash;473 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSidaway, P. MSI-H: a truly agnostic biomarker? Nat. Rev. Clin. Oncol. \u003cb\u003e17\u003c/b\u003e, 68 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYarchoan, M., Hopkins, A. \u0026amp; Jaffee, E. M. Tumor mutational burden and response rate to PD-1 inhibition. N. Engl. J. Med. \u003cb\u003e377\u003c/b\u003e, 2500\u0026ndash;2501 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCristescu, R. \u003cem\u003eet al.\u003c/em\u003e Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy. Science \u003cb\u003e362\u003c/b\u003e, eaar3593 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePanda, A. \u003cem\u003eet al.\u003c/em\u003e Identifying a clinically applicable mutational burden threshold as a potential biomarker of response to immune checkpoint therapy in solid tumors. JCO Precis. Oncol. \u003cb\u003e2017\u003c/b\u003e, PO.17.00146 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKlement, J. D. \u003cem\u003eet al.\u003c/em\u003e Osteopontin blockade immunotherapy increases cytotoxic T lymphocyte lytic activity and suppresses colon tumor progression. Cancers (Basel) \u003cb\u003e13\u003c/b\u003e, 1006 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDeng, G. \u003cem\u003eet al.\u003c/em\u003e BET inhibitor suppresses melanoma progression via the noncanonical NF-kappaB/SPP1 pathway. Theranostics \u003cb\u003e10\u003c/b\u003e, 11428\u0026ndash;11443 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFarrokhi, V., Chabot, J. R., Neubert, H. \u0026amp; Yang, Z. Assessing the feasibility of neutralizing osteopontin with various therapeutic antibody modalities. Sci. Rep. \u003cb\u003e8\u003c/b\u003e, 7781 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePelka, K. \u003cem\u003eet al.\u003c/em\u003e Spatially organized multicellular immune hubs in human colorectal cancer. Cell \u003cb\u003e184\u003c/b\u003e, 4734\u0026ndash;4752.e20 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChe, L. H. \u003cem\u003eet al.\u003c/em\u003e A single-cell atlas of liver metastases of colorectal cancer reveals reprogramming of the tumor microenvironment in response to preoperative chemotherapy. Cell Discov. \u003cb\u003e7\u003c/b\u003e, 80 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSharma, A. \u003cem\u003eet al.\u003c/em\u003e Onco-fetal reprogramming of endothelial cells drives immunosuppressive macrophages in hepatocellular carcinoma. Cell \u003cb\u003e183\u003c/b\u003e, 377\u0026ndash;394.e21 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKi, D. H. \u003cem\u003eet al.\u003c/em\u003e Whole genome analysis for liver metastasis gene signatures in colorectal cancer. Int. J. Cancer \u003cb\u003e121\u003c/b\u003e, 2005\u0026ndash;2012 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJorissen, R. N. \u003cem\u003eet al.\u003c/em\u003e Metastasis-associated gene expression changes predict poor outcomes in patients with dukes stage B and C colorectal cancer. Clin. Cancer Res. \u003cb\u003e15\u003c/b\u003e, 7642\u0026ndash;7651 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSmith, J. J. \u003cem\u003eet al.\u003c/em\u003e Experimentally derived metastasis gene expression profile predicts recurrence and death in patients with colon cancer. Gastroenterology \u003cb\u003e138\u003c/b\u003e, 958\u0026ndash;968 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGaedcke, J. \u003cem\u003eet al.\u003c/em\u003e Mutated KRAS results in overexpression of DUSP4, a MAP-kinase phosphatase, and SMYD3, a histone methyltransferase, in rectal carcinomas. Genes Chromosomes Cancer \u003cb\u003e49\u003c/b\u003e, 1024\u0026ndash;1034 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSkrzypczak, M. \u003cem\u003eet al.\u003c/em\u003e Modeling oncogenic signaling in colon tumors by multidirectional analyses of microarray data directed for maximization of analytical reliability. PLoS One \u003cb\u003e5\u003c/b\u003e, e13091 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarisa, L. \u003cem\u003eet al.\u003c/em\u003e Gene expression classification of colon cancer into molecular subtypes: characterization, validation, and prognostic value. PLoS Med. \u003cb\u003e10\u003c/b\u003e, e1001453 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSheffer, M. \u003cem\u003eet al.\u003c/em\u003e Association of survival and disease progression with chromosomal instability: a genomic exploration of colorectal cancer. \u003cem\u003eProc. Natl. Acad. Sci. U. S. A.\u003c/em\u003e \u003cb\u003e106\u003c/b\u003e, 7131\u0026ndash;7136 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSole, X. \u003cem\u003eet al.\u003c/em\u003e Discovery and validation of new potential biomarkers for early detection of colon cancer. PLoS One \u003cb\u003e9\u003c/b\u003e, e106748 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRyan, B. M. \u003cem\u003eet al.\u003c/em\u003e Germline variation in NCF4, an innate immunity gene, is associated with an increased risk of colorectal cancer. Int. J. Cancer \u003cb\u003e134\u003c/b\u003e, 1399\u0026ndash;1407 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKwon, Y. \u003cem\u003eet al.\u003c/em\u003e Prognosis of stage III colorectal carcinomas with FOLFOX adjuvant chemotherapy can be predicted by molecular subtype. Oncotarget \u003cb\u003e8\u003c/b\u003e, 39367\u0026ndash;39381 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHu, Y. \u003cem\u003eet al.\u003c/em\u003e Colorectal cancer susceptibility loci as predictive markers of rectal cancer prognosis after surgery. Genes Chromosomes Cancer \u003cb\u003e57\u003c/b\u003e, 140\u0026ndash;149 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKamal, Y., Schmit, S. L., Hoehn, H. J., Amos, C. I. \u0026amp; Frost, H. R. Transcriptomic differences between primary colorectal adenocarcinomas and distant metastases reveal metastatic colorectal cancer subtypes. Cancer Res. \u003cb\u003e79\u003c/b\u003e, 4227\u0026ndash;4241 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSzeglin, B. C. \u003cem\u003eet al.\u003c/em\u003e A SMAD4-modulated gene profile predicts disease-free survival in stage II and III colorectal cancer. Cancer Rep. (Hoboken) \u003cb\u003e5\u003c/b\u003e, e1423 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcInnes, L. \u0026amp; Healy, J. UMAP: uniform manifold approximation and projection for dimension reduction. \u003cem\u003earXiv:1802.03426\u003c/em\u003e (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMulder, K. \u003cem\u003eet al.\u003c/em\u003e Cross-tissue single-cell landscape of human monocytes and macrophages in health and disease. Immunity \u003cb\u003e54\u003c/b\u003e, 1883\u0026ndash;1900.e5 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuilliams, M. \u0026amp; Scott, C. L. Does niche competition determine the origin of tissue-resident macrophages? Nat. Rev. Immunol. \u003cb\u003e17\u003c/b\u003e, 451\u0026ndash;460 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKorsunsky, I. \u003cem\u003eet al.\u003c/em\u003e Fast, sensitive and accurate integration of single-cell data with Harmony. Nat. Methods \u003cb\u003e16\u003c/b\u003e, 1289\u0026ndash;1296 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTrapnell, C. \u003cem\u003eet al.\u003c/em\u003e The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. \u003cb\u003e32\u003c/b\u003e, 381\u0026ndash;386 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEfremova, M., Vento-Tormo, M., Teichmann, S. A. \u0026amp; Vento-Tormo, R. CellPhoneDB: inferring cell-cell communication from combined expression of multi-subunit ligand-receptor complexes. Nat. Protoc. \u003cb\u003e15\u003c/b\u003e, 1484\u0026ndash;1506 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWei, J. \u003cem\u003eet al.\u003c/em\u003e Characterization of glycolysis-associated molecules in the tumor microenvironment revealed by pan-cancer tissues and lung cancer single cell data. Cancers (Basel) \u003cb\u003e12\u003c/b\u003e, 1788 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMak, M. P. \u003cem\u003eet al.\u003c/em\u003e A patient-derived, Pan-cancer EMT signature identifies global molecular alterations and immune target enrichment following epithelial-to-mesenchymal transition. Clin. Cancer Res. \u003cb\u003e22\u003c/b\u003e, 609\u0026ndash;620 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWhitfield, M. L., George, L. K., Grant, G. D. \u0026amp; Perou, C. M. Common markers of proliferation. Nat. Rev. Cancer \u003cb\u003e6\u003c/b\u003e, 99\u0026ndash;106 (2006).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoshihara, K. \u003cem\u003eet al.\u003c/em\u003e Inferring tumour purity and stromal and immune cell admixture from expression data. Nat. Commun. \u003cb\u003e4\u003c/b\u003e, 2612 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNewman, A. M. \u003cem\u003eet al.\u003c/em\u003e Robust enumeration of cell subsets from tissue expression profiles. Nat. Methods \u003cb\u003e12\u003c/b\u003e, 453\u0026ndash;457 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, Y. \u003cem\u003eet al.\u003c/em\u003e Prognostic value and clinicopathological roles of phenotypes of tumour-associated macrophages in colorectal cancer. J. Cancer Res. Clin. Oncol. \u003cb\u003e145\u003c/b\u003e, 3005\u0026ndash;3019 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, T. \u003cem\u003eet al.\u003c/em\u003e TIMER: a web server for comprehensive analysis of tumor-infiltrating immune cells. Cancer Res. \u003cb\u003e77\u003c/b\u003e, e108\u0026ndash;e110 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMayakonda, A., Lin, D. C., Assenov, Y., Plass, C. \u0026amp; Koeffler, H. P. Maftools: efficient and comprehensive analysis of somatic variants in cancer. Genome Res. \u003cb\u003e28\u003c/b\u003e, 1747\u0026ndash;1756 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHoshida, Y., Brunet, J. P., Tamayo, P., Golub, T. R. \u0026amp; Mesirov, J. P. Subclass mapping: identifying common subtypes in independent disease data sets. PLoS One \u003cb\u003e2\u003c/b\u003e, e1195 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu, X. \u003cem\u003eet al.\u003c/em\u003e Immune signature-based subtypes of cervical squamous cell carcinoma tightly associated with human papillomavirus type 16 expression, molecular features, and clinical outcome. Neoplasia \u003cb\u003e21\u003c/b\u003e, 591\u0026ndash;601 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiberzon, A. \u003cem\u003eet al.\u003c/em\u003e The molecular signatures database (MSigDB) hallmark gene set collection. Cell Syst. \u003cb\u003e1\u003c/b\u003e, 417\u0026ndash;425 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou, Y. \u003cem\u003eet al.\u003c/em\u003e Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat. Commun. \u003cb\u003e10\u003c/b\u003e, 1523 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanzelmann, S., Castelo, R. \u0026amp; Guinney, J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics \u003cb\u003e14\u003c/b\u003e, 7 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVasaikar, S. V., Straub, P., Wang, J. \u0026amp; Zhang, B. LinkedOmics: analyzing multi-omics data within and across 32 cancer types. Nucleic Acids Res. \u003cb\u003e46\u003c/b\u003e, D956\u0026ndash;D963 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSubramanian, A. \u003cem\u003eet al.\u003c/em\u003e Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. \u003cem\u003eProc. Natl. Acad. Sci. U. S. A.\u003c/em\u003e \u003cb\u003e102\u003c/b\u003e, 15545\u0026ndash;15550 (2005).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-1984416/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-1984416/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAccumulating single-cell studies suggest that SPP1\u0026thinsp;+\u0026thinsp;macrophages are key players in the tumor microenvironment. However, a systematic investigation of SPP1\u0026thinsp;+\u0026thinsp;macrophages in colorectal cancer (CRC) has not been conducted. A total of eight single-cell RNA-seq datasets and 16 bulk RNA-seq datasets were included in this study. On the basis of existing research, we propose the SPP1\u0026thinsp;+\u0026thinsp;macrophage model paradigm, which can explain the clinical features and functional changes of macrophages in CRC better than the M1/M2 polarization theory. We identified four macrophage subsets from CRC myeloid cells: FCN1\u0026thinsp;+\u0026thinsp;macrophages, C1QC\u0026thinsp;+\u0026thinsp;macrophages, SPP1\u0026thinsp;+\u0026thinsp;macrophages, and MKI67\u0026thinsp;+\u0026thinsp;macrophages. Inflammation, phagocytosis, malignancy, and proliferation were identified as the most prominent features related to each of the four macrophage subsets. Our results show that SPP1\u0026thinsp;+\u0026thinsp;macrophages can serve as markers of CRC occurrence, progression, metastasis and a poor prognosis and exhibit enhanced transcription of genes associated with angiogenesis, epithelial-mesenchymal transition, glycolysis, hypoxia, and immunosuppressive signatures. CellPhoneDB analysis further indicated that SPP1 may mediate crosstalk between SPP1\u0026thinsp;+\u0026thinsp;macrophages and other cells via the SPP1-CD44, SPP1-PTGER4 and SPP1-a4b1 complex axes. Additionally, our study suggests that SPP1\u0026thinsp;+\u0026thinsp;macrophages are associated with the benefit of immune checkpoint blockade (ICB) therapy and that immunotherapy targeting SPP1\u0026thinsp;+\u0026thinsp;macrophages is expected to improve the prognosis of CRC patients. Furthermore, anti-CSF1R treatment is more likely to preferentially deplete CSF1R-enriched C1QC\u0026thinsp;+\u0026thinsp;macrophages versus SPP1\u0026thinsp;+\u0026thinsp;macrophages, which may account for the minimal effect of this monotherapy.\u003c/p\u003e","manuscriptTitle":"SPP1+ macrophages: A malignant macrophage subset in the colorectal cancer microenvironment revealed by single-cell sequencing","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2022-09-07 16:30:28","doi":"10.21203/rs.3.rs-1984416/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1bb95819-806b-4cd3-b769-6f406ee02893","owner":[],"postedDate":"September 7th, 2022","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2022-10-26T05:29:11+00:00","versionOfRecord":[],"versionCreatedAt":"2022-09-07 16:30:28","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-1984416","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-1984416","identity":"rs-1984416","version":["v1"]},"buildId":"J0_U0BvcaRcwD8yVFaRlm","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}