Impact of the canine osteosarcoma tumor microenvironment on immune cell composition and gene expression profiles

Impact of the canine osteosarcoma tumor microenvironment on immune cell composition and gene expression profiles | 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 Research Article Impact of the canine osteosarcoma tumor microenvironment on immune cell composition and gene expression profiles Dylan Ammons, Adam Harris, Lyndah Chow, Steven Dow This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5332445/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 11 Feb, 2025 Read the published version in Cancer Immunology, Immunotherapy → Version 1 posted 10 You are reading this latest preprint version Abstract Immune cells play key roles in host responses to malignant tumors. The selective pressure that immune cells elicit on tumors promotes immune escape, while tumor associated modulation of immune cells creates an environment favorable to tumor growth and progression. In this study we used publicly available single-cell RNA sequencing (scRNA-seq) data from the translationally relevant canine osteosarcoma (OS) model to compare tumor infiltrating leukocytes (TILs) to circulating leukocytes. Through computational analysis we investigated the differences in cell type proportions and how the OS TME impacted TIL transcriptomic profiles relative to circulating leukocytes. Differential abundance analysis revealed increased proportions of follicular helper T cells and mature regulatory dendritic cells (mregDCs) in the OS TME. Differential gene expression analysis identified exhaustion markers (LAG3, HAVCR1, PDCD1) to be upregulated in CD4 and CD8 T cells within the OS TME. Comparisons of B cell gene expression profiles revealed an enrichment of protein processing and endoplasmic reticulum pathways, suggesting infiltrating B cells were activated and participating in antigen presentation. Gene expression changes within myeloid cells identified increased expression of immune suppressive molecules (CD274, OSM, MSR1) in the OS TME, supporting their role as immunosuppressors. Comparisons to human literature revealed similar immune modulation in canine and human OS, further supporting the dog as a model for studies investigating novel immunotherapeutics. Overall, the analysis presented here provides new insights into how the OS TME impacts the transcriptional programs of major immune cell populations in dogs. Canine (dog) Osteosarcoma scRNA-seq transcriptomics cancer immunology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Introduction Systemic and local tumor immune responses have direct impacts on the clinical outcomes of cancer patients. For instance, overall survival and disease-free survival have been demonstrated to be positively corelated with effector T cell infiltration across multiple tumor types (1). The infiltration of immune cells into a tumor is dependent on the early recruitment of cells to the tumor, and adequate antitumor adaptive immunity has proven to be fundamental for successful immunotherapy-based intervention (2,3). Although certain tumor microenvironments (TME) recruit and maintain enough immune cells to benefit from immunotherapy, other tumors, such as osteosarcoma (OS), are notoriously defined by poor immune cell infiltration (4). The poor immune infiltration has contributed to the lack of response to immunotherapy in humans and dogs with naturally occurring OS (5). Thus, there is a need to understand how the TME impacts tumor infiltrating leukocytes (TILs) to facilitate the identification of therapeutic approaches to modulate the tumor immune microenvironment in OS. The lack of response to immunotherapy in OS has been attributed to an immune suppressive TME that prevents generation of sufficient antitumor immunity. Specifically, the OS TME has been described to have a notable immune suppressive gene expression pattern consisting of highly expressed programmed cell death protein 1 (PD-L1), indoleamine 2,3-dioxygenase (IDO) and interleukin-10 (IL-10) (6,7). Generally, these features of the TME promote the expansion of immune suppressive myeloid cells, exacerbate T cell dysfunction, and contribute to T cell exhaustion within the TME (6). Despite a basic understanding of underlying immune suppressive mechanisms active in OS across species, there is a need for a deeper characterization of the OS-associated changes to immune populations in dogs with naturally occurring OS. Recent technological advances in single-cell RNA sequencing (scRNA-seq) have enabled the dissection of complex tissues and have aided in the discovery of previously unrecognized mechanisms of immune cell modulation in the TME (8). The robustness of data integration approaches has also made it possible to compare gene expression profiles across tissue types (9–11). Thus, normal tissues, such as circulating leukocytes, can be used as a point of reference to investigate TME associated transcriptomic changes to TILs. Furthermore, canine OS is regarded as a valuable large animal model that enables the study of conserved TME components across species (12,13). As such this study aims to generate a canine-specific resource that documents TME associated transcriptomic changes in immune cells, while also drawing parallels to findings reported in human OS. In the present study we used two previously published scRNA-seq datasets consisting of TILs and circulating leukocytes from dogs with OS (14,15). The analysis presented in this manuscript revealed that effector CD8 T cell and CD4 T cells exhibited broadly upregulated expression of exhaustion markers (HAVCR1, LAG3, IL4I1) and identified increases in the relative proportions of follicular helper and regulatory T cells in the OS TME relative to circulating leukocytes. Consistent with the human literature, we demonstrated that mature regulatory dendritic cells (mregDCs) were found in the TME, but not in circulation (16). Furthermore, we confirmed that, in dogs, intratumoral neutrophils, monocytes, and dendritic cells upregulated immune modulatory molecules (OSM, PD-L1, and IDO1) relative to their circulating leukocyte counterparts. Overall, the analysis presented here provides key insights into how the TME shapes immune cell gene expression patterns and reveals mechanisms of immune modulation that are present within the canine OS TME. Materials and Methods Data acquisition, read mapping, and quantification Raw FASTQ data generated from Ficoll Paque prepared canine whole blood and treatment naïve OS tumor biopsies using the 10x Genomics Chromium platform were obtained from NCBI GEO (GSE225599 and GSE252470) (14,15). The raw data were aligned to the canine genome (CanFam3.1 Ensembl, release 104) and count matrices generated using a Cell Ranger analysis pipeline (version 6.1.2, 10x Genomics) as previously described (14). Annotated datasets from the primary reports were obtained from GSE225599 and from Zenodo (17). Data filtering and integration For each sample, the count matrix was imported into R using the Read10X function then converted to a Seurat object using the CreateSeuratObject function (18). Dead and poor-quality cells were filtered out by only retaining cells that met quality control thresholds. For the tumor dataset thresholds used were: 200 < nFeature_RNA < 5500, percent.mt < 12.5, and 100 < nCount_RNA < 15000. For the blood dataset thresholds used were: 200 < nFeature_RNA < 3500, percent.mt < 20, and 500 < nCount_RNA < 20000. [AH1] [DA2] After initial filtering, DoubletFinder was used to remove putative cell doublets (19). Samples within each tissue were integrated into separate objects using a SCTransform normalization protocol and canonical correlation analysis (CCA) integration workflow. During integration the percent mitochondrial reads was used as a latent variable in a linear regression framework to minimize the impact of mitochondrial reads on dimension reduction and integration. After each tissue type had all samples integrated into its respective object, previous cell type annotations were transferred and any cells lacking annotation were removed. To focus analysis on cell types that are found in circulation and in tumors we excluded non-immune cells from the tumor dataset. Furthermore, mast cells, osteoclasts, IFN T cells, and macrophages were also excluded from analysis due to a lack of a circulating counterpart. The blood dataset was filtered to exclude eosinophils, double negative T cells, γδ T cells, IFN signature CD4 T cells, basophils, and CD34+ unclassified cells. All filtered samples were then integrated into one object using the same approach applied to integrate individual tissues. The top 2500 variably expressed features were used as integration anchors then unsupervised clustering was completed. Ideal clustering parameters were identified using the R package clustree (20). Dimension reduction and visualization was then completed, and the data were presented using 2-dimensional, non-linear uniform manifold approximation and projection (UMAP) plots. Subcluster analysis For each of the major immune cell type populations, the dataset was subset to include only cells from one major population. The subset dataset was then used to identify new highly variable features, then the data were re-integrated and dimension reduction was repeated as described for the full dataset. Cell classification Cell annotations, as previously reported, were transferred to the integrated dataset using the unique cell barcodes associated with each cell (14,15). Unsupervised clustering was completed, then the composition of cell types (based on the transferred classifications) within each cluster was examined. For clusters in which one cell type predominated, [AH3] [DA4] the label was directly transferred. When conflicting cell types fell within a cluster, a new cell identity was assigned to capture the cell partitioning as determined through unsupervised clustering cells included in the current study. The gene signatures of each cluster identified in this manuscript, as determined using the FindAllMarkers function in the Seurat package (test.use = “wilcox [AH5] [DA6] ”, only.pos = TRUE), are provided as supplemental data. Feature visualization Feature expression was visualized using feature plots. Selected features were chosen based on the identification of a feature to be differentially expressed when contrasting tumor-infiltrating and blood leukocytes. Feature plots depict normalized expression for each feature and are presented on variable scales. When visualizing expression between tumor and blood leucocytes in a UMAP embedding, tissues were down sampled to obtain equal representation of each tissue. Cell abundance analysis All cell abundance comparisons were made using the percentage of total cells in the data subset being analyzed. To make statistical inferences on changes in cell abundance two-sided Wilcoxon Rank Sum tests were used. Differences in cell abundances were discussed as over-/under-represented or unique. Relative abundance differences were classified as over-/under-represented if P value < 0.05 and | log2(Fold change) | < 3 when comparing between the two tissues sources. The term “unique” was reserved for changes in which P value 3. The classification scheme was designed to identify unique cell types based on the idea that cell types with low to no representation will result in an exaggerated log2(Fold change), and in turn pass the high-end cutoff. Differential gene expression analysis Differential gene expression analysis (DE) was completed using pseudobulk conversion followed by a DESeq2 pipeline (21). Prior to running DESeq2, low abundance features, defined as features with less than 10 raw counts across all cells sampled, were filtered out. For analysis comparing gene expression between TILs and blood leukocytes, P values were determined by testing the null hypothesis that | log2(Fold change) | < 0.58. Features were then considered to be significantly differentially expressed if the adjusted (FDR) P value was less than 0.01. Any subsequent pathway analysis was completed using lists of upregulated genes and the clusterProfiler package (22). Gene ontology and Reactome gene sets were used (23), and terms were considered enriched if they achieved an adjusted (FDR) P value less than 0.05. Identification of tissue signatures and removal from analysis After completing DE analysis for each major cell population, we observed a bias for tumor-infiltrating immune cells to have increased expression of extracellular matrix associated features. Given previous reports documenting that the release of mRNA during sample processing and subsequent incorporation in cell droplet can result in confounding background tissue signatures, we devised a strategy to identify and remove features associated with background tissue signatures (9,24). To accomplish this, we completed differential expression analysis within each major immune cell population, then evaluated the features for consistent differential expression across all cell types. This revealed 46 features to be upregulated and four to be downregulated (TXNIP, PPBP, STK38, MITD1) across all tumor-infiltrating cell types (Supplemental table 1). For tissue signature estimation, we considered features with a P value less than 0.05 to be significant when testing the null hypothesis that | log2(Fold change) | < 0.58. We excluded the 46 tumor tissue-associated features, the 4 blood associated features, and a list of 108 platelet associated features, then repeated DE analysis as described above (Supplemental table 2) (14). Data and software availability A project specific GitHub page containing all analysis code and software versions used to analyze the data presented in this manuscript is available at https://github.com/dyammons/canine-blood-VS-tils-scrna. The annotated dataset for the integrated dataset and each subset is available for browsing at the UCSC Cell Browser (25). [AH1]Any particular reason these are different? Reviewers might also ask. [DA2]Different datasets, so warrant different settings, easy to defend. Some ppl will set thresholds differently for each sample [AH3]Does the cutoff come from somewhere? [DA4]Arbitrarily set [AH5]Is this the method you mentioned that yields the most false positives? [DA6]Yes, this is prone to false positives, but it does not matter much when it is just being used to identify cell type gene signatures - they are very easy to pick out. I don’t recommend using the test to compare between groups, but it is fine for cell type identification Results Overview of study Annotated data obtained from circulating leukocytes of 10 dogs diagnosed with osteosarcoma (OS) and a subset of OS tumor-infiltrating leukocytes (TILs) from 6 dogs were integrated into one dataset (14,15). We then established consensus annotations by evaluating the percentage of cell types in each cluster and transferred labels that best fit the clustering of the integrated dataset. In instances where none of the original annotations matched the unsupervised clustering results, we assigned new cell type identities to better match the integrated dataset. Our analysis approach consisted of completing subcluster analysis for each major cell population then evaluating changes in the relative cell type proportions and identification of differentially expressed genes (DEGs) between TILs and blood leukocytes ( Figure 1a/b ). When we completed differential gene expression (DE) analysis between TILs and blood leukocytes we identified background tissue signatures that were impacting DE analysis. These tissue signatures likely arose from ambient mRNA present during cell capture (26), so we identified tumor blood, and platelet gene signatures then excluded all the associated features from DE analysis. In total 46 tumor-associated features, 4 blood-associated features, and 108 platelet-associated features were excluded from DE analysis ( Supplemental figure 1a, Supplemental table 1-2 ). Gene set enrichment analysis (GSEA) of the 46 features revealed associations with extracellular matrix pathways, supporting the conclusion that these gene signatures likely originated from non-immune cells and represented background noise, which was subsequently filtered out ( Supplemental figure 1b/c ). Summary of the integrated canine circulating leucocyte and TILs dataset The datasets used in this study consisted of 10 blood leukocyte samples (n = 37,887 cells) from dogs with primary OS and TILs from 6 treatment-naïve dogs with primary OS of the axial skeleton (n = 11,257 cells). Although macrophages, mast cells, and osteoclasts are present in the OS TME, they were excluded from this analysis because a homologous cell type was not found in circulation. Therefore, analysis of the OS-associated impacts on immune cells was limited to six major cell types: neutrophils, monocytes, CD4 T cells, CD8 T cells, B cells, and dendritic cells (DCs), plus a minor cluster of cycling T cells ( Figure 1c/d ). Following label transfer of each cell type, we evaluated if any population was unique or under-/over-represented in the OS TME versus blood ( Figure 1e ). As expected, all the major cell types were found in both tissue sources. However, several cell type abundance changes were observed, including a greater proportion of CD4 T cells in blood and an overrepresentation of DCs in the tumor. Additionally, we found cycling T cells exhibited a marked increase in the relative proportion of cells within the tumor samples (4.52% ± 2.97) relative to blood samples (0.29% ± 0.25) ( Supplemental table 3 ). As such, cycling T cells were classified as unique to the tumor, despite being present at low levels in circulating leukocytes. We then further subset each major cell type and conducted a detailed comparison between TILs and blood leukocytes. Follicular helper and regulatory CD4 T cells are overrepresented in the tumor microenvironment After transferring cell type annotations to the CD4 T cell subset, we identified 6 distinct CD4 + T cell populations in the combined TILs-blood dataset. The cell types consisted of naïve, central memory (TCM), effector memory (TEM), Th1-like TEM, Th2-like TEM, and regulatory/follicular helper T cells (T reg /T fh ) ( Figure 2a, Supplemental data 1 ). The cluster annotated as T reg /T fh did not reach a consensus when transferring cell type labels and was annotated based on the presence of both regulatory and follicular helper T cells ( Supplemental figure 2a ). Of the six different CD4 T cell populations identified, naïve T cells were found to be more abundant in blood compared to the TME (32.13 ± 8.26% [blood] versus 13.54 ± 4.29% [TILs]). Similarly, Th1-like TEM were also more abundant in blood compared to the TME (6.42 ± 2.19% [blood] versus 2.25 ± 1.56% [TILs]). Conversely, Treg/Tfh cells were overrepresented in the tumor, making up 37.66 ± 5.76% in TME compared to 18.2 ± 3.72% in blood ( Figure 2b, Supplemental Table 4 ). To further investigate how the heterogeneity within the T reg /T fh cluster impacted differential abundance analysis, we used CXCL13 expression (a molecule essential for T fh mediated B cell recruitment) as a proxy for T fh cell identification to better understand T fh presence in the TME (27,28). Through evaluation of CXCL13 + cells (normalized count > 0) within the T reg /T fh cluster we found that CXCL13 + T fh were almost exclusively found in TILs (Blood: 2/14583; TILs: 121/2182 CXCL13 + cells) ( Supplemental figure 2b ). Together, this suggests that the overrepresentation of T reg /T fh CD4 T cells in the tumor, was in part due to the increased proportion of CXCL13 + T fh cells. Next, we completed DE analysis within CD4 T cells to compare expression profiles in cells isolated from tumor and blood. The analysis revealed 368 genes to be more highly expressed in CD4 TILs and 112 genes to be overexpressed in CD4 blood leukocytes ( Figure 2c, Supplemental data 2 ). Subsequent GSEA using the gene ontology (GO) database with the genes enriched in the CD4 TILs revealed enrichment for multiple terms associated with leukocyte activation and proliferation ( Figure 2d, Supplemental data 3 ). GSEA using the Reactome database suggested that CD4 TILs were active in interleukin signaling and interferon responses ( Figure 2e, Supplemental data 3 ). Select DEGs were then visualized to identify which clusters were driving differential expression ( Figure 2f, Supplemental figure 3 ). The expression patterns of DEGs revealed that the overabundance of SELL (CD62L) and LEF1 in blood leukocytes was associated with naïve T cells, while exhaustion (HAVCR1, PDCD1, LAG3) and activation (TNFRSF4, TNFRSF18) markers overexpressed in CD4 TILs were largely confined to TEM and T reg /T fh cell types (29–32). Overall, our analysis suggests that in addition to shifts in cell type proportions, tumor-infiltrating CD4 T cells also exhibit altered transcriptional profiles suggestive of activation and exhaustion relative to their circulating CD4 T cell counterparts. Features associated with T cell exhaustion are enriched in tumor-infiltrating effector CD8 T cells We next applied the same workflow to investigate tumor associated changes to CD8 T cells and NK cells. Unsupervised clustering of the integrated dataset resulted in identification of 5 transcriptomically distinct clusters which largely matched with the original cell type annotations ( Figure 3a, Supplemental data 4 ). One cell type, NK cells, could not be resolved as a distinct cluster in the integrated dataset. We found that previously annotated NK cells were interspersed within CD8 effector T cell clusters suggesting a substantial overlap in the gene signatures of CD8 T cells and NK cells ( Supplemental figure 4 ). Differential abundance analysis revealed that naïve CD8 T cells were the only subcluster to exhibit an overrepresentation in the blood (15.92 ± 9.17% [Blood] versus 5.5 ± 2.22% [TILs]) ( Figure 3b, Supplemental table 5 ). When comparing gene expression profiles between tumor-infiltrating and blood derived CD8 T cells, we identified 64 features to be more highly expressed in blood CD8 T cells and 241 more highly expressed in the CD8 TILs. ( Figure 3c, Supplemental data 2 ). Multiple T cell exhaustion markers including LAG3, TNFSF9, and HAVCR1 (TIM-1), were identified to be more abundantly expressed in CD8 TILs (29,30,32). Despite our efforts to filter out features associated with background tissue gene signatures, GSEA indicated that NK/CD8 TIL gene expression profiles were associated with extracellular matrix processes ( Figure 3d, Supplemental data 3 ). This enrichment pattern could be an artifact driven by tissue signatures or it is possible that the analysis revealed biologically relevant changes in CD8 T cells as they transition from circulating to tissue infiltrating T cells. Further analysis of GSEA revealed TILs demonstrated an association with T cell activation and recruitment of mononuclear cells. Reactome pathway analysis identified multiple terms associated with NFkB signaling to be enriched in tumor infiltrating CD8 T cells, which may suggest increased T cell activation ( Figure 3e, Supplemental data 3 ) (33). Visualization of DEGs revealed the increased expression of LEF1 in blood leukocytes was associated with naïve CD8 T cells, while the increased CX3CR1/PTGDR expression was associated with effector CD8 T cells ( Figure 3f, Supplemental figure 5 ). Tumor-infiltrating effector CD8 T cells were identified as drivers of LAG3, TNFSF9 (4-1BB ligand) and HAVCR1 (TIM-1) overexpression, which suggests effector CD8 T cells are activated and exhausted relative to circulating leukocytes (34). Consistent with studies in humans (9), our analysis revealed canine CD8 TILs to be activated and enriched in immune exhaustion markers relative to circulating CD8 T cells. Tumor-infiltrating B cells upregulate FOS and have gene expression patterns suggestive of protein processing aberrations Integration of circulating and tumor-infiltrating B cells revealed the presence of 3 B cell subtypes (immature, naïve, and class switched) and a cluster of plasma cells ( Figure 4a, Supplemental data 5 ). Differential abundance analysis indicated an increase in the relative proportion of immature B cells in blood (6.59 ± 4.44% [Blood] versus 1.09 ± 1.27% [TILs]) and an increase in the relative proportion of plasma cells in TILs (11.43 ± 4.93% [Blood] versus 22.66 ± 5.21% [TILs]) ( Figure 4b, Supplemental table 6 ). Through evaluation of the UMAP embedding and Euclidean distance of each cluster, we found that plasma cells were distantly related to the B cell subtypes, suggesting plasma cells should be treated as a distinct cell type ( Supplemental figure 6a ). As such, we completed DE analysis within plasma cells separately from the B cell subsets. DE analysis within the plasma cells revealed relatively few differentially expressed genes (27 features over expressed in tumor infiltrating plasma cells and 15 features over expressed on circulating plasma cells), implying only subtle tumor associated changes in gene expression ( Supplemental figure 6b ). Completion of DE analysis within in B cell subsets (c0, c1, and c3) between tissue sources revealed 44 features to be more highly expressed in circulating B cells and 222 features to be more highly expressed in B cell TILs ( Figure 4c, Supplemental data 2 ). Top DGEs included the proto-oncogenes, FOS and FOSB, which were identified as two of the most upregulated features in tumor-infiltrating B cells. The expression of FOS family gene members in B cells has been associated with terminal differentiation following interaction with a cognate antigen (35), but has also been assoicated with activation of apoptotic pathways in B cells (36,37). To investigate gene expression patterns in B cells further, we utilized GSEA which revealed enrichment of several terms associated with endoplasmic reticulum-associated degradation (ERAD), protein modification, endoplasmic reticulum activity, and antigen presentation in tumor infiltrating B cells ( Figure 4d/e, Supplemental data 3 ). The enrichment of ERAD and endoplasmic reticulum stress pathways further suggested that B cells within tumor tissues may be undergoing apoptosis through ER stress-induced cell death (38). It is also possible that these pathways were enriched due to increased antigen processing and presentation, as terms associated with antigen presentation indicates were also enriched in tumor infiltrating B cells (39). Further investigation is needed to determine the functional implication of the transcriptomic changes observed within the tumor infiltrating B cells. Localization of DEGs, indicated that IGHM expression was broadly reduced in tumor infiltrating B cells, suggesting the tumor infiltrating B cells are differentiating away from a naïve gene signature ( Figure 4f, Supplemental figure 6a ) (40). Overall, these findings indicate that tumor infiltrating B cells may be playing an active role in shaping T cell mediated immunity though antigen cross presentation, or potentially undergoing apoptosis within the TME. Mature regulatory dendritic cells (mregDCs) are present in the OS tumor microenvironment, but not in circulation. All DC annotations reached a consensus which enabled the direct transfer of cell type labels across datasets ( Figure 5a, Supplemental data 6 ). Of the five DC subtypes identified, we identified plasmacytoid DCs (pDCs) and precursor (pre) DCs to be overrepresented in blood relative to the TME (pDC: 17.96 ± 8.67% [Blood] versus 6.03 ± 3.74% [TILs]; preDC: 15.65 ± 2.85% [Blood] versus 4.54 ± 3.67% [TILs]) ( Figure 5b, Supplemental table 7 ). In contrast, mature regulatory DCs (mregDCs), a recently described immune modulatory population, were the only DC subpopulation to be classified as unique to TILs (Blood: 0.11% ± 0.26; TILs: 11.31% ± 6.69). Lastly, conventional DC1s (cDC1) and cDC2s were determined to have unaltered abundances when comparing blood and tumor leukocytes. ( Figure 5b ). The identification of mregDCs as unique to the TME has been previously documented in human cancers (16,41), with previous reports suggesting mregDC are derived from circulating DC populations following infiltration into the tumor. Although we did not investigate how mregDCs accumulated in the canine OS TME, the marked overrepresentation in suggests a potentially conserved mechanism in mregDC biology between the two species. To investigate how the TME impacted DC gene expression, we completed DE analysis and subsequent GSEA. The analysis revealed 257 features enriched in tumor DCs and 26 enriched in blood DCs ( Figure 5c, Supplemental data 2 ). Gene ontology analysis revealed associations with tumor necrosis factor (TNF) responses and vascular development within tumor infiltrating DC ( Figure 5d, Supplemental data 3 ). GSEA using Reactome terms identified increased interleukin activity of tumor infiltrating DCs with IL4, IL13, and IL10 predicted to elicit the greatest impact on tumor DCs ( Figure 5e, Supplemental data 3 ). The relative expansion of mregDCs impacted DE analysis (as evidenced by CCR7, IDO1, IL4I1, and CD274 upregulation in tumor infiltrating DCs), so we further investigated the localization of DEGs. We identified IL16 and IRF4 to be broadly down regulated in tumor infiltrating DCs, while CXCR4 was broadly upregulated ( Figure 5f, Supplemental figure 6b ). Interestingly, VEGFA, a feature determined to be associated with the tumor tissue signature, was selectively upregulated in cDC2s, suggesting the differential expression of VEGFA may be a biologically relevant change rather than an artifact of tissue bias. In summary, we provide evidence that canine mregDCs are enriched in the TME and that infiltrating DCs may modulate interleukin signaling within the OS TME. Tumor-infiltrating monocytes upregulate chemokine and immunoregulatory molecule expression relative to circulating monocytes Tumor-infiltrating and circulating monocytes integrated uniformly and no changes to the relative proportions of the four monocyte subtypes were identified ( Figure 6a/b, Supplemental data 7, Supplemental table 8 ). Despite uniform integration, DE analysis revealed marked transcriptomic changes induced by tumor infiltration with 356 features enriched in tumor infiltrating monocytes (TIMs) and 69 enriched in blood monocytes ( Figure 6c, Supplemental data 2 ). Of the features enriched in TIMs, several macrophage-associated features (MSR1, CTSK, APOC1) were identified. This suggests that although TIMs were able to be distinguished from tumor associated macrophage (TAMs) when analyzed together in the original publication (15,42), our integrated analysis found TIMs to have a macrophage-like gene profile relative to circulating monocytes. To further investigate the transcriptomic similarities between TAMs, TIMs, and circulating monocytes, we completed hierarchical clustering of the three cell types and found TIMs and circulating monocytes clustered together, while TAMs were on a distinct clade ( Supplemental figure 7 ). Overall, our analysis suggests that TIMs are in an intermediate state, with greater transcriptomic similarities to circulating monocytes than TAMs. Pathway analysis of the features identified to be overexpressed in TIMs relative to blood monocytes revealed TIMs to be enriched in gene ontology terms associated with general immune activation, adhesion, and interleukin signaling. ( Figure 6d/e, Supplemental data 3 ). Visualization of DEGs, demonstrated IL16 and FGL2 (Fibrinogen-like protein 2) to be broadly expressed features whereas LTF (lactotransferin) expression was primarily localized to myeloid-derived suppressor cells (M-MDSCs) ( Figure 6f, Supplemental figure 8 ). Further investigation of DEGs identified TIMs to exhibit increased expression of multiple chemokines (CXCL10, CXCL16, CCL19, CCL5, CCL7, CCL8) ( Supplemental figure 8d ). Relative to TIMs, circulating monocytes exhibited higher levels of CCR2 expression which may interact with CCL7 and CCL8 secreted by TIMs to promote further infiltration (43). Relative to blood monocytes, immune modulatory molecules (IL1A, OSM, CD274, and PTGES) were determined to be broadly upregulated in TIMs. Lastly, the overexpression of C1QC in TIMs relative to blood monocytes may represent that cells are transitioning to macrophage. It is also possible that some true TAMs were unintentionally included in the analysis. Overall, we found marked gene expression changes associated with differentiation toward macrophage and upregulation of immune modulatory molecules in TIMs relative to blood monocytes. Tumor-associated neutrophils increase Oncostatin M and chemokine expression relative to circulating neutrophils Evaluation of tumor-associated neutrophils (TANs) and blood neutrophils revealed no changes in the relative proportions of neutrophils (c0) and PMN-MDSCs (c1); indicating the ratio of PMN-MDSCs to neutrophils is consistent in the blood and tumor of dogs with primary OS ( Figure 7a/b, Supplemental data 8, Supplemental table 9 ). DE analysis identified upregulation of 139 features in TANs, with only 4 features preferentially expressed in the blood ( Figure 7c, Supplemental data 2 ). Subsequent pathway analysis suggested TAN transcriptomic signatures were associated with general neutrophil activation, responses to interleukins, and response to endoplasmic reticulum stress ( Figure 7d/e, Supplemental data 3 ). Together the analysis indicated TANs exhibited a shift toward an activated state with increased enrichment of gene programs associated with cell migration and interleukin signaling. Oncostatin M (OSM), a member of the IL-6 family with reported immune suppressive properties, was found to be broadly upregulated in TANs, relative to blood neutrophils ( Figure 7f, Supplemental figure 9 ) (44). TANs also overexpressed multiple chemokines (CCL5, CCL7, and CXCL8) relative to circulating neutrophils, which suggests that upon infiltration, TAN secrete chemokines to further promote myeloid cell infiltration ( Supplemental figure 9c ). Plasminogen activator urokinase genes, PLAU and PLAUR, were also enriched in TANs. Both PLAU and PLAUR have been associated with neutrophil infiltration and their expression levels carry negative prognostic value in multiple cancer types, suggesting these genes may have prognostic value in canine OS (45,46). IL1R2, a decoy receptor for IL1A/B, was broadly upregulated in TANs which suggests they may function to dampen inflammatory responses (47). Lastly, CD274 (PD-L1) was enriched in TANs relative to blood neutrophils which implicates TANs as potential suppressors of adaptive T cell responses (48). Ultimately, we found that TANs broadly upregulated immune suppressive molecules relative to blood neutrophils, suggesting that TANs may play a role in maintenance of the immune suppressive TME. Discussion Immune suppression and evasion are hallmarks of cancer and understanding the mechanisms in which immune cells are impacted by the TME is foundational for the development of more effective immunotherapies. In the context of OS, a tumor type that is almost uniformly unresponsive to immunotherapy in both humans and dogs, it can be valuable to identify which cell types drive immune suppression to then target them therapeutically (5). In the present study we investigated how the canine OS TME modulates the transcriptomic signatures and relative cell type abundance of TILs. Using circulating immune cells as a point of reference, we identified upregulated expression of exhaustion markers on tumor-infiltrating T cells and found that mregDCs were present in the TME, but not in circulation. Our analysis indicated that tumor-infiltrating monocytes (TIMs) exhibited a macrophage signature relative to blood monocytes, but still retained a monocytic signature relative to tumor associated macrophages, suggesting the TIMs were likely isolated from the tumor microvasculature. Transcriptionally, we identified upregulation of well-known immune regulatory features including PD-L1 (CD274), CD80, and CD86 within tumor infiltrating myeloid cells. Furthermore, Oncostatin M (OSM) was found to be more highly expressed on TIMs and TANs relative to their circulating counterparts, implicating these myeloid cell populations as the primary producers of this immune regulatory cytokine (44). The analysis presented here suggests that many features of tumor-induced immune suppression and exhaustion that have been reported across species are active in canine OS which supports the use of spontaneously occurring canine OS to conduct immunotherapy studies. Through our analysis we identified upregulation of immune suppressive transcripts in tumor infiltrating monocytes, neutrophils, and dendritic cells. Immune suppressive roles of myeloid cells within tumors have been described in a large body of literature, and our analysis provides unique insights into specific features of immune suppression in canine OS (49–51). In particular, our analysis revealed increased expression of various immune modulatory molecules, including CD274 (PD-L1), OSM, CD36, and MSR1, by tumor infiltrating monocytes and neutrophils suggesting an immune suppressive impact on the TME. Additionally, we identified IL1R2, a decoy receptor for IL1A/B, to be upregulated in TANs which suggests immunological checkpoint blockage with IL1R2 antagonists may be of therapeutic interest in canine OS (52,53). Consistent with scRNA-seq data in T cells across multiple tumor types in humans (9,10), we identified increased relative proportions of regulatory and exhausted T cells within the canine OS TME. Our analysis further identified that the relative abundances of naïve and Th1-like TEM T cells were reduced in tumor infiltrating T cells relative to circulating T cell populations. Outside of cell abundance shifts we identified a broad increase in the expression of features associated with exhaustion (TOX, CTLA4, LAG3, TNFRSF9) across CD8 effector T cells and most non-naïve CD4 T cell populations (29–32). As most of these molecules can also be upregulated in activated T cells, and not only exhausted T cells, our analysis cannot fully distinguish between activated and exhausted T cells (54,55). Further experimental investigation of the T cell populations is needed to determine the functional status of the cells described in this study. There is a growing body of literature suggesting intratumoral B cells play a role in antitumoral responses through antigen presentation and by participating in tertiary lymphoid structure (TLS) formation (56,57). Our analysis revealed transcriptomic evidence of profound aberrations to tumor infiltrating B cell protein processing machinery (protein modification, endoplasmic reticulum activity), which could be indicative of B cell dysfunction, antibody production, or antigen presentation. Given that mregDCs, follicular helper T cells, and B cells were all identified in the TIL population, it is possible that these cell types could have been interacting within TLS of the tumor (58–61). Further investigation using spatial transcriptomics is required to determine if mregDCs, B cells, and T fh co-localize within the canine OS TME. Overall, our analysis suggests that B cells are modulated by the canine OS TME and may play a role in shaping adaptive immunity. Although our findings provide insight into how TILs differ from blood leukocytes in dogs, it is possible that some of the differences in gene expression may be due to the reference population being in circulation rather than cells present in normal bone. Because normal bone has minimal leukocytes present, the use of circulating leukocytes represents a best current avenue to explore how infiltrating immune cells are modulated by the OS TME. To minimize the impact of background tissue signatures from ambient mRNA released during sample processing, we identified a tumor tissue signature and excluded those features from DE analysis. In doing so we attempted to filter out the background signal, but we cannot exclude the possibility that some of the features identified may also represent tumor-induced gene expression changes observed across all cell types. The 46 excluded features were associated with extracellular matrix GSEA terms suggesting the genes are implicated in a tissue signature, not immune-associated terms. Future work incorporating lymph nodes or other non-circulating immune cells may help to more effectively distinguish between tissue- and tumor-associated changes. Lastly, the primary aim of this paper was to focus on canine cancer immunology and did not attempt to provide a comprehensive human to canine comparison of TME-associated transcriptomic changes. Future studies could address this question by incorporating scRNA-seq data from human OS tissues and peripheral blood to comprehensively profile the similarities and differences between the two species. The data presented here describe the transcriptomic responses that canine TILs exhibit relative to circulating leukocytes (8–10). Through our analysis we identified dysregulation of immune modulatory features with marked changes across all cells investigated in this study, including B cells which historically have not been studied extensively in OS. From a comparative immune-oncology standpoint, the findings enable connections to be made human literature and provide insights into immune responses within spontaneously occurring canine OS. Overall, our analysis sheds light on the immune suppression and dysfunction that is present in TILs within the canine OS TME. Declarations Author Contribution Conception and design: DTA, RAH, and SD. Data acquisition: DTA. Data analysis: DTA, RAH, LC, and SD. First draft of manuscript: DTA. Manuscript revisions: DTA, RAH, LC, and SD. Final approval of completed manuscript: All authors. Acknowledgement This project was supported by grants from the National Institutes of Health (NIH): U01 CA224182 (to S.D.) and the Shipley Family Foundation (to S.D.). This work utilized the Alpine high performance computing resource at the University of Colorado Boulder. Alpine is jointly funded by the University of Colorado Boulder, the University of Colorado Anschutz, and Colorado State University. Data storage was supported by the University of Colorado Boulder ‘PetaLibrary. Data Availability A project specific GitHub page containing all analysis code and software versions used to analyze the data presented in this manuscript is available at https://github.com/dyammons/canine-blood-VS-tils-scrna. The annotated dataset for the integrated dataset and each subset is available for browsing at the UCSC Cell Browser. References Jochems C, Schlom J. Tumor-infiltrating immune cells and prognosis: the potential link between conventional cancer therapy and immunity. Exp Biol Med. 2011;236(5):567–79. Liu R, Yang F, Yin JY, Liu YZ, Zhang W, Zhou HH. Influence of tumor immune infiltration on immune checkpoint inhibitor therapeutic efficacy: A computational retrospective study. Front Immunol. 2021;12:685370. Yin Y, Sakakibara R, Honda T, Kirimura S, Daroonpan P, Kobayashi M, et al. High density and proximity of CD8+ T cells to tumor cells are correlated with better response to nivolumab treatment in metastatic pleural mesothelioma. Thorac Cancer. 2023;14(20):1991–2000. Fritzsching B, Fellenberg J, Moskovszky L, Sápi Z, Krenacs T, Machado I, et al. CD8+/FOXP3+-ratio in osteosarcoma microenvironment separates survivors from non-survivors: a multicenter validated retrospective study. Oncoimmunology. 2015;4(3):e990800. Wu CC, Beird HC, Andrew Livingston J, Advani S, Mitra A, Cao S, et al. Immuno-genomic landscape of osteosarcoma. Nat Commun. 2020;11(1):1008. Tian H, Cao J, Li B, Nice EC, Mao H, Zhang Y, et al. Managing the immune microenvironment of osteosarcoma: the outlook for osteosarcoma treatment. Bone Res. 2023;11(1):11. Mannheimer JD, Tawa G, Gerhold D, Braisted J, Sayers CM, McEachron TA, et al. Transcriptional profiling of canine osteosarcoma identifies prognostic gene expression signatures with translational value for humans. Commun Biol. 2023;6(1):856. Maier B, Leader AM, Chen ST, Tung N, Chang C, LeBerichel J, et al. A conserved dendritic-cell regulatory program limits antitumour immunity. Nature. 2020;580(7802):257–62. Szabo PA, Levitin HM, Miron M, Snyder ME, Senda T, Yuan J, et al. Single-cell transcriptomics of human T cells reveals tissue and activation signatures in health and disease. Nat Commun. 2019;10(1):4706. Zheng L, Qin S, Si W, Wang A, Xing B, Gao R, et al. Pan-cancer single-cell landscape of tumor-infiltrating T cells. Science (1979). 2021;374(6574):abe6474. Mulder K, Patel AA, Kong WT, Piot C, Halitzki E, Dunsmore G, et al. Cross-tissue single-cell landscape of human monocytes and macrophages in health and disease. Immunity. 2021;54(8):1883–900. Gustafson DL, Duval DL, Regan DP, Thamm DH. Canine sarcomas as a surrogate for the human disease. Pharmacol Ther. 2018;188:80–96. Fenger JM, London CA, Kisseberth WC. Canine osteosarcoma: a naturally occurring disease to inform pediatric oncology. ILAR J. 2014;55(1):69–85. Ammons DT, Harris RA, Hopkins LS, Kurihara J, Weishaar K, Dow S. A single-cell RNA sequencing atlas of circulating leukocytes from healthy and osteosarcoma affected dogs. Front Immunol. 2023;14:1162700. Ammons DT, Hopkins LS, Cronise KE, Kurihara J, Regan DP, Dow S. Single-cell RNA sequencing reveals the cellular and molecular heterogeneity of treatment-naïve primary osteosarcoma in dogs. Commun Biol. 2024;7(1):496. Liu W, Hu H, Shao Z, Lv X, Zhang Z, Deng X, et al. Characterizing the tumor microenvironment at the single-cell level reveals a novel immune evasion mechanism in osteosarcoma. Bone Res. 2023;11(1):4. Ammons D. Canine osteosarcoma single-cell RNA sequencing reference dataset: analysis code and processed data for publication. 2024. Hao Y, Hao S, Andersen-Nissen E, Mauck WM, Zheng S, Butler A, et al. Integrated analysis of multimodal single-cell data. Cell. 2021;184(13):3573–87. McGinnis CS, Murrow LM, Gartner ZJ. DoubletFinder: doublet detection in single-cell RNA sequencing data using artificial nearest neighbors. Cell Syst. 2019;8(4):329–37. Zappia L, Oshlack A. Clustering trees: a visualization for evaluating clusterings at multiple resolutions. Gigascience. 2018;7(7):giy083. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15:1–21. Wu T, Hu E, Xu S, Chen M, Guo P, Dai Z, et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. The innovation. 2021;2(3). Liberzon A, Subramanian A, Pinchback R, Thorvaldsdóttir H, Tamayo P, Mesirov JP. Molecular signatures database (MSigDB) 3.0. Bioinformatics. 2011;27(12):1739–40. Young MD, Behjati S. SoupX removes ambient RNA contamination from droplet-based single-cell RNA sequencing data. Gigascience. 2020;9(12):giaa151. Speir ML, Bhaduri A, Markov NS, Moreno P, Nowakowski TJ, Papatheodorou I, et al. UCSC Cell Browser: visualize your single-cell data. Bioinformatics. 2021;37(23):4578–80. Young MD, Behjati S. SoupX removes ambient RNA contamination from droplet-based single-cell RNA sequencing data. Gigascience. 2020;9(12):giaa151. Lin X, Ye L, Wang X, Liao Z, Dong J, Yang Y, et al. Follicular helper T cells remodel the immune microenvironment of pancreatic cancer via secreting CXCL13 and IL-21. Cancers (Basel). 2021;13(15):3678. Moser B. CXCR5, the defining marker for follicular B helper T (TFH) cells. Front Immunol. 2015;6:296. Wolf Y, Anderson AC, Kuchroo VK. TIM3 comes of age as an inhibitory receptor. Nat Rev Immunol. 2020;20(3):173–85. Yang ZZ, Kim HJ, Villasboas JC, Chen YP, Price-Troska T, Jalali S, et al. Expression of LAG-3 defines exhaustion of intratumoral PD-1+ T cells and correlates with poor outcome in follicular lymphoma. Oncotarget. 2017;8(37):61425. Ward-Kavanagh LK, Lin WW, Šedý JR, Ware CF. The TNF receptor superfamily in co-stimulating and co-inhibitory responses. Immunity. 2016;44(5):1005–19. Li Y, Wang Z, Jiang W, Zeng H, Liu Z, Lin Z, et al. Tumor-infiltrating TNFRSF9+ CD8+ T cells define different subsets of clear cell renal cell carcinoma with prognosis and immunotherapeutic response. Oncoimmunology. 2020;9(1):1838141. Barnes SE, Wang Y, Chen L, Molinero LL, Gajewski TF, Evaristo C, et al. T cell-NF-κB activation is required for tumor control in vivo. J Immunother Cancer. 2015;3(1):1–11. Anderson AC, Joller N, Kuchroo VK. Lag-3, Tim-3, and TIGIT: co-inhibitory receptors with specialized functions in immune regulation. Immunity. 2016;44(5):989–1004. Ohkubo Y, Arima M, Arguni E, Okada S, Yamashita K, Asari S, et al. A role for c-fos/activator protein 1 in B lymphocyte terminal differentiation. The Journal of Immunology. 2005;174(12):7703–10. Inada K, Okada S, Phuchareon J, Hatano M, Sugimoto T, Moriya H, et al. c-Fos induces apoptosis in germinal center B cells. The Journal of Immunology. 1998;161(8):3853–61. Liebermann DA, Gregory B, Hoffman B. AP-1 (Fos/Jun) transcription factors in hematopoietic differentiation and apoptosis. Int J Oncol. 1998;12(3):685–1385. Sano R, Reed JC. ER stress-induced cell death mechanisms. Biochimica et Biophysica Acta (BBA)-Molecular Cell Research. 2013;1833(12):3460–70. Rastogi I, Jeon D, Moseman JE, Muralidhar A, Potluri HK, McNeel DG. Role of B cells as antigen presenting cells. Front Immunol. 2022;13:954936. Morgan D, Tergaonkar V. Unraveling B cell trajectories at single cell resolution. Trends Immunol. 2022;43(3):210–29. Gordon JR, Ma Y, Churchman L, Gordon SA, Dawicki W. Regulatory dendritic cells for immunotherapy in immunologic diseases. Front Immunol. 2014;5:7. Ma RY, Black A, Qian BZ. Macrophage diversity in cancer revisited in the era of single-cell omics. Trends Immunol. 2022;43(7):546–63. Hao Q, Vadgama J V, Wang P. CCL2/CCR2 signaling in cancer pathogenesis. Cell Communication and Signaling. 2020;18:1–13. Wallace PM, MacMaster JF, Rouleau KA, Brown TJ, Loy JK, Donaldson KL, et al. Regulation of inflammatory responses by oncostatin M. The Journal of Immunology. 1999;162(9):5547–55. Shi K, Zhou J, Li M, Yan W, Zhang J, Zhang X, et al. Pan-cancer analysis of PLAU indicates its potential prognostic value and correlation with neutrophil infiltration in BLCA. Biochimica et Biophysica Acta (BBA)-Molecular Basis of Disease. 2024;1870(2):166965. Zeng F, Li G, Liu X, Zhang K, Huang H, Jiang T, et al. Plasminogen activator urokinase receptor implies immunosuppressive features and acts as an unfavorable prognostic biomarker in glioma. Oncologist. 2021;26(8):e1460–9. Dinarello CA. Overview of the IL‐1 family in innate inflammation and acquired immunity. Immunol Rev. 2018;281(1):8–27. Yajuk O, Baron M, Toker S, Zelter T, Fainsod-Levi T, Granot Z. The PD-L1/PD-1 axis blocks neutrophil cytotoxicity in cancer. Cells. 2021;10(6):1510. Su P, Wang Q, Bi E, Ma X, Liu L, Yang M, et al. Enhanced lipid accumulation and metabolism are required for the differentiation and activation of tumor-associated macrophages. Cancer Res. 2020;80(7):1438–50. Gulay KCM, Aoshima K, Maekawa N, Suzuki T, Konnai S, Kobayashi A, et al. Hemangiosarcoma cells induce M2 polarization and PD-L1 expression in macrophages. Sci Rep. 2022;12(1):2124. Christofides A, Strauss L, Yeo A, Cao C, Charest A, Boussiotis VA. The complex role of tumor-infiltrating macrophages. Nat Immunol. 2022;23(8):1148–56. Xia J, Zhang L, Peng X, Tu J, Li S, He X, et al. IL1R2 blockade alleviates immunosuppression and potentiates anti-PD-1 efficacy in triple-negative breast cancer. Cancer Res. 2024; Wu TC, Xu K, Martinek J, Young RR, Banchereau R, George J, et al. IL1 receptor antagonist controls transcriptional signature of inflammation in patients with metastatic breast cancer. Cancer Res. 2018;78(18):5243–58. Gholami MD, Saeedi Y, Heydari S, Garssen J, Falak R. Exhaustion of T lymphocytes in the tumor microenvironment: Significance and effective mechanisms. Cell Immunol. 2017;322:1–14. Rha MS, Shin EC. Activation or exhaustion of CD8+ T cells in patients with COVID-19. Cell Mol Immunol. 2021;18(10):2325–33. Kinker GS, Vitiello GAF, Ferreira WAS, Chaves AS, Cordeiro de Lima VC, Medina T da S. B cell orchestration of anti-tumor immune responses: a matter of cell localization and communication. Front Cell Dev Biol. 2021;9:678127. Laumont CM, Banville AC, Gilardi M, Hollern DP, Nelson BH. Tumour-infiltrating B cells: immunological mechanisms, clinical impact and therapeutic opportunities. Nat Rev Cancer. 2022;22(7):414–30. Withers SS, York D, Choi JW, Woolard KD, Laufer‐Amorim R, Sparger EE, et al. Metastatic immune infiltrates correlate with those of the primary tumour in canine osteosarcoma. Vet Comp Oncol. 2019;17(3):242–52. Ukita M, Hamanishi J, Yoshitomi H, Yamanoi K, Takamatsu S, Ueda A, et al. CXCL13-producing CD4+ T cells accumulate in the early phase of tertiary lymphoid structures in ovarian cancer. JCI Insight. 2022;7(12). Chen L, Oke T, Siegel N, Cojocaru G, Tam AJ, Blosser RL, et al. The immunosuppressive niche of soft-tissue sarcomas is sustained by tumor-associated macrophages and characterized by intratumoral tertiary lymphoid structures. Clinical Cancer Research. 2020;26(15):4018–30. Yan P, Wang J, Yue B, Wang X. Unraveling molecular aberrations and pioneering therapeutic strategies in osteosarcoma. Biochimica et Biophysica Acta (BBA)-Reviews on Cancer. 2024;189171. Additional Declarations No competing interests reported. Supplementary Files 20241025supplementalfigures.pdf supplementaldata1.csv supplementaldata2.csv supplementaldata3.csv supplementaldata4.csv supplementaldata5.csv supplementaldata6.csv supplementaldata7.csv supplementaldata8.csv supplementaltables.xlsx Cite Share Download PDF Status: Published Journal Publication published 11 Feb, 2025 Read the published version in Cancer Immunology, Immunotherapy → Version 1 posted Editorial decision: Revision requested 21 Nov, 2024 Reviews received at journal 18 Nov, 2024 Reviews received at journal 04 Nov, 2024 Reviewers agreed at journal 30 Oct, 2024 Reviewers agreed at journal 30 Oct, 2024 Reviewers agreed at journal 28 Oct, 2024 Reviewers invited by journal 27 Oct, 2024 Editor assigned by journal 26 Oct, 2024 Submission checks completed at journal 26 Oct, 2024 First submitted to journal 25 Oct, 2024 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-5332445","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":373926484,"identity":"7477a3d6-6374-450d-bfc5-5b23ed314165","order_by":0,"name":"Dylan Ammons","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABEklEQVRIie2QP0vDQBiH33DgdOQcL0TqJxBOAkEx+EVcGgKdGnAMGGikEJe6x2/hlDnHQbMcdj1wUfwCgSyFgthToYI56ehwz/b+eXj5vQAWy3+kQcWu6ACOfoxMirNTnAoAA6Dv7X0UhPdRXBXf9tdweXVydy/6KI8wISvedxmMXDUeVDwVz/0KkrSWTxN/upxgr0oQbSQEnkFhKi58DCit1TREaSEwUwgoLyF+NCvzDYaZVoL+XCsrgdb8HWZ/KOX2itAK8x2tNMkB5QWMmSmLfC0vMGu3WWToLb6yhGdySU8f5Mvwx9pEPOPsJq3bRdCt82hECH9TWR4du+3wFTjUffa7T4fXNaQxzywWi8XyyQe7w2ZIybbZxgAAAABJRU5ErkJggg==","orcid":"","institution":"Flint Animal Cancer Center, Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University","correspondingAuthor":true,"prefix":"","firstName":"Dylan","middleName":"","lastName":"Ammons","suffix":""},{"id":373926485,"identity":"7b72c814-87bb-4540-bb5d-08f27af7fe60","order_by":1,"name":"Adam Harris","email":"","orcid":"","institution":"Department of Microbiology, Immunology and Pathology, College of Veterinary Medicine and Biomedical Sciences, Colorado State University","correspondingAuthor":false,"prefix":"","firstName":"Adam","middleName":"","lastName":"Harris","suffix":""},{"id":373926486,"identity":"3996fe02-6ad0-4502-8d68-ec6c9ac1485e","order_by":2,"name":"Lyndah Chow","email":"","orcid":"","institution":"Flint Animal Cancer Center, Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University","correspondingAuthor":false,"prefix":"","firstName":"Lyndah","middleName":"","lastName":"Chow","suffix":""},{"id":373926487,"identity":"dc489732-7744-494d-b06f-456bd898d644","order_by":3,"name":"Steven Dow","email":"","orcid":"","institution":"Flint Animal Cancer Center, Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University","correspondingAuthor":false,"prefix":"","firstName":"Steven","middleName":"","lastName":"Dow","suffix":""}],"badges":[],"createdAt":"2024-10-25 12:38:59","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5332445/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5332445/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s00262-025-03950-3","type":"published","date":"2025-02-11T15:57:50+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":68356738,"identity":"635b3c73-3c8c-4bef-ac54-41bc455298a1","added_by":"auto","created_at":"2024-11-06 11:28:05","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":3445484,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStudy overview and cell type annotations.\u003c/strong\u003e(a/b) Diagram of study design. Tumor infiltrating immune cell and circulating immune cell were integrated into one dataset then transcriptomic and cell type abundances were evaluated. (c) UMAP representation with major cell type annotations of circulating immune cells from 10 dogs with osteosarcoma (n = 42,292 cells) and TILs from 6 treatment-naïve OS dogs (n = 11,431 cells). (d) Feature plots depicting canonical cell type markers. (e) Boxplots depicting distribution of cell type percentages within blood and TILs. Significance determined using two-sided Wilcoxon rank sum test. *** = P value \u0026lt; 0.001; ** = P value \u0026lt; 0.01; * = P value \u0026lt; 0.05.\u003c/p\u003e","description":"","filename":"Slide1.png","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/19f32bf5dea5861406943af3.png"},{"id":68357017,"identity":"86c748f0-2294-459d-9e5e-5da377c7a111","added_by":"auto","created_at":"2024-11-06 11:36:05","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":7097087,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTumor infiltrating cells exhibit higher relative proportions of exhausted CD4 T cell relative to circulating CD4 T cells.\u003c/strong\u003e(a) UMAP representation of 16,765 CD4 T cells colorized by cell type annotation. (b) Boxplots depicting distribution of cell type percentages within blood and TILs. Significance determined using two-sided Wilcoxon rank sum test. *** = P value \u0026lt; 0.001; ** = P value \u0026lt; 0.01; * = P value \u0026lt; 0.05. (c) Volcano plot depicting results of differential gene expression analysis contrasting all tumor infiltrating CD4 T cells versus all circulating CD4 T cells. Lollipop charts depicting log10 transformed adjusted P values from gene set enrichment analysis using genes upregulated in tumor infiltrating CD4 T cells against gene ontology (d) and Reactome (e) gene sets. (f) Feature plots split by tissue source depicting normalized expression of differentially expressed features.\u003c/p\u003e","description":"","filename":"Slide2.png","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/7edaa8537c2c10fc3e0326f6.png"},{"id":68356748,"identity":"56867611-1dd2-4a55-a83e-12557330dba8","added_by":"auto","created_at":"2024-11-06 11:28:05","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":6861155,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTumor infiltrating CD8 T cells express high levels of exhaustion markers.\u003c/strong\u003e (a) UMAP representation of 10,962 CD8 T cells colorized by cell type annotation. (b) Boxplots depicting distribution of cell type percentages within blood and TILs. Significance determined using two-sided Wilcoxon rank sum test. *** = P value \u0026lt; 0.001; ** = P value \u0026lt; 0.01; * = P value \u0026lt; 0.05. (c) Volcano plot depicting results of differential gene expression analysis contrasting all tumor infiltrating CD8 T cells versus all circulating CD8 T cells. Lollipop charts depicting P value from gene set enrichment analysis using genes upregulated in tumor infiltrating CD8 T cells against gene ontology (d) and Reactome (e) gene sets. (f) Feature plots split by tissue source depicting normalized expression of differentially expressed features.\u003c/p\u003e","description":"","filename":"Slide3.png","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/edfd75480cd8e8607a59ef56.png"},{"id":68356743,"identity":"5d908f43-37f2-4d16-baef-0e14d906cf3d","added_by":"auto","created_at":"2024-11-06 11:28:05","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":3208949,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTumor infiltrating B cells exhibit gene expression patterns suggestive of antigen presentation and maturation.\u003c/strong\u003e(a) UMAP representation of unsupervised clustering of 3,108 B cells. (b) Boxplots depicting distribution of cell type percentages within blood and TILs. Significance determined using two-sided Wilcoxon rank sum test. *** = P value \u0026lt; 0.001; ** = P value \u0026lt; 0.01; * = P value \u0026lt; 0.05. (c) Volcano plot depicting results of differential gene expression analysis contrasting all tumor infiltrating B cells versus all circulating B cells. Lollipop charts depicting log10 transformed adjusted P values from gene set enrichment analysis using genes upregulated in tumor infiltrating B cells against gene ontology (d) and Reactome (e) gene sets. (f) Feature plots split by tissue source depicting normalized expression of differentially expressed features.\u003c/p\u003e","description":"","filename":"Slide4.png","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/c1d711534d07bf2a04bf9acc.png"},{"id":68357014,"identity":"93b42dc3-bc52-407f-bc6d-b051ba17a6f9","added_by":"auto","created_at":"2024-11-06 11:36:05","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2623158,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMature regulatory dendritic cells are found in OS tumors, but not in circulation.\u003c/strong\u003e (a) UMAP representation of unsupervised clustering of 1,556 dendritic cells. (b) Boxplots depicting distribution of cell type percentages within blood and TILs. Significance determined using two-sided Wilcoxon rank sum test. *** = P value \u0026lt; 0.001; ** = P value \u0026lt; 0.01; * = P value \u0026lt; 0.05. (c) Volcano plot depicting results of differential gene expression analysis contrasting all tumor infiltrating dendritic cells versus all circulating dendritic cells. Lollipop charts depicting log10 transformed adjusted P values from gene set enrichment analysis using genes upregulated in tumor infiltrating dendritic cells against gene ontology (d) and Reactome (e) gene sets. (f) Feature plots split by tissue source depicting normalized expression of differentially expressed features.\u003c/p\u003e","description":"","filename":"Slide5.png","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/ae2b1d97d549ddd876df0f20.png"},{"id":68356750,"identity":"639183f6-53e5-414a-8ee4-767ac2e15746","added_by":"auto","created_at":"2024-11-06 11:28:05","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":6040646,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTumor infiltrating monocytes upregulate inflammatory and immune modulatory molecules relative to circulating monocytes.\u003c/strong\u003e(a) UMAP representation of unsupervised clustering of 10,004 monocytes. (b) Boxplots depicting distribution of cell type percentages within blood and TILs. Significance determined using two-sided Wilcoxon rank sum test. *** = P value \u0026lt; 0.001; ** = P value \u0026lt; 0.01; * = P value \u0026lt; 0.05. (c) Volcano plot depicting results of differential gene expression analysis contrasting all tumor infiltrating monocytes versus all circulating monocytes. Lollipop charts depicting log10 transformed adjusted P values from gene set enrichment analysis using genes upregulated in tumor infiltrating monocytes against gene ontology (d) and Reactome (e) gene sets. (f) Feature plots split by tissue source depicting normalized expression of differentially expressed features.\u003c/p\u003e","description":"","filename":"Slide6.png","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/54722898425196a73466b052.png"},{"id":68356755,"identity":"8df46b92-2a20-4ee9-a45c-9dc55998def4","added_by":"auto","created_at":"2024-11-06 11:28:06","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":5743892,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTumor infiltrating neutrophils increase expression of activation and inflammatory markers relative to circulating neutrophils.\u003c/strong\u003e (a) UMAP representation of unsupervised clustering of 5,470 neutrophils. (b) Boxplots depicting distribution of cell type percentages within blood and TILs. Significance determined using two-sided Wilcoxon rank sum test. *** = P value \u0026lt; 0.001; ** = P value \u0026lt; 0.01; * = P value \u0026lt; 0.05. (c) Volcano plot depicting results of differential gene expression analysis contrasting all tumor infiltrating neutrophils versus all circulating neutrophils. Lollipop charts depicting log10 transformed adjusted P values from gene set enrichment analysis using genes upregulated in tumor infiltrating neutrophils against gene ontology (d) and Reactome (e) gene sets. (f) Feature plots split by tissue source depicting normalized expression of differentially expressed features.\u003c/p\u003e","description":"","filename":"Slide7.png","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/7fa57dfcc59eeb7e70a46697.png"},{"id":76487623,"identity":"0fa16163-c867-40f5-b28d-624143d7fdc3","added_by":"auto","created_at":"2025-02-17 16:10:09","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":35459563,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/93f4782b-4560-4286-96f9-081834585d97.pdf"},{"id":68356741,"identity":"6bb07453-bb30-4ce2-b202-ff155e57352e","added_by":"auto","created_at":"2024-11-06 11:28:05","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":3100911,"visible":true,"origin":"","legend":"","description":"","filename":"20241025supplementalfigures.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/f2df83be5219e0aba90ec387.pdf"},{"id":68356739,"identity":"cf82f9ed-c20b-4012-a992-111962eba749","added_by":"auto","created_at":"2024-11-06 11:28:05","extension":"csv","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":79154,"visible":true,"origin":"","legend":"","description":"","filename":"supplementaldata1.csv","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/c8fbaeca8986ddda02a0c3fa.csv"},{"id":68358036,"identity":"5a0e203e-6e7b-4ae8-bb58-d92f27b682a5","added_by":"auto","created_at":"2024-11-06 11:44:05","extension":"csv","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":266506,"visible":true,"origin":"","legend":"","description":"","filename":"supplementaldata2.csv","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/0847f15bafec548441f565ab.csv"},{"id":68357015,"identity":"4400915f-5c5f-460d-8f99-278f06de411c","added_by":"auto","created_at":"2024-11-06 11:36:05","extension":"csv","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":373805,"visible":true,"origin":"","legend":"","description":"","filename":"supplementaldata3.csv","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/ab788a4c53c6b3864485f32b.csv"},{"id":68358566,"identity":"3675537c-6219-495f-ad7f-185727d2305e","added_by":"auto","created_at":"2024-11-06 11:52:05","extension":"csv","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":45611,"visible":true,"origin":"","legend":"","description":"","filename":"supplementaldata4.csv","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/22bda0e611cb1469d187aca1.csv"},{"id":68358037,"identity":"295c9512-7539-467c-9cfe-2b7a8afe1a21","added_by":"auto","created_at":"2024-11-06 11:44:05","extension":"csv","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":79337,"visible":true,"origin":"","legend":"","description":"","filename":"supplementaldata5.csv","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/66dbf2846b9454c75ca47e36.csv"},{"id":68356759,"identity":"7cb9ed59-068e-4583-a4f5-92576f5e6b80","added_by":"auto","created_at":"2024-11-06 11:28:06","extension":"csv","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":151148,"visible":true,"origin":"","legend":"","description":"","filename":"supplementaldata6.csv","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/b235e706b2fdb28d180fee04.csv"},{"id":68357022,"identity":"756ad289-e995-46c3-8abe-6852621d0c2c","added_by":"auto","created_at":"2024-11-06 11:36:06","extension":"csv","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":60474,"visible":true,"origin":"","legend":"","description":"","filename":"supplementaldata7.csv","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/1d5c3c700a88251875032304.csv"},{"id":68356752,"identity":"c95a6ebf-a545-4ca5-a92c-f148d6e694e8","added_by":"auto","created_at":"2024-11-06 11:28:05","extension":"csv","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":41246,"visible":true,"origin":"","legend":"","description":"","filename":"supplementaldata8.csv","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/00c884db8f2a94042ab7f728.csv"},{"id":68357020,"identity":"0a53cebf-0330-4d27-b873-362e23fcca40","added_by":"auto","created_at":"2024-11-06 11:36:05","extension":"xlsx","order_by":10,"title":"","display":"","copyAsset":false,"role":"supplement","size":23358,"visible":true,"origin":"","legend":"","description":"","filename":"supplementaltables.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5332445/v1/6fc27fd312a9c37a351032ea.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Impact of the canine osteosarcoma tumor microenvironment on immune cell composition and gene expression profiles","fulltext":[{"header":"Introduction","content":"\u003cp\u003eSystemic and local tumor immune responses have direct impacts on the clinical outcomes of cancer patients. For instance, overall survival and disease-free survival have been demonstrated to be positively corelated with effector T cell infiltration across multiple tumor types (1). The infiltration of immune cells into a tumor is dependent on the early recruitment of cells to the tumor, and adequate antitumor adaptive immunity has proven to be fundamental for successful immunotherapy-based intervention (2,3). Although certain tumor microenvironments (TME) recruit and maintain enough immune cells to benefit from immunotherapy, other tumors, such as osteosarcoma (OS), are notoriously defined by poor immune cell infiltration (4). The poor immune infiltration has contributed to the lack of response to immunotherapy in humans and dogs with naturally occurring OS (5). Thus, there is a need to understand how the TME impacts tumor infiltrating leukocytes (TILs) to facilitate the identification of therapeutic approaches to modulate the tumor immune microenvironment in OS.\u003c/p\u003e \u003cp\u003eThe lack of response to immunotherapy in OS has been attributed to an immune suppressive TME that prevents generation of sufficient antitumor immunity. Specifically, the OS TME has been described to have a notable immune suppressive gene expression pattern consisting of highly expressed programmed cell death protein 1 (PD-L1), indoleamine 2,3-dioxygenase (IDO) and interleukin-10 (IL-10) (6,7). Generally, these features of the TME promote the expansion of immune suppressive myeloid cells, exacerbate T cell dysfunction, and contribute to T cell exhaustion within the TME (6). Despite a basic understanding of underlying immune suppressive mechanisms active in OS across species, there is a need for a deeper characterization of the OS-associated changes to immune populations in dogs with naturally occurring OS.\u003c/p\u003e \u003cp\u003eRecent technological advances in single-cell RNA sequencing (scRNA-seq) have enabled the dissection of complex tissues and have aided in the discovery of previously unrecognized mechanisms of immune cell modulation in the TME (8). The robustness of data integration approaches has also made it possible to compare gene expression profiles across tissue types (9\u0026ndash;11). Thus, normal tissues, such as circulating leukocytes, can be used as a point of reference to investigate TME associated transcriptomic changes to TILs. Furthermore, canine OS is regarded as a valuable large animal model that enables the study of conserved TME components across species (12,13). As such this study aims to generate a canine-specific resource that documents TME associated transcriptomic changes in immune cells, while also drawing parallels to findings reported in human OS.\u003c/p\u003e \u003cp\u003eIn the present study we used two previously published scRNA-seq datasets consisting of TILs and circulating leukocytes from dogs with OS (14,15). The analysis presented in this manuscript revealed that effector CD8 T cell and CD4 T cells exhibited broadly upregulated expression of exhaustion markers (HAVCR1, LAG3, IL4I1) and identified increases in the relative proportions of follicular helper and regulatory T cells in the OS TME relative to circulating leukocytes. Consistent with the human literature, we demonstrated that mature regulatory dendritic cells (mregDCs) were found in the TME, but not in circulation (16). Furthermore, we confirmed that, in dogs, intratumoral neutrophils, monocytes, and dendritic cells upregulated immune modulatory molecules (OSM, PD-L1, and IDO1) relative to their circulating leukocyte counterparts. Overall, the analysis presented here provides key insights into how the TME shapes immune cell gene expression patterns and reveals mechanisms of immune modulation that are present within the canine OS TME.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003e\u003cstrong\u003eData acquisition, read mapping, and quantification\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRaw FASTQ data generated from Ficoll Paque prepared canine whole blood and treatment na\u0026iuml;ve OS tumor biopsies using the 10x Genomics Chromium platform were obtained from NCBI GEO (GSE225599 and GSE252470) (14,15). The raw data were aligned to the canine genome (CanFam3.1 Ensembl, release 104) and count matrices generated using a Cell Ranger analysis pipeline (version 6.1.2, 10x Genomics) as previously described (14). Annotated datasets from the primary reports were obtained from GSE225599 and from Zenodo (17). \u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData filtering and integration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor each sample, the count matrix was imported into R using the \u003cem\u003eRead10X\u003c/em\u003e function then converted to a Seurat object using the \u003cem\u003eCreateSeuratObject\u003c/em\u003e function (18). Dead and poor-quality cells were filtered out by only retaining cells that met quality control thresholds. For the tumor dataset thresholds used were: 200 \u0026lt; nFeature_RNA \u0026lt; 5500, percent.mt \u0026lt; 12.5, and 100 \u0026lt; nCount_RNA \u0026lt; 15000. For the blood dataset thresholds used were: 200 \u0026lt; nFeature_RNA \u0026lt; 3500, percent.mt \u0026lt; 20, and 500 \u0026lt; nCount_RNA \u0026lt; 20000. \u003ca id=\"_anchor_1\" href=\"#_msocom_1\" language=\"JavaScript\" name=\"_msoanchor_1\"\u003e[AH1]\u003c/a\u003e \u003ca id=\"_anchor_2\" href=\"#_msocom_2\" language=\"JavaScript\" name=\"_msoanchor_2\"\u003e[DA2]\u003c/a\u003e After initial filtering, DoubletFinder was used to remove putative cell doublets (19). Samples within each tissue were integrated into separate objects using a SCTransform normalization protocol and canonical correlation analysis (CCA) integration workflow. During integration the percent mitochondrial reads was used as a latent variable in a linear regression framework to minimize the impact of mitochondrial reads on dimension reduction and integration. After each tissue type had all samples integrated into its respective object, previous cell type annotations were transferred and any cells lacking annotation were removed. To focus analysis on cell types that are found in circulation and in tumors we excluded non-immune cells from the tumor dataset. Furthermore, mast cells, osteoclasts, IFN T cells, and macrophages were also excluded from analysis due to a lack of a circulating counterpart. The blood dataset was filtered to exclude eosinophils, double negative T cells, \u0026gamma;\u0026delta; T cells, IFN signature CD4 T cells, basophils, and CD34+ unclassified cells. All filtered samples were then integrated into one object using the same approach applied to integrate individual tissues. The top 2500 variably expressed features were used as integration anchors then unsupervised clustering was completed. Ideal clustering parameters were identified using the R package clustree (20). Dimension reduction and visualization was then completed, and the data were presented using 2-dimensional, non-linear uniform manifold approximation and projection (UMAP) plots.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSubcluster analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor each of the major immune cell type populations, the dataset was subset to include only cells from one major population. The subset dataset was then used to identify new highly variable features, then the data were re-integrated and dimension reduction was repeated as described for the full dataset. \u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCell classification\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCell annotations, as previously reported, were transferred to the integrated dataset using the unique cell barcodes associated with each cell (14,15). Unsupervised clustering was completed, then the composition of cell types (based on the transferred classifications) within each cluster was examined. For clusters in which one cell type predominated,\u003ca id=\"_anchor_3\" href=\"#_msocom_3\" language=\"JavaScript\" name=\"_msoanchor_3\"\u003e[AH3]\u003c/a\u003e \u003ca id=\"_anchor_4\" href=\"#_msocom_4\" language=\"JavaScript\" name=\"_msoanchor_4\"\u003e[DA4]\u003c/a\u003e the label was directly transferred. When conflicting cell types fell within a cluster, a new cell identity was assigned to capture the cell partitioning as determined through unsupervised clustering cells included in the current study. The gene signatures of each cluster identified in this manuscript, as determined using the \u003cem\u003eFindAllMarkers\u003c/em\u003e function in the Seurat package (test.use = \u0026ldquo;wilcox\u003ca id=\"_anchor_5\" href=\"#_msocom_5\" language=\"JavaScript\" name=\"_msoanchor_5\"\u003e[AH5]\u003c/a\u003e \u003ca id=\"_anchor_6\" href=\"#_msocom_6\" language=\"JavaScript\" name=\"_msoanchor_6\"\u003e[DA6]\u003c/a\u003e \u0026rdquo;, only.pos = TRUE), are provided as supplemental data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFeature visualization\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFeature expression was visualized using feature plots. Selected features were chosen based on the identification of a feature to be differentially expressed when contrasting tumor-infiltrating and blood leukocytes. Feature plots depict normalized expression for each feature and are presented on variable scales. When visualizing expression between tumor and blood leucocytes in a UMAP embedding, tissues were down sampled to obtain equal representation of each tissue.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCell abundance analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll cell abundance comparisons were made using the percentage of total cells in the data subset being analyzed. To make statistical inferences on changes in cell abundance two-sided Wilcoxon Rank Sum tests were used. Differences in cell abundances were discussed as over-/under-represented or unique. Relative abundance differences were classified as over-/under-represented if P value \u0026lt; 0.05 and |\u0026thinsp;log2(Fold change)\u0026thinsp;| \u0026lt; 3 when comparing between the two tissues sources. The term \u0026ldquo;unique\u0026rdquo; was reserved for changes in which P value \u0026lt; 0.05 and |\u0026thinsp;log2(Fold change)\u0026thinsp;| \u0026gt; 3. The classification scheme was designed to identify unique cell types based on the idea that cell types with low to no representation will result in an exaggerated log2(Fold change), and in turn pass the high-end cutoff.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDifferential gene expression analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDifferential gene expression analysis (DE) was completed using pseudobulk conversion followed by a DESeq2 pipeline (21). Prior to running DESeq2, low abundance features, defined as features with less than 10 raw counts across all cells sampled, were filtered out. For analysis comparing gene expression between TILs and blood leukocytes, P values were determined by testing the null hypothesis that |\u0026thinsp;log2(Fold change)\u0026thinsp;| \u0026lt; 0.58. Features were then considered to be significantly differentially expressed if the adjusted (FDR) P value was less than 0.01. Any subsequent pathway analysis was completed using lists of upregulated genes and the clusterProfiler package (22). Gene ontology and Reactome gene sets were used (23), and terms were considered enriched if they achieved an adjusted (FDR) P value less than 0.05.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIdentification of tissue signatures and removal from analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAfter completing DE analysis for each major cell population, we observed a bias for tumor-infiltrating immune cells to have increased expression of extracellular matrix associated features. Given previous reports documenting that the release of mRNA during sample processing and subsequent incorporation in cell droplet can result in confounding background tissue signatures, we devised a strategy to identify and remove features associated with background tissue signatures (9,24). To accomplish this, we completed differential expression analysis within each major immune cell population, then evaluated the features for consistent differential expression across all cell types. This revealed 46 features to be upregulated and four to be downregulated (TXNIP, PPBP, STK38, MITD1) across all tumor-infiltrating cell types (Supplemental table 1). For tissue signature estimation, we considered features with a P value less than 0.05 to be significant when testing the null hypothesis that |\u0026thinsp;log2(Fold change)\u0026thinsp;| \u0026lt; 0.58. We excluded the 46 tumor tissue-associated features, the 4 blood associated features, and a list of 108 platelet associated features, then repeated DE analysis as described above (Supplemental table 2) (14). \u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData and software availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA project specific GitHub page containing all analysis code and software versions used to analyze the data presented in this manuscript is available at https://github.com/dyammons/canine-blood-VS-tils-scrna. The annotated dataset for the integrated dataset and each subset is available for browsing at the UCSC Cell Browser (25).\u003c/p\u003e\n\u003cp\u003e[AH1]Any particular reason these are different? Reviewers might also ask.\u003c/p\u003e\n\u003cp\u003e[DA2]Different datasets, so warrant different settings, easy to defend. Some ppl will set thresholds differently for each sample\u003c/p\u003e\n\u003cp\u003e[AH3]Does the cutoff come from somewhere?\u003c/p\u003e\n\u003cp\u003e[DA4]Arbitrarily set\u003c/p\u003e\n\u003cp\u003e[AH5]Is this the method you mentioned that yields the most false positives?\u003c/p\u003e\n\u003cp\u003e[DA6]Yes, this is prone to false positives, but it does not matter much when it is just being used to identify cell type gene signatures - they are very easy to pick out. I don\u0026rsquo;t recommend using the test to compare between groups, but it is fine for cell type identification\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eOverview of study\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAnnotated data obtained from circulating leukocytes of 10 dogs diagnosed with osteosarcoma (OS) and a subset of OS tumor-infiltrating leukocytes (TILs) from 6 dogs were integrated into one dataset (14,15). We then established consensus annotations by evaluating the percentage of cell types in each cluster and transferred labels that best fit the clustering of the integrated dataset. In instances where none of the original annotations matched the unsupervised clustering results, we assigned new cell type identities to better match the integrated dataset. \u003c/p\u003e\n\u003cp\u003eOur analysis approach consisted of completing subcluster analysis for each major cell population then evaluating changes in the relative cell type proportions and identification of differentially expressed genes (DEGs) between TILs and blood leukocytes (\u003cstrong\u003eFigure 1a/b\u003c/strong\u003e). When we completed differential gene expression (DE) analysis between TILs and blood leukocytes we identified background tissue signatures that were impacting DE analysis. These tissue signatures likely arose from ambient mRNA present during cell capture (26), so we identified tumor blood, and platelet gene signatures then excluded all the associated features from DE analysis. In total 46 tumor-associated features, 4 blood-associated features, and 108 platelet-associated features were excluded from DE analysis (\u003cstrong\u003eSupplemental figure 1a, Supplemental table 1-2\u003c/strong\u003e). Gene set enrichment analysis (GSEA) of the 46 features revealed associations with extracellular matrix pathways, supporting the conclusion that these gene signatures likely originated from non-immune cells and represented background noise, which was subsequently filtered out (\u003cstrong\u003eSupplemental figure 1b/c\u003c/strong\u003e). \u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eSummary of the integrated canine circulating leucocyte and TILs dataset\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used in this study consisted of 10 blood leukocyte samples (n = 37,887 cells) from dogs with primary OS and TILs from 6 treatment-na\u0026iuml;ve dogs with primary OS of the axial skeleton (n = 11,257 cells). Although macrophages, mast cells, and osteoclasts are present in the OS TME, they were excluded from this analysis because a homologous cell type was not found in circulation. Therefore, analysis of the OS-associated impacts on immune cells was limited to six major cell types: neutrophils, monocytes, CD4 T cells, CD8 T cells, B cells, and dendritic cells (DCs), plus a minor cluster of cycling T cells (\u003cstrong\u003eFigure 1c/d\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eFollowing label transfer of each cell type, we evaluated if any population was unique or under-/over-represented in the OS TME versus blood (\u003cstrong\u003eFigure 1e\u003c/strong\u003e). As expected, all the major cell types were found in both tissue sources. However, several cell type abundance changes were observed, including a greater proportion of CD4 T cells in blood and an overrepresentation of DCs in the tumor. Additionally, we found cycling T cells exhibited a marked increase in the relative proportion of cells within the tumor samples (4.52%\u0026thinsp;\u0026plusmn;\u0026thinsp;2.97) relative to blood samples (0.29%\u0026thinsp;\u0026plusmn;\u0026thinsp;0.25) (\u003cstrong\u003eSupplemental table 3\u003c/strong\u003e). As such, cycling T cells were classified as unique to the tumor, despite being present at low levels in circulating leukocytes. We then further subset each major cell type and conducted a detailed comparison between TILs and blood leukocytes.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eFollicular helper and regulatory CD4 T cells are overrepresented in the tumor microenvironment\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAfter transferring cell type annotations to the CD4 T cell subset, we identified 6 distinct CD4\u003csup\u003e+\u003c/sup\u003e T cell populations in the combined TILs-blood dataset. The cell types consisted of na\u0026iuml;ve, central memory (TCM), effector memory (TEM), Th1-like TEM, Th2-like TEM, and regulatory/follicular helper T cells (T\u003csub\u003ereg\u003c/sub\u003e/T\u003csub\u003efh\u003c/sub\u003e) (\u003cstrong\u003eFigure 2a, Supplemental data 1\u003c/strong\u003e). The cluster annotated as T\u003csub\u003ereg\u003c/sub\u003e/T\u003csub\u003efh\u003c/sub\u003e did not reach a consensus when transferring cell type labels and was annotated based on the presence of both regulatory and follicular helper T cells (\u003cstrong\u003eSupplemental figure 2a\u003c/strong\u003e). Of the six different CD4 T cell populations identified, na\u0026iuml;ve T cells were found to be more abundant in blood compared to the TME (32.13\u0026thinsp;\u0026plusmn;\u0026thinsp;8.26% [blood] versus 13.54\u0026thinsp;\u0026plusmn;\u0026thinsp;4.29% [TILs]). Similarly, Th1-like TEM were also more abundant in blood compared to the TME (6.42\u0026thinsp;\u0026plusmn;\u0026thinsp;2.19% [blood] versus 2.25\u0026thinsp;\u0026plusmn;\u0026thinsp;1.56% [TILs]). Conversely, Treg/Tfh cells were overrepresented in the tumor, making up 37.66\u0026thinsp;\u0026plusmn;\u0026thinsp;5.76% in TME compared to 18.2\u0026thinsp;\u0026plusmn;\u0026thinsp;3.72% in blood (\u003cstrong\u003eFigure 2b, Supplemental Table 4\u003c/strong\u003e). \u003c/p\u003e\n\u003cp\u003eTo further investigate how the heterogeneity within the T\u003csub\u003ereg\u003c/sub\u003e/T\u003csub\u003efh\u003c/sub\u003e cluster impacted differential abundance analysis, we used CXCL13 expression (a molecule essential for T\u003csub\u003efh\u003c/sub\u003e mediated B cell recruitment) as a proxy for T\u003csub\u003efh\u003c/sub\u003e cell identification to better understand T\u003csub\u003efh\u003c/sub\u003e presence in the TME (27,28). Through evaluation of CXCL13\u003csup\u003e+\u003c/sup\u003e cells (normalized count \u0026gt; 0) within the T\u003csub\u003ereg\u003c/sub\u003e/T\u003csub\u003efh\u003c/sub\u003e cluster we found that CXCL13\u003csup\u003e+\u003c/sup\u003e T\u003csub\u003efh\u003c/sub\u003e were almost exclusively found in TILs (Blood: 2/14583; TILs: 121/2182 CXCL13\u003csup\u003e+\u003c/sup\u003e cells) (\u003cstrong\u003eSupplemental figure 2b\u003c/strong\u003e). Together, this suggests that the overrepresentation of T\u003csub\u003ereg\u003c/sub\u003e/T\u003csub\u003efh\u003c/sub\u003e\u003csup\u003e \u003c/sup\u003eCD4 T cells in the tumor, was in part due to the increased proportion of CXCL13\u003csup\u003e+\u003c/sup\u003e T\u003csub\u003efh\u003c/sub\u003e cells.\u003c/p\u003e\n\u003cp\u003eNext, we completed DE analysis within CD4 T cells to compare expression profiles in cells isolated from tumor and blood. The analysis revealed 368 genes to be more highly expressed in CD4 TILs and 112 genes to be overexpressed in CD4 blood leukocytes (\u003cstrong\u003eFigure 2c, Supplemental data 2\u003c/strong\u003e). Subsequent GSEA using the gene ontology (GO) database with the genes enriched in the CD4 TILs revealed enrichment for multiple terms associated with leukocyte activation and proliferation (\u003cstrong\u003eFigure 2d, Supplemental data 3\u003c/strong\u003e). GSEA using the Reactome database suggested that CD4 TILs were active in interleukin signaling and interferon responses (\u003cstrong\u003eFigure 2e, Supplemental data 3\u003c/strong\u003e). Select DEGs were then visualized to identify which clusters were driving differential expression (\u003cstrong\u003eFigure 2f, Supplemental figure 3\u003c/strong\u003e). The expression patterns of DEGs revealed that the overabundance of SELL (CD62L) and LEF1 in blood leukocytes was associated with na\u0026iuml;ve T cells, while exhaustion (HAVCR1, PDCD1, LAG3) and activation (TNFRSF4, TNFRSF18) markers overexpressed in CD4 TILs were largely confined to TEM and T\u003csub\u003ereg\u003c/sub\u003e/T\u003csub\u003efh\u003c/sub\u003e cell types (29\u0026ndash;32). Overall, our analysis suggests that in addition to shifts in cell type proportions, tumor-infiltrating CD4 T cells also exhibit altered transcriptional profiles suggestive of activation and exhaustion relative to their circulating CD4 T cell counterparts.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eFeatures associated with T cell exhaustion are enriched in tumor-infiltrating effector CD8 T cells\u003c/strong\u003e \u003c/p\u003e\n\u003cp\u003eWe next applied the same workflow to investigate tumor associated changes to CD8 T cells and NK cells. Unsupervised clustering of the integrated dataset resulted in identification of 5 transcriptomically distinct clusters which largely matched with the original cell type annotations (\u003cstrong\u003eFigure 3a, Supplemental data 4\u003c/strong\u003e). One cell type, NK cells, could not be resolved as a distinct cluster in the integrated dataset. We found that previously annotated NK cells were interspersed within CD8 effector T cell clusters suggesting a substantial overlap in the gene signatures of CD8 T cells and NK cells (\u003cstrong\u003eSupplemental figure 4\u003c/strong\u003e). Differential abundance analysis revealed that na\u0026iuml;ve CD8 T cells were the only subcluster to exhibit an overrepresentation in the blood (15.92\u0026thinsp;\u0026plusmn;\u0026thinsp;9.17% [Blood] versus 5.5\u0026thinsp;\u0026plusmn;\u0026thinsp;2.22% [TILs]) (\u003cstrong\u003eFigure 3b, Supplemental table 5\u003c/strong\u003e). \u003c/p\u003e\n\u003cp\u003eWhen comparing gene expression profiles between tumor-infiltrating and blood derived CD8 T cells, we identified 64 features to be more highly expressed in blood CD8 T cells and 241 more highly expressed in the CD8 TILs. (\u003cstrong\u003eFigure 3c, Supplemental data 2\u003c/strong\u003e). Multiple T cell exhaustion markers including LAG3, TNFSF9, and HAVCR1 (TIM-1), were identified to be more abundantly expressed in CD8 TILs (29,30,32). Despite our efforts to filter out features associated with background tissue gene signatures, GSEA indicated that NK/CD8 TIL gene expression profiles were associated with extracellular matrix processes (\u003cstrong\u003eFigure 3d, Supplemental data 3\u003c/strong\u003e). This enrichment pattern could be an artifact driven by tissue signatures or it is possible that the analysis revealed biologically relevant changes in CD8 T cells as they transition from circulating to tissue infiltrating T cells. \u003c/p\u003e\n\u003cp\u003eFurther analysis of GSEA revealed TILs demonstrated an association with T cell activation and recruitment of mononuclear cells. Reactome pathway analysis identified multiple terms associated with NFkB signaling to be enriched in tumor infiltrating CD8 T cells, which may suggest increased T cell activation (\u003cstrong\u003eFigure 3e, Supplemental data 3\u003c/strong\u003e) (33). Visualization of DEGs revealed the increased expression of LEF1 in blood leukocytes was associated with na\u0026iuml;ve CD8 T cells, while the increased CX3CR1/PTGDR expression was associated with effector CD8 T cells (\u003cstrong\u003eFigure 3f, Supplemental figure 5\u003c/strong\u003e). Tumor-infiltrating effector CD8 T cells were identified as drivers of LAG3, TNFSF9 (4-1BB ligand) and HAVCR1 (TIM-1) overexpression, which suggests effector CD8 T cells are activated and exhausted relative to circulating leukocytes (34). Consistent with studies in humans (9), our analysis revealed canine CD8 TILs to be activated and enriched in immune exhaustion markers relative to circulating CD8 T cells.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eTumor-infiltrating B cells upregulate FOS and have gene expression patterns suggestive of protein processing aberrations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIntegration of circulating and tumor-infiltrating B cells revealed the presence of 3 B cell subtypes (immature, na\u0026iuml;ve, and class switched) and a cluster of plasma cells (\u003cstrong\u003eFigure 4a, Supplemental data 5\u003c/strong\u003e). Differential abundance analysis indicated an increase in the relative proportion of immature B cells in blood (6.59\u0026thinsp;\u0026plusmn;\u0026thinsp;4.44% [Blood] versus 1.09\u0026thinsp;\u0026plusmn;\u0026thinsp;1.27% [TILs]) and an increase in the relative proportion of plasma cells in TILs (11.43\u0026thinsp;\u0026plusmn;\u0026thinsp;4.93% [Blood] versus 22.66\u0026thinsp;\u0026plusmn;\u0026thinsp;5.21% [TILs]) (\u003cstrong\u003eFigure 4b, Supplemental table 6\u003c/strong\u003e). Through evaluation of the UMAP embedding and Euclidean distance of each cluster, we found that plasma cells were distantly related to the B cell subtypes, suggesting plasma cells should be treated as a distinct cell type (\u003cstrong\u003eSupplemental figure 6a\u003c/strong\u003e). As such, we completed DE analysis within plasma cells separately from the B cell subsets. DE analysis within the plasma cells revealed relatively few differentially expressed genes (27 features over expressed in tumor infiltrating plasma cells and 15 features over expressed on circulating plasma cells), implying only subtle tumor associated changes in gene expression (\u003cstrong\u003eSupplemental figure 6b\u003c/strong\u003e). Completion of DE analysis within in B cell subsets (c0, c1, and c3) between tissue sources revealed 44 features to be more highly expressed in circulating B cells and 222 features to be more highly expressed in B cell TILs (\u003cstrong\u003eFigure 4c, Supplemental data 2\u003c/strong\u003e). Top DGEs included the proto-oncogenes, FOS and FOSB, which were identified as two of the most upregulated features in tumor-infiltrating B cells. The expression of FOS family gene members in B cells has been associated with terminal differentiation following interaction with a cognate antigen (35), but has also been assoicated with activation of apoptotic pathways in B cells (36,37). \u003c/p\u003e\n\u003cp\u003eTo investigate gene expression patterns in B cells further, we utilized GSEA which revealed enrichment of several terms associated with endoplasmic reticulum-associated degradation (ERAD), protein modification, endoplasmic reticulum activity, and antigen presentation in tumor infiltrating B cells (\u003cstrong\u003eFigure 4d/e, Supplemental data 3\u003c/strong\u003e). The enrichment of ERAD and endoplasmic reticulum stress pathways further suggested that B cells within tumor tissues may be undergoing apoptosis through ER stress-induced cell death (38). It is also possible that these pathways were enriched due to increased antigen processing and presentation, as terms associated with antigen presentation indicates were also enriched in tumor infiltrating B cells (39). Further investigation is needed to determine the functional implication of the transcriptomic changes observed within the tumor infiltrating B cells. Localization of DEGs, indicated that IGHM expression was broadly reduced in tumor infiltrating B cells, suggesting the tumor infiltrating B cells are differentiating away from a na\u0026iuml;ve gene signature (\u003cstrong\u003eFigure 4f, Supplemental figure 6a\u003c/strong\u003e) (40). Overall, these findings indicate that tumor infiltrating B cells may be playing an active role in shaping T cell mediated immunity though antigen cross presentation, or potentially undergoing apoptosis within the TME. \u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eMature regulatory dendritic cells (mregDCs) are present in the OS tumor microenvironment, but not in circulation.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll DC annotations reached a consensus which enabled the direct transfer of cell type labels across datasets (\u003cstrong\u003eFigure 5a, Supplemental data 6\u003c/strong\u003e). Of the five DC subtypes identified, we identified plasmacytoid DCs (pDCs) and precursor (pre) DCs to be overrepresented in blood relative to the TME (pDC: 17.96\u0026thinsp;\u0026plusmn;\u0026thinsp;8.67% [Blood] versus 6.03\u0026thinsp;\u0026plusmn;\u0026thinsp;3.74% [TILs]; preDC: 15.65\u0026thinsp;\u0026plusmn;\u0026thinsp;2.85% [Blood] versus 4.54\u0026thinsp;\u0026plusmn;\u0026thinsp;3.67% [TILs]) (\u003cstrong\u003eFigure 5b, Supplemental table 7\u003c/strong\u003e). In contrast, mature regulatory DCs (mregDCs), a recently described immune modulatory population, were the only DC subpopulation to be classified as unique to TILs (Blood: 0.11%\u0026thinsp;\u0026plusmn;\u0026thinsp;0.26; TILs: 11.31%\u0026thinsp;\u0026plusmn;\u0026thinsp;6.69). Lastly, conventional DC1s (cDC1) and cDC2s were determined to have unaltered abundances when comparing blood and tumor leukocytes. (\u003cstrong\u003eFigure 5b\u003c/strong\u003e). The identification of mregDCs as unique to the TME has been previously documented in human cancers (16,41), with previous reports suggesting mregDC are derived from circulating DC populations following infiltration into the tumor. Although we did not investigate how mregDCs accumulated in the canine OS TME, the marked overrepresentation in suggests a potentially conserved mechanism in mregDC biology between the two species. \u003c/p\u003e\n\u003cp\u003eTo investigate how the TME impacted DC gene expression, we completed DE analysis and subsequent GSEA. The analysis revealed 257 features enriched in tumor DCs and 26 enriched in blood DCs (\u003cstrong\u003eFigure 5c, Supplemental data 2\u003c/strong\u003e). Gene ontology analysis revealed associations with tumor necrosis factor (TNF) responses and vascular development within tumor infiltrating DC (\u003cstrong\u003eFigure 5d, Supplemental data 3\u003c/strong\u003e). GSEA using Reactome terms identified increased interleukin activity of tumor infiltrating DCs with IL4, IL13, and IL10 predicted to elicit the greatest impact on tumor DCs (\u003cstrong\u003eFigure 5e, Supplemental data 3\u003c/strong\u003e). The relative expansion of mregDCs impacted DE analysis (as evidenced by CCR7, IDO1, IL4I1, and CD274 upregulation in tumor infiltrating DCs), so we further investigated the localization of DEGs. We identified IL16 and IRF4 to be broadly down regulated in tumor infiltrating DCs, while CXCR4 was broadly upregulated (\u003cstrong\u003eFigure 5f, Supplemental figure 6b\u003c/strong\u003e). Interestingly, VEGFA, a feature determined to be associated with the tumor tissue signature, was selectively upregulated in cDC2s, suggesting the differential expression of VEGFA may be a biologically relevant change rather than an artifact of tissue bias. In summary, we provide evidence that canine mregDCs are enriched in the TME and that infiltrating DCs may modulate interleukin signaling within the OS TME.\u003c/p\u003e\n\u003cp\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTumor-infiltrating monocytes upregulate chemokine and immunoregulatory molecule expression relative to circulating monocytes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTumor-infiltrating and circulating monocytes integrated uniformly and no changes to the relative proportions of the four monocyte subtypes were identified (\u003cstrong\u003eFigure 6a/b, Supplemental data 7, Supplemental table 8\u003c/strong\u003e). Despite uniform integration, DE analysis revealed marked transcriptomic changes induced by tumor infiltration with 356 features enriched in tumor infiltrating monocytes (TIMs) and 69 enriched in blood monocytes (\u003cstrong\u003eFigure 6c, Supplemental data 2\u003c/strong\u003e). Of the features enriched in TIMs, several macrophage-associated features (MSR1, CTSK, APOC1) were identified. This suggests that although TIMs were able to be distinguished from tumor associated macrophage (TAMs) when analyzed together in the original publication (15,42), our integrated analysis found TIMs to have a macrophage-like gene profile relative to circulating monocytes. To further investigate the transcriptomic similarities between TAMs, TIMs, and circulating monocytes, we completed hierarchical clustering of the three cell types and found TIMs and circulating monocytes clustered together, while TAMs were on a distinct clade (\u003cstrong\u003eSupplemental figure 7\u003c/strong\u003e). Overall, our analysis suggests that TIMs are in an intermediate state, with greater transcriptomic similarities to circulating monocytes than TAMs. \u003c/p\u003e\n\u003cp\u003ePathway analysis of the features identified to be overexpressed in TIMs relative to blood monocytes revealed TIMs to be enriched in gene ontology terms associated with general immune activation, adhesion, and interleukin signaling. (\u003cstrong\u003eFigure 6d/e, Supplemental data 3\u003c/strong\u003e). Visualization of DEGs, demonstrated IL16 and FGL2 (Fibrinogen-like protein 2) to be broadly expressed features whereas LTF (lactotransferin) expression was primarily localized to myeloid-derived suppressor cells (M-MDSCs) (\u003cstrong\u003eFigure 6f, Supplemental figure 8\u003c/strong\u003e). Further investigation of DEGs identified TIMs to exhibit increased expression of multiple chemokines (CXCL10, CXCL16, CCL19, CCL5, CCL7, CCL8) (\u003cstrong\u003eSupplemental figure 8d\u003c/strong\u003e). Relative to TIMs, circulating monocytes exhibited higher levels of CCR2 expression which may interact with CCL7 and CCL8 secreted by TIMs to promote further infiltration (43). Relative to blood monocytes, immune modulatory molecules (IL1A, OSM, CD274, and PTGES) were determined to be broadly upregulated in TIMs. Lastly, the overexpression of C1QC in TIMs relative to blood monocytes may represent that cells are transitioning to macrophage. It is also possible that some true TAMs were unintentionally included in the analysis. Overall, we found marked gene expression changes associated with differentiation toward macrophage and upregulation of immune modulatory molecules in TIMs relative to blood monocytes.\u003c/p\u003e\n\u003cp\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTumor-associated neutrophils increase Oncostatin M and chemokine expression relative to circulating neutrophils\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEvaluation of tumor-associated neutrophils (TANs) and blood neutrophils revealed no changes in the relative proportions of neutrophils (c0) and PMN-MDSCs (c1); indicating the ratio of PMN-MDSCs to neutrophils is consistent in the blood and tumor of dogs with primary OS (\u003cstrong\u003eFigure 7a/b, Supplemental data 8, Supplemental table 9\u003c/strong\u003e). DE analysis identified upregulation of 139 features in TANs, with only 4 features preferentially expressed in the blood (\u003cstrong\u003eFigure 7c, Supplemental data 2\u003c/strong\u003e). Subsequent pathway analysis suggested TAN transcriptomic signatures were associated with general neutrophil activation, responses to interleukins, and response to endoplasmic reticulum stress (\u003cstrong\u003eFigure 7d/e, Supplemental data 3\u003c/strong\u003e). Together the analysis indicated TANs exhibited a shift toward an activated state with increased enrichment of gene programs associated with cell migration and interleukin signaling. \u003c/p\u003e\n\u003cp\u003eOncostatin M (OSM), a member of the IL-6 family with reported immune suppressive properties, was found to be broadly upregulated in TANs, relative to blood neutrophils (\u003cstrong\u003eFigure 7f, Supplemental figure 9\u003c/strong\u003e) (44). TANs also overexpressed multiple chemokines (CCL5, CCL7, and CXCL8) relative to circulating neutrophils, which suggests that upon infiltration, TAN secrete chemokines to further promote myeloid cell infiltration (\u003cstrong\u003eSupplemental figure 9c\u003c/strong\u003e). Plasminogen activator urokinase genes, PLAU and PLAUR, were also enriched in TANs. Both PLAU and PLAUR have been associated with neutrophil infiltration and their expression levels carry negative prognostic value in multiple cancer types, suggesting these genes may have prognostic value in canine OS (45,46). IL1R2, a decoy receptor for IL1A/B, was broadly upregulated in TANs which suggests they may function to dampen inflammatory responses (47). Lastly, CD274 (PD-L1) was enriched in TANs relative to blood neutrophils which implicates TANs as potential suppressors of adaptive T cell responses (48). Ultimately, we found that TANs broadly upregulated immune suppressive molecules relative to blood neutrophils, suggesting that TANs may play a role in maintenance of the immune suppressive TME.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eImmune suppression and evasion are hallmarks of cancer and understanding the mechanisms in which immune cells are impacted by the TME is foundational for the development of more effective immunotherapies. In the context of OS, a tumor type that is almost uniformly unresponsive to immunotherapy in both humans and dogs, it can be valuable to identify which cell types drive immune suppression to then target them therapeutically (5). In the present study we investigated how the canine OS TME modulates the transcriptomic signatures and relative cell type abundance of TILs. Using circulating immune cells as a point of reference, we identified upregulated expression of exhaustion markers on tumor-infiltrating T cells and found that mregDCs were present in the TME, but not in circulation. Our analysis indicated that tumor-infiltrating monocytes (TIMs) exhibited a macrophage signature relative to blood monocytes, but still retained a monocytic signature relative to tumor associated macrophages, suggesting the TIMs were likely isolated from the tumor microvasculature. Transcriptionally, we identified upregulation of well-known immune regulatory features including PD-L1 (CD274), CD80, and CD86 within tumor infiltrating myeloid cells. Furthermore, Oncostatin M (OSM) was found to be more highly expressed on TIMs and TANs relative to their circulating counterparts, implicating these myeloid cell populations as the primary producers of this immune regulatory cytokine (44). The analysis presented here suggests that many features of tumor-induced immune suppression and exhaustion that have been reported across species are active in canine OS which supports the use of spontaneously occurring canine OS to conduct immunotherapy studies.\u003c/p\u003e \u003cp\u003eThrough our analysis we identified upregulation of immune suppressive transcripts in tumor infiltrating monocytes, neutrophils, and dendritic cells. Immune suppressive roles of myeloid cells within tumors have been described in a large body of literature, and our analysis provides unique insights into specific features of immune suppression in canine OS (49\u0026ndash;51). In particular, our analysis revealed increased expression of various immune modulatory molecules, including CD274 (PD-L1), OSM, CD36, and MSR1, by tumor infiltrating monocytes and neutrophils suggesting an immune suppressive impact on the TME. Additionally, we identified IL1R2, a decoy receptor for IL1A/B, to be upregulated in TANs which suggests immunological checkpoint blockage with IL1R2 antagonists may be of therapeutic interest in canine OS (52,53).\u003c/p\u003e \u003cp\u003eConsistent with scRNA-seq data in T cells across multiple tumor types in humans (9,10), we identified increased relative proportions of regulatory and exhausted T cells within the canine OS TME. Our analysis further identified that the relative abundances of na\u0026iuml;ve and Th1-like TEM T cells were reduced in tumor infiltrating T cells relative to circulating T cell populations. Outside of cell abundance shifts we identified a broad increase in the expression of features associated with exhaustion (TOX, CTLA4, LAG3, TNFRSF9) across CD8 effector T cells and most non-na\u0026iuml;ve CD4 T cell populations (29\u0026ndash;32). As most of these molecules can also be upregulated in activated T cells, and not only exhausted T cells, our analysis cannot fully distinguish between activated and exhausted T cells (54,55). Further experimental investigation of the T cell populations is needed to determine the functional status of the cells described in this study.\u003c/p\u003e \u003cp\u003eThere is a growing body of literature suggesting intratumoral B cells play a role in antitumoral responses through antigen presentation and by participating in tertiary lymphoid structure (TLS) formation (56,57). Our analysis revealed transcriptomic evidence of profound aberrations to tumor infiltrating B cell protein processing machinery (protein modification, endoplasmic reticulum activity), which could be indicative of B cell dysfunction, antibody production, or antigen presentation. Given that mregDCs, follicular helper T cells, and B cells were all identified in the TIL population, it is possible that these cell types could have been interacting within TLS of the tumor (58\u0026ndash;61). Further investigation using spatial transcriptomics is required to determine if mregDCs, B cells, and T\u003csub\u003efh\u003c/sub\u003e co-localize within the canine OS TME. Overall, our analysis suggests that B cells are modulated by the canine OS TME and may play a role in shaping adaptive immunity.\u003c/p\u003e \u003cp\u003eAlthough our findings provide insight into how TILs differ from blood leukocytes in dogs, it is possible that some of the differences in gene expression may be due to the reference population being in circulation rather than cells present in normal bone. Because normal bone has minimal leukocytes present, the use of circulating leukocytes represents a best current avenue to explore how infiltrating immune cells are modulated by the OS TME. To minimize the impact of background tissue signatures from ambient mRNA released during sample processing, we identified a tumor tissue signature and excluded those features from DE analysis. In doing so we attempted to filter out the background signal, but we cannot exclude the possibility that some of the features identified may also represent tumor-induced gene expression changes observed across all cell types. The 46 excluded features were associated with extracellular matrix GSEA terms suggesting the genes are implicated in a tissue signature, not immune-associated terms. Future work incorporating lymph nodes or other non-circulating immune cells may help to more effectively distinguish between tissue- and tumor-associated changes.\u003c/p\u003e \u003cp\u003eLastly, the primary aim of this paper was to focus on canine cancer immunology and did not attempt to provide a comprehensive human to canine comparison of TME-associated transcriptomic changes. Future studies could address this question by incorporating scRNA-seq data from human OS tissues and peripheral blood to comprehensively profile the similarities and differences between the two species.\u003c/p\u003e \u003cp\u003eThe data presented here describe the transcriptomic responses that canine TILs exhibit relative to circulating leukocytes (8\u0026ndash;10). Through our analysis we identified dysregulation of immune modulatory features with marked changes across all cells investigated in this study, including B cells which historically have not been studied extensively in OS. From a comparative immune-oncology standpoint, the findings enable connections to be made human literature and provide insights into immune responses within spontaneously occurring canine OS. Overall, our analysis sheds light on the immune suppression and dysfunction that is present in TILs within the canine OS TME.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eConception and design: DTA, RAH, and SD. Data acquisition: DTA. Data analysis: DTA, RAH, LC, and SD. First draft of manuscript: DTA. Manuscript revisions: DTA, RAH, LC, and SD. Final approval of completed manuscript: All authors.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eThis project was supported by grants from the National Institutes of Health (NIH): U01 CA224182 (to S.D.) and the Shipley Family Foundation (to S.D.). This work utilized the Alpine high performance computing resource at the University of Colorado Boulder. Alpine is jointly funded by the University of Colorado Boulder, the University of Colorado Anschutz, and Colorado State University. Data storage was supported by the University of Colorado Boulder \u0026lsquo;PetaLibrary.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eA project specific GitHub page containing all analysis code and software versions used to analyze the data presented in this manuscript is available at https://github.com/dyammons/canine-blood-VS-tils-scrna. The annotated dataset for the integrated dataset and each subset is available for browsing at the UCSC Cell Browser.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eJochems C, Schlom J. Tumor-infiltrating immune cells and prognosis: the potential link between conventional cancer therapy and immunity. Exp Biol Med. 2011;236(5):567\u0026ndash;79. \u003c/li\u003e\n\u003cli\u003eLiu R, Yang F, Yin JY, Liu YZ, Zhang W, Zhou HH. Influence of tumor immune infiltration on immune checkpoint inhibitor therapeutic efficacy: A computational retrospective study. Front Immunol. 2021;12:685370. \u003c/li\u003e\n\u003cli\u003eYin Y, Sakakibara R, Honda T, Kirimura S, Daroonpan P, Kobayashi M, et al. High density and proximity of CD8+ T cells to tumor cells are correlated with better response to nivolumab treatment in metastatic pleural mesothelioma. Thorac Cancer. 2023;14(20):1991\u0026ndash;2000. \u003c/li\u003e\n\u003cli\u003eFritzsching B, Fellenberg J, Moskovszky L, S\u0026aacute;pi Z, Krenacs T, Machado I, et al. CD8+/FOXP3+-ratio in osteosarcoma microenvironment separates survivors from non-survivors: a multicenter validated retrospective study. Oncoimmunology. 2015;4(3):e990800. \u003c/li\u003e\n\u003cli\u003eWu CC, Beird HC, Andrew Livingston J, Advani S, Mitra A, Cao S, et al. Immuno-genomic landscape of osteosarcoma. Nat Commun. 2020;11(1):1008. \u003c/li\u003e\n\u003cli\u003eTian H, Cao J, Li B, Nice EC, Mao H, Zhang Y, et al. Managing the immune microenvironment of osteosarcoma: the outlook for osteosarcoma treatment. Bone Res. 2023;11(1):11. \u003c/li\u003e\n\u003cli\u003eMannheimer JD, Tawa G, Gerhold D, Braisted J, Sayers CM, McEachron TA, et al. Transcriptional profiling of canine osteosarcoma identifies prognostic gene expression signatures with translational value for humans. Commun Biol. 2023;6(1):856. \u003c/li\u003e\n\u003cli\u003eMaier B, Leader AM, Chen ST, Tung N, Chang C, LeBerichel J, et al. A conserved dendritic-cell regulatory program limits antitumour immunity. Nature. 2020;580(7802):257\u0026ndash;62. \u003c/li\u003e\n\u003cli\u003eSzabo PA, Levitin HM, Miron M, Snyder ME, Senda T, Yuan J, et al. Single-cell transcriptomics of human T cells reveals tissue and activation signatures in health and disease. Nat Commun. 2019;10(1):4706. \u003c/li\u003e\n\u003cli\u003eZheng L, Qin S, Si W, Wang A, Xing B, Gao R, et al. Pan-cancer single-cell landscape of tumor-infiltrating T cells. Science (1979). 2021;374(6574):abe6474. \u003c/li\u003e\n\u003cli\u003eMulder K, Patel AA, Kong WT, Piot C, Halitzki E, Dunsmore G, et al. Cross-tissue single-cell landscape of human monocytes and macrophages in health and disease. Immunity. 2021;54(8):1883\u0026ndash;900. \u003c/li\u003e\n\u003cli\u003eGustafson DL, Duval DL, Regan DP, Thamm DH. Canine sarcomas as a surrogate for the human disease. Pharmacol Ther. 2018;188:80\u0026ndash;96. \u003c/li\u003e\n\u003cli\u003eFenger JM, London CA, Kisseberth WC. Canine osteosarcoma: a naturally occurring disease to inform pediatric oncology. ILAR J. 2014;55(1):69\u0026ndash;85. \u003c/li\u003e\n\u003cli\u003eAmmons DT, Harris RA, Hopkins LS, Kurihara J, Weishaar K, Dow S. A single-cell RNA sequencing atlas of circulating leukocytes from healthy and osteosarcoma affected dogs. Front Immunol. 2023;14:1162700. \u003c/li\u003e\n\u003cli\u003eAmmons DT, Hopkins LS, Cronise KE, Kurihara J, Regan DP, Dow S. Single-cell RNA sequencing reveals the cellular and molecular heterogeneity of treatment-na\u0026iuml;ve primary osteosarcoma in dogs. Commun Biol. 2024;7(1):496. \u003c/li\u003e\n\u003cli\u003eLiu W, Hu H, Shao Z, Lv X, Zhang Z, Deng X, et al. Characterizing the tumor microenvironment at the single-cell level reveals a novel immune evasion mechanism in osteosarcoma. Bone Res. 2023;11(1):4. \u003c/li\u003e\n\u003cli\u003eAmmons D. Canine osteosarcoma single-cell RNA sequencing reference dataset: analysis code and processed data for publication. 2024. \u003c/li\u003e\n\u003cli\u003eHao Y, Hao S, Andersen-Nissen E, Mauck WM, Zheng S, Butler A, et al. Integrated analysis of multimodal single-cell data. Cell. 2021;184(13):3573\u0026ndash;87. \u003c/li\u003e\n\u003cli\u003eMcGinnis CS, Murrow LM, Gartner ZJ. DoubletFinder: doublet detection in single-cell RNA sequencing data using artificial nearest neighbors. Cell Syst. 2019;8(4):329\u0026ndash;37. \u003c/li\u003e\n\u003cli\u003eZappia L, Oshlack A. Clustering trees: a visualization for evaluating clusterings at multiple resolutions. Gigascience. 2018;7(7):giy083. \u003c/li\u003e\n\u003cli\u003eLove MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15:1\u0026ndash;21. \u003c/li\u003e\n\u003cli\u003eWu T, Hu E, Xu S, Chen M, Guo P, Dai Z, et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. The innovation. 2021;2(3). \u003c/li\u003e\n\u003cli\u003eLiberzon A, Subramanian A, Pinchback R, Thorvaldsd\u0026oacute;ttir H, Tamayo P, Mesirov JP. Molecular signatures database (MSigDB) 3.0. Bioinformatics. 2011;27(12):1739\u0026ndash;40. \u003c/li\u003e\n\u003cli\u003eYoung MD, Behjati S. SoupX removes ambient RNA contamination from droplet-based single-cell RNA sequencing data. Gigascience. 2020;9(12):giaa151. \u003c/li\u003e\n\u003cli\u003eSpeir ML, Bhaduri A, Markov NS, Moreno P, Nowakowski TJ, Papatheodorou I, et al. UCSC Cell Browser: visualize your single-cell data. Bioinformatics. 2021;37(23):4578\u0026ndash;80. \u003c/li\u003e\n\u003cli\u003eYoung MD, Behjati S. SoupX removes ambient RNA contamination from droplet-based single-cell RNA sequencing data. Gigascience. 2020;9(12):giaa151. \u003c/li\u003e\n\u003cli\u003eLin X, Ye L, Wang X, Liao Z, Dong J, Yang Y, et al. Follicular helper T cells remodel the immune microenvironment of pancreatic cancer via secreting CXCL13 and IL-21. Cancers (Basel). 2021;13(15):3678. \u003c/li\u003e\n\u003cli\u003eMoser B. CXCR5, the defining marker for follicular B helper T (TFH) cells. Front Immunol. 2015;6:296. \u003c/li\u003e\n\u003cli\u003eWolf Y, Anderson AC, Kuchroo VK. TIM3 comes of age as an inhibitory receptor. Nat Rev Immunol. 2020;20(3):173\u0026ndash;85. \u003c/li\u003e\n\u003cli\u003eYang ZZ, Kim HJ, Villasboas JC, Chen YP, Price-Troska T, Jalali S, et al. Expression of LAG-3 defines exhaustion of intratumoral PD-1+ T cells and correlates with poor outcome in follicular lymphoma. Oncotarget. 2017;8(37):61425. \u003c/li\u003e\n\u003cli\u003eWard-Kavanagh LK, Lin WW, \u0026Scaron;ed\u0026yacute; JR, Ware CF. The TNF receptor superfamily in co-stimulating and co-inhibitory responses. Immunity. 2016;44(5):1005\u0026ndash;19. \u003c/li\u003e\n\u003cli\u003eLi Y, Wang Z, Jiang W, Zeng H, Liu Z, Lin Z, et al. Tumor-infiltrating TNFRSF9+ CD8+ T cells define different subsets of clear cell renal cell carcinoma with prognosis and immunotherapeutic response. Oncoimmunology. 2020;9(1):1838141. \u003c/li\u003e\n\u003cli\u003eBarnes SE, Wang Y, Chen L, Molinero LL, Gajewski TF, Evaristo C, et al. T cell-NF-\u0026kappa;B activation is required for tumor control in vivo. J Immunother Cancer. 2015;3(1):1\u0026ndash;11. \u003c/li\u003e\n\u003cli\u003eAnderson AC, Joller N, Kuchroo VK. Lag-3, Tim-3, and TIGIT: co-inhibitory receptors with specialized functions in immune regulation. Immunity. 2016;44(5):989\u0026ndash;1004. \u003c/li\u003e\n\u003cli\u003eOhkubo Y, Arima M, Arguni E, Okada S, Yamashita K, Asari S, et al. A role for c-fos/activator protein 1 in B lymphocyte terminal differentiation. The Journal of Immunology. 2005;174(12):7703\u0026ndash;10. \u003c/li\u003e\n\u003cli\u003eInada K, Okada S, Phuchareon J, Hatano M, Sugimoto T, Moriya H, et al. c-Fos induces apoptosis in germinal center B cells. The Journal of Immunology. 1998;161(8):3853\u0026ndash;61. \u003c/li\u003e\n\u003cli\u003eLiebermann DA, Gregory B, Hoffman B. AP-1 (Fos/Jun) transcription factors in hematopoietic differentiation and apoptosis. Int J Oncol. 1998;12(3):685\u0026ndash;1385. \u003c/li\u003e\n\u003cli\u003eSano R, Reed JC. ER stress-induced cell death mechanisms. Biochimica et Biophysica Acta (BBA)-Molecular Cell Research. 2013;1833(12):3460\u0026ndash;70. \u003c/li\u003e\n\u003cli\u003eRastogi I, Jeon D, Moseman JE, Muralidhar A, Potluri HK, McNeel DG. Role of B cells as antigen presenting cells. Front Immunol. 2022;13:954936. \u003c/li\u003e\n\u003cli\u003eMorgan D, Tergaonkar V. Unraveling B cell trajectories at single cell resolution. Trends Immunol. 2022;43(3):210\u0026ndash;29. \u003c/li\u003e\n\u003cli\u003eGordon JR, Ma Y, Churchman L, Gordon SA, Dawicki W. Regulatory dendritic cells for immunotherapy in immunologic diseases. Front Immunol. 2014;5:7. \u003c/li\u003e\n\u003cli\u003eMa RY, Black A, Qian BZ. Macrophage diversity in cancer revisited in the era of single-cell omics. Trends Immunol. 2022;43(7):546\u0026ndash;63. \u003c/li\u003e\n\u003cli\u003eHao Q, Vadgama J V, Wang P. CCL2/CCR2 signaling in cancer pathogenesis. Cell Communication and Signaling. 2020;18:1\u0026ndash;13. \u003c/li\u003e\n\u003cli\u003eWallace PM, MacMaster JF, Rouleau KA, Brown TJ, Loy JK, Donaldson KL, et al. Regulation of inflammatory responses by oncostatin M. The Journal of Immunology. 1999;162(9):5547\u0026ndash;55. \u003c/li\u003e\n\u003cli\u003eShi K, Zhou J, Li M, Yan W, Zhang J, Zhang X, et al. Pan-cancer analysis of PLAU indicates its potential prognostic value and correlation with neutrophil infiltration in BLCA. Biochimica et Biophysica Acta (BBA)-Molecular Basis of Disease. 2024;1870(2):166965. \u003c/li\u003e\n\u003cli\u003eZeng F, Li G, Liu X, Zhang K, Huang H, Jiang T, et al. Plasminogen activator urokinase receptor implies immunosuppressive features and acts as an unfavorable prognostic biomarker in glioma. Oncologist. 2021;26(8):e1460\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eDinarello CA. Overview of the IL‐1 family in innate inflammation and acquired immunity. Immunol Rev. 2018;281(1):8\u0026ndash;27. \u003c/li\u003e\n\u003cli\u003eYajuk O, Baron M, Toker S, Zelter T, Fainsod-Levi T, Granot Z. The PD-L1/PD-1 axis blocks neutrophil cytotoxicity in cancer. Cells. 2021;10(6):1510. \u003c/li\u003e\n\u003cli\u003eSu P, Wang Q, Bi E, Ma X, Liu L, Yang M, et al. Enhanced lipid accumulation and metabolism are required for the differentiation and activation of tumor-associated macrophages. Cancer Res. 2020;80(7):1438\u0026ndash;50. \u003c/li\u003e\n\u003cli\u003eGulay KCM, Aoshima K, Maekawa N, Suzuki T, Konnai S, Kobayashi A, et al. Hemangiosarcoma cells induce M2 polarization and PD-L1 expression in macrophages. Sci Rep. 2022;12(1):2124. \u003c/li\u003e\n\u003cli\u003eChristofides A, Strauss L, Yeo A, Cao C, Charest A, Boussiotis VA. The complex role of tumor-infiltrating macrophages. Nat Immunol. 2022;23(8):1148\u0026ndash;56. \u003c/li\u003e\n\u003cli\u003eXia J, Zhang L, Peng X, Tu J, Li S, He X, et al. IL1R2 blockade alleviates immunosuppression and potentiates anti-PD-1 efficacy in triple-negative breast cancer. Cancer Res. 2024; \u003c/li\u003e\n\u003cli\u003eWu TC, Xu K, Martinek J, Young RR, Banchereau R, George J, et al. IL1 receptor antagonist controls transcriptional signature of inflammation in patients with metastatic breast cancer. Cancer Res. 2018;78(18):5243\u0026ndash;58. \u003c/li\u003e\n\u003cli\u003eGholami MD, Saeedi Y, Heydari S, Garssen J, Falak R. Exhaustion of T lymphocytes in the tumor microenvironment: Significance and effective mechanisms. Cell Immunol. 2017;322:1\u0026ndash;14. \u003c/li\u003e\n\u003cli\u003eRha MS, Shin EC. Activation or exhaustion of CD8+ T cells in patients with COVID-19. Cell Mol Immunol. 2021;18(10):2325\u0026ndash;33. \u003c/li\u003e\n\u003cli\u003eKinker GS, Vitiello GAF, Ferreira WAS, Chaves AS, Cordeiro de Lima VC, Medina T da S. B cell orchestration of anti-tumor immune responses: a matter of cell localization and communication. Front Cell Dev Biol. 2021;9:678127. \u003c/li\u003e\n\u003cli\u003eLaumont CM, Banville AC, Gilardi M, Hollern DP, Nelson BH. Tumour-infiltrating B cells: immunological mechanisms, clinical impact and therapeutic opportunities. Nat Rev Cancer. 2022;22(7):414\u0026ndash;30. \u003c/li\u003e\n\u003cli\u003eWithers SS, York D, Choi JW, Woolard KD, Laufer‐Amorim R, Sparger EE, et al. Metastatic immune infiltrates correlate with those of the primary tumour in canine osteosarcoma. Vet Comp Oncol. 2019;17(3):242\u0026ndash;52. \u003c/li\u003e\n\u003cli\u003eUkita M, Hamanishi J, Yoshitomi H, Yamanoi K, Takamatsu S, Ueda A, et al. CXCL13-producing CD4+ T cells accumulate in the early phase of tertiary lymphoid structures in ovarian cancer. JCI Insight. 2022;7(12). \u003c/li\u003e\n\u003cli\u003eChen L, Oke T, Siegel N, Cojocaru G, Tam AJ, Blosser RL, et al. The immunosuppressive niche of soft-tissue sarcomas is sustained by tumor-associated macrophages and characterized by intratumoral tertiary lymphoid structures. Clinical Cancer Research. 2020;26(15):4018\u0026ndash;30. \u003c/li\u003e\n\u003cli\u003eYan P, Wang J, Yue B, Wang X. Unraveling molecular aberrations and pioneering therapeutic strategies in osteosarcoma. Biochimica et Biophysica Acta (BBA)-Reviews on Cancer. 2024;189171. \u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"cancer-immunology-immunotherapy","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ciim","sideBox":"Learn more about [Cancer Immunology, Immunotherapy](http://link.springer.com/journal/262)","snPcode":"262","submissionUrl":"https://submission.nature.com/new-submission/262/3","title":"Cancer Immunology, Immunotherapy","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Canine (dog), Osteosarcoma, scRNA-seq, transcriptomics, cancer immunology","lastPublishedDoi":"10.21203/rs.3.rs-5332445/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5332445/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eImmune cells play key roles in host responses to malignant tumors. The selective pressure that immune cells elicit on tumors promotes immune escape, while tumor associated modulation of immune cells creates an environment favorable to tumor growth and progression.\u003c/p\u003e \u003cp\u003eIn this study we used publicly available single-cell RNA sequencing (scRNA-seq) data from the translationally relevant canine osteosarcoma (OS) model to compare tumor infiltrating leukocytes (TILs) to circulating leukocytes. Through computational analysis we investigated the differences in cell type proportions and how the OS TME impacted TIL transcriptomic profiles relative to circulating leukocytes.\u003c/p\u003e \u003cp\u003eDifferential abundance analysis revealed increased proportions of follicular helper T cells and mature regulatory dendritic cells (mregDCs) in the OS TME. Differential gene expression analysis identified exhaustion markers (LAG3, HAVCR1, PDCD1) to be upregulated in CD4 and CD8 T cells within the OS TME. Comparisons of B cell gene expression profiles revealed an enrichment of protein processing and endoplasmic reticulum pathways, suggesting infiltrating B cells were activated and participating in antigen presentation. Gene expression changes within myeloid cells identified increased expression of immune suppressive molecules (CD274, OSM, MSR1) in the OS TME, supporting their role as immunosuppressors. Comparisons to human literature revealed similar immune modulation in canine and human OS, further supporting the dog as a model for studies investigating novel immunotherapeutics.\u003c/p\u003e \u003cp\u003eOverall, the analysis presented here provides new insights into how the OS TME impacts the transcriptional programs of major immune cell populations in dogs.\u003c/p\u003e","manuscriptTitle":"Impact of the canine osteosarcoma tumor microenvironment on immune cell composition and gene expression profiles","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-11-06 11:28:00","doi":"10.21203/rs.3.rs-5332445/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-11-22T03:50:54+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-11-18T15:02:36+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-11-04T14:44:12+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"243012029756862029123008491848435606271","date":"2024-10-31T00:48:18+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"130290576068975974751325452288476831077","date":"2024-10-30T07:32:25+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"291174691192189983548407726892390469408","date":"2024-10-28T22:58:39+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-10-28T02:55:24+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-10-26T04:29:50+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-10-26T04:29:40+00:00","index":"","fulltext":""},{"type":"submitted","content":"Cancer Immunology, Immunotherapy","date":"2024-10-25T12:32:35+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"cancer-immunology-immunotherapy","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ciim","sideBox":"Learn more about [Cancer Immunology, Immunotherapy](http://link.springer.com/journal/262)","snPcode":"262","submissionUrl":"https://submission.nature.com/new-submission/262/3","title":"Cancer Immunology, Immunotherapy","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"3c576823-f2f2-43e9-a724-1cf735384851","owner":[],"postedDate":"November 6th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-02-17T16:03:30+00:00","versionOfRecord":{"articleIdentity":"rs-5332445","link":"https://doi.org/10.1007/s00262-025-03950-3","journal":{"identity":"cancer-immunology-immunotherapy","isVorOnly":false,"title":"Cancer Immunology, Immunotherapy"},"publishedOn":"2025-02-11 15:57:50","publishedOnDateReadable":"February 11th, 2025"},"versionCreatedAt":"2024-11-06 11:28:00","video":"","vorDoi":"10.1007/s00262-025-03950-3","vorDoiUrl":"https://doi.org/10.1007/s00262-025-03950-3","workflowStages":[]},"version":"v1","identity":"rs-5332445","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5332445","identity":"rs-5332445","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}