Single‐cell RNA sequencing reveals immune cell dysfunction in the peripheral blood of patients with highly aggressive gastric cancer

Abstract Highly aggressive gastric cancer (HAGC) is a gastric cancer characterized by bone marrow metastasis and disseminated intravascular coagulation (DIC). Information about the disease is limited. Here we employed single‐cell RNA sequencing to investigate peripheral blood mononuclear cells (PBMCs), aiming to unravel the immune response of patients toward HAGC. PBMCs from seven HAGC patients, six normal advanced gastric cancer (NAGC) patients, and five healthy individuals were analysed by single‐cell RNA sequencing. The expression of genes of interest was validated by bulk RNA‐sequencing and ELISA. We found a massive expansion of neutrophils in PBMCs of HAGC. These neutrophils are activated, but immature. Besides, mononuclear phagocytes exhibited an M2‐like signature and T cells were suppressed and reduced in number. Analysis of cell‐cell crosstalk revealed that several signalling pathways involved in neutrophil to T‐cell suppression including APP‐CD74, MIF‐(CD74+CXCR2), and MIF‐(CD74+CD44) pathways were increased in HAGC. NETosis‐associated genes S100A8 and S100A9 as well as VEGF, PDGF, FGF, and NOTCH signalling that contribute to DIC development were upregulated in HAGC too. This study reveals significant changes in the distribution and interactions of the PBMC subsets and provides valuable insight into the immune response in patients with HAGC. S100A8 and S100A9 are highly expressed in HAGC neutrophils, suggesting their potential to be used as novel diagnostic and therapeutic targets for HAGC.


| INTRODUCTION
Gastric cancer (GC) is the fifth most prevalent cancer worldwide, with a notably high incidence in East Asia. 1 In certain instances, GC is presented with bone marrow metastasis (BMM) and disseminated intravascular coagulation (DIC), and is diagnosed as highly aggressive GC (HAGC).In contrast to normal advanced gastric cancer (NAGC) which lacks DIC, HAGC has a very poor prognosis and is relatively understudied. 2C is the primary clinical symptom and the main cause of fatality in patients with HAGC. 2 DIC is a clinicopathological syndrome that arises in multiple diseases due to coagulation dysfunction.In cancer patients, tissue necrosis resulting from malignant solid tumours or their contiguous tissues can induce the release of diverse molecular factors, such as cancer procoagulant and tissue factor, which promote blood coagulation and disrupt the balance between coagulation and anticoagulation, resulting in DIC. 3 Additionally, blood abnormalities have also been implicated in the pathogenesis of DIC.Damage to mononuclear phagocytes responsible for phagocytosing or clearing procoagulant substances contributes to DIC. 4 Neutrophil dysfunction also plays a significant role in DIC.Upon stimulation, neutrophils release nuclear components into the extracellular matrix where they form a net-like structure called the neutrophil extracellular traps (NETs). NTs can act as scaffolds to retain and activate platelets, which lead to thrombus formation and DIC. 5 Nevertheless, the etiology of DIC in patients with HAGC remains unclear.
Tumour cells exert a significant influence on the immune system of patients, and peripheral blood cells can mediate the systemic immune response to tumour development. 6Research indicates that changes in peripheral blood mononuclear cells (PBMCs) are associated with the progression of tumour malignancy and patient prognosis. 6Therefore, PBMCs can potentially help assess the disease progression and immune status of patients.
In order to avoid subjecting HAGC patients, who are often in poor physical condition, to undue harm, we chose to focus solely on collecting blood and studying PBMCs to gain insight into the pathogenesis of HAGC.
HAGC is highly heterogeneous and therapeutically challenging.While traditional bulk RNA-seq may offer insights into macro-level gene expression from tens of thousands of cells in samples, it does not allow us to account for each individual cell's effects, thus losing critical data.][9][10] Using single-cell sequencing technology that provides unparalleled depth and precision, we may be able to discover new diagnostic and treatment targets for HAGC.
Here, using the microwell-seq platform, 11,12 we constructed a high-quality single-cell atlas of PBMCs from HAGC patients and NAGC patients, as well as healthy individuals.We demonstrated that HAGC was associated with substantial changes in the composition and activity of PBMCs.The neutrophil population was significantly increased in HAGC with a preponderance of immature and activated

| Single-cell RNA sequencing of PBMCs from HAGC patients reveals disturbances in cell type composition
A total of 18 peripheral blood samples were collected from seven HAGC patients, six NAGC patients and five healthy individuals for single-cell RNA sequencing using microwell-seq platform (Figure 1A and Table S1).
In total, 51,631 PBMCs passed the quality control, and each cell expressed approximately 469 genes with a mean read count of 934 (Supplementary Figure S1).A roughly equivalent contribution from the three conditions to the final number of cells was observed (Supplementary Figure S1).We projected the resulting PBMC gene expression using UMAP and identified clusters corresponding to typical PBMC cell types (Figure 1B).The identified cell types included mononuclear phagocytes (MPs), which highly express CD14, LYZ and LST1; T cells, which highly express CD3D and CD3E; NK cells, which highly express CCL5 and NKG7 and plasma cells (B cells), which highly express JCHAIN, IGHA1, IGLC2 and IGLL5 (Figure 1B, C, Table S2 and S3).Three dominant subpopulations of neutrophils were identified based on their differential gene expression profiles, which included DEFA (high expression of DEFA3 and DEFA1), S100 (high expression of S100A8 and S100A9), and Ery-like (high expression of HBB and HBA1) subpopulation (Figure 1B, C and Supplementary Figure S1).The presence of high levels of pro-inflammatory factors-defensins (DEFA1/3 genes) and S100 in neutrophils, [13][14][15] suggests a strong proinflammatory systemic response in patients with HAGC.
Comparing the PBMC compositions in NAGC patients and healthy individuals, PBMCs of HAGC patients did not exhibit any unique cell type.However, there were significant alterations in the proportion of PBMCs.HAGC patient samples showed a remarkable increase in neutrophils, whereas the proportions of MPs, T cells, and NK cells were significantly reduced as compared to NAGC patients and healthy individuals (Figure 1D, E and Supplementary Figure S1).These findings were consistent with the clinical blood test result, which demonstrated increased neutrophils and decreased lymphocytes in HAGC patients compared to NAGC patients (Figure 1F and Table S1).Additionally, the single-cell RNA-sequencing analysis showed approximately 10% neutrophils in NAGC PBMCs, whereas the clinical blood test reported around 60% neutrophils.This disparity was attributed to the fact that the clinical blood test encompasses all blood cells, whereas our single-cell RNA-sequencing specifically detects PBMCs with a single round nucleus.Hence, this finding suggests that approximately one-sixth of the neutrophils in NAGC resemble the round nucleus neutrophils found in HAGC, whereas the remaining neutrophils in NAGC patients exhibit normal polymorphonuclear characteristics.

| PBMCs of HAGC patients consist primarily of immature neutrophils
As neutrophils expanded significantly in HAGC, we then isolated and re-clustered the neutrophils from the PBMC data.This process identified seven new subclusters, with subclusters 3, 5, and 6 corresponding to DEFA-expressing neutrophils; subclusters 0, 2, and 4 corresponding to Ery-like neutrophils; and subcluster 1 corresponding to S100-expressing neutrophils (Figure 2A, Supplementary Figure S2A and Table S4).Each sub-cluster exhibited characteristic gene expression patterns.Subcluster 0 showed high expression of HBA2 and HBB, which are specific to the red blood cell lineage, while subcluster 1 showed high expression of LTF and CAMP.Subcluster 2 showed high expression of IFI30 and AC007192.1 l.Subclusters 3 and 5 showed high expression of DEFA1, DEFA1B and DEFA3.Subcluster 4 showed high expression of L1PA6 and TMSB10.Subcluster 6 showed high expression of ELANE, PRTN3 and MPO (Figure 2B and Supplementary Figure S2B).Interestingly, HAGC patients had a unique subcluster of neutrophils (subcluster 0).These cells expressed HBA2 and HBB, which Moreover, our investigation revealed that HAGC neutrophils expressed genes associated with the early stages of neutrophil differentiation, 16 such as LTF, PRTN3 and ELANE (Supplementary Figure S2B).This intriguing finding suggests the existence of potential maturation abnormalities in HAGC neutrophils.In order to delve deeper into this phenomenon, we employed CytoTRACE to estimate the pseudo-time trajectory of neutrophils. 17The analysis results  S1.And statistical significance was determined using two-sided Mann-Whitney U test.2E and Supplementary Figure S2C).Additional support for these observations was obtained through trajectory analysis using scVelo, 18 which further confirmed the lower differentiation status of cells from clusters 2 and 4 (Supplementary Figure S2D).To gain more insights into the maturation issue, we compared HAGC neutrophils with bone marrow neutrophils of healthy individuals. 19Interestingly, CytoTRACE revealed that HAGC neutrophils exhibited lower levels of differentiation when compared to fully developed neutrophils in the bone marrow (Supplementary Figure S2E, F).In addition, representative granule genes are expressed during the process of neutrophil maturation. 20Therefore, we also investigated the expression of granule genes in each subcluster of HAGC neutrophils.Notably, sub-cluster 6 exhibited high expression of primary granule genes, whereas sub-clusters 0 and 1 displayed increased expression of secondary granule genes.The expression of tertiary and mature granule genes was quite low across all sub-clusters, indicating the immaturity of overall HAGC neutrophils (Figure 2F; Supplementary Figure S2G).
Next, we examined gene sets related to various biological categories, including activation, aging, apoptosis, chemotaxis, NETosis, and pro-tumour responses in neutrophils. 16,21,22Interestingly, we observed that only genes associated with activation and NETosis were significantly upregulated in HAGC neutrophils (Figure 2G and Supplementary Figure S2H).Collectively, these results suggest that HAGC is characterized by a significant expansion of the neutrophils, which, although partially activated, retain an immature state.Furthermore, since NETs are involved in the pathophysiology of DIC, upregulation of NETosis-associated genes in neutrophils may contribute to DIC in patients with HAGC.

| Mononuclear phagocytes of HAGC present an M2-like polarization signature
In addition to neutrophils, MPs within myeloid cells also play influential roles in response to tumour. 23Notably, the MPs in HAGC patients did not form any discrete clusters.However, they constituted only a minor fraction of the total PBMCs when compared to MPs from NAGC patients and healthy individuals (Figure 1D, E).
Microphages, when activated, can polarize into two main subsets: classically activated M1 macrophages and alternatively activated M2 macrophages. 24M1 macrophages exhibit intrinsic phagocytic abilities and enhanced inflammatory response, thus behaving as anti-tumour cells.On the other hand, M2 macrophages possess a diverse array of tumour-promoting capabilities, including immunosuppression, angiogenesis and neovascularization.Intriguingly, analysis of M1-and M2-specific gene expression revealed a significantly elevated M2 score and a decreased M1 score in HAGC MP cells, resulting in a biased overall M2-like signature (Figure 3A).This finding was exemplified by the downregulation of the M1 marker gene CCL5 and the expression of M2-specific genes, such as MMP9 and CTSA/B/C (Figure 3B, C).RNA velocity analysis further supported the transition of MPs towards an M2-like state in HAGC (Figure 3D), suggesting the immunosuppressive properties of HAGC MPs.

| Suppressed T cell activity in HAGC patients
Considering the immunosuppressive potential of MPs in patients with HAGC, we next investigated the responses of lymphoid cells.Compared to patients with NAGC and healthy individuals, B cell responses appeared relatively unaffected in HAGC patients (Figure 1E and Supplementary Figure S3A, Table S5).Gene ontology (GO) analysis of differentially expressed genes revealed that genes associated with type I interferon response, adaptive immune response, positive regulation of immune response and innate immune response were decreased in T cells of HAGC patients compared to NAGC patients (Figure 4A), suggesting the suppression of HAGC T cells.Indeed, the expression of genes associated with T cell activation and interferon was significantly lower in HAGC patients than in NAGC patients and healthy individuals (Figure 4B and Supplementary Figure S4A, B).T cell effector genes GZMA and GZMB exhibited low expression in HAGC patients too (Supplementary Figure S4C).These findings, combined with the reduced lymphocyte proportions in HAGC PBMCs (Figure 1E, F and Table S1), support the suppression of T cells in HAGC patients.However, further examination revealed that gene sets associated with cell exhaustion were expressed at a low level in HAGC, unlike tumour-infiltrating T cells. 9,10,25Additionally, gene sets associated with senescence, including ribosome and histone genes, showed no significant differences between HAGC and NAGC, although there was a slight reduction compared to healthy controls (Figure 4D and Supplementary Figure S4C).Therefore,    T cells in PBMCs of HAGC patients are suppressed but do not display signs of exhaustion or senescence.

| HAGC myeloid cells suppress T cells
Mononuclear phagocytes and granulocytes are crucial groups of innate immune cells that have essential roles in defending against pathogens.
In disease settings, some of these cells display an immature morphology and exhibit immunosuppressive functions, which have been associated with negative prognostic outcomes and treatment responses.These cells are referred to as myeloid-derived suppressor cells (MDSCs) and can be further categorized into granulocyte-like MDSCs (G-MDSC) and mononuclear phagocyte-like MDSCs (M-MDSCs). 26,27The granulocyte and the mononuclear phagocyte are primarily composed of neutrophil and monocyte.Considering that HAGC PBMCs were primarily composed of active yet immature neutrophils, and MP cells exhibited an M2 immunosuppressive phenotype, we next investigated whether neutrophils and M2-polarised MPs expressed G-MDSC and M-MDSC markers, respectively.Remarkably, neutrophil subsets 0, 1, and 5 of HAGC patients exhibited expression of G-MDSC-related markers (Figure 5A).M-MDSC markers were also upregulated in M2-like MPs in HAGC patients (Figure 5A).These findings suggest that myeloid cells in HAGC patients acquire the characteristics of MDSCs with immunosuppressive activity.
To elucidate the mechanism involved in T cell suppression, we explore the predicted crosstalk between different types of cells by analysing receptor-ligand pairs.Correlation analysis indicated a close relationship among MPs, S100-expressing neutrophils, Ery-like neutrophils, T cells, and NK cells in HAGC patients (Figure 5B).Interestingly, we observed a higher number of predicted crosstalk interactions in HAGC patients compared to both NAGC patients and healthy controls (Supplementary Figure S5A).To further investigate cell-cell communication between myeloid cells and T cells in HAGC patients, we utilized CellChat to identify potential ligand-receptor pairs. 28Elevated levels of immunosuppressive signallings exerted from HAGC neutrophils to T cells and NK cells were observed, including APP-CD74, MIF-(CD74+CXCR2), and MIF-(CD74+CD44) (Figure 5C and Supplementary Figure S5B). 29In addition, the signalling pathways of RETN-CAP1 and ANXA1-FPR1 from MPs to neutrophils, known to trigger cytokine storms, 30 were significantly upregulated in HAGC patients, which is consistent with the hyper-inflammatory status in HAGC (Supplementary Figure S5B-D).
To verify the reliability of scRNA-seq results and gain further insights into HAGC, we conducted bulk RNA-seq using PBMCs.
When compared to PBMCs of NAGC, we identified 252 upregulated genes and 24 downregulated genes in HAGC PBMCs.Notably, markers associated with MDSC, including MMP9, LCN2, S100A8, S00A9, S100A12 and S100P were significantly upregulated in HAGC PBMCs as compared to NAGC PBMCs (Figure 5D, Table S6).In addition, gene ontology (GO) analysis of the differentially expressed genes demonstrated a strong association of upregulated genes with the biological process of neutrophil-mediated immunity.Enrichment of genes related to the response to foreign entities such as bacteria, fungi and toxic substances was observed in HAGC, suggesting the activation of innate immunity (Supplementary Figure S5E).Furthermore, gene set enrichment analysis (GSEA) indicated that genes involved in Th17 cell differentiation and T cell receptor signalling pathways were downregulated in HAGC compared to NAGC (Figure 5E), suggesting the repression of T cell function in HAGC.
Collectively, our findings support the notion that myeloid cell populations in HAGC patients effectively suppress T cell activity and induce systemic inflammatory responses.This validation through bulk RNA-seq provides enhanced confidence in the scRNA-seq results and facilitates a deeper understanding of the molecular characteristics of HAGC.

| Aberrant cell-cell communication and S100 family gene expression in neutrophils contributes to DIC development in HAGC
Furthermore, we explored the potential mechanisms underlying DIC development in HAGC patients.Our analysis revealed a substantial activation of signalling pathways associated with angiogenesis and fibrosis, including VEGF, PDGF, FGF and NOTCH signalling pathways 31,32 (Figure 6A).Centrality analysis identified Neutrophil_Ery cells as the primary source of the ligands of these pathways (Supplementary Figure S6A), which may be involved in DIC pathogenesis by inducing the formation of microvessels and fibrous tissue (Supplementary Figure S6B).
Previous studies have implicated S100A8 and S100A9 in neutrophil NET formation, which has been further linked to thrombosis in DIC. 33,34In our scRNA-seq study of HAGC PBMCs, we observed a significant upregulation of S100A8 and S100A9 genes, particularly in  neutrophils (Supplementary Figure S6C).This finding aligns with the observed DIC among HAGC patients.Additionally, our bulk RNA-seq analysis comparing PBMCs from HAGC and NAGC patients confirmed the enhanced expression of S100A8 and S100A9 in HAGC (Figure 5D, Table S6).GSEA revealed an upregulation of genes associated with complement and coagulation cascades as well as NET formation, in HAGC patients (Figure 6B).To further validate these findings, we performed an ELISA assay using blood plasma and observed a significant increase in S100A8 protein levels in HAGC.
This suggests that S100A8 may serve as a potential biomarker for HAGC diagnosis (Figure 6C).

| DISCUSSION
In this study, we employed scRNA-seq to examine the gene expression of HAGC PBMCs.In comparison with NAGC and healthy controls, we observed a substantial increase in neutrophil population in HAGC PBMCs, and these neutrophils were immature but active.MPs G-MDSCs in HAGC patients. 26These neutrophils suppressed T cell activity through multiple signalling pathways.This phenomenon could potentially be attributed to the extensive metastasis of tumour cells in the bone marrow which leads to complex systemic pro-inflammatory myeloid cells coupled with the immunosuppression of lymphoid cells.
Moreover, DIC is a major pathological feature observed in HAGC patients, profoundly impacting their overall health. 2 Previous studies unrelated to HAGC have demonstrated a correlation between DIC and higher levels of NET formation in patients with conditions such as COVID-19 or sepsis. 35,36Interestingly, genes vital for NET formation, particularly S100A8 and S100A9, showed a substantial upregulation in HAGC patients.Recent studies have shown that S100A8 and S100A9 serve as markers of NETosis and can induce aberrant activation of neutrophils. 37,38Strikingly, the blood of COVID-19 patients with thrombosis contains high levels of circulating NETs, S100A8 and S100A9, along with an abundance of immature neutrophils in the bloodstream. 39 The analysis reveals a significant association of these genes with Th17 differentiation and T cell receptor signalling pathway.
cells.Mononuclear phagocytes (MPs) showed signs of bias toward an M2-like immunosuppressive signature.Conversely, T/NK cells displayed signs of suppression of activation, although not exhaustion.Additionally, our data suggested several signalling pathways that mediate the suppression of T cells, including APP-CD74, MIF-(CD74+CXCR2), and MIF-(CD74+CD44) in HAGC.Meanwhile, S100A8 and S100A9 as well as VEGF, PDGF, FGF, and NOTCH signalling pathways that are implicated in the development of DIC were highly upregulated in HAGC neutrophils.Our study highlights significant alterations in the distribution and interactions of PBMC subsets in HAGC, reflecting a substantial remodelling of the patient's immune landscapes.This information offers a novel theoretical framework that could provide valuable insights into the precise diagnosis and treatment of HAGC, representing a significant advancement in our understanding of this disease.
are normally associated with red blood cell development.Their expression in the neutrophils of HAGC patients might reflect the influence of metastatic cancer cells on the bone marrow microenvironment, leading to abnormal neutrophil development.Additionally, the proportion of neutrophils in subcluster 6 was significantly increased in HAGC patients compared to NAGC patients and healthy individuals, while the proportions of other subclusters remained relatively unchanged (Figure 2C, D).Neutrophils in subcluster 6 expressed NETosis genes such as MPO, ELANE and PRTN3, indicating increased neutrophil NET formation in HAGC patients.

1
Single-cell RNA-seq analysis reveals abnormal cell type compositions in PBMCs from HAGC patients.(A) Study design overview: PBMCs were collected from patients with HAGC (n = 7), NAGC (n = 6), and healthy controls (n = 5).PBMCs were subjected to the microwellseq platform for scRNA-seq library preparation.(B) UMAP visualization of 51,631 single cells for all combined samples.Cells were clustered using the Leiden algorithm (resolution = 0.9).(C) Bubble plot illustrating selected marker genes for each cluster.The color scheme represents the scaled mean expression from (0 = white, 1 = red) in the group.The bubble size indicates the fraction of cells expressing marker genes in the group.(D) Split UMAP distribution displaying the clusters in PBMCs and coloured by the sample origin.(E) Scaled bar chart showing the proportion of cell types in each sample.(F) Clinical blood test data plots representing the absolute numbers (top) and fractions (bottom) of neutrophils (left panel), monocytes (middle panel), and lymphocytes (right panel) in HAGC and NAGC patients.Each dot represents a single patient.The detailed data are listed in Table

F
I G U R E 2 Legend on next page.revealed that clusters 2, 6, and 4 displayed lesser degrees of differentiation, while clusters 0 and 1 demonstrated intermediate differentiation, and clusters 5 and 3 exhibited more advanced differentiation (Figure Immature neutrophils in PBMCs of HAGC patients.(A) UMAP visualization of the neutrophil subclusters after re-clustering using the Leiden algorithm (resolution = 0.8).(B) Bubble plot displaying specific genes for each neutrophil category.The color scheme represents the scaled mean expression (0 = white, 1 = red) in the group.The bubble size indicates the fraction of cells expressing marker genes.(C) Split UMAP visualization of neutrophils, highlighting their origin (marked in red).(D) Boxplot demonstrating the proportions of neutrophil subtypes in each group.Statistical significance was determined using a two-sided unpaired Welch's t-test.(E) UMAP visualization of neutrophils coloured according to CytoTRACE scores.The color scheme represents the predicted order of differentiation, with 1 (red) indicating less differentiation and 1 (blue) indicating more differentiation.(F) Boxplot showing primary, secondary and tertiary granule scores, as well as the maturation score based on the gene sets listed in U R E 3 M2-polarization characteristics of HAGC mononuclear phagocytes and reduced T cell function in HAGC patients.(A) Boxplot illustrating the M1 and M2 scores, calculated based on the gene sets listed in Table S2, for the MP fraction of cells across different conditions.Statistical significance was determined using a two-sided unpaired Welch's t-test.(B) Bubble plot displaying the expression level of selected M1/M2-signature genes in myeloid cells from HAGC patients, NAGC patients, and healthy individuals.The color scheme represents the scaled mean expression in the group.The bubble size indicates the fraction of cells expressing marker genes under different conditions.(C) UMAP visualization demonstrating the expression patterns of a selected M1-signature gene, CCL5 (top), and an M2-signature gene, MMP9 (bottom), in each group.The color scheme represents the levels of gene expression.The UMAP plots are categorized based on the origin of disease or control.(D) scVelo RNA velocity analysis, projecting the velocity fields onto the UMAP distribution, to estimate macrophage polarization under different conditions.

4
Legend on next page.
function in HAGC patients.(A) Heatmap showing the enrichment of gene ontology terms associated with biological process in differentially expressed genes of T cells among the indicated groups.The color bar represents the -log10 transformed P value generated by Metascape.(B) Boxplots demonstrating the activation and IFN scores, calculated based on gene sets for T/NK cell populations, under different conditions.Statistical significance was determined using a two-sided unpaired Welch's t-test.(C) UMAP showing the expression of selected T cell-activation genes GZMA (top) and GZMB (bottom) across the different conditions.The color scheme indicates the expression level.(D) Boxplots illustrating the exhaustion and senescence scores, calculated based on gene sets for T/NK cell populations, under different conditions.Statistical significance was calculated using a two-sided unpaired Welch's t-test.

F
of G-MDSC markers in neutrophil subclusters Expression of M-MDSC markers in MP subclusters Fraction of cells in group I G U R E 5 Legend on next page.
exhibited an M2-like state.Together with MPs, neutrophils in HAGC patients suppress T/NK cell activity, at least partially by forming G-MDSCs and M-MDSCs.APP and MIF signalling pathways mediate this effect.Our data also shed light on the potential mechanisms underlying DIC development in HAGC.The activation of angiogenesis and fibrosis-associated signalling pathways, along with the upregulation of S100A8 and S100A9 genes in neutrophils, provides insights into the pathogenesis of DIC in HAGC (Figure 6D).It is worth noting that no neutrophil expansion was observed in previous scRNA-seq studies of stomach tissues and PBMCs in non-HAGC gastric cancer patients. 7-10Neutrophil dysfunction has not previously been acknowledged as a characteristic of GC.Neutrophils play a crucial role as the first line of defense against infectious pathogens in the human body.However, clinical examination of HAGC patients revealed no signs of infection, suggesting that neutrophils may respond to other inflammatory or immunostimulatory signals resulting from HAGC.Notably, our study identified a significant increase in the ratio of immature neutrophils resembling

4 | METHODS 4 . 1 |
Inhibition of S100A8/A9 by the inhibitor paquinimod has demonstrated effective amelioration of the immune disorder caused by COVID-19. 38Considering the similar innate immune system abnormalities observed in HAGC and severe COVID-19 patients, S100A8 and S100A9 may be promising diagnostic and therapeutic targets for DIC in HAGC patients.In summary, our study provides important insights into the molecular and cellular mechanisms involved in HAGC patient immune response.We demonstrated the induction of activated but aberrant immature neutrophils in HAGC, which contribute to the immunosuppression of lymphoid cells and the activation of signalling pathways associated with DIC development.Our findings highlight the role of innate immunity in the pathogenesis of HAGC.Additionally, the identification of S100A8 and S100A9 as potential diagnostic and therapeutic targets further emphasizes their clinical significance in HAGC management.Collection of samples and clinical data This study was approved by the Clinical Research Ethics Committee in the Sixth Affiliated Hospital, Sun Yat-sen University (Number 2021ZSLYEC-090).Informed consent was obtained from all patients before the study.The study included patients newly diagnosed with HAGC according to the diagnostic criteria for HAGC. 2 NAGC is defined as unresectable, advanced, or recurrent GC according to the World Health Organization (WHO) criteria.HAGC was identified as gastric cancer characterized by BMM and DIC, with DIC being defined by the criteria established by the International Society on Thrombosis F I G U R E 5 Suppressive crosstalk of myeloid cells on T cells.(A) Bubble plots showing the expression of G-MDSC marker genes in neutrophil subclusters (top) and expression of M-MDSC marker genes in MP clusters (bottom) across different conditions.The color scheme is based on the scaled mean expression in the respective group, ranging from 0 (white) to 1 (red).The bubble size indicates the fraction of cells expressing marker genes in each group.(B) Heatmap illustrating the count of interacting ligand-receptor pairs between each cell type in HAGC cells, generated using CellChat.The color scheme represents the number of interacting receptor-ligand pairs.(C) Bubble plot displaying ligand-receptor pair putative interactions between neutrophils and T/NK cells, divided by diseased/healthy conditions.The color bar indicates the communication probability of each pair in each group.The size of the bubbles corresponds to the P-value.(D) Volcano plot depicting the differentially expression of genes in PBMCs between HAGC and NAGC patients.The plot was generated based on fold change >10 and q-value < 0.05 calculations obtained from bulk RNA-seq.(E) Gene Set Enrichment Analysis (GSEA) results for differentially expressed genes between PBMCs of HAGC and NAGC patients.

F
I G U R E 6 Enhanced fibrosis and angiogenesis signalling pathways and S100 expression in HAGC patients.(A) Circle plot showing the inferred VEGF/PDGF/FGF/NOTCH signalling pathway network strength in the HAGC group.(B) Gene set enrichment analysis (GSEA) results for differentially expressed genes between PBMCs of HAGC and NAGC patients.The analysis reveals a significant association of these genes with complement and coagulation cascades and neutrophil extracellular trap formation.(C) ELISA assay reveals that S100A8 protein is significantly upregulated in PBMCs of HAGC patients as compared to NAGC patients and healthy individuals.Statistical significance was determined using one-way ANOVA.(D) Schematic model depicting the underlying mechanism by which HAGC contributes to the development of a complex systemic DIC coupled with immunosuppression of lymphoid cells.