Serum amyloid A 1 induces suppressive neutrophils through the Toll‐like receptor 2–mediated signaling pathway to promote progression of breast cancer

Abstract Immune inflammation plays a key role in breast cancer development, progression, and therapeutic efficacy. Neutrophils are crucial for the regulation of the suppressive tumor microenvironment and are associated with poor clinical survival. However, the mechanisms underlying the activation of suppressive neutrophils in breast cancer are poorly understood. Here, we report that breast cancer cells secrete abundant serum amyloid A 1 (SAA1), which is associated with the accumulation of suppressive neutrophils. High expression of SAA1 in breast cancer induces neutrophil immunosuppressive cytokine production through the activation of Toll‐like receptor 2 (TLR2)‐mediated signaling pathways. These include the TLR2/myeloid differentiation primary response 88 (MYD88)‐mediated PI3K/nuclear factor‐κB signaling pathway and p38 MAPK‐associated apoptosis resistance pathway, which eventually promote the progression of breast cancer. Our study shows a mechanistic link between breast cancer cell secretion of SAA1 and suppressive neutrophils that potentiate tumor progression. These findings provide potential therapeutic targets for breast cancer.


| INTRODUC TI ON
Breast cancer is the most commonly diagnosed malignancy and the leading cause of cancer death among women worldwide. 1 Although the death rate from breast cancer has slightly declined due to advances in diagnosis and treatment, a considerable number of patients still have a poor prognosis due to breast cancer heterogeneity. 2,3 Therefore, it is critical to develop innovative strategies for breast cancer therapy.
Known as a hallmark of cancer, immune inflammation plays a predominant role in tumor development and progression.
Immunotherapeutic agents, most notably those triggering immune checkpoint blockade, definitely improve the treatment benefit in various solid tumors. Therefore, these could also be useful in in breast cancer. [4][5][6][7] Increasing evidence has shown that the immune system is constrained in the immunosuppressive tumor microenvironment (TME), which contains multiple immune cells participating in a dynamic and complex cross-talk with cancer cells. [8][9][10] In the TME, neutrophils are called tumor-associated neutrophils (TANs), which are important inflammatory immune cells, and have an immunosuppressive phenotype that promotes carcinogenesis and cancer cell proliferation, metastasis, and immune evasion. 11,12 Recent studies have reported the potential mechanisms by which TANs participate in tumor promotion and development. 13 We previously reported that the prognosis of luminal breast cancer correlates with immune responses. 14 Hence, neutrophils could play potential roles as predictive biomarkers and be used for therapeutic intervention in breast cancer. However, the mechanisms that govern the activation of suppressive neutrophils in breast cancer and how these cells perform immunosuppressive activities remain unclear.
During inflammation, cancer cells undergo adaptive transformation and acquire immunoediting characteristics to maintain sustained survival signals and escape the immune response. [15][16][17] The acquisition and maintenance of the neutrophil immunosuppressive phenotype is accompanied by the secretion of cytokines, chemokines, and adhesion molecules by coexisting tumor cells, emphasizing the complex interplay between cancers and neutrophils. 18,19 Serum amyloid A (SAA) is one of the most prominent acute phase proteins secreted by cancer cells and confers cytokine-like properties in response to inflammation. [20][21][22] Serum amyloid A levels are significantly elevated in several malignancies 23 and correlate with poor prognosis. Previous studies have reported that SAA is a critical mediator of pro-inflammatory cytokine production and recruitment of immune cells, and thereby promotes tumor development by facilitating angiogenesis, tumor invasion, and immunosuppression. 24,25 Recent evidence has shown that SAA mediates systemic neutrophil function in specific contexts of inflammation, suggesting that it might influence neutrophilmediated tumor development. 26 Here, we investigated whether SAA was elevated in the breast cancer TME and its effects on neutrophils, thus unfolding the mechanism underlying neutrophil involvement in cancer, inflammation, and immunomodulatory processes.

| Bioinformatics analysis
Gene expression datasets were obtained from The Cancer Genome Atlas (TCGA) data portal (https://portal.gdc.cancer.gov). One thousand two hundred seventeen TCGA samples were grouped into breast cancer and healthy control groups. R implementation of the DESeq2 package was used for differential expression analysis.
Differentially expressed genes (DEGs) are sorted according to the adjusted p value threshold (false discovery rate [FDR] method) of 0.05. Functional enrichment analysis was carried out using DAVID (https://david.ncifc rf.gov) with an adjusted p value (FDR correction) with cut-off of 0.05.
Relevance of SAA1 expression to neutrophils from the TCGA database was analyzed using the TIMER 2.0 algorithm (http://timer. cistr ome.org/). The protein-protein interaction (PPI) network related to SAA1 gene and biological analyses were undertaken using the STRING online database. Finally, the network was imported to Cytoscape for visualization, and MCODE was used to identify functional submodules in PPI.

| Sample collection
This study enrolled patients pathologically diagnosed with breast cancer in Harbin Medical University Cancer Hospital and 38 healthy controls between January 1, 2017 and June 30, 2020.
The characteristics of the breast cancer patients are summarized in Table 1. All patients were diagnosed by complete clinical and pathologic data, received adjuvant standard chemotherapy with or without anti-human epithelial receptor 2 (HER2) therapy. None of the patients received granulocyte/macrophage colony-stimulating factor treatment for neutropenia. Peripheral blood samples (5 ml) were collected from each breast cancer patient and healthy control.
For each sample, 1 ml was centrifuged to collect plasma for SAA1 analysis by ELISA; others were used for neutrophil and leukocyte isolation and cell culture. This study protocol was approved by the

Institutional Review Board of Harbin Medical University Cancer
Hospital. Written informed consent was obtained from the patients and healthy controls.

| Neutrophils and leukocytes isolation
Neutrophils were separated from 3 ml peripheral blood samples using Ficoll-Hypaque density gradient centrifugation, according to the manufacturer's instructions (GE Healthcare). The purity of neutrophils was within acceptable parameters (higher than 95%).
Peripheral blood samples (1 ml) were lysed twice with RBC lysis buffer (BD Biosciences) to obtain leukocytes.

| Enzyme-linked immunosorbent assay
The expression levels of SAA1 in plasma collected from breast cancer patients or breast cancer cell lines, as well as interleukin-10 (IL-10), arginase (Arg), and inducible nitric oxide synthase (iNOS) from the supernatant of neutrophils, were quantified with the appropriate Quantikine Kit (R&D Systems) according to the manufacturer's instructions.

| Flow cytometry
The leukocytes isolated from breast cancer patients and healthy con-

| Immunofluorescence
For immunofluorescent staining the neutrophils treated with or without SAA1 and TLR2-blocking Ab were fixed in 4% paraformaldehyde and then permeabilized with 0.5% Triton X-100.

| Western blot analysis
Breast cancer cells and neutrophils were each homogenized in 0.5 ml RIPA buffer. Polyacrylamide gels (10%) were used for protein testing. Membranes were incubated with anti-SAA1 Ab (1:3000 dilution, ab2302; Abcam). The specific bands were visualized using ECL reagent. The bands were analyzed using Bio-Rad Chemi EQ densitometer and Bio-Rad Quantity One software (Bio-Rad Laboratories).

| Statistical analysis
The correlations among clinicopathologic factors and the expression of SAA1 were analyzed by using Fisher's exact test and the χ 2 test. Data are expressed as mean ± SD. p Values <0.05 were considered statistically significant with an unpaired two-tailed Student's t test and one-way ANOVA test. All datasets were tested for normal distribution.

| Serum amyloid A 1 is secreted by breast cancer cells, and its expression is increased in highgrade breast cancer patients
To investigate how gene expression alteration contributes to pathogenesis of breast cancer, we analyzed the DEGs of breast cancer patients and healthy controls using the RNA sequencing data from the TCGA database. Analysis results of gene expression profile data were visualized with volcano plots ( Figure 1A) and nonsupervised hierarchical clustering ( Figure 1B and Table S1). To gain more information about molecular functions and pathways that are affected by DEGs, the DAVID database was used to perform the functional enrichment analysis of all genes ( Figure 1C). The DEGs identified in breast cancers were involved in a wide range of molecular functions associated with immune response, such as signaling receptor activator activity, receptor ligand activity, cytokine activity, and chemokine receptor binding ( Figure 1C).
Secreted by cancer cells, SAA is an acute-phase protein involved in many malignancies. 23,25 The human SAA family comprises two SAA isoforms, SAA1 and SAA2. First, we detected the SAA1 isoform that was mainly expressed in human breast cancer cells, but not in the nontumorigenic cell line MCF-10 ( Figure S1A). To evaluate whether the expression of SAA1 was upregulated in breast cancer, we collected peripheral blood samples from 178 breast cancer patients and 38 healthy controls to detect the concentrations of SAA1 in plasma. As shown in Figure 1D, SAA1 expression was significantly elevated in breast cancer patients. Furthermore, to investigate the potential role of SAA1 in breast cancer, we analyzed the clinical data of breast cancer patients (Table 1) and classified the patients into high and low SAA1 groups (optimal cut-off value, 7.68) according to the receiver operating characteristic curve ( Figure S1B).
We found that patients with SAA1 ≥7.68 often had a larger tumor size, lymphatic metastasis, and higher incidence of advanced clinical stages ( In addition, we found a high expression of SAA1 in breast cancer tissue samples and a positive correlation between SAA1 expression and disease staging ( Figure 1E), which was consistent with the results for SAA1 levels in peripheral blood samples from patients with different stages of breast cancer ( Figure S1C). Next, survival curves were evaluated based on the SAA1 levels and indicated that patients with higher SAA1 expression had shorter OS and DFS than those with low SAA1 levels ( Figure S1D). Together, these data indicate that SAA1 expression is upregulated and is correlated with higher disease grade and poor prognosis in breast cancer patients.

| Immunosuppressive neutrophils are associated with SAA1 secretion in breast cancer
Neutrophils are the most abundant inflammatory immune cells in the TME and usually exert an immunosuppressive phenotype that promotes cancer progression; 12,27,28 neutrophils are regulated by signals produced by cancer cells. 11,29 Thus, we hypothesized that an abundant expression of SAA1 could regulate neutrophils in breast cancer.
To address this possibility, we used the online platform TIMER 2.0 to analyze the correlation between SAA1 expression level and the infiltration level of neutrophils in breast cancer (Figure 2A). The results showed that the expression level of SAA1 was positively correlated with the infiltration level of neutrophils in breast cancer. To evaluate whether the level of neutrophils was increased in breast cancer, we also analyzed the positive rate of infiltration of neutrophils, which was significantly higher in breast cancers than in healthy controls ( Figure 2B). The CD15 + and CD11b + populations of neutrophils were more susceptible to SAA treatment as reported previously, 23 then total leukocytes were extracted from breast cancer patients and healthy controls, and CD11b + CD15 + neutrophil levels were determined. We observed that the levels of CD11b + CD15 + neutrophils were significantly increased in patients with breast cancer than in healthy controls ( Figure 2C). Additionally, the immunosuppressive cytokines IL-10, Arg, and iNOS, were significantly higher in neutrophils from breast cancer patients than in those from the control group ( Figure 2D), suggesting that the neutrophils in breast cancer presented an immunosuppressive phenotype. CD66b + neutrophils are reported to be an independent prognostic factor of cancer, 30,31 thus we undertook IHC staining to evaluate SAA1 and CD66b + neutrophils in tissue samples from breast cancer patients. As shown in Figure 2E, elevated SAA1 expression was associated with abundant neutrophil infiltration in breast cancer tissue samples (p = 0.0056).
In addition, the SAA1 expression levels in plasma were positively correlated with neutrophils (Table S2) as well as the expression levels of IL-10, Arg, and iNOS, which were mainly secreted by neutrophils ( Figure 2F). These results suggest that SAA1 potentially contributes to the accumulation of immunosuppressive neutrophils in breast cancer.

| Serum amyloid A 1 facilitates neutrophil secretion of immunosuppressive cytokines through the TLR2 receptor
To determine the mechanism by which SAA1 induces immunosuppressive neutrophils, STRING database was searched to extract the PPI network related to the SAA1 gene ( Figure 3A). Results showed that SAA1 was connected to genes with important functions, such Interestingly, we found that the expression levels of both FPRL1 and TLR2 were elevated in SAA1-treated neutrophils ( Figure 3C), but whether these two receptors played a role in regulating the secretion of immunosuppressive cytokines was unknown. To identify the effects of FPRL1 and TLR2 in SAA1-treated neutrophils, the neutrophils were incubated with FPRL1-and TLR2-blocking Abs.
As shown in Figure 3D, the FPRL1-blocking Ab had no effect on the levels of inflammatory cytokines. In contrast, the addition of a  Figure 3D). These results were further confirmed by immunofluorescence ( Figure 3E). Together, these data indicate that SAA1 facilitates the production of immunosuppressive cytokines in neutrophils through the TLR2 receptor.

| Serum amyloid A 1 stimulates neutrophils to exert immunosuppressive functions through the TLR2/MYD88-mediated PI3K/NF-κB signaling pathway
Next, we examined the impact of TLR2 on neutrophils activated by SAA1. Toll-like receptor 2 is thought to contribute to signal transduction in a variety of immune cells, but it is not known whether TLR2 is required for the activation of downstream signals in neutrophils. [34][35][36] We found that the expression level of MYD88, which is an adaptor protein for TLR2, was significantly increased in SAA1-treated neutrophils ( Figure 4A). In addition, SAA1 promoted the expression of PI3K, p-AKT, and p-p65/NF-κB in neutrophils, and the addition of a TLR2-blocking Ab, a PI3K inhibitor, and an NF-κB inhibitor repressed the expression of these proteins ( Figure 4A). These data suggested that SAA1 could trigger the TLR2/MYD88-mediated PI3K and NF-κB pathway in neutrophils. To determine whether this signaling pathway influenced neutrophil cytokine production, we evaluated the synthesis of IL-10, Arg, and iNOS. As shown in Figure 4B

| Serum amyloid A 1 induces neutrophil apoptosis resistance through the p38 MAPK pathway
In addition to the effects of SAA1 on neutrophils mentioned above, we also observed that SAA1-treated neutrophils showed appreciably lower apoptotic levels than untreated cells ( Figure 4C), indicating that SAA1 plays a crucial role in mediating neutrophil apoptosis.
It has been reported that the p38 MAPK pathway is involved in TLR2-mediated apoptosis. 37,38 Therefore, the ratio of apoptotic cells in SAA1-treated neutrophils was investigated with and without a TLR2-blocking Ab and p38 inhibitor. We observed that the suppression of the TLR2 and p38 MAPK pathways stimulated the apoptotic process ( Figure 4C), suggesting that SAA1 might induce neutrophil apoptosis resistance through the p38 MAPK pathway. To further evaluate potential apoptotic signals, we detected the expression of several apoptotic proteins in SAA1-treated and untreated neutrophils. As shown in Figure 4D, SAA1 promoted the expression of p-p38, Bcl-2, and cleaved caspase-3 and activated TLR2 and MYD88; this promotion was inhibited by TLR2 and p38 MAPK suppression, further confirming the ability of SAA1 to promote neutrophil apoptosis resistance. We next determined whether SAA1-mediated apoptosis resistance could influence the production of IL-10, Arg, and iNOS in neutrophils. As expected, the addition of a p38 inhibitor prevented SAA1-treated cytokine secretion by neutrophils from patients with breast cancer and healthy controls ( Figure 4E).

| Growth-promoting effects of SAA1 expression and immunosuppressive neutrophils in breast cancer cells are recapitulated in vivo
Finally, we sought to validate the correlation of SAA1 expression and immunosuppressive neutrophils in breast cancer cells in vivo.
Towards this, we established implanted xenografts of 4T1 cells and compared the growth of mice treated with or without anti-SAA1 and anti-Ly6G Ab to block SAA1 and neutrophils ( Figure 5A).
Tumor growth in animals with anti-SAA1 or anti-Ly6G blocking was significantly inhibited ( Figure 5A,B). The lowest rate of growth was observed in animals treated with anti-SAA1 and anti-Ly6G Ab ( Figure 5A,B). Additionally, mice treated with control IgG showed a notable increase in SAA1 and CD66b expression, whereas animals treated with anti-SAA1 Ab showed decreased SAA1 as well as CD66b ( Figures 5C and S1H). Furthermore, to investigate that immunosuppressive function of neutrophils is regulated by the SAA1/neutrophil blockade, infiltration of CD3 + T cells were also evaluated ( Figure   S1G,H). The results showed that after anti-SAA1 and anti-Ly6G Ab treatment, the average numbers of CD3 + cells increased, which is probably due to the release of neutrophil inhibition, supporting our belief that SAA1 might be associated with immunosuppressive neutrophils in breast cancer. Taken together, these findings indicate that SAA1 could promote neutrophils to produce immunosuppressive but also by mediating apoptosis resistance ( Figure 6).

| DISCUSS ION
In this study, we detected the high level of SAA expression in breast cancer and investigated its effects on the regulation of immunosuppressive neutrophils in the TME. Our data showed that SAA1  42 Neutrophil immunosuppression can occur by the release of a variety of suppressive inflammatory cytokines in solid tumors. 43 Several mechanisms have been proposed to explain how neutrophils are a substantial obstacle to successful immunotherapy. 13 Neutrophils elicited from cancer cells can prevent T cell proliferation through PD-1/PD-L1 signaling and inhibit natural killer cell functions, leading to tumor cell proliferation, survival, and metastasis. 44 Several suppressive factors, such as Arg and iNOS, produced by neutrophils can inhibit immune responses and infiltration of CD8 + T lymphocytes into the TME. 45 There are many markers that can be expressed by neutrophils, such as CD11b, CD66b, CD15, and MPO as reported previously. 31 Consistent with previous observations, we found here that CD11b + CD15 + neutrophils was more susceptible to SAA treatment 23 and had an increased ability to secrete immunosuppressive molecules, including IL-10, Arg, and iNOS, confirming the immunosuppressive role of neutrophils in breast cancer.
However, the intrinsic mechanisms that mediate the acquisition of immunosuppressive characteristics by neutrophils remain unknown.
There is emerging experimental and clinical evidence indicating that the regulation of the phenotype and function of neutrophils requires the participation of a series of inflammatory mediators.
Serum amyloid A is an apolipoprotein considered to be a biomarker of inflammation and tumor progression. The levels of SAA appeared to increase in malignancies, such as breast cancer, 46 lung cancer, 25 and melanoma. 23 As SAA regulates the immune system and immunosuppressive neutrophils accumulate in the breast cancer TME, we hypothesized that there might be a strong link between SAA1 and immunosuppressive neutrophils. As expected, SAA1 secreted by cancer cells was found to contribute to the promotion of the immunosuppressive phenotype of neutrophils, resulting in tumor progression. These data suggest a possible feedback loop between breast cancer cells and immunosuppressive neutrophils through SAA.
Inhibition of neutrophils or SAA1-associated TLR2 on neutrophils by blocking Abs thus provided us a potential intervention approach in breast cancer treatment.
We then sought to determine the potential mechanisms by which SAA induces immunosuppression of neutrophils. Toll-like both expressed on SAA1-treated neutrophils in our study. Although both of these receptors are responsible for SAA-mediated biological activities, TLR2 seems to play a more prominent role in SAA-treated neutrophils and is significant in recognizing a wide range of molecules required for downstream signal transduction. 49,50 We found that MYD88, which was reported to induce the Ras-mediated signaling pathway and engage in protumorigenic function in many immune cells [51][52][53] was the most important adaptor protein for TLR2 in neutrophils. We confirmed that cytokine production by immunosup- SAA promotes neutrophil apoptosis resistance, which decreases the life span of neutrophils in the TME. Both of these effects of SAA on immunosuppressive neutrophils result in the progression of breast cancer ( Figure 6).
There are some limitations of our study. For example, the substantial experiment should be conducted to support SAA1 derived from cancer cell, and the other cytokines released from tumor cell or TME could have somewhat impact on immunosuppressive neutrophils, the exploration for interaction between SAA1 and other cytokines would meaningful in the future.
In conclusion, our study describes a previously unknown mechanism underlying the effect of SAA on immunosuppressive neutrophils in breast cancer. These findings provide new insights into the mechanisms by which neutrophils are regulated in the TME and provide potential therapeutic targets for breast cancer patients.

D I SCLOS U R E
The authors have no conflict of interest. Secreted by breast cancer cells, SAA1 activates the Toll-like receptor 2 (TLR2)/myeloid differentiation primary response 88 (MYD88) regulatory pathway in neutrophils. Activation of TLR2/MYD88 stimulates the PI3K/nuclear factor-κB (NF-κB) signaling pathway; however, SAA1 bound to TLR2 receptor induces neutrophil apoptosis resistance through the p38 MAPK pathway. Both effects of SAA1 on neutrophils promote the secretion of interleukin-10 (IL-10), arginase (Arg), and inducible nitric oxide synthase (iNOS) from neutrophils, which exert an immunosuppressive phenotype in breast cancer. Thus, there is a potential positive feedback loop between breast cancer cells and immunosuppressive neutrophils through SAA1