Tumoral IL-10-activated SHP2 in macrophages promotes mammary carcinoma progression.

1
Introduction
The Src homology 2 domain-containing tyrosine phosphatase 2 (SHP2), a ubiquitously expressed cytoplasmic protein tyrosine phosphatase, plays an important role in tumor immunity [1]. The tyrosine phosphorylation of SHP2 results in the release of the autoinhibitory interface and provides the active phosphatase site for substrate recognition [2]. Activated SHP2 is involved in the negative regulation of immune responses through the impact on JAK-STAT activation and inhibitory receptor signaling pathways. Despite its function in the inhibitory pathways, T cell specific-deletion SHP2 failed to improve anti-tumor immunity but instead had a detrimental effect on tumor progression [3,4]. Conversely, conditional knockout of SHP2 in myeloid cells diminished tumor growth and enhanced anti-tumor responses [5]. Thus, SHP2 functions in a cellular context-dependent manner in tumor immunity.
Macrophages are one of the major constituents of tumor stroma in many solid tumors, including breast cancer, and act as a key orchestrator in shaping tumor microenvironment [6,7]. As a heterogeneous and plastic group of cells, macrophages have different functions, morphologies and phenotypic properties depending on the specific tissue microenvironment [8]. Once recruited to neoplatic tissue, macrophages are present at all stages of tumor progression and exhibit heterogeneous responses that lead to both pro- and anti-tumor properties [9,10]. Tumor-infiltrating macrophages have been characterized as an immunostimulatory phenotype and produce proinflammatory cytokines in the early-stage carcinogenesis, while there are an increased number of immunosuppressive macrophages at the late stage [11,12]. Tumor cells can condition macrophages in the tumor microenvironment so that these cells are more permissive for tumor growth and metastasis. However, the underlying mechanisms by which their crosstalk dictates the phenotypic and functional properties of macrophages remain largely unclear. Understanding how tumor cells co-opt macrophage plasticity to facilitate tumor progression will allow for the identification of pharmacological targets.
Our previous study demonstrated that SHP2 inhibition enhanced pro-inflammatory activities of tumor-associated macrophages (TAMs) and improved anti-tumor immune responses in murine models of colorectal cancer [13,14]. Interestingly, recent findings have shown that SHP2 inhibitors facilitated TAMs into immunosuppressive programming and enhanced the migration and invasion abilities of colorectal cancers [15]. To resolve the perplexing effects of SHP-2 on TAMs, the present study describes the crosstalk between macrophages and mammary carcinoma cells in tumor microenvironments, wherein macrophages triggered the induction of IL-10 in the tumor cells. Reciprocally, tumoral IL-10 activated SHP2 and rendered macrophages immunosuppressive phenotype and function.

2
Materials and methods/experiment
2.1
Reagents
PHPS-1 sodium salt hydrate (a tyrosine phosphatase inhibitor of SHP2), thioglycollate broth and collagenase IV were purchased from Sigma–Aldrich (St. Louis, MO). DNase I was purchased from Roche (Basel, Switzerland). The inhibitors PP2, UNC2025 and Piceatannol were purchased from MedChem Express (Monmouth Junction, NJ). Fluorescently labeled antibodies were purchased from eBioscience (CD11b, F4/80, CD45, CD3, CD4, CD8, IL-10, IL-4 and IFN-γ, Santa Cruz, CA) or BD Biosciences (pSHP2 (Tyr542), CD68, CD206, CD86 and CD11c, San Jose, CA). Anti-pSHP2 (Tyr542), anti-p-STAT1 (Tyr701), anti-STAT and anti-IL-10 antibodies were purchased from Cell Signaling Technology (Beverly, MA). Anti-SHP2 and anti-p-Scr antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-Src and anti-IL-10RA antibodies were from Proteintech (Chicago, IL). Anti-β-Actin and anti-tublin antibodies were from Abmart (Shanghai, China). Recombinant mouse IL-10 and IFN-α were from BioLegend (San Diego, CA). ELISA Kits for mouse IL-6, TNF-α and IL-10 were purchased from Dakewe Biotech Co. Ltd (Shenzhen, China). Anti-F4/80, anti-CD4, and anti-CD8 magnetic beads were from Miltenyi (Bergisch Gladbach, Germany). All other chemicals were purchased from Sigma-Aldrich.

2.2
Mice
Six-week-old C57BL/6 mice were obtained from Model Animal Research Center of Nanjing University (Nanjing). SHP2flox/flox mice on a C57BL/6 background were bred with Lyz2-Cre+/− on the same background to yield Cre+/- SHP2△Mac mice (SHP2lyz-/−, KO) and Cre−/- SHP2flox/flox littermate control mice (WT). Genotyping and sequence confirmation were performed by PCR analyses of tail genomic DNA as previously described [13].

2.3
Cell culture
Murine macrophage RAW264.7, murine colon adenocarcinoma cell MC38, murine lung carcinoma cell LLC and murine melanoma cell B16BL6 were purchased from the Cell Bank Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China), who provided authentication certificates. Murine breast adenocarcinoma cell E0771 was kindly provided by professor Siying Wang (Anhui Medical University, China). Peritoneal macrophages (PMs) elicited by thioglycollate broth were harvested by lavage of the peritoneal cavity. Bone marrow-derived macrophages (BMDMs) were differentiated with M-CSF and bone marrow-derived dendritic cells (BMDCs) were differentiated with GM-CSF as previously described [16,17]. Cells were maintained in DMEM or RPMI-1640 medium (Life Technologies, Grand Island, NY) supplemented with 10% fetal bovine serum (GIBCO, Grand Island), 100 U/mL penicillin and 100 µg/mL streptomycin at 37 °C in a humidified atmosphere containing 5% CO2.

2.4
Mammary tumor implantation and treatment
E0771 cells (2.5 × 105 cells in 20 µL PBS) were subcutaneously injected near the fat pad of the fourth mammary gland in the lower abdomen of C57BL/6 mice [18]. The tumor volumes were measured and calculated using the following formula: 0.5236 × L1 × (L2)2, where L1 and L2 are the long and short axes of the tumor, respectively. At the end of the experiments, the lung tissues were examined to identify metastases. In some experiment, tumors were resected on days 21, 28 and 35 after the day of tumor cell inoculation. Lung tissues with metastases were either photographed for gross morphology or further analyzed by histology. In another animal experiment, female WT and SHP2lyz-/- mice were inoculated with E0771 cells. After 7 days, the mice were randomly distributed into 4 groups. Two of the groups were treated with PBS or PHPS-1 (1 mg/kg/day) intraperitoneally for another 21 days. Animal welfare and experimental procedures were performed in compliance with the guidelines “The Detailed Rules and Regulations of Medical Animal Experiments Administration and Implementation” (Order No. 1998–55, Ministry of Public Health, China) and approved by the Science and Technology Ethics Committee of Nanjing University (IACUC-2210009). All efforts were made to minimize animal suffering and to reduce the number of animals used.

2.5
Isolation of tumor-infiltrating macrophages
Mammary tumor tissues were minced and digested with 0.5 mg/mL collagenase IV and 0.01 mg/mL DNase I in RPMI 1640 supplemented with 5% FCS for 1 h at 37 °C. The cell suspension was then filtered through a 70-µm nylon mesh, layered on a Percoll gradient (30%−70%) and centrifuged at 800 × g for 20 min. The separated tumor-infiltrating lymphocyte (TIL) fraction was further selected using anti-F4/80 magnetic beads.

2.6
Flow cytometry
Tumor-infiltrating lymphocytes and TAMs were incubated with specific surface-binding antibodies for 15–20 min at room temperature. For intracellular IL-4, IL-10 or pSHP2 staining, cells were pretreated with Cytofix/Cytoperm Kit reagents (BD Biosciences). Samples were analyzed by flow cytometry on a FACScan flow cytometer (Becton Dickinson, Franklin Lakes, NJ).

2.7
Western blotting and immunoprecipitation
Western blot and immunoprecipitation analysis were performed as described previously [19]. The densitometry of immunoblots was quantified with Image J software (NIH, Bethesda, MD) and normalized with loading controls.

2.8
Histology and immunostaining
Human breast cancer tissue microarray (#HBreD090CS01) was purchased from Shanghai Outdo Biotech Co., LTD (Shanghai), containing 45 pairs of primary breast invasive ductal carcinoma tissues and their para-cancerous tissues. For the immunofluorescence analyses, paraffin-embedded sections were stained with fluorescently labeled antibodies as previously described [4]. Positive expression was based upon the score (0–300) that was a product of intensity of stain. Low expression was characterized as +/-. High expression (+/+) had scores between 150 and 300. At least three random measurements were taken per tissue section, and the average of the scores was calculated. Mouse tumor and lung tissues were gradually dehydrated in alcohol, embedded in paraffin and cut into 5 µm of sections. Paraffin-embedded sections were deparaffinized and successively stained with hematoxylin and eosin (H&E). For pulmonary nodule enumeration, the number of metastatic foci in H&E-stained lung sections was counted in a blinded fashion.

2.9
Cytokine analysis by ELISA
Sera obtained from tumor-bearing mice or culture supernatant were stored at −80 °C until use. The levels of IL-6, TNF-α, and IL-10 were measured using specific ELISA kits according to the manufacturer's instructions.

2.10
Quantitative RT-PCR
As described previously [17], total RNA was extracted from cells and reverse transcribed to cDNA and subjected to quantitative PCR, which was performed with the BioRad CFX96 ouchTM Real-Time PCR Detection System (BioRad, CA) using iQTMSYBR1Green supermix (BioRad), and threshold cycle numbers were obtained using BioRad CFX manager software. The program for amplification was 1 cycle of 95 °C for 2 min followed by 40 cycles of 95 °C for 10 s, 60 °C for 30 s, and 72 °C for 30 s. The primer sequences used in this study were listed in Supplemental Table 1.

2.11
Adoptive transfer experiments
E0771 cells (2.5 × 105 cells in 20 µL PBS) were subcutaneously injected near the fat pad of the fourth mammary gland of WT or SHP2lyz-/− mice. After 28 days, F4/80+ macrophages were enriched from the tumors using specific magnetic beads. The isolated TAMs (1 × 105 / mouse) were co-injected with E0771 cells (2 × 105 / mouse) into female C57BL/6 mice, and tumor growth was monitored.

2.12
RNA sequencing
PMs form WT or SHP2lyz-/− mice were co-cultured with E0771 cells in a transwell for 24 h. The total RNA was extracted from PMs using Trizol Reagent (Thermo Fisher Scientific, Waltham, MA). RNA-seq analysis was carried out on Illumina HiSeq platform by Novogene Co., Ltd. (Beijing, China) as described previously [20]. The genes with adjusted p-values less than 0.05 were chose to draw heat map. ClusterProfiler R package was applied to examine the statistical enrichment of differential expression genes in KEGG pathways. Gene Set Enrichment Analysis (GSEA) was performed across the Molecular Signatures Database to identify the molecular pathways in PMs co-cultured with E0771 cells.

2.13
Transfection
Chemically synthesized 21-nucleotide siRNA was obtained from Genescript (Nanjing). Murine IL-10R siRNA sequences were 5′-CCUGAGCAUCUUAGUCAUA-3′. Luciferase siRNA was used as a negative control. BMDM cells were electroporated with IL-10R siRNA using Gemini X2 Electroporation System (BTX, Holliston, MA). ShRNA-IL-10–1 and shRNA-IL-10–2 were designed and cloned into pLKD-CMV-G&PR-U6-shRNA vectors. Stably E0771 cells with IL-10 knockdown were selected with 1 µg/mL puromycin and maintained in the culture medium with 0.5 µg/mL puromycin.

2.14
Statistical analysis
Data are presented as the means ± SD of at least three independent experiments. Statistical significance was calculated using two-tailed Student's t-tests for comparisons between two groups and one-way ANOVA followed by post-hoc Tukey HSD (Honestly Significant Difference) test for comparisons among more than two groups. P < 0.05 was considered significant.

3
Results
3.1
SHP2 activation is enhanced in TAMs of breast cancer
In order to assess the status of SHP2 activation in TAMs from breast cancer tissues, we used an orthotopic-grafted breast cancer model generated by injecting murine breast adenocarcinoma cell E0771 into female C57BL/6 mice. Its orthotopic implantation represents a breast cancer model [18,21]. The levels of phosphorylated SHP2 (pSHP2) at Tyr542 were monitored during mammary carcinoma progression and greatly increased in a time-dependent manner in CD11b+F4/80+ TAMs from the TIL cell populations (Figs. 1a, S1). Similar results were confirmed in enriched F4/80+TAMs by western blotting (Fig. 1b). pSHP2 levels increased progressively, even though the total SHP2 expression did not change significantly. To test whether SHP2 activation in TAMs is attributed to breast cancer microenvironment, we cultured mouse BMDMs with syngeneic E0771 cells. Remarkably, the admixed tumor cells promoted the elevation of pSHP2 levels in BMDMs in a time-dependent manner (Fig. 1c). Co-culture of both cells in the presence of a transwell induced SHP2 activation in BMDMs to a comparable extent. Similar results were obtained after co-culture of PMs or Raw264.7 cells with E0771 cells (Fig. 1d and e). These results indicate that tumor cells could activate SHP2 in macrophages through secreting soluble mediators. When co-cultured in the presence of a transwell with other murine carcinoma cells, including colon adenocarcinoma cell MC38, lung carcinoma cell LLC and melanoma cell B16BL6, the elevation of pSHP2 levels was also observed in Raw264.7 cells (Fig. S2). In parallel to those in co-culture system, immunofluorescence detection of CD68+ macrophages in human breast cancer tissues revealed an increase in pSHP2 expression compared with those in adjacent tissues (Fig. 1f). The percentage of the samples with pSHP2 high (+/+) expression level is 57.8% in tumors, while 37.8% in adjacent tissues (Table 1). These results implicate a link between the activation of SHP2 in TAMs and breast cancer microenvironment, through which the microenvironment manipulates TAM properties.

3.2
Loss of SHP2 in macrophages is protective against mammary carcinoma progression
To determine the role of SHP2 in macrophages in fostering breast cancer development, we investigated the growth and metastasis of mammary carcinomas using macrophage-specific SHP2 knockout (SHP2lyz-/−) mice (Fig. 2a) [13]. Both tumor volume and weight were remarkably reduced in SHP2lyz-/− mice compared with WT mice (Fig. 2b and c). Selective deletion of SHP2 in macrophages also resulted in significant attenuation of pulmonary metastasis, as proved by the reduced number of metastatic nodes in the lungs 28 days after E0771 cell inoculation (Fig. 2d). Moreover, increased levels of IL-6 and TNF-α, but decreased IL-10 levels were detected in the sera from SHP2lyz-/− mice compared with those from WT mice (Fig. 2e-g), indicative of a possible anti-tumor immune microenvironment. To verify that SHP-2 activity is required for potentiating mammary carcinoma progression, we utilized PHPS1, a specific inhibitor of SHP-2 enzyme, to the E0771 tumor-bearing mice. Administering PHPS1 to the WT mice resulted in significant reduction in tumor volume and weight (Fig. 2h, 2i). The number of metastatic nodes in the lungs was greatly less in PHPS1-treated WT mice than that in the PBS-treated WT controls (Fig. 2j). H&E staining also showed attenuated pulmonary metastasis in these mice (Fig. 2k). Notably, no gross histopathological or quantitative difference was observed between SHP2lyz-/− mice with or without PHS1 treatment when evaluated for tumor burden and metastasis (Fig. 2h-k). These results suggest that SHP2 in macrophages is essential for promoting mammary carcinoma progression through its phosphatase activity.

3.3
Loss of SHP2 facilitates immunostimulatory TAM differentiation
Given that SHP2 activation in TAMs paralleled mammary carcinoma progression, we addressed the possibility that SHP2 might foster the malignancy of breast cancer by regulating the phenotype or effector functions of TAMs. In breast cancer model generated by injecting E0771 cells into female C57BL/6 mice, immunofluorescence showed that the infiltration of CD11c+M1-like macrophages was temporally decreased, while that of CD206+M2-like macrophages was increased in tumors during malignant progression (Fig. S3a). Moreover, the change in the levels of IFN-γ and IL-10 in tumor tissues were in a consistent pattern (Fig. S3b), suggesting the formation of immunosuppressive tumor microenvironment in the late-stage mammary carcinogenesis. When compared between WT and SHP2lyz-/− mice after 28 days of E0771 implantation, more CD45+TILs and CD45+F4/80+ TAMs were detected by flow cytometry from SHP2lyz-/− mice than those from WT mice (Fig. 3a). Based on the expression of costimulatory marker CD86, the percentage of CD86+CD206− TAMs was significantly increased, while that of CD86−CD206+ TAMs was decreased in the tumors from SHP2lyz-/− mice (Fig. 3b). In addition, more CD4+ T and CD8+ T cells infiltrated into the tumors from SHP2lyz-/− mice (Fig. S4). To further confirm differentiation/maturation and activation status of the TAMs, we analyzed the expression of the related cytokines and other factors using quantitative RT-PCR. TAMs of SHP2lyz-/− mice expressed significantly elevated levels of proinflammatory cytokines (IL-1β, IL-6, and Tnf-α) and inos mRNA, while there was no significant difference in alternatively activated (M2)-related factors (Arg-1, Cd206, Fizz-1 and Ym-1), indicative of a prevalent anti-tumor TAM phenotype (Fig. 3c). Co-injection of E0771 cells with TAMs isolated from tumor-bearing SHP2lyz-/− mice significantly inhibited tumor growth compared with the co-injection with TAMs from WT tumor-bearing mice (Fig. 3d). These results suggest that selective SHP2 depletion renders TAMs a proinflammatory phenotype with improved immunostimulatory and anti-tumor capacity in the late-stage mammary carcinogenesis.

3.4
SHP2 activation deafens type I IFN response in TAMs
To gain sight into the mechanism underlying which SHP2 activation modulates TAM differentiation and function in breast cancer environment, we used RNA sequencing to comprehensively compare the gene expression profile in the PMs derived from SHP2lyz-/− mice with WT controls following E0771 co-culture in separate chamber. 493 genes with differential expression were detected according to prespecified criteria (with log2 fold change > 0.3, adjusted p value < 0.05, 277 upregulated and 216 downregulated) between two groups (Figs. 4a, S5, Supplementary data 1). The differentially upregulated genes (lfit2, Cmpk2, lfit3, stat1, stat2, Traf2, Eif2ak2, Tyk2 and Cxcl10) in SHP2lyz-/− PMs were strongly linked to type I interferon (IFN) signaling pathway that is associated with anti-tumor immunity. In contrast, the typically downregulated genes, including Ptgs2, Csf3, Mmps and Args, were associated with immunosuppression, metastasis and angiogenesis [22,23]. The SHP2lyz-/− PMs also exhibited higher expression of the immunostimulatory molecules CD40 and Siglec1, and lower expression of Havcr2 (encoding for Tim3) that has a detrimental immunosuppressive role in anti-tumor immunity (Fig. S5, Supplementary data 1) [5,24]. Pathway enrichment analysis showed that the SHP2lyz-/− PMs were enriched for the genes of Toll-like receptor, TNF and NF-κB signaling, Jak-STAT signaling, virus infection, cytosolic DNA-sensing, chemokine signaling, antigen progressing and presentation (Fig. 4b), all of which are associated with anti-tumor properties of macrophages. Some of the related genes were confirmed using quantitative RT-PCR (Fig. 4c). Gene set enrichment analysis (GSEA) characterized active response to IFN-α in the SHP2lyz-/− PMs (FDR < 0.001, Fig. 4d). As shown in Fig. 4e, co-culture with E0771 cells resulted in an attenuated response to IFN-α in BMDMs from WT mice, supported by the decrease in the phosphorylation of STAT1 at Y701. SHP2 deficiency robustly increased their susceptibility to IFN-α even if following co-culture. Similar results were obtained in the expression of IFN-α responsive antiviral genes Oas1a and Eif2ak2 (Fig. 4e), suggesting that SHP2 is essential for deafening TAMs to type I IFN signaling in the breast cancer microenvironment.

3.5
IL-10 derived from breast cancer cells stimulates SHP2 activation in macrophages
To identify the soluble mediators secreted by breast cancer cells that activate SHP2 in macrophage upon their interaction, we incubated E0771 cells with the PMs from C57BL/6 mice in separate chambers and found over 140-fold elevation of Il-10 gene level in E0771 cells (Fig. 5a). Meanwhile, Pdgfβ and Il-4 expression were increased by about 7 folds. ELISA assay further confirmed a 2.5-fold increase of IL-10 in the co-culture supernatant compared with the culture supernatant alone (Fig. 5b). It is noteworthy that co-culture with only enriched PMs, but not CD8+ T, CD4+ T cells or BMDCs, enhanced Il-10 gene expression in E0771 cells (Fig. 5c). Using flow cytometry, we found that the production of IL-10, but not IL-4, was more in co-cultured E0771 cells than those cultured alone (Fig. 5d). When PMs were stimulated with recombinant IL-10, the levels of pSHP2 were increased in a time-dependent manner (Fig. 5e). In contrast, the elevation of pSHP2 levels induced by co-culture was attenuated by anti-IL-10 neutralizing antibody (Fig. 5f). Knockdown of IL-10R in BMDMs also greatly decreased the elevation of pSHP2 levels induced by co-culture (Figs. 5g, S6). These results suggest that tumoral IL-10 serves as an activator of SHP2 in TAMs upon reciprocal cell interaction. SHP2 is rapidly activated after cytokine stimulation, and its tyrosine residues are phosphorylated by activated protein tyrosine kinases to create binding sites for the substrates [1,25]. We found that IL-10-induced SHP2 phosphorylation was attenuated in Raw264.7 cells by the Src family kinase inhibitor PP2, while neither the Mer/Flt3 inhibitor UNC2025 nor the Syk inhibitor Piceatannol showed any effect (Fig. 5h). SHP2 immunoprecipitation followed by Src immunoblot detected a robust SHP2/Src interaction in Raw264.7 cells treated with recombinant IL-10 (Fig. 5i).

3.6
Tumoral IL-10 deficiency induces mammary carcinoma regression via SHP2 inactivation in TAMs
To further demonstrate the effects of tumor-derived IL-10 on SHP2-modulated TAM differentiation, we silenced IL-10 in E0771 cells using shRNA (Fig. S7). Co-culture with the resulting E0771 cells failed to induce SHP2 activation in BMDMs (Fig. 6a). Knockdown of IL-10 decreased the growth rate of tumor cells in vivo, reduced tumor weight and attenuated pulmonary metastasis (Fig. 6b-e). Among CD45+F4/80+ TAMs from E0771 tumors with IL-10 knockdown, the percentage of CD11c+CD206− macrophages was increased, while that of CD11c−CD206+macrophages was decreased (Fig. 6f). In addition, TAMs from E0771 tumors with IL-10 knockdown showed elevated levels of Tnf-α and Ifn-γ expression and reduced levels of Il-10, Cd206, Ym-1 and Ppar-γ than those from E0771 tumor controls (Fig. 6g), suggestive of proinflammatory programming. Consistent with in vitro results, markedly lower levels of pSHP2 were detected in TAMs from E0771 tumors with IL-10 knockdown (Fig. 6h). Taken together, SHP2 plays an important role when breast cancer cells condition macrophages. Macrophages trigger the induction of IL-10 in the tumor cells. And tumoral IL-10 renders macrophages an immunosuppressive phenotype by activating SHP2 and in turn deafening macrpophagic response to type I IFN, which can be suppressed by genetic or pharmacological SHP2 inhibition (Fig. 7).

4
Discussion
Accumulating evidence demonstrates that SHP2 influences the tumor microenvironment and is identified as a potential therapeutic target for cancer immunotherapy [[26], [27], [28]]. However, it is still elusive for the role of SHP2 in manipulating tumor-immune microenvironment to promote malignant progression. Here, we demonstrate that SHP2 is a key regulator during the crosstalk between TAMs and mammary carcinoma cells in tumor microenvironments. Reciprocal cell interaction triggered IL-10 production in tumor cells, and tumoral IL-10 activated SHP2 and rendered TAMs immunosuppressive phenotypes and functions. These findings highlight that targeting SHP2 in macrophages could be a therapeutic approach to improve anticancer therapy.
In this study, we found that SHP2 activation in TAMs paralleled mammary carcinoma progression. TAMs isolated from tumor-bearing SHP2lyz-/− mice exhibited an immunostimulatory phenotype and significantly inhibited tumor growth. Meanwhile, there were increased fractions of CD4+ and CD8+T cells in TILs isolated from SHP2lyz-/− mice. Contradictory to our findings, Li et al. reported that inhibition of SHP2 using PHPS1 increased the protein expression of Arginase-1 (Arg-1, a marker for M2 macrophage), IL-10 as well as p-PI3K and p-AKT in IL-4-treated THP1 cells [15]. Ke et al. found that IL-4 enhanced the protein expression and activity of Arg-1 in both PHPS1-treated Raw264.7 cells and BMDMs with SHP2 knockout [29]. As such, SHP2 inhibition seems to switch the macrophages towards a M2-like phenotype. But in our in vivo experiment, PHPS1 was administered to the WT mice bearing E0771 tumors. After 21-day treatment, PHPS1 remarkably inhibited tumor progression. The mRNA expression of proinflammatory cytokines (IL-1β, IL-6, Tnf-α, IFN-γ) was increased, while that of M2-related factors (Arg-1, Cd206, Ym-1) was decreased in enriched TAMs (Fig. S8). TAMs could not be simply represented by LPS/IFN-γ polarized M1 type or IL-4 polarized M2 type macrophages, since monocyte/macrophage differentiation is driven by complex environmental factors [30]. In fact, more anti-tumor macrophages were also observed to infiltrate in the tumors after another SHP2 allosteric inhibitor SHP099 was administered to the murine colorectal cancer model [13].
In the present study, RNA sequencing analysis showed high activation of type I IFN response in SHP2-depleted macrophages after co-culture with E0771 cells. Similar to this finding, single-cell transcriptomics analyses have indicated that STING-TBK1-IRF3-mediated type I IFN signaling is highly activated by SHP099 in infiltrated myeloid cells in murine colorectal cancer model [13]. It has been recently recognized that tumors with high levels of chromosomal instability can release cytosolic nucleic acids to activate cGAS-STING-mediated type I IFN signaling at risk of enhancing anti-tumor responses [31,32]. And the activation of cytosolic nucleic-acid-sensing pathways enforces tumor-antigen presentation on antigen presenting cells, leading to T cell response [33]. We found that SHP2 deficiency significantly increased the expression of type I interferon-stimulated genes (ISGs), including Ifit1, Ifit2, Ifit3, Cxcl10, Ifi44, stat1 [5,13], and the expression of the immunostimulatory molecules CD40 and Siglec1. It is possible that SHP2 negatively regulates type I IFN signaling in TAMs. This negative regulation of SHP2 might be helpful for tumors to inhibit type I IFN signaling and drive tumorigenic programs. On the other hand, SHP2 deafened type I IFN response in TAMs. The SHP2lyz-/− macrophages showed higher responsiveness to IFN-α than WT controls, as proven by higher level of phosphorylated STAT1. SHP2 activation induced by co-culture attenuated STAT1 phosphorylation. Therefore, SHP2 facilitates immunosuppressive TAM differentiation by downregulating type I IFN signaling and deafening their response to type I IFN. The mechanism underlying which SHP2 negatively regulates type I IFN signaling is still unclear.
The crosstalk between macrophage and tumor cells can render macrophage phenotype reprogrammable [9]. Tumor-derived factors such as cytokines and chemokines switch TAMs toward a pro-tumor phenotype [34]. For instance, CSF1 and CCL2 produced by breast cancer cells specifically promote the pro-tumor phenotype of macrophage [35]. In this study, we observed that IL-10 expression levels in E0771 cells were greatly elevated upon tumor-macrophage interaction. Breast tumor cells are capable of expressing IL-10 that exhibits potent immunosuppressive properties [36,37]. In consistent, deletion of IL-10 in E0771 cells caused mammary carcinoma regression. Notably, coculture with macrophages other than T cells or BMDCs enhanced Il-10 gene expression in E0771 cells. This finding suggests that macrophages actively participate in a positive feedback loop with tumors and tumors take advantage of macrophage plasticity to promote malignant progression. In fact, mesenchymal-like breast cancer cells activate TAMs to a pro-tumor phenotype by GM-CSF, and CCL18 from TAMs reciprocally induces cancer cell epithelial-mesenchymal transition [7]. Our study reveals that macrophages triggered the induction of IL-10 in mammary carcinoma cells. Reciprocally, tumoral IL-10 rendered macrophages protumorigenic properties. However, the drivers of Il-10 expression by macrophages remains to be investigated.
Tumoral IL-10 induced SHP2 activation in macrophages, which was inhibited by a neutralizing anti-IL-10 antibody or IL-10R depletion. A plethora of evidence has demonstrated that SHP-2 is rapidly activated by protein tyrosine kinases upon stimulation with cytokines and growth factors [[38], [39], [40]]. Indeed, IL-10 stimulation resulted in the increase of p-Src levels and the binding of SHP2 to Src in macrophages. IL-10-induced SHP2 activation was attenuated by a Src inhibitor. Previous studies indicate that IL-10 stimulation results in the recruitment of SHP2 to STAT3 [41]. Src activity is required for tyrosine phosphorylation of STAT3 by proinflammatory macrophages [42]. These results suggest that Src activity could be also responsible for tyrosine phosphorylation of SHP2 in macrophages upon IL-10 stimulation.

5
Conclusion
In summary, our study provides insight into the role of SHP2 in the crosstalk between macrophage and mammary carcinoma cells and suggests SHP2 in macrophages as a potential therapeutic target for breast cancer therapy.

Declaration of competing interest
The authors declare that they have no conflicts of interest in this work.