Signal regulatory protein α is associated with tumor-polarized macrophages phenotype switch and plays a pivotal role in tumor progression

Authors

  • Yu-fei Pan,

    1. International Cooperation Laboratory on Signal Transduction, Eastern Hepatobiliary Surgery Institute, Second Military Medical University, Shanghai, P.R. China
    Search for more papers by this author
    • Supported by grants from National Natural Science Foundation of China (81001075, 91229205), the Funds for Creative Research Groups of China (81221061), the State Key Project for Liver Cancer (2012ZX10002-009, 2013ZX10002-010) and the Key Project for Military (BWS11J036).

  • Ye-xiong Tan,

    1. International Cooperation Laboratory on Signal Transduction, Eastern Hepatobiliary Surgery Institute, Second Military Medical University, Shanghai, P.R. China
    Search for more papers by this author
    • Supported by grants from National Natural Science Foundation of China (81001075, 91229205), the Funds for Creative Research Groups of China (81221061), the State Key Project for Liver Cancer (2012ZX10002-009, 2013ZX10002-010) and the Key Project for Military (BWS11J036).

  • Min Wang,

    1. International Cooperation Laboratory on Signal Transduction, Eastern Hepatobiliary Surgery Institute, Second Military Medical University, Shanghai, P.R. China
    Search for more papers by this author
  • Jian Zhang,

    1. International Cooperation Laboratory on Signal Transduction, Eastern Hepatobiliary Surgery Institute, Second Military Medical University, Shanghai, P.R. China
    Search for more papers by this author
  • Bo Zhang,

    1. International Cooperation Laboratory on Signal Transduction, Eastern Hepatobiliary Surgery Institute, Second Military Medical University, Shanghai, P.R. China
    Search for more papers by this author
  • Chun Yang,

    1. International Cooperation Laboratory on Signal Transduction, Eastern Hepatobiliary Surgery Institute, Second Military Medical University, Shanghai, P.R. China
    Search for more papers by this author
  • Zhi-wen Ding,

    1. International Cooperation Laboratory on Signal Transduction, Eastern Hepatobiliary Surgery Institute, Second Military Medical University, Shanghai, P.R. China
    Search for more papers by this author
  • Li-wei Dong,

    Corresponding author
    1. International Cooperation Laboratory on Signal Transduction, Eastern Hepatobiliary Surgery Institute, Second Military Medical University, Shanghai, P.R. China
    • Address reprint requests to: Hong-yang Wang, M.D., or Li-wei Dong, Ph.D., International Cooperation Laboratory on Signal Transduction, Eastern Hepatobiliary Surgery Institute, Second Military Medical University, 225 Changhai Road, 200438, Shanghai, China. E-mail: hywangk@vip.sina.com or donliwei@126.com; fax: +86 21 6556 6851.

    Search for more papers by this author
  • Hong-yang Wang

    Corresponding author
    1. International Cooperation Laboratory on Signal Transduction, Eastern Hepatobiliary Surgery Institute, Second Military Medical University, Shanghai, P.R. China
    2. State Key Laboratory of Oncogenes and Related Genes, Shanghai Cancer Institute, Renji Hospital, Shanghai Jiaotong University School of Medicine, P.R. China
    • Address reprint requests to: Hong-yang Wang, M.D., or Li-wei Dong, Ph.D., International Cooperation Laboratory on Signal Transduction, Eastern Hepatobiliary Surgery Institute, Second Military Medical University, 225 Changhai Road, 200438, Shanghai, China. E-mail: hywangk@vip.sina.com or donliwei@126.com; fax: +86 21 6556 6851.

    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

  • These authors contributed equally to this work.

Abstract

Macrophages (Mψ) are the major component of infiltrating leukocytes in tumors and exhibit distinct phenotypes according to the microenvironment. We have recently found that signal regulatory protein α (SIRPα), the inhibitory molecule expressed on myeloid cells, plays a critical role in controlling innate immune activation. Here, we identify that SIRPα is down-regulated on monocytes/Mψ isolated from peritumoral areas of hepatocellular carcinoma (HCC) samples, while its level is moderately recovered in intratumor Mψ. In vitro assays demonstrate that SIRPα expression is significantly reduced on Mψ when cocultured with hepatoma cells. This reduction is partly due to the soluble factors in the tumor microenvironment. Knockdown (KD) of SIRPα prolongs activation of nuclear factor kappa B (NF-κB) and PI3K-Akt pathways as Mψ encounter tumor cells, leading to an increased capacity of Mψ for migration, survival, and proinflammatory cytokine production. Enhanced Stat3 and impaired Stat1 phosphorylation are also observed in tumor-exposed SIRPα-KD Mψ. Adoptive transfer with SIRPα-KD Mψ accelerates mouse hepatoma cells growth in vivo by remolding the inflammatory microenvironment and promoting angiogenesis. SIRPα accomplishes this partly through its sequestration of the signal transducer Src homology 2-containing phosphotyrosine phosphatase (SHP2) from IκB kinase β (IKKβ) and PI3K regulatory subunit p85 (PI3Kp85). Conclusion: These findings suggest that SIRPα functions as an important modulator of tumor-polarized Mψ in hepatoma, and the reduction of SIRPα is a novel strategy used by tumor cells to benefit their behavior. Therefore, SIRPα could be utilized as a potential target for HCC therapy. (Hepatology 2013;58:680–691)

Abbreviations
BMDM

bone marrow derived macrophage

COX2

cyclooxygenase 2

CSF1

colony stimulating factor 1

HCC

hepatocellular carcinoma

HIF1α

hypoxia-inducible factor 1α

HRE

hypoxia transcriptional response element

IL

interleukin

KD

knockdown

macrophages

MCP-1

monocyte chemotactic protein-1

NFAT

nuclear factor of activated T-cells

NOS2

nitric oxide synthase 2 (inducible)

SHP2

the Src homology 2-containing phosphotyrosine phosphatase

SIRPα

signal regulatory protein α

VEGF

vascular endothelial growth factor.

Hepatocellular carcinoma (HCC) is the fifth most common cancer worldwide, and the second leading cause of cancer death in China. HCC is usually present in inflamed fibrotic and/or cirrhotic liver with extensive leukocyte infiltration. Thus, the immune status at different tumor sites is tightly associated with the biological behavior of HCC.[1]

Macrophages (Mψ) are the most prominent component of the infiltrated leukocytes in tumors. These cells are derived from circulating monocytes and recruited into tumor by cytokines and chemokines, such as CSF1 and MCP-1.[2, 3] Mψ have remarkable plasticity and can acquire special phenotypic characteristics with diverse functions in response to environmental signals.[4, 5] Tumor-associated Mψ (TAMs), closely associated with M2, can suppress antitumor immunity and promote tumor progression.[6] Evidence from clinical and epidemiological studies have shown a strong association between TAMs density and poor prognosis in several types of cancer, including HCC.[7] However, some studies demonstrated that Mψ in tumor stroma were activated and displayed a human leukocyte antigen (HLA)-DRhigh phenotype. These cells can also facilitate tumor progression.[8-10] Taken together, these results indicate that tumors can take advantage of either immune suppression or activation status of Mψ at distinct tumor sites to promote tumor progression. Currently, the precise mechanism of how tumors educate Mψ to accomplish specific tasks has not been fully elucidated.

Signal regulatory protein α (SIRPα) is a cell-surface protein mainly expressed on myeloid cells, including Mψ and dendritic cells.[11, 12] The extracellular region of SIRPα is heavily glycosylated and comprised of three immunoglobin superfamily (IgSF) domains, which are similar to TCR and BCR, suggesting that SIRPα may have a pivotal role in immune regulation.[13] The intracellular region contains two immunoreceptor tyrosine-based inhibition motifs (ITIM) with four tyrosine residues that are phosphorylated in response to a variety of stimuli.[12] The phosphorylation allows SHP1 or SHP2 recruitment to SIRPα that, in turn, dephosphorylates specific substrates involved in various physiological effects.[14, 15] SIRPα can bind to either widely expressed transmembrane ligand CD47 or soluble ligands, such as the surfactant proteins A and D.[16] It is suggested that the SIRPα/CD47 signaling axis is important in tumor therapy.[17, 18] Our previous work has shown that SIRPα negatively regulate Toll-like receptor (TLR) signaling in Mψ.[16, 19] However, it is still unknown whether SIRPα expression on tumor-polarized Mψ can act on tumor progression.

We demonstrate here that SIRPα expression is reduced on Mψ obtained from peritumoral tissues of HCC patients. Down-regulated SIRPα expression is coincident with transiently activated Mψ during the early stage of exposure to tumor. Moreover, adoptive transfer of SIRPα-KD Mψ could promote tumor growth in vivo. These findings provide a new role of SIRPα on tumor-polarized Mψ and tumor progression.

Patients and Methods

Patient Samples and Mononuclear Cell Isolation

Peripheral blood samples of healthy donors (n = 20) and untreated HCC patients (n = 22) as well as HCC tumor samples (n = 25) were obtained. Tumor tissues were collected from the areas of tumor nest, while the peritumoral samples were obtained near the tumor tissues (0.5-1 cm from tumor margin). The patients were pathologically confirmed as HCC at the Eastern Hepatobiliary Surgery Hospital, Shanghai, China. Detailed information about the patients and their tumors are shown in Supporting Table. 1. Written informed consent was obtained and the protocols were approved by the Review Board of the Eastern Hepatobiliary Surgery Hospital. The circulating mononuclear cells were obtained by Ficoll density gradient centrifugation. The infiltrated leukocytes were isolated according to the following protocols: specimens were cut into small pieces and digested with 0.05% collagenase IV, 0.002% DNase I (Sigma-Aldrich), and 20% fetal bovine serum (FBS) (Gibco) at 37°C for 1 hour. The dissociated cells were filtered through 150-μm mesh and separated by Percoll centrifugation. The obtained cells were washed for the fluorescent-activated cell sorter (FACS) analysis.

Animals

Male C57BL/6 mice and Balb/c mice (6-8 weeks old) were obtained from the Chinese Science Academy, Shanghai, China, and maintained under pathogen-free conditions. All animals received humane care according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals, prepared by the National Institutes of Health.

Statistical Analysis

Experiments were performed repeatedly and representative data are shown. Continuous variables were compared with the Student t test, ordinal variables with the Mann-Whitney U test. P 0.05 was considered significant. Data analysis was performed with SPSS 16.0 for Windows (Chicago, IL).

A detailed description of Patients and Methods can be found in the online Supporting Information.

Results

SIRPα Expression Is Down-Regulated on Tumor-Polarized Mψ

To evaluate the potential role of SIRPα in tumor immunopathology, we used flow cytometry to analyze SIRPα expression on circulating leukocytes from healthy donors (n = 20) and HCC patients (n = 22), and on infiltrated leukocytes freshly isolated from paired samples of tumor and peritumor tissues from HCC patients (n = 25). Detailed information of the patients and the pattern of SIRPα expression are shown in Supporting Table 1 and Supporting Fig. 1A. In all of the samples analyzed, 83 ± 12% of the SIRPα-positive cells were identified as CD14high monocytes/Mψ (Supporting Fig. 1B,C). As shown in Fig. 1A, there was no significant difference of SIRPα protein expression between the circulating monocytes isolated from HCC patients and healthy donors. However, the monocytes/Mψ obtained from HCC tissues had dramatically decreased SIRPα levels. Surprisingly, the level of SIRPα on monocytes/Mψ located in peritumoral tissues was much lower than that in tumor nests. Consistent with the observations from HCC patients, when examined in mice models bearing hepatoma, SIRPα expression was also reduced on monocytes/Mψ isolated from tumor tissues derived from Hepa1-6 cells compared with that in circulating leukocytes (Fig. 1B; Supporting Fig. 1D,E). The same results were found in mice bearing H22 hepatoma cells (Fig. 1C). Collectively, these data indicate that SIRPα is down-regulated on monocyte/Mψ from tumor tissues both in humans and mice.

Figure 1.

SIRPα expression is reduced on monocytes/Mψ in the tumor microenvironment. (A) FACS assay of SIRPα expression on CD14high mononuclear cells freshly isolated from peripheral blood and HCC tissues. (B,C) Hepa1-6 cells were subcutaneously injected into syngeneic C57BL/6 mice (B) and H22 cells were intraperitoneally injected into Balb/c mice (C), CD11bhigh mononuclear cells were obtained from the tumor-bearing mice and the expression of SIRPα was detected. (D) BMDMs were cocultured with complete medium, primary mouse hepatocytes, or Hepa1-6 cells. SIRPα expression was detected by FACS. Data are representative of at least three individual experiments. (E) BMDMs were left untreated (gray) or cocultured with Hepa1-6 cells for 1 day (blank) or 5 days (black), SIRPα and MHC II expression levels were detected. Data represent three individual experiments. (F) BMDMs left untreated or cocultured with Hepa1-6 cells for the indicated times; cytokine production pattern was determined by enzyme-linked immunosorbent assay (ELISA). Representative data are from at least three independent experiments. All the data are shown as mean ± standard error of the mean (SEM); *P < 0.05; **P < 0.01; ***P < 0.001.

SIRPα Expression Is Regulated by Tumor-Derived Factors and Is Associated With Mψ Phenotype Switch

It is well known that Mψ in tumor niches can be educated to cooperatively support tumor progression.[20] However, it remains unknown whether SIRPα expression on Mψ is involved in these mechanisms. Mouse hepatoma cells Hepa1-6 or primary hepatocytes were cocultured with mouse bone marrow-derived Mψ (BMDMs) without direct cell-cell contact. As shown in Fig. 1D, Hepa1-6 cells induced a dramatic decrease of SIRPα expression on BMDMs, reaching a minimum within 24 hours and returning to the basal levels 120 hours post-coculture. In contrast, normal hepatocytes had only a marginal effect on SIRPα expression on Mψ. Meanwhile, the SIRPα messenger RNA (mRNA) level also transiently decreased in response to tumor, indicating an inhibition role of tumor cells in SIRPα transcript (Supporting Fig. 2A).

Recent studies suggested that the tumor environment affects Mψ activation.[8, 16] To confirm this in HCC, we examined the immune status of Mψ after coculture with Hepa1-6 cells for 1 or 5 days. As illustrated in Fig. 1E, BMDM was transiently activated, together with SIRPα decline and MHC II elevation 1 day post-coculture with Hepa1-6 cells; however, after 5 days coculture SIRPα expression was recovered and MHC II was decreased. These results imply that the status of Mψ can be altered gradually by tumor cells, and SIRPα expression level may represent the different stages during this process.

Furthermore, Mψ treated with TNFα, H2O2 and hypoxia in vitro resulted in a significant decrease of SIRPα expression on BMDMs (Supporting Fig. 2B), indicating that these factors existing in the tumor microenvironment may regulate SIRPα expression on Mψ. Finally, cytokines such as IL1β, TNFα, IL6, IL12, and IL10 were markedly elevated 1 day post-coculture (Fig. 1F).

Knockdown of SIRPα Alters the Phenotype of Tumor-Polarized Mψ

To address whether SIRPα plays a role in the phenotype switch of Mψ, SIRPα expression in BMDMs was suppressed by small interfering RNA (siRNA) transfection (si-KD) or by lentivirus infection (LV-KD) (Supporting Fig. 3A,B). Compared with the control cells, SIRPα knockdown in BMDMs increased production of IL1β, IL6, and TNFα upon coculture with Hepa1-6 cells in vitro (Fig. 2A). However, targeting SIRPα increased production of immunosuppressive cytokine IL10 while reducing IL12 expression (Fig. 2B). Furthermore, SIRPα-depleted Mψ exhibited elevated expression of arginase-1 (Arg1) and decreased nitric oxide synthase 2 (inducible) (NOS2) expression (Fig. 2C). These results indicate that SIRPα plays a pivotal role in regulating the phenotype of Mψ upon tumor exposure.

Figure 2.

Knockdown of SIRPα alters the phenotype of Mψ in response to tumor. (A,B) SIRPα-siRNA (Si-KD) or Control-siRNA (Si-Ctrl), lentivirus-shSIRPα (LV-KD), or Control (LV-NC) was introduced into BMDMs for 36 hours. Then the cells were cocultured with Hepa1-6 cells for 24 (A) or 48 (B) hours. The cell-free supernatant was collected and analyzed by ELISA. Data are shown as mean ± SEM from three individual experiments. (C) In parallel, total RNA was isolated from cocultured Mψ for real-time polymerase chain reaction (PCR) analysis of Arg1 and NOS2. (D) BMDMs were cocultured with Hepa1-6 cells for the indicated times and the proteins were extracted and measured by immunoblotting analysis with GAPDH as a loading control. Data are representative of three independent experiments. (E) Luciferase assays of NF-κB and NOS2 reporter activity were performed on cocultured Mψ at the indicated times. Representative data are shown as mean ± SEM of at least four independent experiments. *P < 0.05.

Since NF-κB and Stat3 are considered essential transcription factors in Mψ linking inflammation and cancer,[21, 22] we then analyzed whether SIRPα could modulate their activation in Mψ when exposed to tumor cells. As shown in Fig. 2D, SIRPα-KD BMDMs showed increased serine phosphorylation of IκBα, together with elevated NF-κB activation upon coculture with Hepa1-6 cells (Fig. 2E). Tyrosine phosphorylation of Stat3 was also increased, while p-Stat1 (Tyr701) declined in SIRPα-KD Mψ than the control group, which was correlated with decreased NOS2 expression (Fig. 2D,E). Together, these results suggest that the function of SIRPα on Mψ may be partly mediated by way of the modulation of NF-κB and Stat3 activation.

SIRPα Negatively Regulates Mψ Recruitment to Tumor Cells

Since TAMs are derived from circulating leukocytes, we then investigated whether SIRPα could affect Mψ migration during tumor exposure. The results from transwell assay showed that BMDMs were recruited to Hepa1-6 tumor cells, and the migration ability was significantly increased when SIRPα expression on Mψ was silenced (Fig. 3A). To test the effects of SIRPα silencing on BMDMs infiltration in vivo, CellTracker Green CMFDA-labeled SIRPα-KD and Control BMDMs were intravenously injected into Hepa1-6-bearing mice, followed by examining CMFDA-labeled cells in tumor tissues. As illustrated in Fig. 3B, the number of SIRPα-KD BMDMs infiltrated into tumor nests was higher than that of control cells (Fig. 3B), indicating that SIRPα impairs the migration capacity of BMDMs toward tumor.

Figure 3.

SIRPα inhibits Mψ migration in the tumor microenvironment. (A) Transwell assay for the capacity of Mψ migration. SIRPα-LV-KD and LV-NC Mψ were plated on the upper layer of the chamber with a 5-μm filter coated on the underside with 20 μg/mL mouse fibronectin peptides; Hepa1-6 cells were seeded on the lower chambers. The migrated Mψ were fixed and stained. One representative of four individual experiments is shown (left). The histogram represents the numbers of migrated cells (right). (B) Infiltration of circulating cells into tumor. CellTracker Green CMFDA-labeled SIRPα-KD and control Mψ (5 × 106) were intravenously injected into Hepa1-6-bearing mice, the tumor tissues were surgically excised 24 hours after the injection, and frozen sections were prepared and analyzed by fluorescence microscopy (200×, left). CMFDA-Mψ numbers were averaged from 15 random fields each of the groups (right). Data represent five individual experiments. (C) Transwell assay of migrated Mψ in response to CSF1 (50 ng/mL) or MCP-1 (20 ng/mL). Representative data are shown from at least three experiments. (D) BMDMs were treated with 50 ng/mL CSF1 (upper) for the indicated times. The proteins were extracted and detected by immunoblotting. In response to MCP-1 (lower), the activity form of Rac (Rac-GTP) was precipitated by PAK-PDB and immunoblotted by Rac antibody. The relative intensity of the indicated bands is shown as the average of three independent experiments. The data are all shown as mean ± SEM. *P < 0.05; **P < 0.01.

MCP-1 and CSF1 were found expressed more in Hepa1-6 cells than in primary mouse liver cells, while expression of chemokine CCL5 saw no change between these two cell types (Supporting Fig. 4A). Silencing MCP-1 or CSF1 in Hepa1-6 significantly inhibited Mψ migration toward tumor cells (Supporting Fig. 4B). In addition, knockdown of SIRPα expression on Mψ dramatically accelerated migration in response to MCP-1 and CSF1 (Fig. 3C), consistent with the inhibitory role of SIRPα in Mψ migration toward tumors, as mentioned above. We also found that the activities of PLCγ1, Akt, and Rac1, which were involved in Mψ migration, were elevated in SIRPα-KD Mψ in response to CSF1 or MCP-1 (Fig. 3D). These data suggest that tumor-derived factors induce down-regulation of SIRPα expression on Mψ, followed by promoting their migration to the tumor; on the other hand, the recruited Mψ gradually restore SIRPα under long-term education by tumor environment, and weaken the ability of migration out of the nest.

Knockdown of SIRPα on Mψ Triggers Survival Signal

To investigate whether SIRPα was involved in regulation of Mψ survival in response to tumor, we treated SIRPα-KD and Control BMDMs with proapoptotic factors (such as TNFα and TRAIL) existing in the tumor microenvironment. TNFα treatment following cycloheximide (CHX) preincubation significantly induced Mψ apoptosis. Compared with the control group, SIRPα-KD BMDMs displayed delayed activation of effector caspase3, together with lower levels of cleaved poly (ADP-ribose) polymerase (PARP) (Fig. 4A). The ratio of apoptotic cells (annexin-V positive) was also lower in SIRPα-KD BMDMs (Fig. 4B). A similar pattern of Mψ apoptosis was also observed in response to TRAIL (Fig. 4A,B). In accordance with this, the activities of prosurvival pathways, such as Akt and NF-κB, were also increased in SIRPα-KD BMDMs when cocultured with tumor (Figs. 2D, 4C). These results demonstrate that SIRPα decreases the threshold for Mψ to undergo apoptosis in an adverse environment.

Figure 4.

SIRPα inhibits tumor-induced Mψ apoptosis. (A,B) LV-NC or LV-KD infected BMDMs were treated with TNFα (25 ng/mL) plus CHX (1 ng/mL) or TRAIL (25 ng/mL) for the indicated times. Immunoblotting was performed and cleaved caspase-3 and PARP were measured (A). The apoptotic cells were labeled by Annexin-V staining after treatment (B, left), and the ratio of Annexin-V-positive cells was determined (B, right). The representative data from three independent experiments are shown as mean ± SEM. *P < 0.05. (C) Serine phosphorylation of Akt of cocultured BMDMs was measured by immunoblotting.

Adoptive Transfer SIRPα-KD Mψ Promotes Tumor Progression in Mice

Since SIRPα had an important role in regulating the phenotype of Mψ and cell migration as well as cell survival upon tumor exposure, we wondered whether mice adoptive transfer with SIRPα-KD Mψ could affect tumor progression. We incised tumor samples derived from Hepa1-6 in C57BL/6 mice into 1 × 1 mm pieces, and loaded one piece per mouse under the liver capsule of healthy C57BL/6 mice. Since GdCl3 could selectively deplete circulating mononuclear cells of a monocyte/Mψ lineage,[16, 23] we intravenously injected GdCl3 into the tumor-loaded mice and then adoptively transferred SIRPα-LV-KD or SIRPα-si-KD Mψ by tail vein injection. Tumors were assessed 15 days later. Transfer of SIRPα-KD BMDMs into tumor-bearing mice led to a significant increase of tumor burden when compared with the control group (Fig. 5A). Transfer of SIRPα-targeted Mψ into mice with subcutaneously bearing Hepa1-6 also accelerated tumor growth (Fig. 5B).

Figure 5.

SIRPα knockdown on Mψ promotes tumor progression in vivo. (A) Mouse hepatoma tissues derived from Hepa1-6 cells were seeded under the liver capsule of C57BL/6 male mice. After 5 days, Mψ were depleted by intravenous injection of 10 mg/kg GdCl3, then 1 × 107 SIRPα-targeted or Control Mψ were adoptively transferred into the tumor-bearing mice by tail vein injection. Fifteen days later the livers were excised and are shown (upper, white arrow). Tumor burden was analyzed by measuring the diameter of tumor (lower). Representative data are shown as mean ± SEM (n = 4-6 mice per group) of two independent experiments. (B) 5 × 105 Hepa1-6 cells were subcutaneously injected into C57BL/6 mice. Five to 7 days later, WT-, SIRPα-KD, or Control Mψ were adoptively transferred by intravenous injection. After 16 days, tumors were surgically excised and shown (upper). Tumor volumes were measured every 4 days after intravenous injection (middle) and tumor mass was measured and shown (lower). Data are representative of three independent experiments (n = 5 mice per group) and shown as mean ± SEM. (C,D) Mice hepatoma cell H22 was intraperitoneally injected into Balb/c mice. WT-, SIRPα-KD, or Control Mψ derived from Balb/c mice and were then intraperitoneally injected into the tumor-bearing mice. The tumor burden was determined by tumor cell count in the ascites of mice (n = 6 mice per mice) and shown as mean ± SEM (C). In parallel, the survival time of the mice receiving the same treatment was monitored, each group containing six mice (D). *P < 0.05; **P < 0.01.

To further determine the relationship between SIRPα on Mψ and tumor progression, another mouse hepatoma cell line H22 (Balb/c mice-derived) was employed for further investigation. H22 cells were intraperitoneally injected into the syngeneic Balb/c mice. WT-, SIRPα-KD and control Mψ were then adoptively transferred into the established tumors by intraperitoneal injection. About 7 days later, the tumors were examined and the ascites of the tumor-bearing mice were collected. As shown in Fig. 5C, transfer of SIRPα-KD Mψ led to a significant increase in tumor burden when compared with control Mψ. In another parallel experiment, mice adoptively transferred with SIRPα-KD Mψ showed a shorter survival time than the control group (Fig. 5D). Collectively, these results indicate that knockdown of SIRPα on Mψ promotes tumor progression in vivo.

SIRPα-KD Mψ Support Tumor Growth by Remodeling the Tumor Microenvironment

The above results demonstrated that tumor-exposed SIRPα-KD Mψ produced larger amounts of proinflammatory cytokines than control cells in vitro. Here, we found that adoptive transfer of SIRPα-KD Mψ also increased expression of tumor promoting cytokines in both Hepa1-6 and H22-derived tumor tissues (Fig. 6A,B). Furthermore, immunostaining assays showed that the density of CD31+ endothelial cells, considered the marker of microvessel neogenesis, was higher in Hepa1-6 tumors receiving an intravenous injection of SIRPα-KD Mψ than that of control Mψ (Fig. 6C). Meanwhile, the KD group also showed an increased expression of vascular endothelial growth factor (VEGF) (Supporting Fig. 5). Interestingly, it was shown that the stromal cells, including Mψ, were the major source of VEGF production other than tumor cells, indicating an important role of Mψ in tumor neovascularization (Supporting Fig. 5). HIF1α, whose stability is associated with the activation of Akt and NF-κB, is essential for angiogenesis.[24] Luciferase reporter gene assay showed that the activity of hypoxia transcriptional response element (HRE) was increased in SIRPα-KD Mψ upon exposure to Hepa1-6 cells, and the protein level of HIF1α was also increased in SIRPα-KD Mψ (Fig. 6D,E). In accordance with this, the reporter activity of the downstream molecules, such as NFAT, COX2, and VEGF, was increased by 2-4-fold compared with control Mψ (Fig. 6D). These results indicate that SIRPα negatively regulates the stability of HIF1α on Mψ in response to tumor, suggesting that SIRPα plays an important role in tumor angiogenesis.

Figure 6.

SIRPα-KD Mψ support tumor growth by remodeling the tumor microenvironment. (A) Total RNA was extracted from the tumor tissues of Fig. 5B; IL6, TNFα, VEGF, and COX2 mRNA levels were measured by real-time PCR. (B) Cytokine production in cell-free ascitic fluid of the H22 tumor-bearing mice in Fig. 5C was measured by ELISA. Each group contained six mice. (C) Immunohistochemistry staining of CD31-positive cells in paraffin-embedded tumor tissues obtained from Fig. 5B (upper, red arrows). The density of CD31+ cells was measured and shown (lower, n = 5). (D) Mψ were transfected with HRE, NFAT, VEGF, or COX2 reporter plasmids together with pRL-TK. The cells were then cocultured with Hepa1-6 cells and the Luciferase activity was detected and normalized with Renilla luciferase activity. Representative data from at least three individual experiments are shown. (E) BMDMs were exposed to tumor cells. HIF1α protein level was detected by immunoblotting. Data shown as means ± SEM. *P < 0.05.

SIRPα Negatively Regulates Tumor Progression Through Sequestration of SHP2 From IκB Kinase β (IKKβ) and PI3Kp85

SIRPα is a cell surface protein containing the ITIM motif domains which are known to exert an inhibitory function through recruitment of phosphatase enzyme SHP2 to its phosphorylated tyrosine residues. We analyzed whether SIRPα phosphorylation was increased when cocultured with tumor cells. As shown in Fig. 7A, tyrosine phosphorylation of SIRPα was increased in response to Hepa1-6 cells, together with enhanced binding to SHP2. SHP2 was constitutively associated with SIRPα even when SIRPα had the undetectable phosphorylation level at the basal time. Moreover, knockdown of SHP2 by siRNA transfection significantly decreased phosphorylation of IκBα and Akt compared with control Mψ when cocultured with tumor (Fig. 7B). As expected, the amount of IL6 and TNFα production was about 2-fold lower in SHP2-KD Mψ than control (Fig. 7C).

Figure 7.

SIRPα sequestrates SHP2 from IKKβ and PI3Kp85 when cocultured with tumor cells. (A) Coimmunoprecipitation was performed on tumor-exposed Mψ by specific anti-SIRPα antibody. Tyrosine phosphorylation of SIRPα, PI3Kp85, and SHP2 were detected by immunoblotting. Representative data are shown from at least three independent experiments. (B,C) SHP2 was specifically knocked down in Mψ by siRNA transfection. The cells then cocultured with Hepa1-6 cells for the indicated times. Proteins were extracted and detected by immunoblotting analysis with GAPDH as a loading control (B); the cell-free supernatant was collected and measured by ELISA (C, mean ± SEM, *P < 0.05); data represent three individual experiments. (D) Coimmunoprecipitation was performed on tumor-activated SIRPα-KD and Control Mψ by anti-SHP2 antibody. PI3Kp85 and IKKβ were detected. Data represent at least three experiments. (E) The mechanistic schematic model of the role of SIRPα in tumor-polarized Mψ. Briefly, SIRPα is tyrosine phosphorylated and sequestrates SHP2 from IKKβ and PI3K regulatory subunit PI3Kp85, resulting in affecting PI3K-Akt and NF-κB pathways in the tumor microenvironment. Therefore, knockdown of SIRPα expression on Mψ enhanced Akt and NF-κB activation, promoting Mψ migration, survival, cytokines production, as well as angiogenesis in tumor sites.

To investigate how SHP2 was involved in the regulation of NF-κB and Akt signaling pathways, we performed coimmunoprecipitation experiments by targeting SHP2 on SIRPα-KD and control Mψ upon exposure to Hepa1-6. Tumor cells induced an interaction of SHP2 with IKKβ and PI3K regulatory subunit p85 (PI3Kp85) in Mψ, which was critical in the activation of the NF-κB and Akt pathway, respectively (Fig. 7D). Compared with control Mψ, SHP2 in SIRPα-KD Mψ showed enhanced association with IKKβ and PI3Kp85, which led to an increased activation of NF-κB and Akt (Fig. 7D). These results suggest that SIRPα may compete with IKKβ and PI3Kp85 in binding to SHP2, thus inhibiting NF-κB and Akt activation in Mψ, leading to negative regulation of tumor progression (Fig. 7E).

Discussion

HCC is tightly associated with chronic inflammation.[25] Although immune cells, the major source of inflammatory mediators, are thought to have tumoricidal activity, many reports have shown that they may be educated and become the “accomplice” of tumors.[20] The present study demonstrates that SIRPα is down-regulated on tumor-polarized Mψ and plays a negative role in tumor progression, which may represent a new link between the proinflammatory response and tumor progression.

Human tumor and peritumor tissues are classified according to the anatomic locations. The compositions and properties are different between them.[9, 26] In HCC samples, Mψ infiltrated in the tumor tissues always exhibit an immunosuppressive phenotype like M2. However, the peritumor tissues contain a larger amount of M1-like cells.[9] In this study, we observed that Mψ in HCC tumor tissues expressed a lower level of SIRPα compared with the circulating monocytes, while SIRPα proteins in peritumor showed the lowest staining. In accordance with this, in vitro coculture assay also showed a dynamic expression of SIRPα, together with the alteration of Mψ immune status, suggesting that SIRPα plays an essential role in regulating the Mψ phenotype switch. Moreover, statistical analysis indicated that relatively lower levels of SIRPα expression on Mψ in peritumor tissues was correlated with tumor size, indicating that tumors may take advantage of the Mψ with SIRPα reduction to benefit themselves (Supporting Tables 2, 3).

SIRPα is a cell-surface glycoprotein considered an inhibitory molecular due to the ITIMs domains in the intracellular region. As expected, tyrosine phosphorylation of SIRPα is increased in Mψ when cocultured with tumor cells, resulting in SHP2 recruitment to the cell surface. Surprisingly, SIRPα knockdown increases association of SHP2 with IKKβ when cocultured with tumor cells, leading to activation of NF-κB. It is reported that IKKβ could inhibit Stat1 activation in tumor-polarized Mψ.[21] We found that Stat1 activation was inhibited in SIRPα-KD Mψ, which may partly be due to the increased functional form of IKKβ. Since Stat1 is an essential transcription factor in immune response against tumor, this result indicates that the tumoricidal activity of SIRPα-KD Mψ may be decreased as well. We also found that the association of SHP2 with PI3Kp85, the regulatory subunit of PI3K, was increased in SIRPα-KD Mψ when cocultured with tumor cells. This interaction increases the activity of the catalytic subunit PI3Kp110, leading to Akt activation. Recently, the transcription factor HIF1α, which is essential in angiogenesis, was reported to be stabilized and activated by Akt.[27] This may explain the proangiogenesis potential by SIRPα-KD Mψ in tumor-bearing models. Moreover, it is reported that SIRPα is required in T- and NK-cells homeostasis in vivo, indicating an important role of SIRPα in immunomodulation.[28, 29] Nevertheless, further investigation is required to determine the role of SIRPα in Mψ crosstalking with other immune cells in the tumor microenvironment.

The mechanisms contributing to the reduction of SIRPα on monocytes/Mψ in response to tumor are not yet well elucidated. Although soluble factors in the tumor microenvironment, such as TNFα, could decrease SIRPα expression on Mψ, specific anti-TNFα neutralization antibody only partially reversed Hepa1-6-induced reduction of SIRPα with a concentration of antibody that effectively neutralized TNFα in the coculture system (Supporting Fig. 6A). To determine whether tumor cell-induced SIRPα reduction was attributable to an increased rate of protein degradation, we examined the effects of inhibitors of protein degradation by lysosomes or the proteasome. MG132, a reagent that blocks proteasome activity, had no marked effect on the ability of tumor cells to suppress SIRPα expression. In contrast, preincubation of macrophages with dexamethasone, chloroquine, or NH4Cl, all of which inhibit lysosomal function, markedly attenuated the effect of tumor cell-induced SIRPα reduction (Supporting Fig. 6B). Confocal assay showed that, in response to Hepa1-6 stimulation, SIRPα proteins translocated from cell membrane to the cytosolic fraction and mainly localized on lysosomes (Supporting Fig. 6C). Taken together, these data suggest that, in addition to transcriptional repression of SIRPα expression (Supporting Fig. 2A), tumor cells also induce SIRPα protein degradation mainly by lysosome, which contributes to the loss of SIRPα proteins in macrophages after tumor cells stimulation.

SIRPα-CD47 interactions serve as a “don't eat me” and “self-recognized” signaling.[30] Many reports have demonstrated that the SIRPα-CD47 signaling pathway could be a therapeutic target for human tumors.[18, 30, 31] This may due to enhanced phagocytic capability of Mψ against tumor cells when treated with anti-CD47 antibodies. One problem with these series researches is that CD47-IgG antibody may not only disrupt SIRPα-CD47 interaction but also target the tumor cells by antibody-dependent cellular cytotoxicity (ADCC) or antibody-dependent phagocytosis.[32] Another problem is that most of the experiments were performed on a xenograft tumor model, which may not exactly meet the role of SIRPα in syngeneic models. It was demonstrated that additional activating signals, such as Fcγ or complement-receptor ligation, are required for Mψ-mediated phagocytosis of target cells in the absence of SIRPα-dependent inhibition.[32] In our study, we found that knockdown of SIRPα expression on syngeneic Mψ did not significantly enhance contact/phagocytosis of Hepa1-6 cells even when they were activated by IFNγ/LPS (Supporting Fig. 7A), indicating that the SIRPα-CD47 interaction may not involve tumor immunosurveillance against Hepa1-6 cells in syngeneic mice. In contrast, knockdown of SIRPα on Mψ promoted Hepa1-6 cell proliferation even without cell-cell direct interaction, suggesting that the content released by Mψ may have an important role in tumor progression (Supporting Fig. 7B-E).

In summary, our results suggest that there is a fine-tuned collaborative action between SIRPα expression on Mψ and tumor progression. Mψ with SIRPα-KD have the powerful potential to migrate and survive in tumor sites. Soluble factors derived from tumors trigger transient activation of newly recruited Mψ and reduce SIRPα expression, thereby inducing these cells to produce a large amount of cytokines, in turn leading to the down-expression of SIRPα on Mψ and ultimately create an inflammatory environment supporting tumor progression. Our findings provide new insight into the importance of SIRPα in tumor progression, which may be helpful for new antitumor drug design.

Acknowledgment

We thank Dr. Bin Gao (Laboratory of Liver Diseases, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD) and Dr. G.S. Feng (School of Medicine, Section of Molecular Biology, Division of Biological Sciences, University of California, San Diego) for helpful discussion and suggestions.

Ancillary

Advertisement