Accumulation of tumor-associated macrophages (TAM) correlates with malignant progression, immune suppression, and poor prognosis. In this study, we defined a critical role for the cell-surface guidance molecule SEMA3A in differential proliferative control of TAMs. Tumor cell–derived SEMA3A restricted the proliferation of protumoral M2 macrophages but increased the proliferation of antitumoral M1, acting through the SEMA3A receptor neuropilin 1. Expansion of M1 macrophages in vivo enhanced the recruitment and activation of natural killer (NK) cells and cytotoxic CD8+ T cells to tumors, inhibiting their growth. In human breast cancer specimens, we found that immunohistochemical levels of SEMA3A correlated with the expression of genes characteristic of M1 macrophages, CD8+ T cells, and NK cells, while inversely correlating with established characters of malignancy. In summary, our results illuminate a mechanism whereby the TAM phenotype is controlled and identify the cell-surface molecule SEMA3A as a candidate for therapeutic targeting. Cancer Res; 76(11); 3166–78. ©2016 AACR.

Over the past few decades, it has become evident that tumor-associated macrophages (TAM) positively regulate tumor progression [reviewed in (1)]. Macrophages (Mφ) in normal tissue are highly plastic and can present tumor-associated antigens, lyse tumor cells, and stimulate antitumor functions in T cells and natural killer (NK) cells [reviewed in (2)]. However, most TAMs lack these antitumoral properties (3–5). In fact, high levels of TAM infiltration are associated with poor clinical outcome in most cancer types [reviewed in (6)]. During the course of tumor progression, the TAM population shifts from being predominantly an M1-like antitumoral phenotype toward an M2-like protumoral phenotype [reviewed in (7)]. Protumoral M2-Mφs suppress cytotoxic immune cells by secreting immune suppressive factors, such as TGFβ and IL10 (8). On the other hand, M1-Mφs are thought to restrict tumor growth by secreting proinflammatory factors, such as IFNα/β and IFNγ, TNFα, CXCL9, and CXCL10. These cytokines activate and recruit immune cells with antitumoral properties, such as cytotoxic T cells and NK cells [reviewed in (9)]. Therefore, altering TAMs to acquire an antitumoral phenotype is an attractive anticancer strategy.

In this study, we investigated the pleiotropic role of the tumor suppressor SEMA3A in tumor immunity. SEMA3A binds to its receptor, neuropilin 1 (NP1), which in turn leads to activation of the Plexin A family (Plexin A1-4) signaling complex, which transmits SEMA3A-mediated downstream signals regulating cell survival, proliferation, and migration (10, 11). We and others have shown that SEMA3A restricts tumor growth mainly by hampering tumor angiogenesis [reviewed in (12)]. However, the role of SEMA3A in tumor immunity remains poorly defined. On one hand, SEMA3A is reported to recruit a subset of resident NP1-expressing monocytes (defined as CD11b+Gr1) that possess antitumoral properties (13). On the other hand, Casazza and colleagues found that SEMA3A is specifically increased in hypoxic areas, where it entraps NP1-expressing Mφs, resulting in accumulation of protumoral Mφs that drive tumor progression (14). Altogether, these contradictory data prompted us to further explore the role of SEMA3A in tumor immunity.

We show here that SEMA3A does not skew Mφs toward a specific phenotype, rather it restricts M2-Mφ proliferation while enhancing M1-Mφ proliferation. The differential regulation of M1- and M2-Mφ proliferation in SEMA3A-treated tumors correlated with increased activity and recruitment of cytotoxic lymphocytes and NK cells, which consequently led to tumor growth inhibition. Until recently, it was believed that the majority of Mφs involved in homeostasis and pathologic conditions arose from recruited bone marrow (BM)–derived monocytes (15). However, recent evidence suggests that under pathologic conditions, resident Mφs may proliferate, increasing the pool of tissue Mφs as well as influencing their phenotype, thus dictating the outcome of the disease (16–18). Therefore, we propose here that altering the capacity of M1 and M2-Mφ for proliferation is a novel mechanism whereby the overall TAM phenotype may be regulated in the context of cancer.

Cell culture

4T1 tumor cells were grown in complete (supplemented with l-glutamine, penicillin/streptomycin, and 10% FBS) RPMI-1640 (Gibco/Life Technologies). 4T1 mouse mammary carcinoma cells were originally derived from a single spontaneous tumor that arose in a BALB/cfC3H mouse, and were purchased from the Karmanos Cancer Institute at Wayne State University.

CD8+ T-cell activation and NK-cell cytotoxicity assay

CD8+ T cells were cultured alone (control; CTR) or cocultured with M1-/M2-Mφs ± recombinant murine SEMA3A (rmSEMA3A) or 4T1-CTR/4T1-3A+ tumor–derived Mφs overnight (O/N), stained for IFNγ, and analyzed by flow cytometry. NK cells were stimulated with conditioned medium from M1-/M2-Mφs ± rmSEMA3A or were incubated in Transwell inserts above 4T1-CTR/4T1-3A+ tumor–derived Mφs. The NK cells were then cocultured with 51Cr-labeled 4T1 target cells, and Cr release was measured after 4 hours.

Lentiviral vectors

Tumor cells were transduced with an empty lentivirus vector or a murine SEMA3A-encoding vector, as described in ref. 19.

Animals and tumor models

Female BALB/c mice (4–6 weeks old) were purchased from Charles River Laboratory. For SEMA3A overexpression tumorigenesis experiments, 2 × 105 4T1 cells were injected in a volume of 50 μL into the mammary fat pad of anesthetized animals. For Matrigel tumorigenesis, 1 × 106 4T1 cells mixed with 1 μg rmSEMA3A in 2 mg/mL Matrigel (Corning) were injected in 100 μL subcutaneously of anesthetized animals. Tumor size was measured externally using calipers, and tumor volumes were estimated using the following equation: V = 4/3Π × (d/2)2 × D/2, where d is the minor tumor axis and D is the major tumor axis. Mice were typically euthanized 2 to 3 weeks after tumor cell injection, and tumors were weighed after dissection. All ethical permits were obtained from the Swedish Board of Agriculture.

NK-cell/CD8+ T-cell/macrophage depletion

Mice were depleted of NK cells by i.p. injection of 25 μL anti-asialo GM1 (Wako Chemicals) twice a week for the duration of the experiment. BALB/c mice were depleted of CD8+ T cells by i.p. injection with 15 mg/kg body weight of a rat anti-CD8a antibody (clone 53.6.72; BioXCell) twice a week. Mφs were depleted by i.p. injection with 20 mg/body weight of a rat anti-CSF1 antibody (clone 5A1; BioXCell) twice a week. As a control in all experiments, mice were given injections of an equivalent dose of an isotype control antibody.

Immunofluorescence

Tumor sections (10 μm) were stained for detection of NK cells (anti-mouse CD49b; eBioscience) or CD8+ T cells (anti-mouse CD8a; BD Phamingen) to detect cytotoxic lymphocytes. In order to detect MAPK phosphorylation in M1 or M2-Mφs, tumor sections where stained for F4/80 (Serotec), MRC1 (R&D Systems) and p-MAPK (Cell Signaling Technology) or F4/80, CD11c (both from Serotec), and p-MAPK (Cell Signaling Technology). Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; Molecular Probes/Life Technologies). The sections were blocked with 10% donkey serum and incubated with primary antibodies for 1.5 hours at room temperature. After washing, the slides were incubated with the secondary antibodies for 1 hour at room temperature. Images were acquired using an LSM 700 confocal microscope and analyzed and quantified using ImageJ software.

Mφ proliferation

BM-derived Mφs were isolated as described below and plated on 8-well chamber glass slides, polarized, and treated with rmSEMA3A O/N as described below. BrdUrd was added to cultures 2 hours prior to fixation with paraformaldehyde (PFA). Glass slides were treated with 1.5 mol/L HCl and 0.1 mol/L borate buffer before staining. Glass slides were blocked with 10% donkey serum and incubated with antibodies for BrdUrd (BD Biosciences) and F4/80 (Serotec) for 1.5 hours at room temperature. After washing, the slides were incubated with the appropriate secondary antibodies for 1 hour at room temperature. Images were acquired using an LSM 700 confocal microscope and analyzed and quantified using ImageJ software.

BM-derived Mφs

BM precursors were acquired by flushing the BM from the femur and tibia of 6- to 8-week-old female BALB/c mice. Cells were cultured in complete RPMI-1640 medium supplemented with 50 ng/mL M-CSF for 6 to 10 days (M0-Mφs). M0-BM–derived Mφs were treated for 4 hours with 100 ng/mL LPS and 200 units/mL IFNγ to induce an M1-Mφ phenotype, or 20 ng/mL IL4 and 10 ng/mL TGFβ for 4 or 12 hours to induce an M2-Mφ phenotype. For analysis of the effect of SEMA3A on gene expression in M0-, M1-, and M2-Mφs, Mφs were incubated ±rmSEMA3A O/N. Silencing of NP1 was performed by lentivirus-mediated gene transfer with the shCTR and shNP1 vectors (shNP1: NM_008737.2-943s21c1; Sigma), as described in Supplementary Data.

Immunohistochemistry

For immunohistochemistry, tissue microarrays (TMA) with dual 1 mm cores were generated from 83 primary breast cancers diagnosed and resected from 2011–2013 at Karolinska University Hospital, Stockholm, Sweden. The establishment of a TMA platform was approved by Ethics Committee at the Karolinska Institutet, and only patients who gave informed consent for biobanking were included. The TMA block was sectioned into 4-μm sections. Sections were treated with xylene O/N to remove paraffin, washed in ethanol (EtOH) and water, and then treated with an antigen retrieval agent before blocking nonspecific signals with PBS, 1% Triton, 10% donkey serum, and 1% BSA for 1 hour. Slides were then incubated with a specific antibody for human SEMA3A (HPA029447, Human Protein Atlas, Uppsala University; SciLife) O/N followed by incubation with the appropriate horseradish peroxidase (HRP)–conjugated secondary antibody (DAKO). The slides were developed with diaminobenzidine (DAB) solution (Vector Laboratories) and counterstained with hematoxylin. Slides were mounted using a xylene-based mounting media [CytoSeal XYL (Thermo Fisher Scientific)].

Flow cytometry

To prevent nonspecific binding of irrelevant antibodies, single-cell suspensions of tumors were preincubated with anti-CD16/32 mAb (BioLegend) on ice for 15 minutes before a 30-minute incubation on ice with specific antibodies. Cells were stained using the following antibodies for extracellular markers: CD11b (M1/70), CD45 (30-F11), Ly6C (AL-21), Ly6G (1A8), CD86 (GL1), CD11c (HL3), CD3 (500A2/145-2c11), CD8 (53-6.7), CD4 (RM4-5), CD69 (H1.2F3), CD49b (HM ALPHA2), NK 1.1 (PK136), MHC class II (I-A/I-E), MHC class I (H-2K(d)), CD115 and CD80 (16-10A1; all from BD Bioscience), F4/80 (BM8) and MRC1 (C068C2; BioLegend), and NP1, CCR2, and Cx3CR1 (R&D Systems). For IFNγ analysis, single-cell suspensions were stained with extracellular markers and then treated with Cytoperm/Cytofix reagents (BD Biosciences) and stained for IFNγ according to the manufacturer's instructions. For BrdUrd incorporation analysis, staining was performed with the BrdUrd flow kit (BD Biosciences) according to the manufacturer's instructions. For in vivo analysis, BrdUrd was injected i.p. 12 hours before sacrifice. The viability of cells was verified using 7AAD or the Live/Dead fixable dead cell stain (Life Technologies). Samples were acquired with a LSR II (BD Biosciences) and analyzed using FlowJo software (Tree Star).

Quantitative PCR

qPCR was performed using TaqMan Universal MasterMix II and TaqMan Gene Expression Assays (Applied Biosystems) in a total volume of 10 μL. The polymerase was activated at 95°C for 10 minutes, and the PCR was run in 40 cycles of denaturation at 95°C for 15 seconds and annealing/extension at 60°C for 60 seconds. TaqMan probes are listed in Supplementary Table S1. Gene expression was normalized to levels of β-actin.

Western blot

4T1-3A+ and 4T1-CTR tumor cells were lysed in Nonidet P-40 (NP-40) lysis buffer with freshly added SIGMAFAST protease inhibitors (Sigma-Aldrich). BM-derived M1-Mφs and M2-Mφs were incubated ± rmSEMA3A O/N and stimulated with 10 or 50 ng/mL CSF1 for 10 minutes. Cells were lysed in lysis buffer [20 mmol/L Tris (pH 7.5), 150 mmol/L NaCl, 10 mmol/L EDTA, 0.5% Triton X-100, and 0.5% sodium deoxycholate] supplemented with protease inhibitors [20 μg/mL aprotinin, 1 mmol/L sodium orthovanadate, and 1 mmol/L phenylmethylsulfonyl fluoride (PMSF); Sigma-Aldrich]. Protein concentration was determined by the bicinchoninic acid method (Thermo Fisher Scientific) according to the manufacturer's instructions. Total lysates were subjected to electrophoresis on 10% or 4% to 12% Bis-Tris gels and transferred onto Hybond-C extra membranes (Amersham Biosciences) or PVDF membranes (Merck Millipore). Membranes were probed with the indicated primary antibodies against Sema3A (MBS472057; MyBioSource), HSP90 (SC-13119; Santa Cruz Biotechnology), pAKT (9271; Cell Signaling Technology), total AKT (9272; Cell Signaling Technology), p-p44/42 MAPK (9101; Cell Signaling Technology), and total p44/42 MAPK (9102; Cell Signaling Technology), followed by HRP-conjugated secondary antibodies (Life Technologies). Membranes were then washed and visualized with an enhanced chemiluminescence detection system (GE Healthcare) and ImageQuant LAS 4000 (GE Healthcare Life Science). Pixels were quantified using ImageJ software (NIH).

SEMA3A expression is downregulated with the increasing degree of malignancy in human breast cancer

SEMA3A expression is downregulated in various human cancers, including epithelial ovarian, breast, gastric, and non–small cell lung cancer compared with the respective normal tissue (20–23). Immunohistochemical staining of 83 breast cancer patient samples with varying grades of malignancy (grade I and III ductal breast cancer accounted for 39 patient samples) revealed that SEMA3A expression is significantly downregulated in grade III ductal breast carcinoma compared with grade I (Fig. 1A).

Figure 1.

SEMA3A reduces tumor growth and induces accumulation of Mφs. A, immunohistochemistry for SEMA3A protein expression on TMAs consisting of samples from 39 patients with grade I or III ductal breast cancer. SEMA3A scoring was performed on a scale of 0 to 3, where 3 indicates the highest expression of SEMA3A and 0 indicates the lowest expression of SEMA3A. Results from all patients are shown (*, P < 0.05). Scale bars, 50 μm. B and C, 4T1 tumor cells were transduced with SEMA3A or a control vector. SEMA3A expression was quantified by qPCR (B) or Western blot analysis (C). qPCR data are normalized to β-actin and Western blot to HSP90. Data are presented as the fold change relative to control for both cases (**, P < 0.01). D, 4T1-3A+ and 4T1-CTR tumor cells were cultured for 3 days under normal condition. An XTT proliferation assay was performed to evaluate tumor cell proliferation. Each time point represents the mean of six replicates + the SD (****, P < 0.0001). E and F, 4T1-3A+ and 4T1-CTR tumor cells were injected into the mammary fat pad of BALB/c mice. Graphs display the tumor volume (E) and tumor weight (n = 7; *, P < 0.05; ***, P < 0.001; ****, P < 0.0001; F). G and H, 4T1-3A+ and 4T1-CTR tumors were subjected to flow cytometry. G, gating strategies for F4/80+ Mφs. Numbers indicate the percentage of cells in the gate out of all CD11b+ cells. H, graphs showing the quantification of F4/80+ cells from 4T1-3A+ and 4T1-CTR tumors (n = 6; *, P < 0.05). All data represent one out of three independent experiments and are presented as the mean + SD.

Figure 1.

SEMA3A reduces tumor growth and induces accumulation of Mφs. A, immunohistochemistry for SEMA3A protein expression on TMAs consisting of samples from 39 patients with grade I or III ductal breast cancer. SEMA3A scoring was performed on a scale of 0 to 3, where 3 indicates the highest expression of SEMA3A and 0 indicates the lowest expression of SEMA3A. Results from all patients are shown (*, P < 0.05). Scale bars, 50 μm. B and C, 4T1 tumor cells were transduced with SEMA3A or a control vector. SEMA3A expression was quantified by qPCR (B) or Western blot analysis (C). qPCR data are normalized to β-actin and Western blot to HSP90. Data are presented as the fold change relative to control for both cases (**, P < 0.01). D, 4T1-3A+ and 4T1-CTR tumor cells were cultured for 3 days under normal condition. An XTT proliferation assay was performed to evaluate tumor cell proliferation. Each time point represents the mean of six replicates + the SD (****, P < 0.0001). E and F, 4T1-3A+ and 4T1-CTR tumor cells were injected into the mammary fat pad of BALB/c mice. Graphs display the tumor volume (E) and tumor weight (n = 7; *, P < 0.05; ***, P < 0.001; ****, P < 0.0001; F). G and H, 4T1-3A+ and 4T1-CTR tumors were subjected to flow cytometry. G, gating strategies for F4/80+ Mφs. Numbers indicate the percentage of cells in the gate out of all CD11b+ cells. H, graphs showing the quantification of F4/80+ cells from 4T1-3A+ and 4T1-CTR tumors (n = 6; *, P < 0.05). All data represent one out of three independent experiments and are presented as the mean + SD.

Close modal

SEMA3A overexpression reduces tumor growth and increases accumulation of intratumoral M1-like Mφs

Because SEMA3A expression was lost in advanced breast tumors (grade III), the effect of SEMA3A on TAMs and breast cancer progression was examined in an experimental murine breast cancer model using a gain-of-function approach. 4T1 mammary tumor cells were transduced with a murine SEMA3A vector (4T1-3A+ tumor cells), whereas cells transduced with an empty vector served as controls (4T1-CTR tumor cells). Transduction of the SEMA3A vector resulted in a 15-fold increase in SEMA3A mRNA (Fig. 1B) and a 13-fold increase in protein levels (Fig. 1C). Overexpression of SEMA3A in 4T1 tumor cells increased their proliferation rate in vitro, compared with CTR cells (Fig. 1D), which was associated with increased glycolytic rate (Supplementary Fig. S1A and S1B) but similar oxygen consumption (Supplementary Fig. S1C and S1D). However, upon transplantation into the mammary fat pad of BALB/c mice, 4T1-3A+ tumors displayed a 61% decrease in volume (Fig. 1E) and a 60% decrease in weight (Fig. 1F) compared with 4T1-CTR tumors.

We next investigated the intratumoral myeloid composition, as tumor-infiltrating myeloid cells could be responsible for the SEMA3A-mediated tumor growth inhibition. Flow cytometry analysis revealed that SEMA3A overexpression enhanced the recruitment of F4/80+ Mφs into 4T1 tumors (Fig. 1G and H). Because expansion of TAMs often correlates with increased tumor growth [reviewed in (1)], we reasoned that SEMA3A might inhibit tumor growth by altering the protumoral TAM phenotype. M2-Mφs are typically defined as CD11b+Ly6GLy6ClowMHCIIlow myeloid cells, whereas antitumoral M1-Mφs are defined as CD11b+Ly6GLy6ClowMHCIIhigh myeloid cells (24, 25). Indeed, flow cytometry analysis showed a decrease in M2-like Mφs (Fig. 2A, left lower quadrant, and B) and an increase in proinflammatory M1-like Mφs (Fig. 2A, right lower quadrant, and C) within the CD11b+Ly6G myeloid cell population in 4T1-3A+ tumors compared with controls. Consistent with the above results, flow cytometry analysis revealed that F4/80+ Mφs derived from 4T1-3A+ tumors displayed enhanced levels of M1-Mφ markers involved in costimulation and antigen presentation [reviewed in (26)], including CD11c, CD86, MHC class I, and CD80 (Fig. 2D). In addition, the M2-Mφ marker mannose receptor, c type 1 (MRC1; ref. 24), was decreased in F4/80+ Mφs derived from 4T1-3A+ tumors compared with 4T1-CTR tumors (Fig. 2D). The SEMA3A-mediated antitumoral Mφ phenotype was confirmed by qPCR analysis of flow-sorted CD11b+F4/80+ TAMs from 4T1-3A+ and 4T1-CTR tumors. TAMs from 4T1-3A+ tumors displayed enhanced transcript levels of cytokines and chemokines associated with TH1 immune responses, including CXCL9 and CXCL10, IFNγ, TNFα, IL12a, and IL6 (Fig. 2E; ref. 9), and a decrease in levels of immune suppressive cytokines, such as TGFβ1, CCL17, and IL10 (Fig. 2E; ref. 2). Importantly, 80% of CD11b+F4/80+ TAMs expressed NP1, and the expression of NP1 was not modified by SEMA3A overexpression (data not shown).

Figure 2.

4T1-3A+ tumors display an increase of M1-like Mφs. A–C, 4T1-3A+ and 4T1-CTR tumors were analyzed by flow cytometry. A, dot plots display (1) MHCIIlowLy6Chigh monocytes, (2) immature MHCIIlowLy6Cintermediate, (3) immature MHCIIhighLy6Chigh/intermediate, (4) protumoral M2- (Ly6ClowMHCIIlow), and (5) antitumoral M1- (Ly6ClowMHCIIhigh) Mφs gated from CD11b+Ly6G cells from representative 4T1-3A+ and 4T1-CTR tumors. Graphs display quantification of M2-Mφs (B) and M1-Mφs (n = 7; *, P < 0.05; **, P < 0.01; C). D, CD11b+F4/80+ Mφs from 4T1-3A+ and 4T1-CTR tumors were analyzed by flow cytometry for the antigen-presentation activation surface markers CD11c, CD86, MHC class I, and CD80 and the M2-marker MRC1 (n = 7; *, P < 0.05; **, P < 0.01). E and F, total RNA was extracted from flow-sorted CD11b+F4/80+ Mφs (E) and CD11b+Ly6C+ monocytes (F). Expression of the indicated cytokines and chemokines was measured by qPCR. All levels were normalized to β-actin and are presented as the fold change in expression compared with CTR (n = 7; *, P < 0.05; **, P < 0.01; ***, P < 0.001). G, GFP+ tumor cells were flow-sorted from 4T1-3A+-GFP+ tumors and 4T1-CTR-GFP+ tumors, and gene expression was evaluated by qPCR. Expression levels were normalized to β-actin and presented as the fold change compared with CTR (n = 5; *, P < 0.05). H and I, Affymetrix probes corresponding to the SEMA3A, CD80, and CD86 genes were extracted from the publicly available Uppsala dataset (n = 253), and a Spearman rank correlation was performed. H, SEMA3A vs. CD80; Spearman rho = 0.42, P < 0.012. I, SEMA3A vs. CD86; Spearman rho = 0.24, P < 0.001. All data represent one out of three independent experiments and are presented as the mean + SD.

Figure 2.

4T1-3A+ tumors display an increase of M1-like Mφs. A–C, 4T1-3A+ and 4T1-CTR tumors were analyzed by flow cytometry. A, dot plots display (1) MHCIIlowLy6Chigh monocytes, (2) immature MHCIIlowLy6Cintermediate, (3) immature MHCIIhighLy6Chigh/intermediate, (4) protumoral M2- (Ly6ClowMHCIIlow), and (5) antitumoral M1- (Ly6ClowMHCIIhigh) Mφs gated from CD11b+Ly6G cells from representative 4T1-3A+ and 4T1-CTR tumors. Graphs display quantification of M2-Mφs (B) and M1-Mφs (n = 7; *, P < 0.05; **, P < 0.01; C). D, CD11b+F4/80+ Mφs from 4T1-3A+ and 4T1-CTR tumors were analyzed by flow cytometry for the antigen-presentation activation surface markers CD11c, CD86, MHC class I, and CD80 and the M2-marker MRC1 (n = 7; *, P < 0.05; **, P < 0.01). E and F, total RNA was extracted from flow-sorted CD11b+F4/80+ Mφs (E) and CD11b+Ly6C+ monocytes (F). Expression of the indicated cytokines and chemokines was measured by qPCR. All levels were normalized to β-actin and are presented as the fold change in expression compared with CTR (n = 7; *, P < 0.05; **, P < 0.01; ***, P < 0.001). G, GFP+ tumor cells were flow-sorted from 4T1-3A+-GFP+ tumors and 4T1-CTR-GFP+ tumors, and gene expression was evaluated by qPCR. Expression levels were normalized to β-actin and presented as the fold change compared with CTR (n = 5; *, P < 0.05). H and I, Affymetrix probes corresponding to the SEMA3A, CD80, and CD86 genes were extracted from the publicly available Uppsala dataset (n = 253), and a Spearman rank correlation was performed. H, SEMA3A vs. CD80; Spearman rho = 0.42, P < 0.012. I, SEMA3A vs. CD86; Spearman rho = 0.24, P < 0.001. All data represent one out of three independent experiments and are presented as the mean + SD.

Close modal

We next investigated whether SEMA3A altered the phenotype of intratumoral or circulating monocytes. Intriguingly, even though SEMA3A overexpression caused a slight but nonsignificant increase in total intratumoral CD11b+Ly6C+ monocytic cells (Supplementary Fig. S2A), it did not affect the gene signature of these cells (Fig. 2F). Further, SEMA3A did not modify the frequency of circulating nonclassical (CX3CR1+CD115+CD11b+; Supplementary Fig. S2C) or classical (CCR2+CD115+CD11b+; Supplementary Fig. S2B) monocytes. In sum, these data indicate that SEMA3A did not significantly alter the representation of monocyte subsets in either circulation or tumors; therefore, effects on monocytes are unlikely to be responsible for the observed accumulation of M1-Mφs in 4T1-3A+ tumors. In addition, SEMA3A overexpression did not detectably alter the tumor cell–derived expression of cytokines or chemokines that could be responsible for the accumulation of M1-like Mφs, as shown by qPCR on flow-sorted GFP+ tumor cells from 4T1-3A+ and 4T1-CTR tumors (Fig. 2G). Consistently, rmSEMA3A imbedded in Matrigel together with 4T1 tumor cells resulted in hampering of tumor volume (Supplementary Fig. S2F) and weight (Supplementary Fig. S2G), decreased accumulation of M2-like Mφs (Supplementary Fig. S2H and S2I), and increased accumulation of M1-like Mφs (Supplementary Fig. S2H and S2J) similar to what was observed for 4T1-3A+ tumors.

Further evidence that SEMA3A may affect the Mφ phenotype was obtained in human breast cancer using gene expression data from the previously published Uppsala breast cancer cohort of 253 patients. In line with the above mouse data, SEMA3A gene expression indeed correlated with both CD80 (Fig. 2H) and CD86 gene expressions (Fig. 2I).

Accumulation of cytotoxic lymphocytes regulates SEMA3A-mediated tumor growth inhibition

Because Mφs residing within 4T1-3A+ tumors expressed cytokines associated with recruitment and activation of cytotoxic lymphocytes, we evaluated the effect of SEMA3A overexpression on tumor infiltration by cytotoxic lymphocytes. Flow cytometry analysis showed that 4T1-3A+ tumors displayed an increase in CD3+ T cells (Fig. 3A) and CD49b+ NK cells (Fig. 3B) compared with control tumors. A high ratio of CD8+ T cells to CD4+ T cells correlates with increased survival in breast cancer and other tumor types, including colon cancer (27, 28). Accordingly, flow cytometry analysis showed an increased proportion of CD8+ T cells (Fig. 3C and E) within the total CD3+ T-cell population, whereas the proportion of CD4+ T cells (Fig. 3D and E) decreased in 4T1-3A+ tumors compared with controls. The increase in total CD3+ T cells in response to SEMA3A was due to enhanced accumulation of CD8+ T cells within the CD45+ population (Supplementary Fig. S3A). CD4+ T cells were unaffected by SEMA3A expression (Supplementary Fig. S3B). Immunofluorescence staining of tumor sections confirmed the increase in CD8+ T cells (Fig. 3F and G) and CD49b+ NK cells (Fig. 3H and I) in 4T1-3A+ tumors compared with controls. Also, rmSEMA3A imbedded in Matrigel together with 4T1 tumor cells ensued a high ratio of CD8+ T cells to CD4+ T cells (Supplementary Fig. S3C) and accumulation of CD49b+ NK cells (Supplementary Fig. S3D). The correlation between SEMA3A expression and CD8+ T-cell and NK-cell infiltration was next examined in the human breast cancer cohort described above. Consistent with the data obtained in the experimental model, SEMA3A gene expression correlated with CD8+ T- and NK-cell–specific genes in the human breast cancer cohort of 253 patients, including CD8A (Fig. 3J), CD8B (Fig. 3K), KLRD1 (CD94; Fig. 3L), NCR2 (CD336; Fig. 3M), NCR3 (CD337; Fig. 3N), and CRTAM (CD335; Fig. 3O).

Figure 3.

4T1-3A+ tumors show accumulation of CD8+T cells and CD49b+NK cells. A and B, flow cytometry analyses for the percentage of CD3+ T cells (A) and CD49b+ NK (B) cells out of all CD45+ cells in 4T1-3A+ and 4T1-CTR tumors (n = 7 and n = 6, respectively; **, P < 0.01). C and D, 4T1-3A+ and 4T1-CTR tumors were analyzed for CD8+ T-cell (C) and CD4+ T-cell (D) accumulation by flow cytometry (n = 7; **, P < 0.01). Numbers indicate the percentage of cells in the gate out of all CD3+ cells. E, graph shows the ratio between CD8+ T cells and CD4+ T cells out of the total CD3+ T cell population in 4T1-3A+ and 4T1-CTR tumors (n = 7; **, P < 0.01). F–I, immunofluorescent staining of 4T1-3A+ and 4T1-CTR tumor sections for CD8+ T cells (red) and nuclei (blue; F) and CD49b+ NK cells (n = 5; *, P < 0.05 and n = 4; ****, P < 0.0001, respectively; H). Scale bars, 200 and 50 μm, respectively. J–O, Affymetrix probes corresponding to SEMA3A, to the T-cell–specific genes CD8A and CD8B, and to the NK-cell–specific genes KLRD1, NCR2, NCR3, and CRTRAM (CD94, CD136, CD137, and CD135, respectively) were extracted from the publicly available Uppsala dataset (n = 253), and a Spearman rank correlation was performed. J–O, SEMA3A vs. CD8A (Spearman rho = 0.25, P < 0.001; J), SEMA3A vs. CD8B (rho = 0.4, P < 0.001; K), SEMA3A vs. KLRD1 (rho = 0.42, P < 0.001; L), SEMA3A vs. NCR2 (rho = 0.46, P < 0.001; M), SEMA3A vs. NCR3 (rho = 0.35, P < 0.001; N), and SEMA3A vs. CRTAM (rho = 0.46, P < 0.001; O). Data represent one out of three (A–E) or two (F–I) independent experiments and are presented as the mean + SD.

Figure 3.

4T1-3A+ tumors show accumulation of CD8+T cells and CD49b+NK cells. A and B, flow cytometry analyses for the percentage of CD3+ T cells (A) and CD49b+ NK (B) cells out of all CD45+ cells in 4T1-3A+ and 4T1-CTR tumors (n = 7 and n = 6, respectively; **, P < 0.01). C and D, 4T1-3A+ and 4T1-CTR tumors were analyzed for CD8+ T-cell (C) and CD4+ T-cell (D) accumulation by flow cytometry (n = 7; **, P < 0.01). Numbers indicate the percentage of cells in the gate out of all CD3+ cells. E, graph shows the ratio between CD8+ T cells and CD4+ T cells out of the total CD3+ T cell population in 4T1-3A+ and 4T1-CTR tumors (n = 7; **, P < 0.01). F–I, immunofluorescent staining of 4T1-3A+ and 4T1-CTR tumor sections for CD8+ T cells (red) and nuclei (blue; F) and CD49b+ NK cells (n = 5; *, P < 0.05 and n = 4; ****, P < 0.0001, respectively; H). Scale bars, 200 and 50 μm, respectively. J–O, Affymetrix probes corresponding to SEMA3A, to the T-cell–specific genes CD8A and CD8B, and to the NK-cell–specific genes KLRD1, NCR2, NCR3, and CRTRAM (CD94, CD136, CD137, and CD135, respectively) were extracted from the publicly available Uppsala dataset (n = 253), and a Spearman rank correlation was performed. J–O, SEMA3A vs. CD8A (Spearman rho = 0.25, P < 0.001; J), SEMA3A vs. CD8B (rho = 0.4, P < 0.001; K), SEMA3A vs. KLRD1 (rho = 0.42, P < 0.001; L), SEMA3A vs. NCR2 (rho = 0.46, P < 0.001; M), SEMA3A vs. NCR3 (rho = 0.35, P < 0.001; N), and SEMA3A vs. CRTAM (rho = 0.46, P < 0.001; O). Data represent one out of three (A–E) or two (F–I) independent experiments and are presented as the mean + SD.

Close modal

In some circumstances, cytotoxic lymphocytes accumulate in the tumor but are inactivated by factors within the tumor microenvironment, as evidenced by their reduced expression of IFNγ and CD69. The effect of SEMA3A expression on the suppression of intratumoral lymphocytes was therefore examined. Both CD8+ T cells (Fig. 4A, B, and E–G) and NK cells (Fig. 4C, D, and H–J) from 4T1-3A+ tumors displayed enhanced levels of IFNγ and increased expression of the activation marker CD69 (Fig. 4A–D and E–J, respectively) compared with controls. To determine if these cytotoxic lymphocytes could account for SEMA3A-mediated tumor growth inhibition, anti-CD8 (depletes CD8+ T cells) or anti-asialo GM1 (depletes NK cells) antibodies were administrated to 4T1-3A+ and 4T1-CTR tumor–bearing mice. Depletion of CD8+ T cells (Fig. 4K) or NK cells (Fig. 4L) increased the 4T1-3A+ tumor burden by 2.11-fold and 1.7-fold, respectively, compared with 4T1-3A+ tumors treated with isotype control antibodies (Fig. 4K and L). By contrast, 4T1-CTR tumors treated with anti-CD8 (Fig. 4K) or anti-asialo GM1 (Fig. 4L) increased only by 1.10-fold compared with 4T1-CTR tumors treated with isotype control antibodies (Fig. 4K and L). Together, these results indicate that both cytotoxic lymphocytes and NK cells are involved in SEMA3A-mediated tumor growth inhibition.

Figure 4.

Accumulation of activated cytotoxic lymphocytes regulates SEMA3A-mediated tumor growth inhibition. A and B, IFNγ in CD8+ T cells was analyzed by flow cytometry. A, representative dot plots. B, the quantification of CD8+ T cells that express IFNγ in 4T1-3A+ and 4T1-CTR tumors (n = 7; *, P < 0.05; **, P < 0.01). C and D, IFNγ in CD49b+ NK cells was analyzed by flow cytometry. C, representative dot plots. D, the quantification of CD49b+ NK cells that express IFN-γ from 4T1-3A+ and 4T1-CTR tumors (n = 7; *, P < 0.05). E–G, CD8+ T cells were analyzed for the activation marker CD69 by flow cytometry. E, the mean fluorescence intensity (MFI) for CD69 expression on CD8+ T cells in a representative histogram. Graphs show the MFI for CD69 expression (F) and the percentage of CD8+ T cells that express CD69 (n = 7; *, P < 0.05; **, P < 0.01; G). H–J, the activation marker CD69 on CD49b+ NK cells was analyzed by flow cytometry. The histogram (H) depicts representative MFI and the graph (I) shows quantification for all the tumors. J, the graph depicts the percentage of CD49b+ cells that express CD69 (n = 7; ***, P < 0.001). K, BALB/c mice were implanted with 4T1-3A+ and 4T1-CTR tumor cells and treated with a depleting antibody for CD8+ T cells (53-6.72) or an isotype control antibody twice per week. The graph depicts tumor volume over time (n = 6; ***, P < 0.001 for 3A vehicle vs. 3A anti-CD8; and ####, P < 0.0001 for CTR vehicle vs. 3A vehicle). L, BALB/c mice were implanted with 4T1-3A+ or 4T1-CTR tumor cells in the mammary fat pad and treated with anti-asialo GM1 or PBS twice per week. Tumor volume is depicted in the graph (n = 6; ****, P < 0.0001 for 3A vehicle vs. 3A anti-asialo and ####, P < 0.0001 for CTR vehicle vs. 3A vehicle). Data represent one out of three (A–J) or two (K and L) independent experiments and are presented as the mean + SD.

Figure 4.

Accumulation of activated cytotoxic lymphocytes regulates SEMA3A-mediated tumor growth inhibition. A and B, IFNγ in CD8+ T cells was analyzed by flow cytometry. A, representative dot plots. B, the quantification of CD8+ T cells that express IFNγ in 4T1-3A+ and 4T1-CTR tumors (n = 7; *, P < 0.05; **, P < 0.01). C and D, IFNγ in CD49b+ NK cells was analyzed by flow cytometry. C, representative dot plots. D, the quantification of CD49b+ NK cells that express IFN-γ from 4T1-3A+ and 4T1-CTR tumors (n = 7; *, P < 0.05). E–G, CD8+ T cells were analyzed for the activation marker CD69 by flow cytometry. E, the mean fluorescence intensity (MFI) for CD69 expression on CD8+ T cells in a representative histogram. Graphs show the MFI for CD69 expression (F) and the percentage of CD8+ T cells that express CD69 (n = 7; *, P < 0.05; **, P < 0.01; G). H–J, the activation marker CD69 on CD49b+ NK cells was analyzed by flow cytometry. The histogram (H) depicts representative MFI and the graph (I) shows quantification for all the tumors. J, the graph depicts the percentage of CD49b+ cells that express CD69 (n = 7; ***, P < 0.001). K, BALB/c mice were implanted with 4T1-3A+ and 4T1-CTR tumor cells and treated with a depleting antibody for CD8+ T cells (53-6.72) or an isotype control antibody twice per week. The graph depicts tumor volume over time (n = 6; ***, P < 0.001 for 3A vehicle vs. 3A anti-CD8; and ####, P < 0.0001 for CTR vehicle vs. 3A vehicle). L, BALB/c mice were implanted with 4T1-3A+ or 4T1-CTR tumor cells in the mammary fat pad and treated with anti-asialo GM1 or PBS twice per week. Tumor volume is depicted in the graph (n = 6; ****, P < 0.0001 for 3A vehicle vs. 3A anti-asialo and ####, P < 0.0001 for CTR vehicle vs. 3A vehicle). Data represent one out of three (A–J) or two (K and L) independent experiments and are presented as the mean + SD.

Close modal

SEMA3A-mediated induction of antitumor CD8+ T cells and NK cells is dependent on Mφs

To determine whether SEMA3A has a direct effect on activation of CD8+ T cells and NK cells or if the observed effects are dependent on SEMA3A-mediated accumulation of intratumoral M1-Mφs, a set of ex vivo coculture experiments was performed. M1 (LPS- and IFNγ-treated) or M2 (IL4- and TGFβ-treated) BMDMs exposed to rmSEMA3A were cocultured with CD8+ T cells or CD49b+ NK cells isolated from the spleens of naïve mice. The effect of SEMA3A on CD8+ T cells was investigated first. Coculture of M1-BMDMs with CD8+ T cells increased the number of CD8+ T cells that produced IFNγ, which was further increased upon addition of rmSEMA3A (Fig. 5A). By contrast, M2-polarized BMDMs suppressed the ability of CD8+ T cells to produce IFNγ; SEMA3A treatment partially reversed this inhibition (Fig. 5A). Importantly, SEMA3A had no direct effect on CD8+ T-cell activation. To determine whether intratumoral Mφs derived from SEMA3A-overexpressing tumors could similarly potentiate CD8+ T-cell activation, F4/80+ Mφs were flow-sorted from 4T1-CTR and 4T1-3A+ tumors and cocultured with splenic CD8+ T cells from the same tumor-bearing mice. Flow-sorted F4/80+ Mφs from 4T1-3A+ tumors did indeed increase the proportion of CD8+ T cell that expressed IFNγ. In contrast, flow-sorted F4/80+ Mφs from control tumors reduced the CD8+ T cells that expressed IFNγ (Fig. 5B).

Figure 5.

SEMA3A-mediated activation of CD8+ T cells and NK cells is dependent on Mφs. A, CD8+ T cells purified from the spleens of naïve mice were cocultured with or without BMDMs polarized to an M1 or an M2 phenotype ±rmSEMA3A. IFNγ-producing CD8+ T cells were analyzed by flow cytometry (n = 3; *, P < 0.05; **, P < 0.01). B, CD8+ T cells were purified from the spleens of 4T1-CTR or 4T1-3A+ tumor–bearing mice and cultured alone or cocultured with Mφs flow-sorted from respective tumor. IFNγ-producing CD8+ T cells were analyzed by flow cytometry (n = 3; *, P < 0.05). C and D, splenic NK cells were seeded in 0.4 μm transwell inserts placed in a 12-well plate seeded with M1 (C)- or M2 (D)-polarized BMDMs ± rmSEMA3A. Tumor cell lysis was analyzed by measuring the chromium release with a gamma counter (n = 3; *, NK+M1/M2 vs. NK+3A M1/M2; #, NK+3A vs. NK+3A M1/M2; and +, NK vs. NK+M1/M2). E, NK cells purified from the spleens of naïve mice were cultured alone or cocultured with F4/80+Mφs flow-sorted from 4T1-CTR or 4T1-3A+ tumors as described above. Tumor cell lysis was analyzed by measuring the chromium release with a gamma counter (n = 3; *, P < 0.05 for NK cells + CTR-Mφs vs. NK cells + 3A-Mφs; ##, P < 0.01 for NK cells vs. NK cells + 3A-Mφs). F, NK cells purified from the spleens of naïve mice were cultured with 4T1-CTR or 4T1-3A+ tumor cells. Tumor cell lysis was analyzed by measuring the chromium release with a gamma counter (n = 3). G–I, 4T1-3A+ and 4T1-CTR tumor–bearing mice were treated with a neutralizing antibody against CSF1 (5A1) to deplete Mφs. G, the graph displays tumor volume over time (n = 7; ****, P < 0.0001 for 3A vehicle vs. 3A anti-CSF1; ####, P < 0.0001 for CTR vehicle vs. 3A vehicle). Histograms show the flow cytometric quantification of CD8+ T cells (H) and NK cells (I) in 4T1-3A+ and 4T1-CTR tumor–bearing mice treated with 5A1 (n = 7; *, CTR vehicle vs. 3A vehicle; #, 3A vehicle vs. 3A anti-CSF1). Data represent one out of three (A, C, D, and F) or two (B, E, and G–I) independent experiments and are presented as the mean + SD.

Figure 5.

SEMA3A-mediated activation of CD8+ T cells and NK cells is dependent on Mφs. A, CD8+ T cells purified from the spleens of naïve mice were cocultured with or without BMDMs polarized to an M1 or an M2 phenotype ±rmSEMA3A. IFNγ-producing CD8+ T cells were analyzed by flow cytometry (n = 3; *, P < 0.05; **, P < 0.01). B, CD8+ T cells were purified from the spleens of 4T1-CTR or 4T1-3A+ tumor–bearing mice and cultured alone or cocultured with Mφs flow-sorted from respective tumor. IFNγ-producing CD8+ T cells were analyzed by flow cytometry (n = 3; *, P < 0.05). C and D, splenic NK cells were seeded in 0.4 μm transwell inserts placed in a 12-well plate seeded with M1 (C)- or M2 (D)-polarized BMDMs ± rmSEMA3A. Tumor cell lysis was analyzed by measuring the chromium release with a gamma counter (n = 3; *, NK+M1/M2 vs. NK+3A M1/M2; #, NK+3A vs. NK+3A M1/M2; and +, NK vs. NK+M1/M2). E, NK cells purified from the spleens of naïve mice were cultured alone or cocultured with F4/80+Mφs flow-sorted from 4T1-CTR or 4T1-3A+ tumors as described above. Tumor cell lysis was analyzed by measuring the chromium release with a gamma counter (n = 3; *, P < 0.05 for NK cells + CTR-Mφs vs. NK cells + 3A-Mφs; ##, P < 0.01 for NK cells vs. NK cells + 3A-Mφs). F, NK cells purified from the spleens of naïve mice were cultured with 4T1-CTR or 4T1-3A+ tumor cells. Tumor cell lysis was analyzed by measuring the chromium release with a gamma counter (n = 3). G–I, 4T1-3A+ and 4T1-CTR tumor–bearing mice were treated with a neutralizing antibody against CSF1 (5A1) to deplete Mφs. G, the graph displays tumor volume over time (n = 7; ****, P < 0.0001 for 3A vehicle vs. 3A anti-CSF1; ####, P < 0.0001 for CTR vehicle vs. 3A vehicle). Histograms show the flow cytometric quantification of CD8+ T cells (H) and NK cells (I) in 4T1-3A+ and 4T1-CTR tumor–bearing mice treated with 5A1 (n = 7; *, CTR vehicle vs. 3A vehicle; #, 3A vehicle vs. 3A anti-CSF1). Data represent one out of three (A, C, D, and F) or two (B, E, and G–I) independent experiments and are presented as the mean + SD.

Close modal

We next examined the effect of SEMA3A on NK-cell–mediated killing of 4T1 cancer cells using a chromium release assay. Consistent with the above findings, M1-BMDMs enhanced NK-cell killing of 4T1 tumor cells; this effect was further augmented by pretreatment of M1-BMDMs with SEMA3A (Fig. 5C). In contrast, M2-BMDMs suppressed NK-cell killing (Fig. 5D), and pretreatment of M2-BMDMs with SEMA3A reduced their ability to suppress NK-cell–mediated killing of 4T1 cells (Fig. 5D). In addition, F4/80+ Mφs sorted from 4T1-3A+ tumors significantly boosted the capacity of naïve NK cells to kill 4T1 cells compared with NK cells cultured alone or cocultured with Mφs sorted from 4T1-CTR tumors (Fig. 5E). Of note, NK-cell killing of 4T1-3A+ tumor cells was not increased compared with NK-cell killing of 4T1-CTR tumor cells in vitro (Fig. 5F). Altogether, these results indicate that SEMA3A, either added as a soluble factor or overexpressed by tumor cells, enhances the capacity of Mφs to activate CD8+ T cells and NK cells.

To further investigate whether the accumulation of T and NK cells in 4T1-3A+ tumors was regulated by Mφs, 4T1-3A+ and 4T1-CTR tumor–bearing mice were depleted of Mφs with an α-CSF1 antibody (Supplementary Fig. S4A; ref. 27). Administration of the α-CSF1 antibody to tumor-bearing mice increased 4T1-3A+ tumor growth by 2.5-fold compared with 4T1-3A+ tumors treated with an isotype control (Fig. 5G), whereas depletion of Mφs did not significantly affect control tumor growth, as previously shown (27). Interestingly, depletion of Mφs in 4T1-3A+ tumors blocked the effect of SEMA3A on CD8+ T-cell (Fig. 5H) and NK-cell (Fig. 5I) infiltration into the tumor. By contrast, CD8+ T-cell and NK-cell infiltration of 4T1-CTR tumors was unaffected. These results indicate that the M1-like Mφ population within 4T1-3A+ tumors induces tumor infiltration by lymphocytes and enhances their ability to kill tumor cells.

SEMA3A differentially regulates M1- and M2-Mφ proliferation

We next investigated whether SEMA3A had a direct effect on Mφ education toward an M1 phenotype. Surprisingly, SEMA3A did not affect the differentiation of unpolarized (M0) BMDMs toward an M1- or M2-like phenotype, as determined by qPCR for typical M1 and M2 genes (Supplementary Fig. S5A–S5L). SEMA3A also failed to affect the gene signature of M1- or M2-educated BMDMs ex vivo (Supplementary Fig. S5A–S5L). We next performed a wound-healing assay using M0-, M1-, or M2-skewed Mφs to determine whether SEMA3A affected their migration, which might account for the observed accumulation of M1-like Mφs in 4T1-3A+ tumors. Interestingly, SEMA3A did stimulate M0-BMDM migration but did not affect the migration of M1- or M2-BMDMs (Fig. 6A and B). In sum, these results suggest that SEMA3A does not control Mφ differentiation toward an M1 phenotype or influence the migratory capacity of M1-Mφs.

Figure 6.

SEMA3A enhances M1-Mφ proliferation, whereas M2-Mφ proliferation is decreased. A, pictograms display M0-BMDMs cultured in a monolayer ± rmSEMA3A, and the migration of cells into a scratch was measured after 24 and 48 hours (n = 3). B, graphs display quantification of migrated M0-, M1-, and M2-BMDMs ± rmSEMA3A. Scale bars, 200 μm. C, M1- and M2-polarized BMDMs ± rmSEMA3A were stained for BrdUrd (green) and F4/80 antigen (red) to measure Mφ proliferation. Quantification is shown in D (n = 4; *, P < 0.05). Scale bars, 50 μm. E, BrdUrd incorporation in M1- (CD11b+Ly6GLy6ClowMHCIIhigh) and M2- (CD11b+Ly6GLy6ClowMHCIIlow) Mφs in 4T1-3A+ and 4T1-CTR tumors analyzed by flow cytometry (n = 7; *, P < 0.05). All data represent one out of three independent experiments and are presented as the mean + SD.

Figure 6.

SEMA3A enhances M1-Mφ proliferation, whereas M2-Mφ proliferation is decreased. A, pictograms display M0-BMDMs cultured in a monolayer ± rmSEMA3A, and the migration of cells into a scratch was measured after 24 and 48 hours (n = 3). B, graphs display quantification of migrated M0-, M1-, and M2-BMDMs ± rmSEMA3A. Scale bars, 200 μm. C, M1- and M2-polarized BMDMs ± rmSEMA3A were stained for BrdUrd (green) and F4/80 antigen (red) to measure Mφ proliferation. Quantification is shown in D (n = 4; *, P < 0.05). Scale bars, 50 μm. E, BrdUrd incorporation in M1- (CD11b+Ly6GLy6ClowMHCIIhigh) and M2- (CD11b+Ly6GLy6ClowMHCIIlow) Mφs in 4T1-3A+ and 4T1-CTR tumors analyzed by flow cytometry (n = 7; *, P < 0.05). All data represent one out of three independent experiments and are presented as the mean + SD.

Close modal

Intriguingly, recent data indicate that local expansion of resident Mφs or BMDMs controls the phenotype of Mφs at inflammatory sites, including tumors (16, 17, 29). We then assessed the impact of SEMA3A on M1-Mφ proliferation. In fact, treatment with rmSEMA3A increased proliferation of M1-BMDMs compared with controls when analyzed for BrdUrd incorporation (Fig. 6C and D). By contrast, SEMA3A-treated M2-BMDMs displayed reduced BrdUrd incorporation compared with controls (Fig. 6C and D). This differential regulation of M1- and M2-BMDM proliferation by SEMA3A was further observed in vivo, as indicated by decreased BrdUrd incorporation in Ly6ClowMHCIIlow Mφs (M2-Mφs) of 4T1-3A+ tumors compared with controls, and increased BrdUrd incorporation in Ly6ClowMHCIIhigh Mφs (M1-Mφs) of 4T1-3A+ tumors compared with controls (Fig. 6E).

We next investigated the ability of SEMA3A to modulate signal transduction pathways that are important in regulating cellular proliferation. Western blot analysis revealed that SEMA3A enhanced CSF1-mediated phosphorylation of Akt (Fig. 7A) and MAPK (Fig. 7B) in M1-BMDMs. By contrast, SEMA3A inhibited CSF1-mediated phosphorylation of Akt (Fig. 7A) and MAPK (Fig. 7B) in M2-BMDMs. Immunofluorescence staining of tumor sections confirmed the increased MAPK phosphorylation in M1-Mφs (F4/80+CD11c+; Fig. 7C and D) and decreased MAPK phosphorylation in M2-Mφs (F4/80+MRC1+; Fig. 7E and F) in 4T1-3A+ tumors compared with 4T1-CTR tumors.

Figure 7.

SEMA3A-mediated differential regulation of M1-Mφ and M2-Mφ proliferation is NP1-dependent. A and B, Western blot analysis of Akt (A) and MAPK phosphorylation of M1- and M2-polarized BMDMs treated ±rmSEMA3A and ±CSF1 (B). Data are normalized to β-actin and presented as the fold change relative to CSF1-stimulated CTR-treated Mφs. C–F, 4T1-3A+ and 4T1-CTR tumor sections were stained for M1-Mφs (F4/80:blue, CD11c: red and phospho-MAPK: green; C) or M2-Mφs (F4/80:blue, MRC1:red and phospho-MAPK: green; E). Graphs display quantifications of MAPK phosphorylation in M1- (D) and M2-Mφs (n = 6; **, P < 0.01; ****, P < 0.0001; F). Scale bars, 20 μm. G–J, BMDMs were transduced with a control shRNA (shCTR) or an shRNA against NP1 (shNP1). Western blot analysis of Akt (G and I) and MAPK phosphorylation (H and J) of M1- (G and H) and M2-polarized BMDMs treated ±rmSEMA3A and ±CSF1 (I and J). Data are normalized to β-actin and presented as the fold change relative to shCTR CSF1-stimulated CTR-treated BMDMs. All data represent one out of two (C–F) or three (A, B, and G–J) independent experiments and are presented as the mean + SD.

Figure 7.

SEMA3A-mediated differential regulation of M1-Mφ and M2-Mφ proliferation is NP1-dependent. A and B, Western blot analysis of Akt (A) and MAPK phosphorylation of M1- and M2-polarized BMDMs treated ±rmSEMA3A and ±CSF1 (B). Data are normalized to β-actin and presented as the fold change relative to CSF1-stimulated CTR-treated Mφs. C–F, 4T1-3A+ and 4T1-CTR tumor sections were stained for M1-Mφs (F4/80:blue, CD11c: red and phospho-MAPK: green; C) or M2-Mφs (F4/80:blue, MRC1:red and phospho-MAPK: green; E). Graphs display quantifications of MAPK phosphorylation in M1- (D) and M2-Mφs (n = 6; **, P < 0.01; ****, P < 0.0001; F). Scale bars, 20 μm. G–J, BMDMs were transduced with a control shRNA (shCTR) or an shRNA against NP1 (shNP1). Western blot analysis of Akt (G and I) and MAPK phosphorylation (H and J) of M1- (G and H) and M2-polarized BMDMs treated ±rmSEMA3A and ±CSF1 (I and J). Data are normalized to β-actin and presented as the fold change relative to shCTR CSF1-stimulated CTR-treated BMDMs. All data represent one out of two (C–F) or three (A, B, and G–J) independent experiments and are presented as the mean + SD.

Close modal

Given that NP1 expression is enhanced in M2-Mφs compared with M1-Mφs (Supplementary Fig. S6A), we hypothesized that NP1 might be implicated in the differential SEMA3A-mediated signaling observed in M1-Mφs and M2-Mφs. Indeed, SEMA3A did not have any obvious effect on CSF1-mediated MAPK and Akt phosphorylation in M1-BMDMs (Fig. 7G and H) or M2-BMDMs (Fig. 7I and J) in which NP1 was silenced by lentiviral-mediated knockdown (shNP1). Interestingly, in CSF1-stimulated M1-BMDMs, NP1 silencing alone enhanced Akt (Fig. 7G) and MAPK (Fig. 7H) phosphorylation by 2.2-fold and 3.0-fold, respectively. Reciprocally, NP1 silencing in M2-BMDMs mimicked the effects of SEMA3A on Akt (Fig. 7I) and MAPK (Fig. 7J) phosphorylation. Thus, SEMA3A augmented MAPK signaling, Akt signaling, and proliferation of M1-BMDMs, but restricted MAPK signaling, Akt signaling, and proliferation of M2-BMDMs.

Compelling evidence links accumulation of TAMs in breast cancer to poor prognosis, malignancy, and tumor progression (27, 30, 31). Preclinical studies depleting TAMs as an anticancer strategy have shown promising results in murine models of spontaneous breast cancer and orthotopic murine breast carcinoma (3, 27, 32). However, many questions remain as to how the tumor-promoting phenotype of TAMs is generated and maintained.

Recent evidence suggests that most tissue-resident Mφs throughout the body originate during embryonic/fetal development rather than being derived from adult circulating monocytes, and are in part maintained through local proliferation (15). In pathologic conditions such as atherosclerosis, local Mφ proliferation is essential to renew the Mφ population and may alter the Mφ phenotype (16). For instance, in a model of rodent filarial nematode Litomosoides sigmodontis infection, an overall TH2-inflammatory response mediates the local self-renewal of M2-Mφs via IL4 (17). In addition, tumor cell–derived CSF1 production fosters the local proliferation of intratumoral Mφs, which accounts for the expansion of TAMs in breast carcinoma (18). Interestingly, SEMA3A boosted CSF1-mediated M1-BMDM proliferation in vitro, which resulted in increased MAPK and Akt phosphorylation. However, in M2-BMDMs, SEMA3A restricted CSF1-regulated proliferation and reduced the activation of MAPK and Akt. This effect was regulated by NP1, since SEMA3A lost its effect on MAPK and Akt phosphorylation in M1- or M2-BMDMs in which NP1 had been silenced. Intriguingly, NP1 silencing per se mimicked the effect of SEMA3A on Akt and MAPK phosphorylation in both M1- and M2-BMDMs. It is possible that SEMA3A blocks different NP1 receptor complexes in M1- and M2-Mφs, which partially depend on the NP1 expression levels, thereby resulting in the differential regulation of M1- and M2-BMDM proliferation.

The SEMA3A-mediated differential regulation of M1-and M2-Mφ proliferation and regulation of MAPK phosphorylation observed in vitro was also observed in the experimental 4T1 mammary tumor model. Intratumoral Mφs derived from SEMA3A+ tumors displayed a potentially immunostimulatory/antitumoral phenotype, as evidenced by increased expression of surface markers that are present on activated antigen-presenting cells, including MHC class II, MHC class I, CD11c, CD86, and CD80. The correlation between SEMA3A expression and accumulation of M1-Mφs was also evident in patient material, where the SEMA3A gene expression was correlated with the gene expression of CD80 and CD86: markers associated with M1-Mφs. Moreover, SEMA3A-derived intratumoral Mφs displayed enhanced levels of the proinflammatory cytokines IFNγ, IL6, TNFα, CXCL9, and CXCL10 and decreased levels of the immune suppressive factors TGFβ and CCL17. CXCL9 and CXCL10 are unswervingly described as antitumoral because of their ability to activate and recruit NK cells and CD8+ T cells in both murine models (33–35) as well as in human patients; therefore, their expression correlates with good prognosis (28, 36). Even though IL6 and TNFα are often associated with tumor progression (37, 38), they are also considered classical proinflammatory cytokines that can promote CD8+ T-cell proliferation and activation (39–41). All considered, these M1-like Mφs displayed features that could potentially activate and recruit CD8+ T and NK cells. Indeed, by depleting Mφs, we could show that SEMA3A-mediated tumor growth inhibition was dependent on the intratumoral Mφ phenotype, which in turn controlled the recruitment of cytotoxic lymphocytes. Importantly, the depletion of Mφ cells leading to reduced recruitment of CD8+ T cells and NK cells also negated the inhibitory effect of SEMA3A on tumor growth. Consistent with previous work, depletion of Mφs did not significantly affect tumor growth (3, 27, 32). One explanation could be that the depletion of macrophages with the ant-CSF1 is not sufficient in order to activate cytotoxic T cells and NK cells that will in turn kill the tumor cells. Importantly, SEMA3A did not have a direct effect on neither CD8+ T- nor NK-cell activation, in line with that only 1% of intratumoral and splenocytic CD8+ T and NK cells express NP1 (data not shown).

Finally, direct injection of SEMA3A into the tumor microenvironment efficiently mimicked the effects of tumor cell–derived overexpression of SEMA3A on the microenvironment in terms of restricting tumor growth accompanied with increased accumulation of M1-Mφs, cytotoxic T cells and NK cells, indicating that SEMA3A could potentially be used as theoretical anticancer tool.

In conclusion, we propose that SEMA3A-mediated selective expansion of M1-like Mφs could overcome the immune suppressive features of the tumor microenvironment and stimulate activation and recruitment of cytotoxic lymphocytes. Supporting SEMA3A signaling thus appears as a novel potential cancer treatment.

No potential conflicts of interest were disclosed.

Conception and design: M. Wallerius, T. Wallmann, C. Rolny

Development of methodology: M. Wallerius, T. Wallmann, J. Östling, P. Pangigadde, J. Hartman, J. Bergh, C. Rolny

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Wallerius, T. Wallmann, M. Bartish, J. Östling, N.P. Tobin, E. Nygren, P. Pangigadde, P. Pellegrini, M.L. Squadrito, F. Pontén, J. Hartman, J. Bergh, A. De Milito, M. De Palma, J. Andersson, C. Rolny

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Wallerius, T. Wallmann, J. Östling, A. Mezheyeuski, N.P. Tobin, P. Pangigadde, P. Pellegrini, A. De Milito, A. Östman, J. Andersson, C. Rolny

Writing, review, and/or revision of the manuscript: M. Wallerius, T. Wallmann, M. Bartish, N.P. Tobin, P. Pangigadde, J. Hartman, M. De Palma, A. Östman, J. Andersson, C. Rolny

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Bergh, C. Rolny

Study supervision: C. Rolny

The authors thank Inger Bodin for technical support of histology lab service CCK core facility.

This study was supported by the grants from the Swedish Cancer Society (2013/852), Swedish Scientific Council (2013-5982), Gustaf V:th Radiumhemmets forskningsfonder (134122) foundation and BRECT (to C. Rolny). T. Wallmann and M. Bartish were supported by the KI PhD foundation. A. Östman received support from Swedish Cancer Society, Radiumhemmets forskningsfonder, and Swedish Research Council (STARGET Linneaus-grant). J. Hartman was supported by grants from the Swedish Society for Medical Research. A. De Milito received grants from Swedish Cancer Society (2012/415). J. Andersson was supported by the Swedish Cancer Society (2013/771) and Stiftelsen Olle Engkvist Byggmästare (2012/596). Jonas Bergh's research group is supported by grants from the Swedish Cancer Society, the Swedish Research Council-Starget-Linné, the research funds at Radiumhemmet, the KI-AZ support, KI-Stockholm County Council BRECT, ALF and FoUU and KAW.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Hanahan
D
,
Coussens
LM
. 
Accessories to the crime: Functions of cells recruited to the tumor microenvironment
.
Cancer Cell
2012
;
21
:
309
22
.
2.
Coussens
LM
,
Zitvogel
L
,
Palucka
AK
. 
Neutralizing tumor-promoting chronic inflammation: A magic bullet?
Science
2013
;
339
:
286
91
.
3.
Rolny
C
,
Mazzone
M
,
Tugues
S
,
Laoui
D
,
Johansson
I
,
Coulon
C
, et al
HRG inhibits tumor growth and metastasis by inducing macrophage polarization and vessel normalization through downregulation of PlGF
.
Cancer Cell
2011
;
19
:
31
44
.
4.
Andreu
P
,
Johansson
M
,
Affara
NI
,
Pucci
F
,
Tan
T
,
Junankar
S
, et al
FcRgamma activation regulates inflammation-associated squamous carcinogenesis
.
Cancer Cell
2010
;
17
:
121
34
.
5.
Germano
G
,
Frapolli
R
,
Belgiovine
C
,
Anselmo
A
,
Pesce
S
,
Liguori
M
, et al
Role of macrophage targeting in the antitumor activity of trabectedin
.
Cancer Cell
2013
;
23
:
249
62
.
6.
De Palma
M
,
Coukos
G
,
Hanahan
D
. 
A new twist on radiation oncology: Low-dose irradiation elicits immunostimulatory macrophages that unlock barriers to tumor immunotherapy
.
Cancer Cell
2013
;
24
:
559
61
.
7.
Mantovani
A
,
Allavena
P
. 
The interaction of anticancer therapies with tumor-associated macrophages
.
J Exp Med
2015
;
212
:
435
45
.
8.
De Palma
M
,
Lewis
CE
. 
Macrophage regulation of tumor responses to anticancer therapies
.
Cancer Cell
2013
;
23
:
277
86
.
9.
Galdiero
MR
,
Garlanda
C
,
Jaillon
S
,
Marone
G
,
Mantovani
A
. 
Tumor associated macrophages and neutrophils in tumor progression
.
J Cell Physiol
2013
;
228
:
1404
12
.
10.
Tamagnone
L
,
Artigiani
S
,
Chen
H
,
He
Z
,
Ming
GI
,
Song
H
, et al
Plexins are a large family of receptors for transmembrane, secreted, and GPI-anchored semaphorins in vertebrates
.
Cell
1999
;
99
:
71
80
.
11.
Sakurai
A
,
Doci
CL
,
Gutkind
JS
. 
Semaphorin signaling in angiogenesis, lymphangiogenesis and cancer
.
Cell Res
2012
;
22
:
23
32
.
12.
Eissler
N
,
Rolny
C
. 
The role of immune semaphorins in cancer progression
.
Exp Cell Res
2013
;
319
:
1635
43
.
13.
Carrer
A
,
Moimas
S
,
Zacchigna
S
,
Pattarini
L
,
Zentilin
L
,
Ruozi
G
, et al
Neuropilin-1 identifies a subset of bone marrow Gr1- monocytes that can induce tumor vessel normalization and inhibit tumor growth
.
Cancer Res
2012
;
72
:
6371
81
.
14.
Casazza
A
,
Laoui
D
,
Wenes
M
,
Rizzolio
S
,
Bassani
N
,
Mambretti
M
, et al
Impeding macrophage entry into hypoxic tumor areas by Sema3A/Nrp1 signaling blockade inhibits angiogenesis and restores antitumor immunity
.
Cancer Cell
2013
;
24
:
695
709
.
15.
Epelman
S
,
Lavine
KJ
,
Randolph
GJ
. 
Origin and functions of tissue macrophages
.
Immunity
2014
;
41
:
21
35
.
16.
Robbins
CS
,
Hilgendorf
I
,
Weber
GF
,
Theurl
I
,
Iwamoto
Y
,
Figueiredo
JL
, et al
Local proliferation dominates lesional macrophage accumulation in atherosclerosis
.
Nat Med
2013
;
19
:
1166
72
.
17.
Jenkins
SJ
,
Ruckerl
D
,
Cook
PC
,
Jones
LH
,
Finkelman
FD
,
van Rooijen
N
, et al
Local macrophage proliferation, rather than recruitment from the blood, is a signature of TH2 inflammation
.
Science
2011
;
332
:
1284
8
.
18.
Tymoszuk
P
,
Evens
H
,
Marzola
V
,
Wachowicz
K
,
Wasmer
MH
,
Datta
S
, et al
In situ proliferation contributes to accumulation of tumor-associated macrophages in spontaneous mammary tumors
.
Eur J Immunol
2014
;
44
:
2247
62
.
19.
Follenzi
A
,
Naldini
L
. 
HIV-based vectors. Preparation and use
.
Methods Mol Med
2002
;
69
:
259
74
.
20.
Jiang
H
,
Qi
L
,
Wang
F
,
Sun
Z
,
Huang
Z
,
Xi
Q
. 
Decreased semaphorin 3A expression is associated with a poor prognosis in patients with epithelial ovarian carcinoma
.
Int J Mol Med
2015
;
35
:
1374
80
.
21.
Tang
C
,
Gao
X
,
Liu
H
,
Jiang
T
,
Zhai
X
. 
Decreased expression of SEMA3A is associated with poor prognosis in gastric carcinoma
.
Int J Clin Exp Pathol
2014
;
7
:
4782
94
.
22.
Zhou
H
,
Wu
A
,
Fu
W
,
Lv
Z
,
Zhang
Z
. 
Significance of semaphorin-3A and MMP-14 protein expression in non-small cell lung cancer
.
Oncol Lett
2014
;
7
:
1395
400
.
23.
Mishra
R
,
Thorat
D
,
Soundararajan
G
,
Pradhan
SJ
,
Chakraborty
G
,
Lohite
K
, et al
Semaphorin 3A upregulates FOXO 3a-dependent MelCAM expression leading to attenuation of breast tumor growth and angiogenesis
.
Oncogene
2015
;
34
:
1584
95
.
24.
Murray
PJ
,
Allen
JE
,
Biswas
SK
,
Fisher
EA
,
Gilroy
DW
,
Goerdt
S
, et al
Macrophage activation and polarization: Nomenclature and experimental guidelines
.
Immunity
2014
;
41
:
14
20
.
25.
Laoui
D
,
Van Overmeire
E
,
Di Conza
G
,
Aldeni
C
,
Keirsse
J
,
Morias
Y
, et al
Tumor hypoxia does not drive differentiation of tumor-associated macrophages but rather fine-tunes the M2-like macrophage population
.
Cancer Res
2014
;
74
:
24
30
.
26.
Murray
PJ
,
Wynn
TA
. 
Protective and pathogenic functions of macrophage subsets
.
Nat Rev Immunol
2011
;
11
:
723
37
.
27.
DeNardo
DG
,
Brennan
DJ
,
Rexhepaj
E
,
Ruffell
B
,
Shiao
SL
,
Madden
SF
, et al
Leukocyte complexity predicts breast cancer survival and functionally regulates response to chemotherapy
.
Cancer Discov
2011
;
1
:
54
67
.
28.
Galon
J
,
Angell
HK
,
Bedognetti
D
,
Marincola
FM
. 
The continuum of cancer immunosurveillance: Prognostic, predictive, and mechanistic signatures
.
Immunity
2013
;
39
:
11
26
.
29.
Bleriot
C
,
Dupuis
T
,
Jouvion
G
,
Eberl
G
,
Disson
O
,
Lecuit
M
. 
Liver-resident macrophage necroptosis orchestrates type 1 microbicidal inflammation and type-2-mediated tissue repair during bacterial infection
.
Immunity
2015
;
42
:
145
58
.
30.
Mahmoud
SM
,
Lee
AH
,
Paish
EC
,
Macmillan
RD
,
Ellis
IO
,
Green
AR
. 
Tumour-infiltrating macrophages and clinical outcome in breast cancer
.
J Clin Pathol
2012
;
65
:
159
63
.
31.
Ruffell
B
,
Au
A
,
Rugo
HS
,
Esserman
LJ
,
Hwang
ES
,
Coussens
LM
. 
Leukocyte composition of human breast cancer
.
Proc Natl Acad Sci U S A
2012
;
109
:
2796
801
.
32.
Lin
EY
,
Nguyen
AV
,
Russell
RG
,
Pollard
JW
. 
Colony-stimulating factor 1 promotes progression of mammary tumors to malignancy
.
J Exp Med
2001
;
193
:
727
40
.
33.
Wong
JL
,
Berk
E
,
Edwards
RP
,
Kalinski
P
. 
IL-18-primed helper NK cells collaborate with dendritic cells to promote recruitment of effector CD8+ T cells to the tumor microenvironment
.
Cancer Res
2013
;
73
:
4653
62
.
34.
Wennerberg
E
,
Kremer
V
,
Childs
R
,
Lundqvist
A
. 
CXCL10-induced migration of adoptively transferred human natural killer cells toward solid tumors causes regression of tumor growth in vivo
.
Cancer Immunol Immunother
2015
;
64
:
225
35
.
35.
Saudemont
A
,
Jouy
N
,
Hetuin
D
,
Quesnel
B
. 
NK cells that are activated by CXCL10 can kill dormant tumor cells that resist CTL-mediated lysis and can express B7-H1 that stimulates T cells
.
Blood
2005
;
105
:
2428
35
.
36.
Curtis
C
,
Shah
SP
,
Chin
SF
,
Turashvili
G
,
Rueda
OM
,
Dunning
MJ
, et al
The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups
.
Nature
2012
;
486
:
346
52
.
37.
Taniguchi
K
,
Karin
M
. 
IL-6 and related cytokines as the critical lynchpins between inflammation and cancer
.
Semin Immunol
2014
;
26
:
54
74
.
38.
Grivennikov
SI
,
Karin
M
. 
Inflammatory cytokines in cancer: tumour necrosis factor and interleukin 6 take the stage
.
Ann Rheum Dis
2011
;
70
Suppl 1
:
i104
8
.
39.
Fisher
DT
,
Appenheimer
MM
,
Evans
SS
. 
The two faces of IL-6 in the tumor microenvironment
.
Semin Immunol
2014
;
26
:
38
47
.
40.
Ding
X
,
Yang
W
,
Shi
X
,
Du
P
,
Su
L
,
Qin
Z
, et al
TNF receptor 1 mediates dendritic cell maturation and CD8 T cell response through two distinct mechanisms
.
J Immunol
2011
;
187
:
1184
91
.
41.
Calzascia
T
,
Pellegrini
M
,
Hall
H
,
Sabbagh
L
,
Ono
N
,
Elford
AR
, et al
TNF-alpha is critical for antitumor but not antiviral T cell immunity in mice
.
J Clin Invest
2007
;
117
:
3833
45
.