Blocking CD47 may slow endometrial cancer growth by helping repair-type immune cells eat tumor cells

All human samples were obtained from the International Peace Maternity and Child Health Hospital after receiving patients' informed consent.

The paraffin-embedded tissues were sectioned into 4 μm, then deparaffinized and rehydrated with xylene and graded alcohol. Antigen retrieval was used with EDTA. Sections were incubated with mouse anti-human CD68 antibody (1 : 200; Abcam), mouse anti-human CD163 antibody (1 : 1000; Bio-Rad), and anti-human CD47 antibody (1 : 1000; GeneTex) at 4°C overnight. EXPOSE Mouse and Rabbit Specific HRP/DAB Detection IHC Kit (Abcam) was used for the following steps according to the manufacturer's protocol. All the samples were assessed by two pathologists in 10 different high-power fields (HPFs). The number of CD68⁺ cells and CD163⁺ cells were counted and the average taken. The staining intensity of CD47 was scored as 0 (no staining), 1 (weak staining), 2 (intermediate staining), or 3 (dark staining). The percentage of staining cells was scored as 0 (0–5%), 1 (1–25%), 2 (26–50%), 3 (51–75%), or 4 (76–100%). The product of the two scores were considered as the CD47 IHC score. Samples were classified as low CD47 expression (IHC score ≤ 4) or high CD47 expression (IHC score > 4).

Sections or coculture cell (mouse macrophages + EC cells) dishes were blocked with 10% calf serum and incubated with rabbit anti-human CD68 antibody (1 : 100), mouse anti-human CD163 antibody (1 : 1000), and anti-mouse F4/80 antibody (1 : 200; Abcam) at 4°C overnight. The sections were then incubated with Alexa Fluor 488 donkey anti-mouse IgG, Alexa Fluor 594 donkey anti-mouse IgG, and Alexa Fluor 594 donkey anti-rabbit IgG (Life Technologies) at room temperature for 2 hours followed by nuclear counterstaining with DAPI (Abcam). The samples were detected by confocal microscopy.

The tissues used for RNA extraction were ground by a TissueLyser. Total RNA was extracted using the TRIzol (Invitrogen) method. cDNA was synthesized from 1 μg total RNA using a reverse transcription kit (Tiangen Biotech (Beijing) Co. Ltd., China). Real-time PCR was performed using the SYBR Green Master Mix (Takara Bio Inc.) on a 7500 Real-Time PCR System (Applied Biosystems). The 2^−ΔΔCt method was used to calculate fold changes in the gene expression normalized to GAPDH. The primers that were used are shown Table 1.

From May 2017 to October 2017, 27 patients who underwent hysterectomy for EC or other benign diseases were recruited into this study. Clinical endometrial tissues were obtained from the patients after getting their informed consent. Fresh endometrial tissue specimens were transported on ice to the laboratory, cut into small pieces of 2–4 mm, and enzymatically dissociated with the Tumor Dissociation Kit (Miltenyi Biotec). ACK Lysing Buffer (Thermo Fisher Scientific) was used to remove erythrocytes. Cells were then washed twice with PBS and filtered through a 70 μm filter.

Single-cell suspensions were stained with FITC-conjugated anti-CD47 (eBioscience). 7-AAD (eBioscience) and antibodies targeted to CD45 (BD Biosciences) and CD31 (eBioscience) were used to exclude dead, nontumor cells. Flow cytometry analyses were performed on BD FACS Canto II.

Murine bone marrow cells were collected from 8-week old NOD/SCID/IL2γ^null mice (NSG, Beijing Biocytogen Co. Ltd.). 1 × 10⁶ murine bone marrow cells were planted per well in a 24-well plate and cultured with Dulbecco's Modified Eagle Medium (DMEM, Gibco) supplied with recombinant mouse macrophage colony-stimulating factor (M-CSF; 10 ng/mL, R&D Systems) for 7 days. The macrophages at this state were considered as M0 macrophages. Purity was verified by flow cytometry using F4/80 and CD11b. M1 MΦ were obtained by further treatment on day 7 with recombinant mouse interferon-gamma (IFN-γ, 20 ng/mL) and lipopolysaccharide (LPS; 100 ng/mL, Sigma-Aldrich) for 24 hours. M2 MΦ were obtained by further treatment on day 7 with IL-4 (20 ng/mL) for 24 hours. All cytokines were purchased from PeproTech Inc. unless otherwise stated.

We designed a shRNA to target the human CD47 gene (NM_198793.2↗). The CD47-shRNA and control sequences are as follows: CD47-shRNA (5′-CCGGGCACAATTACTTGGACTAGTTCTCGAGAACTAGTCCAAGTAATTGTGCTTTTT-3′), scramble-shRNA (5′-CCGGTTCTCCGAACGTGTCACGTTTCAAGAGAACGTGACACGTTCGGAGAA TTTTTG-3′). The shRNA was cloned into a lentiviral vector (pL-TO-IRES-LUC) to knockdown the expression of CD47 in EC cells (GeneChem Biotech, Shanghai, China). 1 × 10⁵ Ishikawa cells or KLE cells were transfected with 2 × 10⁶ TU shRNA-encoding lentivirus in the presence of polybrene (5 μg/mL) for 12 h. Then, the EC cells were cultured in DMEM with 10% FBS for 1 week. Puromycin (1 μg/mL) was used to select the cells that were successfully transfected. After 2 weeks, the CD47 protein expression on EC cells was detected by flow cytometry.

For the in vitro phagocytosis assay, Ishikawa cells were labeled with 1 μM CFSE using the CellTrace CFSE Cell Proliferation Kit (Invitrogen). Macrophages were incubated with 1 × 10⁶ CFSE-labeled Ishikawa cells in serum-free medium in the presence of IgG control (10 μg/mL, eBioscience) or anti-CD47 antibodies (10 μg/mL, eBioscience) for 2 h. Then, the plate was washed for 3 times with warm PBS to remove unphagocytosed Ishikawa cells.

For the immunofluorescence assay, the cocultured cells were observed through a fluorescence microscope to investigate the phagocytosis of EC cells by macrophages. For the flow cytometry assay, the cocultured cells were digested with 0.25% Trypsin-EDTA (Gibco). A single-cell suspension was incubated with a mAb specific for mouse CD16/CD32 to prevent nonspecific binding against FcγR, and it was then incubated with F4/80 antibodies (BioLegend) for 30 min at 4°C and washed twice with 2% FBS in PBS. Stained cells were subjected to flow cytometry, and data was analyzed with FlowJo software. The phagocytic index was calculated as the percentage of CFSE⁺ macrophages.

The EC cells were treated with anti-CD47 antibody (B6H12; 10 μg/mL) or control antibody (10 μg/mL) for 2 h or 12 h. Then, the EC cells were digested and the apoptosis was measured by flow cytometry using Annexin V-FITC and PI (BD Biosciences).

Twenty 7-week old female NSG mice were obtained from the Beijing Biocytogen Co. Ltd. Animal research was carried out in strict accordance with the Guideline for the Care and Use of Laboratory Animals of China. 10⁷ CD47-knockdown Ishikawa cells or control Ishikawa cells were injected subcutaneously to the left flank of NSG mice, and tumor growth was monitored. The tumor volume and body weight were measured per week. After four weeks, the mice were sacrificed by cervical dislocation and tumor bulk was removed from the animals.

Statistical analyses were performed using GraphPad Prism 6.0. Data were analyzed by unpaired Student's t-test or one-way ANOVA and were presented as the mean ± SEM. P values < 0.05 were considered statistically significant. All experiments were repeated three times.

To study the distribution of macrophages, immunohistochemistry was used to evaluate the infiltration of TAMs in normal endometrium, endometrial atypical hyperplasia (EAH), and EC (Figure 1(a)). CD68 and CD163 are relatively commonly accepted markers for total macrophages and M2 macrophages, respectively [19]. There were more macrophages infiltrating in EC than in the normal endometrium (Figures 1(b)-1(c)), which were mainly M2 macrophages (Figures 1(d)-1(e)). Besides, there was a progressive upregulation of M2 macrophages from the normal endometrium and EAH to EC (Figure 1(c)).

On the basis of the findings that macrophage infiltration was correlated with patient survival, we hypothesized that macrophage infiltration might play roles in the progression of EC. The relationship between total or M2 macrophages and clinicopathological features was analyzed (Table 2). The high number of macrophages, particularly M2 macrophages in EC, was strongly correlated with unfavorable prognostic factors, such as high pathological grade (P = 0.0118), high FIGO stage (P < 0.0001), lymph node metastasis (P = 0.0008), and lymphovascular space involvement (P = 0.0031). Our results showed that M2 TAM infiltration was closely associated with the progression of EC.

Previous researches have reported that CD47 was overexpressed in various tumors [23–28]. We found that CD47 mRNA was highly expressed in EC samples using The Cancer Genome Atlas Research Network (TCGA) database (Figure 2(a)). CD47 was highly expressed in EC tissue when analyzed by qPCR (Figure 2(b)). CD47 protein expression level was increased in EC tissues by immunochemistry (Figures 2(c)-2(d)). Considering CD47 that expressed on the cell surface interacted with SIRPα, we evaluated the CD47 expression on freshly isolated cells from EC tissue and normal endometrium by flow cytometry. Although CD47 protein was detectable on all specimens, it was significantly overexpressed in tumor tissue compared with normal tissue (Figure 2(e)). Besides, CD47 was detectable in all EC cell lines that we tested (Supplementary Figure 1).

To directly study the inhibitory effect of the interaction between CD47 and SIRPα, we performed phagocytosis assays in vitro. The majority of the NSG or C57BL/6 mouse bone marrow-derived macrophages (BMDMs) were CD11b⁺F4/80⁺ macrophages which suggested successful cultivation (Supplementary Figure 2A). NSG mouse BMDMs were cocultured with EC cells (Ishikawa cells or KLE cells) with or without the anti-CD47 antibody. Phagocytosis was evaluated by the percentage of macrophages engulfing EC cells. CD47 blockade with the anti-CD47 blocking antibody (B6H12) resulted in a significant increase in phagocytosis of Ishikawa cells by NSG and C57BL/6 mouse macrophages, while this effect was not observed with the anti-CD47 nonblocking antibody (2D3), which was specific to CD47 but did not interrupt the interaction between CD47 and SIRPα (Figures 3(a)-3(b), Supplementary Figures 2B-2C). CD47 blockade could increase phagocytosis of KLE cells by NSG BMDMs (Figure 3(c)). Some studies showed that anti-CD47 antibodies might directly induce the apoptosis of tumor cells [29, 30]. However, our results showed that the soluble anti-CD47 antibody (B6H12) could not promote the apoptosis of EC cells (Supplementary Figure 3).

We performed a CD47-knockdown experiment in Ishikawa cells and KLE cells using a lentiviral-based approach and confirmed that the successful establishment of CD47-knockdown EC cells (Figures 4(a), 4(c)). There was an increase in phagocytosis to both CD47-knockdown Ishikawa cells and CD47-knockdown KLE cells by NSG mouse BMDMs in phagocytosis assays (Figures 4(b), 4(d)).

To study whether the CD47 blockade can influence the phagocytosis of the macrophages with different states, we induced NSG mouse BMDMs into different phenotypes (Supplementary Figure 4). Then, the polarized macrophages were cocultured with Ishikawa cells in the presence of anti-CD47 blocking, nonblocking antibodies, and control IgG antibody. In the control group, we found that M1 macrophages had a greater ability of phagocytosis, compared to M2 macrophages (Figures 5(a)-5(c)). However, we found that M2 macrophages display a larger phagocytic response towards EC cells than M1 macrophages when treated with anti-CD47 blocking antibody (Figures 5(a)-5(c)). In other words, CD47⁺ Ishikawa cells mainly inhibited themselves from being engulfed by M2 macrophages rather than by M1 macrophages.

To examine whether CD47 knockdown contributed to the growth of EC, we found that the size of the tumors formed by shCD47 clones in NSG mice were smaller, when compared with those in the control group (Figures 6(a), 6(c)), although there was no difference in mouse body weight between the two groups (Figure 6(b)). More importantly, we observed that there were more macrophages which were mainly M1 macrophages in xenografted tumors formed by shCD47 clones than in the control group (Figures 6(d)-6(f)). These results suggested that CD47 knockdown inhibited the growth of the EC tumors in vivo and promoted the infiltration of macrophages which might play an important role in antitumor activity.

The TCGA database analyzed during the current study is available in the Gepia website (http://gepia.cancer-pku.cn↗). Other data used to support the findings of this study can be accessed by requesting the corresponding author through email (sunxiao000304@163.com).