Breast Cancer Research and Treatment https://doi.org/10.1007/s10549-018-4833-8

CLINICAL TRIAL

Syntenin1/MDA-9 (SDCBP) induces immune evasion in triple-negative breast cancer by upregulating PD-L1 Jing Liu1   · Yanfang Yang1 · Hongwei Wang2 · Bin Wang3 · Kaili Zhao2 · Wenna Jiang4 · Weiwei Bai2 · Jun Liu1 · Jian Yin1 Received: 16 February 2018 / Accepted: 19 May 2018 © Springer Science+Business Media, LLC, part of Springer Nature 2018

Abstract Purpose  Syntenin1/SDCBP (syndecan binding protein), also known as melanoma differentiation associated gene-9 (MDA9), is a PDZ domain-containing molecule, which was initially identified as a key oncogene in melanoma. However, the role of syntenin1 in triple-negative breast cancer (TNBC), especially in suppression of antitumour immune response, remains unknown. Methods and Results  One hundred TNBC tissues were obtained after radical resection and used for analysis. High syntenin1 expression was associated with increased tumour size (r = 0.421, P < 0.001), presence of lymph node metastasis (r = 0.221, P = 0.044) and poor overall survival (P = 0.01) and recurrence-free survival (P = 0.007). Syntenin1 overexpression significantly promoted 4T1 tumour growth and lung metastasis in BALB/c mice by affecting ­CD8+ T cells. Western blot and flow cytometry analyses demonstrated that syntenin1 induced C ­ D8+ T cell apoptosis in vitro and in vivo through upregulating PD-L1. Western blot demonstrated that syntenin1 upregulated PD-L1 expression by inducing Tyr705 stat3 phosphorylation, which was further confirmed by stat3 inhibition study. The correlation between syntenin1 and PD-L1 was further confirmed using tumour tissues derived from patients with TNBC (r = 0.509, P < 0.001). Efficacy studies indicated that 4T1-scramble tumour benefitted from anti-PD-L1 therapy (P < 0.001); however, 4T1-syntenin1-KD demonstrated no response to anti-PDL1 treatment (P = 0.076). Conclusions  Syntenin1 exhibits a profound function in mediating T cells apoptosis by upregulating PD-L1 and thus could be used as a prognostic biomarker of TNBC. Tumoural syntenin1 expression corelated with anti-PD-L1 treatment efficacy. Targeting syntenin1-mediated T-cell suppression could be a potential strategy for improving the prognosis of patients with TNBC. Keywords  Triple negative breast cancer · Syntenin1 · PD-L1 · CD8+ T cell apoptosis · Immune suppression

Introduction Jing Liu, Yanfang Yang, Hongwei Wang and Bin Wang shared co-first authorship. Electronic supplementary material  The online version of this article (https​://doi.org/10.1007/s1054​9-018-4833-8) contains supplementary material, which is available to authorized users. * Jing Liu [email protected] * Jun Liu [email protected] * Jian Yin [email protected] Extended author information available on the last page of the article

Triple-negative breast cancer (TNBC) accounts for 20% of all breast cancer cases and is clinically defined by the absence of oestrogen receptors, progesterone receptors and HER2 overexpression. TNBC is highly heterogeneous and aggressive; as such, no approved targeted treatments are available for it, and the only treatment option is cytotoxic chemotherapy [1–3]. The interaction between programmed death 1 (PD1) on T cells and its ligand PD-L1 on tumour cells inhibits the activation, expansion and functions of antigen-specific ­CD8+ T cells and helps cancer cells evade immune destruction [4]. Among breast cancer types, the PD-L1 expression is the highest in TNBC (20–60%), the lowest in luminal A tumours

13

Vol.:(0123456789)



(3–33%) and moderate in HER2-enriched tumours (20–34%) [5]. Long-term clinical benefits of immunotherapy have been observed for various malignancies, including melanoma, metastatic non-small-cell lung cancer, renal cell cancer and ovarian cancer [6, 7]. Reducing the expression of PD-L1 is a promising method to change the immune suppression status and improve the survival rate of patients with TNBC [8]. Syntenin1 (syndecan binding protein; SDCBP) is a member of the PSD95/Dlg/ZO-1 (PDZ) domain-containing family. Syntenin1 was initially identified while screening for genes that were differentially expressed in human melanoma cells reprogrammed to terminally differentiate [9]. In this regard, syntenin1 is also known as melanoma differentiation associated gene-9 (MDA-9), which plays a causative role in the progression of melanoma [10–13]. Studies implicated syntenin1 as a key gene involved in cancer stem cell growth and survival in different cancer types. However, the role of syntenin1 in inducing tumour immune suppression has never been reported. In the current study, we investigated the role and mechanism of syntenin1 in inducing immune suppression and tumour progression by upregulating PD-L1. Syntenin1 is a negative prognostic marker in patients with TNBC and could be a promising predictor of the outcomes of anti-PD-L1 therapy.

Materials and methods Patients and tissue samples A total of 100 tissues of TNBC were collected from patients who received modified radical surgery with a histological diagnosis of TNBC at the Tianjin Medical University Cancer Institute and Hospital, China from January 2005 to January 2007. Data on clinicopathological characteristics were retrospectively collected. Postoperative follow-up of the patients was conducted by telephone or outpatient records. The last follow-up date was October 2017. The usage of these specimens and patient information was approved by the Ethics Committee of the Tianjin Medical University Cancer Institute and Hospital.

Immunohistochemistry (IHC) IHC was used to detect the expression of syntenin1 (Abnova, NBP1-31136, 1:200) and PD-L1 (Abcam, ab205921 1:200) in human tumour tissues. Stained sections were reviewed independently by two pathologists. In all of the cases, the pathologists were blinded to the clinical outcome. The final scores were calculated by multiplying the scores of the staining intensity with those of the percentage of specifically positively stained TNBC cells. The intensity of the staining was evaluated using the following criteria: 0, negative; 1, low;

13

Breast Cancer Research and Treatment

2, medium and 3, high. The percentage of specifically positively stained TNBC cells was scored as 0, 0% stained; 1, 1–25% stained; 2, 26–50% stained and 3, 51–100% stained. Five random fields (100× magnification) were evaluated under a light microscope.

Cell culture and stable cell line establishment Human TNBC cell lines MDA-MB-231 and CAL-51, and murine TNBC cell line 4T1 were obtained from the Type Culture Collection Committee of the Chinese Academy of Sciences (Shanghai, China). The cell lines were authenticated through short tandem repeat analysis method. Mycoplasma contamination was excluded in these cell lines. The cells were cultured in DMEM and RPMI-1640 supplemented with 10% foetal bovine serum at 37 °C in a humidified atmosphere of 95% air and 5% ­CO2. The complete coding sequence of the human SDCBP gene was cloned into pLV-EF1- MCS- IRES-Bsd vectors (Biosettia). Lentiviruses were produced in HEK293T cells for the stable transfection of the cell lines in accordance with the manufacturer’s instruction. An empty vector was transfected as control. A total of 1 × 105 tumour cells in 1 mL of medium with 8 µg/mL polybrene were infected with 1 mL of lentivirus supernatant. After 48 h, blasticidin (InvivoGen) was added for selection. For the cell lines with stable knockdown, shRNA sequences were designed with Biosettia’s shRNA designer (http://biose​ttia.com/suppo​r t/ shrna​-desig​ner). Three recommended sequences for each of the SDCBP genes were synthesised and cloned into the pLV-hU6-EF1α-puro or pLV-mU6-EF1α-puro vectors (Biosettia). Lentiviruses were produced in HEK293T cells. The scrambled sequences were transfected as controls. Of the three stable cell lines, the most efficient was used for further experiments (see Supplementary Method for primers and shRNA sequences).

Western blot analysis Protein lysates (20  µg) were separated by SDS–PAGE, and target protein was detected by Western blot analysis with the following antibodies: primary antibodies: syntenin1 (Abnova, NBP1-31136, 1:1000), PD-L1 (Abcam, ab205921,1:1000), phosphor-stat3 (Tyr705) (Cell Signalling Technology, 1:1000), t-stat3 (Cell Signalling Technology, 1:1000), GAPDH (Abmart,1:5000) and β-tubulin (Abmart,1:5000); secondary antibodies: goat anti-rabbit or mouse antibody at 1:5000 dilution (Abmart).

CD8+ T cell isolation and co‑culture study Human peripheral blood mononuclear cells (PBMCs) were isolated from healthy donors. C ­ D8+ T cells were purified

Breast Cancer Research and Treatment

from PBMCs by using the ­CD8+ T cell isolation kit (Mtenyi Biotec). The purity of the isolated cells reached > 98%. The purified cells were cultured in the presence of 10 µg/ mL anti-CD3 and 2 µg/ml anti-CD28 (eBiosciences) and IL-2 (20 U/ml; R&D Systems). At 3 days post-stimulation, the cells were used for further experiments. For ­CD8+ T cell apoptosis, isolated C ­ D8+T cells were co-cultured with tumour cells at 5:1 for 12 h. CD8 (Biolegend, 300906) and annexin V-PE/7-AAD (KeyGEN BioTECH) were labelled, and CD8 was gated.

Efficacy study of anti‑PD‑L1 treatment

In vitro inhibition studies

For mouse tumour tissues, samples were prepared into single-cell suspension. CD8 (Biolegend, 140404) and annexin V-PE/7-AAD (KeyGEN BioTECH) were labelled to test ­CD8+T cell apoptosis, and ­CD8+ was gated. For PD-L1 in mouse tumour tissues, total PD-L1 (Biolegend, 124308) was analysed. For intracellular IFN-γ (Biolegend, 505810) staining, 1 × 106 cells were stimulated with 5 ng/mL phorbol 12-myristate 13-acetate and 500 ng/mL ionomycin (SigmaAldrich) in the presence of GolgiStop (BD Biosciences) for 4 h, followed by surface and intracellular staining. CD8 (Biolegend, 140404), Tim3 (Biolegend, 134008) and PD1 (Biolegend, 135206) were labelled to test ­CD8+ T cell exhaustion. For in  vitro study, PD-L1 (Biolegend, 329705) was labelled to test for PD-L1 expression in different TNBC cell lines. Data analysis was performed using FCS Express version 6.

Cryptotanshinone (Selleck, 3 µM) was used for stat3 inhibition. Anti-PD-L1 (R&D system, AF156, 1 µg/mL) was used for in vitro PD-L1 neutralisation.

In vivo experiments All animal studies were approved by the Ethics Committee of Tianjin Medical University Cancer Institute and Hospital and conducted by skilled experimenters under an approved protocol in accordance with the principles and procedures outlined in the NIH Guide for the Care and Use of Laboratory Animals. Female 4–6-week-old BALB/c mice were obtained from Beijing HFK Bioscience Co., Ltd. (Beijing, People’s Republic of China). The mice were maintained in group housed in specific pathogen-free conditions. Tumour cells were harvested by trypsinisation, washed in PBS and resuspended at 1 × 106 cells/mL in Matrigel. Subsequently, 1 × 10 5 4T1-scramble or 4T1-syntenin1-KD cells were injected in the mammary fat pad of each mouse. Tumour size was determined with electronic calipers by two independent experimenters blinded to the experimental group through the following formula: volume = 1/2 L1 × (L2)2, where L1 is the long axis and L2 is the short axis of the tumour. The lungs were stored in Bouin’s solution (Sigma Aldrich, St. Louis, MO). Lung metastasis was confirmed by HE staining and calculated by counting the total number of visible nodules on the lungs.

In vivo CD8 depletion study To evaluate syntenin1 induce tumour progression through affecting ­CD8+T cells, 1 × 105 4T1-scramble or 4T1-syntenin1-KD cells were injected in the mammary fat pad of each mouse. Anti-mouse CD8α (Bioxcell, clone 53–6.7) or isotype IgG (Rat IgG2a, κ) was injected intraperitoneally at 200 µg per mouse one day before and twice a week after tumour inoculation. CD8 depletion efficacy was tested and determined to reach > 95% depletion.

To evaluate the efficacy of anti-PD-L1 therapy in 4T1-scramble tumour and 4T1-syntenin1-KD tumour, 1 × 105 4T1-scramble or 4T1-syntenin1-KD cells were injected in the mammary fat pad of each mouse. PD-L1 antibody (Bioxcell, clone 10F.9G2) or isotype IgG (Rat IgG2b, κ) was injected intraperitoneally at 200 µg per mouse twice a week from day 7 after tumour inoculation.

Flow cytometry

Statistical analyses Statistical analyses were performed with SPSS 21.0 software (SPSS Inc., Chicago, IL, USA) and GraphPad prism version 7.0 for Windows (GraphPad Software, San Diego, CA, USA). Mean ± SD values were calculated from three independent experiments. Paired/unpaired Student’s t test, one-way or two-way ANOVA test followed by appropriate post-hoc analysis, Spearman rank correlation coefficient test, χ2 tests, Kaplan–Meier survival curve, log-rank test and Cox proportional hazards regression model were used. Tumour growth curves were analysed by repeated measure two-way ANOVA (time × tumour volume), followed by Bonferroni post-hoc analysis. Each experiment was conducted independently at least three times. P < 0.05 was considered significant.

Results Syntenin1 is a negative prognostic marker in TNBC To determine whether syntenin1 expression in primary tumour tissues is associated with the survival and

13



Breast Cancer Research and Treatment

clinicopathological characteristics of patients with TNBC, we performed IHC staining of syntenin1 by using samples from 100 patients with TNBC. Syntenin1 expression was divided into two grades: average IHC score < 6: low staining group; and average IHC score ≥ 6: high staining group. Patients with TNBC and high syntenin1 levels showed significantly poorer overall survival (OS) (P = 0.01) and recurrence-free survival (RFS) (P = 0.007) than those with low syntenin1 expression levels (Fig. 1; Table 1). Variables with a P value of < 0.1 in the univariate analysis were included in multivariate analysis. In the Cox proportional hazard regression model for OS and RFS, the multivariate model suggested that syntenin1 and lymph node metastasis

are independent prognostic factors (Table 2). High syntenin1 expression was significantly correlated with tumour size (r = 0.421, P < 0.001), tumour cells Ki-67 positive rate (r = 0.397, P < 0.001) and lymph node metastasis (r = 0.221, P = 0.044) (Table 3).

Syntenin1 facilitates 4T1 growth and lung metastasis partially through ­CD8+ T cell‑mediated immune modulation To determine whether syntenin1 regulates antitumor immunity, we implanted 4T1 murine TNBC cell 4T1-scramble and 4T1-syntenin1-KD into the mammary fat pad of

Fig. 1  Syntenin1 is a negative prognostic marker for TNBC. Overall survival (a) and recurrence free survival (b) of TNBC patients in syntenin1-high group and syntenin1-low group. P < 0.001 by log rank test

Table 1  Univariate analysis of clinicopathological factors for overall survival and relapse-free survival

Variables

N

Overall survival rate (%) 3-year

Age  < 55 67 91.0  ≥ 55 33 90.9 Histological grade  I, I ~ II 45 95.6  II, II ~ III, III 55 87.3 T  T1 28 96.4  T2,T3 72 88.9 N  N0 50 92.0  N+ 50 90.0 Tumour cells Ki-67+ (%)  Low 56 91.1  High 44 90.9 Syntenin1  Low 45 95.6  High 55 87.3

13

5-year

10-year

74.6 78.8

49.3 56.4

77.8 74.5

P

Recurrence free survival rate (%)

P

3-year

5-year

10-year

0.209

80.6 81.8

59.7 63.1

38.7 56.3

0.136

57.3 46.6

0.089

84.4 78.2

66.4 56.4

54.9 35.2

0.095

85.7 72.2

67.9 44.8

0.041

92.9 76.4

75.0 55.3

60.7 37.6

0.026

86.0 66.0

69.5 33.3

< 0.001

88.0 74.0

76.0 45.7

61.4 27.0

< 0.001

80.4 70.5

62.2 37.4

0.003

87.5 81.8

75.0 42.5

56.9 27.5

0.002

91.1 63.6

75.6 31.0

< 0.001

95.6 69.1

86.7 39.5

66.7 25.4

< 0.001

Breast Cancer Research and Treatment Table 2  Multivariate analysis of clinicopathological factors for overall survival and relapse-free survival

Variables

OS Exp(B) (95% CI)

P

RFS Exp(B) (95% CI)

P

Histological grade Tumor size LN metastasis Tumour cells Ki-67+(%) Syntenin1

1.502 (0.849–2.657) 1.131 (0.543–2.357) 2.779 (1.560–4.949) 1.652(0.931–2.931) 2.418 (1.236–4.729)

0.162 0.742 0.001 0.086 0.010

1.484 (0.844–2.609) 1.205 (0.575–2.522) 2.742 (1.550–4.851) 1.672 (0.952–2.939) 2.533 (1.295–4.952)

0.170 0.621 0.001 0.074 0.007

Table 3  Correlation of Syntenin-1 expression to clinicopathological features in TNBC Variables

χ2

Syntenin1 Low

Age(years)  < 55 28 (41.8)  ≥ 55 17 (51.5) Histological grade  I, I ~ II 23 (51.5)  II, II ~ III, III 22 (40.0) Tumor size  T1 22 (78.6)  T2, T3 23 (31.9) LN metastasis  − 28 (56.0)  + 17(34.0) Tumour cells Ki-67+ (%)  ≤ 45 35 (62.5)  > 45 10 (22.7)

r

P

High 39 (58.2) 16 (48.5)

0.845

− 0.092

0.358

22 (48.9) 33 (60.0)

1.235

0.111

0.267

6 (21.4) 17.709 49 (68.1)

0.421 < 0.001

22 (44.0) 33 (66.0)

0.221

4.889

21 (37.5) 15.748 34 (77.3)

0.044

0.397 < 0.001

immunocompetent WT BABL/c mice (Supplementary Fig. 1a). Tumour volume in 4T1-scramble tumour-bearing mice was significantly larger than that in 4T1-syntenin1KD tumour-bearing mice, as detected by repeated measure two-way ANOVA (P < 0.001) (Fig. 2a). On day 34, when mice were sacrificed, lung metastasis was calculated by counting the total number of visible nodules present on the lungs, and lung metastases were further confirmed by HE staining (Fig. 2b). The mean count of metastasis nodules for 4T1-scramble group was 87.29 ± 13.06, which was significantly larger than that in mice of the 4T1-syntenin1-KD group (25.71 ± 6 0.02, P < 0.001) (Fig. 2c). Percentages of tumours infiltrating C ­ D8+ T cells were further compared. The proportions of C ­ D45+ leucocytes in 4T1-scramble and 4T1-syntenin1-KD tumours were comparable (P = 0.937) (Supplementary Fig.  1b), while the relative frequencies of ­CD8+T cells in ­CD45+ leukocytes were evaluated. The ­CD8+T cell proportion significantly decreased in 4T1-scramble tumour compared with that in 4T1-syntenin1-KD tumour (4T1-scramble: 1.21% ± 0.26%, 4T1-syntenin1-KD: 6.79% ± 1.23%, P < 0.001) (Fig. 2d). The 4T1-syntenin1-KD tumour tissues exhibited decreased C ­ D8+ T cell apoptosis

rates (4T1-scramble: 43.7% ± 10.5%, 4T1-syntenin1-KD: 20.4% ± 6.8%, P < 0.001) (Fig.  2e), higher percentages of IFN-γ+ ­CD8+ T cells (4T1-scramble: 2.00% ± 0.61%, 4T1-syntenin1-KD: 8.11% ± 2.61%, P < 0.001) (Fig. 2f) and lower percentages of P ­ D1+Tim3+CD8+T cells (4T1-scramble: 14.42% ± 2.57%, 4T1-syntenin1-KD: 3.53% ± 1.24%, P < 0.001) (Fig.  2g) compared with the 4T1-scramble tumour tissues. Total PD-L1 expression between 4T1-scramble tumour and 4T1-syntenin1-KD tumour was measured. The 4T1-scramble tumour exhibited a higher MFI of PD-L1 than 4T1-syntenin1-KD tumour (4T1-scramble: 1541.23 ± 129.55, 4T1-syntenin1-KD: 896.81 ± 113.93, P < 0.001) (Fig. 2h). Tumour implant experiments were also performed in immunocompromised BABL/c nude mice in in parallel with those in immunocompetent WT BABL/c mice (supplementary Fig. 1c). In BABL/c nude mice, tumour volume of 4T1-scramble was also larger than that of 4T1-syntenin1KD (P < 0.01 by repeated measure two-way ANOVA). The experiment was performed in triplicate, tumour volume decrease rate induced by syntenin1-KD in immunocompetent mice was significantly larger than that in nude mice (53.54%±9.76%v.s. 32.81%±3.94%,P = 0.026 by paired Student’s t test)(supplementary Fig. 1d).To further determine if syntenin1 affects tumour burden through C ­ D8+ T cell-mediated immunity, we implanted 4T1-scramble and 4T1-syntenin1-KD in wild type BALB/c mice with ­CD8+T cell depletion or isotype IgG. After ­CD8+T cell depletion, no significant difference was observed in tumour growth between 4T1-scramble-CD8 depletion and 4T1-syntenin1KD- CD8 depletion groups (P = 0.07, repeated measure two-way ANOVA main effect of group) (Fig. 2i). Tumour volumes analysed using one-way ANOVA followed by Bonferroni post-hoc analysis for each time points were shown in supplementary Fig. 2. Moreover, no significant difference was observed in lung metastasis count between 4T1-scramble-CD8 depletion and 4T1-syntenin1-KD- CD8 depletion groups. (P = 0.36) (Fig. 2j).

Tumour cells expressing high syntenin1 level induce ­CD8+ T cell apoptosis in vitro To determine if tumour syntenin1 induces C ­ D8+T cell + apoptosis in vitro, we sorted ­CD8 T cells from human

13



PBMCs and cultured them with anti-CD3, anti-CD28 antibodies and IL2 for 3 days. Flow cytometry was performed 12  h after co-cultured with tumour cells using annexin V and 7-AAD to detect ­C D8 + T cell apoptosis. The ­C D8 + T cell apoptosis rate was significantly increased in TNBC-syntenin1 compared with that in the TNBC-vector (MDA-MB-231-vector: 35.32% ± 4.43%, MDA-MB-231-syntenin1: 68.16% ± 8.02%, P = 0.0041;

13

Breast Cancer Research and Treatment

CAL-51-vector: 38.58% ± 4.35%, CAL-51-syntenin1: 65.19% ± 6.96%, P = 0.019). By contrast, the ­CD8+T cell apoptosis rate was significantly decreased in TNBC-syntenin1-KD compared with that in TNBC-scramble (MDAMB-231-scramble: 36.16% ± 4.23%, MDA-MB-231-syntenin1-KD: 16.74% ± 2.82%, P = 0.01; CAL-51-scramble: 33.78% ± 5.0%, CAL-51-syntenin1-KD: 15.52% ± 0.81%, P = 0.024) (Fig. 3a, b). The data of IFNγ secretion of C ­ D8+T

Breast Cancer Research and Treatment ◂Fig. 2  Syntenin1 facilitates 4T1 growth and lung metastasis through

­CD8+ T cell-mediated immune modulation. a The effect of syntenin1 on TNBC growth in  vivo. Immunocompetent WT BABL/c mice were inoculated with 4T1-scramble or 4T1-syntenin1-KD. Tumour volumes were measured every 3 days from day 7 to 34 after inoculation. Tumour growth curves for 4T1-scramble groups and 4T1-syntenin1-KD groups were shown. (n = 7 per group) ***In black P < 0.001 by repeated measure two-way ANOVA, (time × tumour volume, main effect of group), *In blue P < 0.05, **In blue P < 0.01, ***In blue P < 0.001 by one-way ANOVA followed by Bonferroni post-hoc analysis for each time points. b Representative HE staining of lung metastasis nodules of mice in 4T1-scramble group and mice in 4T1-syntenin-KD group c Statistical analysis of the count of lung metastasis nodules. d Representative dot plots and statistical analysis of the frequency of tumour infiltrating ­CD8+T cells, ­CD45+ cells were gated. (e-g)Representative dot plots (left) and statistical analysis (right) of the frequency of tumour infiltrating C ­ D8+T cell apoptosis (e), IFNγ+CD8+T cell (f) and C ­ D8+T cells exhaustion(g), ­CD8+T cells were gated.(h) Representative histogram (left) and statistical analysis (right) of total PDL-1 expression. ***P < 0.001 by nonpaired Student’s t test (c–h). i Tumour growth curve of 4T1-scramble and 4T1-syntenin1-KD in immunocompetent WT BABL/c mice after receiving ­CD8+ cells depletion or isotype IgG. Tumour volumes were measured every 3 days from day 7 to 28 after inoculation. (n = 10 per group) **P < 0.01,*** P < 0.001 and ns not significant, by repeated measure two-way ANOVA followed by Bonferroni post-hoc analysis (time × tumour volume, main effect of group). j Statistical analysis of lung metastasis nodules for the four groups. Two-way ANOVA with 4T1 tumor phenotype and CD8 depletion was made followed by Holm-Sidak post-hoc analyses, * P < 0.05, ** P < 0.01, *** P < 0.001

cells after co-cultured with different tumour cell lines are shown in Supplementary Fig. 3a. The proportion of activated caspase-3+CD8+T cells and CFSE dilution of ­CD8+T cells after co-cultured with different tumour cell lines are presented in Supplementary Fig. 3b and Supplementary Fig. 3c.

Tumoural Syntenin1 induced ­CD8+T cell apoptosis can be abolished by PD‑L1 neutralizing in vitro To further understand the mechanism through which syntenin1 induces ­CD8+ T cell apoptosis, we cocultured isolated ­CD8+ T cell and MDA-MB-231-syntenin1/MDA-MB231-vector with anti-PD-L1 added to the co-culture system. After PD-L1 was neutralised, no significant differences were observed in ­CD8+ T cell apoptosis rate between MDA-MB231-syntenin1 and MDA-MB-231-vector (Fig. 3c).

Syntenin1 induces PD‑L1 expression in TNBC cell lines through stat3 activation Tumours with high syntenin1 expression significantly induced ­CD8+ T cell apoptosis in vitro. To identify the underlying mechanism, we performed Western blot and flow cytometry analyses for assessing changes in PD-L1 expression in the TNBC-vector, TNBC-syntenin1, TNBC-scramble and TNBC-syntenin1-KD cell lines. The results showed that syntenin1 significantly induced PD-L1 expression in TNBC

(Fig. 4a, b). Statistical analysis of the band intensities of syntenin1 and PD-L1 relative to β-tubulin is shown in Supplementary Fig. 4. The MFI of PD-L1 in TNBC-syntenin1 was significantly higher than that in TNBC-vector (MDAMB-231-vector: 1193.19 ± 211.11 MDA-MB-231-syntenin1: 2319.68 ± 406.07, P = 0.028; CAL-51-vector: 1380.54 ± 304.91 CAL-51-syntenin1: 2579.16 ± 283.78, P = 0.017). By contrast, the MFI of PD-L1 in TNBCsyntenin1-KD was significantly lower than that in TNBCscramble (MDA-MB-231-scramble: 1081.03 ± 171.89. MDA-MB-231-syntenin1-KD: 497.02 ± 30.07, P = 0.025; CAL-51-scramble: 1207.55 ± 149.84, CAL-51-syntenin1KD: 502.02 ± 108.85, P = 0.014) (Fig. 4b). Among several kinases, Tyr705 P-stat3 plays an important role in the intracellular signal pathway of PD-L1 production. To clarify the intracellular signalling pathways involved in the regulation of PD-L1 by syntenin1, we investigated the change in Tyr705 stat3 phosphorylation through western blot analysis when expressing or downregulating syntenin1. Syntenin1 significantly increased Tyr705 stat3 phosphorylation (Fig. 4a). Statistical analysis of the band intensities of syntenin1, Tyr705 stat3 phosphorylation and t-stat3 relative to β-tubulin are shown in Supplementary Fig. 4. We treated MDA-MB-231-vector and MDA-MB-231-syntenin1 with stat3 inhibitor cryptotanshinone and found that PD-L1 cannot be regulated by syntenin1 after inhibition (Fig. 4c).

Syntenin1 was positively correlated with PD‑L1 expression in human TNBC tissues To further explore the relationship between syntenin1 and PD-L1, we performed IHC staining on consecutive sets of human TNBC tissues. The expression of PD-L1 was divided into two grades: average IHC score < 4: low staining group; average IHC score ≥ 4: high staining group. Syntenin1 expression co-localized with PD-L1 expression in consecutive sections of TNBC specimens. Importantly, syntenin1 expression in TNBC specimens was highly positively correlated with PD-L1 expression (r = 0.509, P < 0.001). (Fig. 5a–b).

Syntenin1 expression is related to the efficacy of anti‑PD‑L1 therapy for TNBC Repeated measure two-way ANOVA followed by Bonferroni post-hoc analysis was used to study the efficacy of antiPD-L1 therapy in 4T1-scramble tumours and in 4T1-syntenin1-KD tumours. Anti-PD-L1 therapy significantly inhibited the growth of 4T1-scramble tumours (P < 0.001, two-way ANOVA main effect analysis); however, the efficacy of anti-PD-L1 therapy in 4T1-syntenin1-KD tumour was not significant (P = 0.076, two-way ANOVA main effect analysis). (Fig.  6a) Data analysed using one-way

13



Breast Cancer Research and Treatment

Fig. 3  Tumoural Syntenin1 induced C ­D8+T cell apoptosis can be abolished by PD-L1 neutralizing in  vitro. a Representative dot plots and statistics analyses of C ­ D8+T cell apoptosis cocultured with MDA-MB-231-vector, MDA-MB-231-syntenin1, MDA-MB231-scramble and MDA-MB-231-syntenin1-KD. b Representative dot plots and statistics analyses of C ­ D8+T cell apoptosis cocultured with CAL-51-vector, CAL-51-syntenin1, CAL-51-scramble and

CAL-51-syntenin1-KD. *P < 0.05, **P < 0.01 by paired Student’s t-test. c Representative dot plots and statistics analyses of ­CD8+T cell apoptosis cocultured with MDA-MB-231-vector, MDA-MB-231-syntenin1 with PD-L1 neutralization or isotype IgG. Two-way ANOVA with MDA-MB-231 tumour phenotype and PD-L1 neutralization was made followed by Holm-Sidak post-hoc comparison of the four groups. *P < 0.05, ***P < 0.001

ANOVA followed by Bonferroni post-hoc analysis of the four groups for each time points were shown in Supplementary Fig. 5. Tumour weight was significantly lower in 4T1- scramble- anti-PD-L1 group compared with that in 4T1-scramble-control group (4T1-scramble-anti-PD-L1 group: 1851.27 ± 526.95; 4T1-scramble-control group: 1076.45 ± 152.29, P < 0.001) (Fig. 6b). When mice were sacrificed, lung metastasis was calculated. Anti-PD-L1 significantly reduced the number of lung metastasis nodules in 4T1-scramble tumour (4T1-scramble-control: 91.43 ± 19.86,

4T1-scramble-anti-PD-L1: 22.86 ± 7.22, P < 0.001) (Fig. 6c). However, no significant difference was observed in tumour weight and lung metastasis between 4T1-syntenin1-KD-control group and 4T1-syntenin1-KD-anti-PDL1 group. (Fig. 6b-c). Tumours were prepared into single cell suspension, and tumour infiltrating ­CD8+ T cells proportion, function and apoptosis rate were tested by flow cytometry. The proportion of C ­ D45 + cells between groups were comparable (Supplementary Fig.  6), while the relative

13

Breast Cancer Research and Treatment

Fig. 4  Syntenin1 upregulates PD-L1 expression through promoting Tyr705 stat3 phosphorylation. a Western-blots of syntenin1, PD-L1, Tyr705-p-stat3, t-stat3, βtublin in indicated cell lines. b Representative histograms and the MFI of PD-L1 expression in indicated cell lines (paired Student’s t test) c PD-L1 expression in MDA-MB-

231-vector and MDA-MB-231-syntenin1 after PD-L1 neutralization. Two-way ANOVA with MDA-MB-231 tumor phenotype and antiPD-L1 treatment was made followed by Holm-Sidak post-hoc comparison of the four groups. *P < 0.05, **P < 0.01, ***P < 0.001 and ns not significant

frequencies of C ­ D8+ T cells in C ­ D45+ leukocytes were evaluated. In 4T1-scramble tumours, anti-PD-L1 therapy significantly elevated the proportion of ­CD8+ T cells (4T1-scramble-control: 1.14 ± 0.16, 4T1-scramble-antiPD-L1: 6.92 ± 0.77, P < 0.001). Moreover, 4T1-scramble-anti-PD-L1 group exhibited decreased ­CD8+ T cell apoptosis rates (4T1-scramble-control: 40.33% ± 6.28%;

4T1-scramble-anti-PD-L1: 20.52% ± 4.70%, P < 0.001, Fig. 6e), and higher percentages of IFN-γ+ ­CD8+T cells (4T1-scramble-control: 2.41% ± 0.23%; 4T1-scramble-antiPD-L1: 5.92% ± 0.66%, P < 0.001, Fig. 6f) compared with 4T1-scramble-control group. However, no significant difference was observed between 4T1-syntenin1-KD-control group and 4T1-syntenin1-KD-anti-PD-L1 group in ­CD8+

13



Breast Cancer Research and Treatment

Fig. 5  Syntenin1 expression is positively correlated with PD-L1 expression in human TNBC tissues. a Two consecutive sets of TNBC tissues were stained with syntenin1 and PD-L1, representative IHC

figures were shown. b Correlation analyses (Spearman coefficients) and ANOVA were performed

cell proportion (Fig.  6d), IFNγ + ­CD8+ cell proportion (Fig. 6f)and apoptotic ­CD8+ cell proportion (Fig. 6e). No significant difference in ­CD8+ T cell proportion and function was observed between 4T1-scramble-anti-PD-L1 and 4T1-syntenin1-KD- anti-PD-L1. (Fig. 6d-f).

PD-L1 expression may be induced in response to inflammatory cytokines, such as IFN-γ secreted by immune cells within the tumour microenvironment, or can be driven by tumour-cell-intrinsic mechanisms. Some intrinsic pathways have been reported to upregulate PD-L1; such pathways include oncogenic signalling through PI3K, STAT3, HIF-1α, TAZ, YAP1 and EGFR, etc. [17–21] The current study found another intrinsic mechanism that induces PD-L1 upregulation. Syntenin1, also known as MDA-9, is a PDZ domain-containing molecule that exhibits a large number of interacting ligands. Syntenin1 may be an important factor in inducing malignant phenotypes in many cancers [22]. In head and neck squamous cell carcinoma, syntenin1 is a valuable biomarker for lymph node metastasis or a potential target for therapeutic intervention [23]. In human brain glioma cells, syntenin1 overexpression can activate FAK-JNK and FAK-Akt signalling and enhance migration capacity [24]. In urothelial cell carcinoma, syntenin1 promotes tumour formation through interplay with EGFR [25]. The role of syntenin1 in TNBC has attracted much attention. Paola et al. reported that syntenin1 plays an important role in NF-κB activation, and the downregulation of syntenin1 strongly increases the expression of Raf kinase inhibitor protein in TNBC [26]. Yang et al. reported that syntenin1 activates integrin β1 and ERK1/2 and mediates the migration and invasion of breast cancer cells [27]. Qian reported that

Discussion TNBCs occur predominantly in the premenopausal period of young women and show aggressive behaviour with a high potential for metastasis. Approximately 200,000 cases of TNBC are diagnosed each year, accounting for almost 20% of all breast cancers worldwide. Patients with TNBC tend to exhibit reduced OS compared with other types of breast cancer mainly because of the lack of targeted therapies. Immune therapy may be a promising therapeutic strategy for TNBC. Several monoclonal antibodies are being studied for TNBC, including those that target PD-L1, such as atezolizumab, durvalumab and avelumab, and those that target PD-1, such as nivolumab and pembrolizumab [4]. The reported response rate in anti PD1/PD-L1 therapy ranges between 3% and 38%. [14–16] The overall response rate of anti PD1/PD-L1 therapy varied from 5 to 30% in heavily pre-treated TNBC. The addition of pembrolizumab to neoadjuvant chemotherapy increases pCR in TNBC patients from 19.3 to 71.4% [5].

13

Breast Cancer Research and Treatment

Fig. 6  Syntenin1 expression is related to the efficacy of anti-PD-L1 therapy for TNBC. a Tumour growth curve of 4T1-scramble and 4T1-syntenin1-KD in immunocompetent WT BABL/c mice after receiving anti-PD-L1 treatment or isotype IgG. (n = 7 per group) Tumour volumes were measured every 3 days from day 7 to 28 after inoculation. Repeated measure two-way ANOVA was used (time × tumour volume, main effect of group) followed by Bonferroni post-hoc analysis of the four groups. b, c Statistical analysis

of tumour weight (b) and lung metastasis nodules (c) for the four groups. d Statistical analysis of the frequency of tumour infiltrating ­CD8+T cells for the four groups. e, f Statistical analysis of the frequency of tumour infiltrating ­ CD8+T cell apoptosis (e) and IFNγ+CD8+T cell (f), ­CD8+T cells was gated. Two-way ANOVA with 4T1 tumor phenotype and anti-PD-L1 treatment was made followed by Holm-Sidak post-hoc comparison of the four groups (b–f). *P < 0.05, **P < 0.01, ***P < 0.001 and ns not significant

syntenin1 expression is negatively correlated with oestrogen receptor expression [28]. In our current study, syntenin1 was also a negative prognostic marker, which is consistent with previous [27–29]. The role of syntenin1 in inducing tumor proliferation has already been reported in various studies [22, 25, 30]. As a result, it not surprising that, in our study, syntenin1-KD also decrease 4T1 tumour volume in BALB/c nude mice. All the functions of syntenin1 which has been reported previously are T cell-independent effects. The difference of tumour volume between 4T1-scramble and 4T1-syntenin1-KD was more obvious in immunocompetent host, which aroused our interest in studying the T cell-dependent function of syntenin1. We found that tumoural syntenin1 induced tumour infiltration ­CD8+T cells exhaustion and apoptosis in vivo. Our study is the first one that indentify the role of syntenin1 in ­CD8+ T cell regulation. Another interesting problem we discovered during our experiment is 4T1 tumors grow faster in immunocompetent wild type BALB/c mice than in nude mice. This indicated that T cells may not only play a role in suppressing tumors; however, during tumor progression, there may be a large number of different subpopulations of

T cells, for example, Treg cells, which have promoting effect on tumour growth. When we used wild type BALB/c with ­CD8+T cell depletion model to compare tumour growth of 4T1-scramble and 4T1-syntenin-KD, no difference of tumour volume between 4T1-scramble and 4T1-syntenin1KD could be observed after C ­ D8+T cells depletion, however, the significance was marginally significant (P = 0.07, by overall two-way ANOVA main effect analysis). After ­CD8+ T depletion, wild type BALB/c mice still have ­CD4+T cells; however, we haven’t test if syntenin1 has any effect on ­CD4+T cells including Treg cells. As a result, further studies were still needed to verified the correlation of syntenin1 and tumour microenvironment. What’s more, T cell independent function of syntenin1 may be observed in CD8 depletion model if we prolonged the observation period after tumour implanted. Syntenin1 is most commonly investigated in melanoma. Syntenin1 is very highly expressed in melanoma cells and is considered as the gatekeeper of melanoma metastasis [11]. Melanoma is also a malignancy that highly expresses PD-L1 and is the first type of tumour for which anti-PD-1/ PD-L1 therapy can result in remarkable anti-tumour

13



efficacy and durability of response. Nivolumab was approved by the FDA in 2014 for treatment of patients with unresectable or metastatic melanoma [31]. This report led us to explore whether syntenin1 can affect ­CD8+ T cells by upregulating PD-L1. We conducted an in vitro study and found that syntenin1 upregulated PD-L1 by inducing the increase in Tyr705 stat3 phosphorylation. A stat3 inhibition study by cryptotanshinone was also conducted and verified that syntenin1 upregulated PD-L1 through stat3 activation. In vitro coculture and an inhibition study confirmed that syntenin1 induces ­CD8+ T cell apoptosis by upregulating PD-L1. In the in vivo study, no significant difference in ­CD8+ T cell proportion and function could be observed after PD-L1 blocking between 4T1-scramble and 4T1-syntenin1-KD. The positive correlation between the expression of syntenin1 and PD-L1was also confirmed in human TNBC tissues. In 1998, Fisher PB et al. identified that syntenin1 is an IFN-γ-inducing gene. The treatment of human SV40-immortalised normal melanoma cells with the immune interferon INF-gamma induces growth suppression and enhances MDA-9 expression without inducing terminal differentiation [9]. Stier S et al. identified TNF-inducible genes in human umbilical arterial endothelial cells using the suppression subtractive hybridisation (SSH) method. Sequencing of the enriched cDNAs identified 12 differentially expressed genes, including vascular cell adhesion molecule-1, monocyte chemoattractant protein-1, IL-8 and IkappaBalpha. Interestingly, syntenin1, a PDZ motif-containing protein that binds to the cytoplasmic domain of syndecans, was identified by SSH [32]. PD-L1 may be expressed due to tumour cell-intrinsic processes but may also be expressed secondary to a robust anti-cancer immune response through IFN-γ and TNFα. [33, 34] We verified in the current study that syntenin1 intrinsically upregulates PD-L1. Syntenin1 may also play an important role in IFN-γ- and TNFα-induced PD-L1 expression. In vivo anti-PD-L1 efficacy study was also conducted in 4T1-scramble and 4T1-syntenin-KD tumour. Our preliminary result indicated that high syntenin1 expression may be a predictive marker for an improved response to anti-PD-L1 treatment. We also proposed that drugs that target syntenin1 may also reverse ­CD8+ T cells function by reducing PD-L1 and prevent tumour growth and metastasis. However, further studies are still required. In conclusion, our study is the first to verify syntenin1, as an oncogene in TNBC, plays a critical role in ­CD8+T cell-mediated tumour immune modulation. Our results demonstrate that the direct inhibitory effect of syntenin1 on ­CD8+ T cells serves as a key mechanism of syntenin1mediated immune evasion. We also suggest that syntenin1 induces ­CD8+ T cell apoptosis through PD-L1 by increasing Tyr705 STAT3 phosphorylation. syntenin1 may be a promising marker from anti-PD-L1treatment. Targeting

13

Breast Cancer Research and Treatment

syntenin1-mediated T-cell suppression will be a candidate strategy to improve the prognosis of patients with TNBC. Funding  The Science &Technology Development Fund of Tianjin Education Commission for Higher Education. (2017KJ198); The Foundation of Tianjin Medical University (2016KYZQ17).

References 1. Khosravi-Shahi P, Cabezon-Gutierrez L, Custodio-Cabello S (2018) Metastatic triple negative breast cancer: Optimizing treatment options, new and emerging targeted therapies. Asia Pac J Clin Oncol 14(1):32–39 2. Collignon J, Lousberg L, Schroeder H, Jerusalem G (2016) Triplenegative breast cancer: treatment challenges and solutions. Breast Cancer (Dove Med Press) 8:93–107 3. Li X, Yang J, Peng L, Sahin AA, Huo L, Ward KC, O’Regan R, Torres MA, Meisel JL (2017) Triple-negative breast cancer has worse overall survival and cause-specific survival than non-triplenegative breast cancer. Breast Cancer Res Treat 161(2):279–287 4. Kythreotou A, Siddique A, Mauri FA, Bower M, Pinato DJ (2017) Pd-L1. J Clin Pathol 71(3):189–194 5. Polk A, Svane IM, Andersson M, Nielsen D (2017) Checkpoint inhibitors in breast cancer—current status. Cancer Treat Rev 63:122–134 6. Jelinek T, Mihalyova J, Kascak M, Duras J, Hajek R (2017) PD-1/ PD-L1 inhibitors in haematological malignancies: update 2017. Immunology 152(3):357–371 7. Gong J, Chehrazi-Raffle A, Reddi S, Salgia R (2018) Development of PD-1 and PD-L1 inhibitors as a form of cancer immunotherapy: a comprehensive review of registration trials and future considerations. J Immunother Cancer 6(1):8 8. Polk A, Svane IM, Andersson M, Nielsen D (2018) Checkpoint inhibitors in breast cancer—current status. Cancer Treat Rev 63:122–134 9. Lin JJ, Jiang H, Fisher PB (1998) Melanoma differentiation associated gene-9, mda-9, is a human gamma interferon responsive gene. Gene 207(2):105–110 10. Boukerche H, Su ZZ, Prevot C, Sarkar D, Fisher PB (2008) mda-9/ Syntenin promotes metastasis in human melanoma cells by activating c-Src. Proc Natl Acad Sci U S A 105(41):15914–15919 11. Das SK, Bhutia SK, Kegelman TP, Peachy L, Oyesanya RA, Dasgupta S, Sokhi UK, Azab B, Dash R, Quinn BA et al (2012) MDA-9/syntenin: a positive gatekeeper of melanoma metastasis. Front Biosci (Landmark Ed) 17:1–15 12. Das SK, Bhutia SK, Sokhi UK, Azab B, Su ZZ, Boukerche H, Anwar T, Moen EL, Chatterjee D, Pellecchia M et al (2012) Raf kinase inhibitor RKIP inhibits MDA-9/syntenin-mediated metastasis in melanoma. Cancer Res 72(23):6217–6226 13. Peinado H, Aleckovic M, Lavotshkin S, Matei I, Costa-Silva B, Moreno-Bueno G, Hergueta-Redondo M, Williams C, GarciaSantos G, Ghajar C et al (2012) Melanoma exosomes educate bone marrow progenitor cells toward a pro-metastatic phenotype through MET. Nat Med 18(6):883–891 14. Katz H, Alsharedi M (2017) Immunotherapy in triple-negative breast cancer. Med Oncol 35(1):13 15. Dirix LY, Takacs I, Jerusalem G, Nikolinakos P, Arkenau HT, Forero-Torres A, Boccia R, Lippman ME, Somer R, Smakal M et al (2017) Avelumab, an anti-PD-L1 antibody, in patients with locally advanced or metastatic breast cancer: a phase 1b JAVELIN solid tumor study. Breast Cancer Res Treat 167(3):671–686 16. Nanda R, Chow LQ, Dees EC, Berger R, Gupta S, Geva R, Pusztai L, Pathiraja K, Aktan G, Cheng JD et al (2016) Pembrolizumab

Breast Cancer Research and Treatment

17.

18.

19.

20.

21.

22.

23.

24.

25.

in patients with advanced triple-negative breast cancer: phase Ib KEYNOTE-012 study. J Clin Oncol 34(21):2460–2467 Concha-Benavente F, Srivastava RM, Trivedi S, Lei Y, Chandran U, Seethala RR, Freeman GJ, Ferris RL (2016) Identification of the cell-intrinsic and -extrinsic pathways downstream of EGFR and IFNgamma that induce PD-L1 expression in head and neck cancer. Cancer Res 76(5):1031–1043 Kumar S, Davra V, A EO, Geng K, T LW, Birge RB (2017) Crk adaptor protein promotes PD-L1 expression, EMT and immune evasion in a murine model of triple-negative breast cancer. Oncoimmunology 7(1):e1376155 Janse van Rensburg HJ, Azad T, Ling M, Hao Y, Snetsinger B, Khanal P, Minassian LM, Graham CH, Rauh MJ, Yang X: The Hippo pathway component TAZ promotes immune evasion in human cancer through PD-L1. Cancer Res 2018 Marzec M, Zhang Q, Goradia A, Raghunath PN, Liu X, Paessler M, Wang HY, Wysocka M, Cheng M, Ruggeri BA et al (2008) Oncogenic kinase NPM/ALK induces through STAT3 expression of immunosuppressive protein CD274 (PD-L1, B7-H1). Proc Natl Acad Sci U S A 105(52):20852–20857 Noman MZ, Desantis G, Janji B, Hasmim M, Karray S, Dessen P, Bronte V, Chouaib S (2014) PD-L1 is a novel direct target of HIF-1alpha, and its blockade under hypoxia enhanced MDSCmediated T cell activation. J Exp Med 211(5):781–790 Kashyap R, Roucourt B, Lembo F, Fares J, Carcavilla AM, Restouin A, Zimmermann P, Ghossoub R (2015) Syntenin controls migration, growth, proliferation, and cell cycle progression in cancer cells. Front Pharmacol 6:241 Cui L, Cheng S, Liu X, Messadi D, Yang Y, Hu S (2016) Syntenin-1 is a promoter and prognostic marker of head and neck squamous cell carcinoma invasion and metastasis. Oncotarget 7(50):82634–82647 Zhong D, Ran JH, Tang WY, Zhang XD, Tan Y, Chen GJ, Li XS, Yan Y (2012) Mda-9/syntenin promotes human brain glioma migration through focal adhesion kinase (FAK)-JNK and FAKAKT signaling. Asian Pac J Cancer Prev 13(6):2897–2901 Dasgupta S, Menezes ME, Das SK, Emdad L, Janjic A, Bhatia S, Mukhopadhyay ND, Shao C, Sarkar D, Fisher PB (2013) Novel role of MDA-9/syntenin in regulating urothelial cell

26. 27.

28.

29.

30. 31. 32.

33.

34.

proliferation by modulating EGFR signaling. Clin Cancer Res 19(17):4621–4633 Poma P, Labbozzetta M, D’Alessandro N, Notarbartolo M (2017) NF-kappaB Is a potential molecular drug target in triple-negative breast cancers. OMICS 21(4):225–231 Yang Y, Hong Q, Shi P, Liu Z, Luo J, Shao Z (2013) Elevated expression of syntenin in breast cancer is correlated with lymph node metastasis and poor patient survival. Breast Cancer Res 15(3):R50 Qian XL, Li YQ, Yu B, Gu F, Liu FF, Li WD, Zhang XM, Fu L (2013) Syndecan binding protein (SDCBP) is overexpressed in estrogen receptor negative breast cancers, and is a potential promoter for tumor proliferation. PLoS One 8(3):e60046 Qian XL, Zhang J, Li PZ, Lang RG, Li WD, Sun H, Liu FF, Guo XJ, Gu F, Fu L (2017) Dasatinib inhibits c-src phosphorylation and prevents the proliferation of triple-negative breast cancer (TNBC) cells which overexpress syndecan-binding protein (SDCBP). PLoS One 12(1):e0171169 Liu X, Zhang X, Lv Y, Xiang J, Shi J (2014) Overexpression of syntenin enhances hepatoma cell proliferation and invasion: potential roles in human hepatoma. Oncol Rep 32(6):2810–2816 Foley K, Kim V, Jaffee E, Zheng L (2016) Current progress in immunotherapy for pancreatic cancer. Cancer Lett 381(1):244–251 Stier S, Totzke G, Grunewald E, Neuhaus T, Fronhoffs S, Sachinidis A, Vetter H, Schulze-Osthoff K, Ko Y (2000) Identification of syntenin and other TNF-inducible genes in human umbilical arterial endothelial cells by suppression subtractive hybridization. FEBS Lett 467(2–3):299–304 Li N, Wang J, Zhang N, Zhuang M, Zong Z, Zou J, Li G, Wang X, Zhou H, Zhang L et al (2018) Cross-talk between TNF-alpha and IFN-gamma signaling in induction of B7-H1 expression in hepatocellular carcinoma cells. Cancer Immunol Immunother 67(2):271–283 Hartley G, Regan D, Guth A, Dow S (2017) Regulation of PD-L1 expression on murine tumor-associated monocytes and macrophages by locally produced TNF-alpha. Cancer Immunol Immunother 66(4):523–535

Affiliations Jing Liu1   · Yanfang Yang1 · Hongwei Wang2 · Bin Wang3 · Kaili Zhao2 · Wenna Jiang4 · Weiwei Bai2 · Jun Liu1 · Jian Yin1 1

2





Department of Breast Oncoplastic Surgery, Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center for Cancer, Key Laboratory of Cancer Prevention and Therapy, Tianjin, Tianjin’s Clinical Research Center for Cancer, Key Laboratory of Breast Cancer Prevention and Therapy, Tianjin Medical University, Ministry of Education, Huan‑Hu‑Xi Road Ti‑Yuan‑Bei, Hexi District, Tianjin 300060, China

3

College of Management and Economics, Tianjin University, Tianjin 300072, China

4

Department of Laboratory Medicine, Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center for Cancer, Key Laboratory of Cancer Prevention and Therapy, Tianjin Medical University, Ministry of Education, Tianjin 300060, China



Tianjin Medical University, Ministry of Education, Tianjin 300060, China

13