Hippo/YAP signaling choreographs the tumor immune microenvironment to promote triple negative breast cancer progression via TAZ/IL-34 axis.

Growing evidence suggests that the bidirectional interactions between cancer cells and their surrounding environment, namely the tumor microenvironment (TME), contribute to cancer progression, metastasis, and resistance to treatment. Intense investigation of the Hippo pathway, which controls multiple central cellular functions in tumorigenesis, was focused on cancer cells. However, the role of the Hippo pathway in modulating tumor-stromal interactions in triple-negative breast cancer remains largely unknown. Therefore, this study focused on revealing the effects of Hippo-YAP/TAZ signaling on the immune microenvironment. Our findings reveal that the activity of the Hippo pathway is associated with worse disease outcomes in TNBC and could increase TAM infiltration through the TAZ/IL-34 axis, leading to an immunosuppressive microenvironment and impairing the treatment efficacy of anti-PD-L1. Thus, the TAZ/IL-34 axis may serve as a novel target for TNBC patients.


Background
Triple-negative breast cancer (TNBC), accounting for 15-20% of all breast cancers [1], represents the most aggressive subtype and indicates a bad prognosis [2].Due to the aggressive nature and lack of targeted therapies, numerous attempts have been made to develop viable targets for TNBC.Recent advances in immunotherapy have greatly improved cancer treatment by inhibiting immune checkpoints, such as programmed death 1 (PD-1) and its ligand programmed death ligand 1 (PD-L1) [3].The IMpassion130 trial showed the clinical success of the anti-PD-L1 drug atezolizumab in treating PD-L1positive metastatic TNBC; the drug was approved as the rst immune checkpoint inhibitor in breast cancer treatment [4].However, the response rate of PD-1/PD-L1 antagonists has been unsatisfactory, mainly due to the immunosuppressive microenvironment of TNBC [5][6][7].Therefore, it is important to reverse the immunosuppressive phenotype and enhance the antitumor e cacy of immune checkpoint blockade in TNBC patients.
Extensive studies have illustrated that interplay of cancer cells and the tumor microenvironment (TME) plays a signi cant role in ineffective treatment and poor prognosis of cancer.It's reported that cancer cells escape from immunosurveillance by producing immunosuppressive factors and recruiting suppressive immune cells-an important process known as the invasion-metastasis cascade [8].Among suppressive immune cells, tumor-associated macrophages (TAMs) play a pivotal role in cancer metastasis and are predictive of poor prognosis in TNBC [9,10].Previous studies have also shown that immunosuppressive factors from cancer cells, such as C-C motif chemokine ligand 2, colony-stimulating factor 1 (CSF-1) and interleukin 4 (IL-4), promote TAM polarization and activation, thus forming an immunosuppressive microenvironment for cancer metastasis [11].However, the mechanisms of TNBCmediated TAM activation are not well understood.
The Hippo pathway, an evolutionarily conserved signaling, controls multiple signi cant cellular function and its dysregulation contributes to tumorigenesis.The Hippo pathway can promote tumor growth, stemness, and chemoresistance in human cancers including breast cancer [12].Our group also found that yes-associated protein (YAP) modulated the crosstalk between breast cancer cells and their microenvironment [13].However, the function and mechanism of the Hippo pathway in regulating the TNBC microenvironment remains unclear.
This study therefore aimed to explore the role of the Hippo pathway in the TNBC microenvironment and its in uence on the antitumor effects of anti-PD-L1.Here, we identi ed the Hippo pathway remodeled the tumor immune microenvironment through mediating the proliferation and migration of TAMs and inhibiting T cell in ltration.Mechanistically, TAZ, the key regulators of Hippo signaling, facilitated TAM recruitment by directly modulating interleukin 34 (IL-34) transcription and decreased T cell in ltration by regulating PD-L1 expression in TNBC.Furthermore, we also identi ed an important positive feedback loop between TAZ and TAMs in the TNBC microenvironment.Thus, targeting TAMs through the IL-34/colonystimulating factor 1 receptor (CSF-1R) inhibitor could enhance the antitumor e cacy of immune checkpoint blockade, making it a promising immunotherapy strategy for TNBC.

Patients and follow-up
Human breast cancer samples were collected from consecutive patients with TNBC who underwent surgery from 2009 to 2012 at the Comprehensive Breast Health Center, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine (Shanghai, China).All patients were histologically diagnosed by expert pathologists.A total of 120 patients diagnosed with TNBC were included in the study (Ruijin cohort), with a median follow-up of 84.6 months (range: 44.3-121.3months).Overall survival (OS) was de ned as the interval between the surgery date and the date of death or the last follow-up.Disease-free survival (DFS) was de ned as the interval between the surgery date and the date of a reported event such as death, locoregional recurrence, contralateral breast cancer, distant metastasis, or second malignancy, or the date of the last follow-up.Patients' clinical characteristics are shown in Tables 1 and 2.  The present study was performed in accordance with the Declaration of Helsinki.Approval for the use of human subjects was obtained from the research ethics committee of Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, and informed consent was obtained from each patient.

Immunohistochemistry
Immunohistochemistry (IHC) staining was applied using the avidin-biotin-peroxidase complex method, as previously described [14].Brie y, after depara nization, rehydration, and antigen retrieval, the slides were incubated with the primary antibodies overnight at 4°C.The slides were then incubated with horseradish peroxidase-conjugated secondary antibodies (Abcam, UK) for 1 h at 37°C.After rinsing with phosphate buffer saline (PBS), the slides were stained with 3,3′-diaminobenzidine and counterstained with hematoxylin.The primary antibodies used are listed in Table 3.

Cell lines and animals
Mouse breast cancer cell lines 4T1 (ATCC, USA) and E0771 (CH3 BioSystems, USA) and human TNBC lines SUM1315 (Asterand, UK) and BT549 (ATCC) were used in this research.Human cell line HEK 293T and mouse macrophage cell line RAW264.7 (ATCC) were also used, as previously described [15].These cell lines were routinely maintained in our laboratory.
Female BALB/c and C57BL/6 mice (5-6 weeks old; Shanghai SLAC Laboratory Animal Company, China) were used and housed under pathogen-free conditions.All animal experiments were approved by the Animal Ethics Committee of Ruijin Hospital and performed according to the "Guide for the Care and Use of Laboratory Animals" by the National Academy of Sciences (Washington, DC, USA).

Protein extraction and WB
The cells and tissues were lysed with radioimmunoprecipitation assay buffer (RIPA) (Beyotime, China), and the protein level in the lysates was quanti ed using an enhanced bicinchoninic acid assay (BCA) kit (Thermo Scienti c, USA).WB was performed as described previously [17].The primary antibodies used are listed in Table 3.

RT-qPCR and RT 2 pro ler PCR array
Xenograft tumors and cultured cells were lysed with TRIzolTM reagent (Invitrogen, USA), and RNA was isolated according to the manufacturer's instructions.RT-qPCR was performed using SYBR Green Realtime PCR Master Mix (Takara, Japan) in the ABI PRISM 7900 Sequence Detection System (Applied Biosystems, USA).Primer sequences for RT-qPCR analysis are shown in Table 5.
Table 5 Primers for RT-qPCR used in this study.Immune genes in xenograft tumors or TNBC cells with stable TAZ silencing and their control transfectants were screened using the RT 2 Pro ler PCR Array System (RayBiotech, USA) and SYBR Green Realtime PCR Master Mix (Takara) in the ABI PRISM 7900 Sequence Detection System (Applied Biosystems), according to the manufacturer's instructions.

Macrophage proliferation assay
Macrophage proliferation was analyzed using Cell Counting Kit-8 (Dojindo, Japan).Cultured cells (RAW264.7) were dispensed into a 96-well plate at a density of 1000 cells/well, with or without a different level of IL-34 protein (Sino Biological, China) or conditioned medium (CM) from TNBC cells.To assess the IL-34-induced p38 and mTOR signaling pathways, the cells were treated with the CSF-1R inhibitor pexidartinib (AbMole, USA), the p38 inhibitor SB203580 (MedChemExpress, USA), or the mTOR inhibitor AZD8055 (MedChemExpress).At the indicated time points, the absorbance at 450 nm was measured to determine the number of viable cells in each well.

Macrophage chemotactic migration assay
Macrophage chemotactic migration assay was performed using a 24-well Transwell system, with upper and lower culture chambers separated by 8µm pore polycarbonate membranes (Corning, USA), as previously described [18].The bottom chamber was lled with IL-34 protein in Dulbecco's modi ed eagle's medium (DMEM) (Hyclone, USA) containing 10% fetal bovine serum (Gibco, USA) or CM from TNBC cells.
In the upper chambers, macrophages (RAW264.7, 5 × 10 4 cells/well) suspended in DMEM were seeded and then incubated for 24 h.Macrophages that migrated to the lower surface of the membrane were xed in 4% paraformaldehyde, stained with Giemsa, and then photographed and counted with a light microscope (Leica).To assess the IL-34-induced p38 and mTOR signaling pathways, the cells were treated with selective inhibitors.The chemotactic index was calculated as the number of macrophages that migrated to IL-34 of the CM from TNBC cells divided by the number of macrophages that migrated to DMEM alone [19].

Macrophage isolation
For TAM isolation, xenograft tumors were collected and digested into single-cell suspensions as previously described [20], and macrophages were isolated using Anti-F4/80 MicroBeads (Miltenyi Biotec, Germany) according to the manufacturer's instructions.

Co-culture assay
A 6-well Transwell system with 0.4µm pore polycarbonate membranes (Corning) was used for the coculture assay.TAMs (1 × 10 5 cells/well) isolated from xenograft tumors were seeded in the upper chambers and co-cultured with 4T1 or E0771 cells (1 × 10 5 cells/well) in the lower chambers for 48 h.The 4T1 or E0771 cells were re-plated and cultured, and after 12 h, supernatants were collected for further analysis.

Flow cytometry analysis
Xenograft tumors were dissected into small pieces and further digested into single cell suspensions.Cells were stained with antibodies and analyzed by ow cytometry (Epics Altra, USA) according to the manufacturer's instructions.The antibodies used are listed in Table 3.

Enzyme-linked immunosorbent assay and cytokine array analysis
IL-34 levels in cell culture supernatants were assessed using the mouse or human IL-34 enzyme-linked immunosorbent assay (ELISA) kit (RayBiotech) according to the manufacturer's instructions.Brie y, a sample solution was added to each well and incubated with an IL-34 conjugate.A substrate solution was then added, and the absorbance was analyzed using a microplate spectrophotometer (Bio-Rad, USA).A curve of the absorbance according to the IL-34 level in the standard wells was formulated.
The sample solution was collected from the cell culture supernatants of TAMs and then incubated with a biotinylated detection antibody cocktail and Cy3 equivalent dye-conjugated streptavidin from a mouse cytokine array (RayBiotech).The signals were visualized with a microarray laser scanning system (GenePix, USA).

Phospho-kinase array analysis
The IL-34-induced downstream signaling pathways were analyzed using the mouse phosphorylation pathway array (RayBiotech).The protein lysate was incubated with the array membrane, and signaling proteins were visualized using a chemical uorescence detection system (Bio-Rad) according to the manufacturer's instructions.

In vivo tumor growth and metastasis assay
For the xenograft tumor growth model, 4T1 or E0771 cells were orthotopically implanted into the mammary fat pad of BALB/c or C57BL/6 mice (1 × 10 5 cells/mouse).Tumor growth was monitored every 3 days, and tumor volume was calculated using the following formula: volume = (width) 2 × length ÷ 2.
Tumor growth was visualized with a bioluminescence-based IVIS imaging system (Caliper Life Sciences, USA).Mice were sacri ced after 4 weeks, and tumor weight was then assessed and xed in 4% paraformaldehyde solution for hematoxylin-eosin (H&E) and IHC staining.
For the metastatic model, 4T1 or E0771 cells were injected into the tail veins of BALB/c or C57BL/6 mice (5 × 10 4 cells/mouse).Four weeks later, the mice were sacri ced and the lungs were xed in 4% paraformaldehyde solution for H&E staining.The number and the largest size of lung metastatic nodules were evaluated under a microscope (Leica).

Immuno uorescence
Brie y, slides were xed in 4% paraformaldehyde for 15 min and then blocked with PBS containing 1% bovine serum albumin and 0.15% glycine for 1 h at room temperature.The slides were incubated with primary antibodies overnight at 4°C, and then incubated with Alexa Fluor secondary antibodies (Thermo Scienti c) for 2 h.After rinsing with PBS, the slides were counterstained with 4′,6-diamidino-2phenylindole and evaluated by uorescence microscopy (Leica).

Luciferase reporter assay
Cells with a stable knockdown of TAZ or S89A mutant TAZ and their control transfectants were plated into 24-well plates and transfected with pGL3-luciferase reporter plasmid and SV40 Renilla luciferase plasmid (Promega, USA).Cells were lysed and the luciferase activity was detected with a dual luciferase assay (Promega) according to the manufacturer's instructions.

Chromatin immunoprecipitation
Cells were crosslinked with 1% formaldehyde and incubated for 10 min at room temperature.Glycine was added into the medium for 10 min at room temperature to stop the crosslinking.After that, the cells were washed, lysed, resuspended, and sonicated to an average length of 500 base pairs.After centrifugation, the samples were incubated with primary antibodies (anti-TAZ, #4883, Cell Signaling Technology, UK; IgGcoated Dynabeads Protein A, #10001D, Invitrogen) overnight at 4°C, and then incubated with protein A/G beads at 4°C for 2 h.Magnetic beads were washed with lysis buffer twice, high salt buffer twice, LiCl buffer once, and TE buffer once successively on ice, and eluted with elution buffer.The eluents were reverse-crosslinked by adding 2 µL of 10 mg/mL RNase A at 37 for 1 h and 2 µL of 20 mg/mL protease K at 50°C for 1 h.De-crosslinked DNA was puri ed with a DNA puri cation kit (Qiagen, Germany) and eluted in nucleic acid-free water.RT-qPCR analysis was used to measure the enrichment of speci c promoter regions.The RT-qPCR primers used for chromatin immunoprecipitation (ChIP) are shown in Table 6.Bioinformatics and statistical analysis RNA-seq data for TNBC patients and corresponding clinical information were downloaded from The Cancer Genome Atlas (TCGA) dataset.Meanwhile, the immune scores and the abundance of immune cells were estimated by Immune Cell Abundance Identi er (ImmuCellAI) [21].According to the immune in ltration score, TNBC samples were divided into high-and low-in ltration group to analyze the activity of conserved YAP signature by GSVA algorithm [22].The cBioPortal for Cancer Genomics, a web-based tool for analyzing RNA-Seq data from The Cancer Genome Atlas (TCGA), was used to evaluate the correlation between the mRNA expression of TAZ and cytokines/chemokines in the TCGA cohort [23].The Kaplan-Meier Plotter, a web-based application, was used to assess the impact of TAZ expression on the prognosis of TNBC patients [24].
Statistical analysis was performed using GraphPad Prism 7 (GraphPad, USA) and SPSS 22.0 (SPSS Inc., USA).Quantitative data were presented as means and standard errors of the mean, and qualitative values were shown as numbers.Unpaired Student's t test, one-way analysis of variance (ANOVA), or two-way ANOVA was performed for quantitative data as appropriate.The chi-squared test or Fisher's exact test was applied to analyze the correlation between qualitative variables, and Pearson's correlation test to analyze the correlation between quantitative variables.Kaplan-Meier curves were used to assess the impact of prognostic factors on patient survival.Two-sided P < 0.05 was considered statistically signi cant and was marked as * P < 0.05, ** P < 0.01, *** P < 0.001.

Results
High Hippo pathway activity was positively correlated with the immunosuppressive microenvironment and poor prognosis in TNBC patients The nding that nuclear translocation of YAP mediating resistance to anti-PD-1 immunotherapy [25] lead us to assess the relationship between YAP/TAZ signaling and immune phenotype in TNBC.We rstly discovered higher level of conserved YAP signature was signi cantly correlated with higher immune in ltration score (Fig. 1A) and worse patient outcome (Fig. 1B).Noteworthily, the in ltration of CD8 + T-cell was more in the group with lower level of conserved YAP signature, while the in ltration of macrophage was quite contrary ((Fig.1C), thus forming an immunosuppressive microenvironment.Then we detected the levels of TAZ, CD8 + T cells, and CD68 + TAMs in TNBC patient samples by IHC staining (Fig. 1D).
Kaplan-Meier curves showed that both 5-year OS and DFS in TNBC patients with high TAZ expression were signi cantly lower than those of patients with low TAZ expression (OS: 85.1% vs 98.1%; DFS: 80.6% vs 96.2%) (Fig. 1G).Similarly, the Kaplan-Meier Plotter analysis also revealed that TNBC patients with high TAZ expression showed an unfavorable prognosis compared to those with low TAZ expression (Figure S1A-B).The 5-year OS and DFS in TNBC patients with an increased number of intratumoral CD8 + T cells were 98.4% and 95.3%, respectively, which were signi cantly higher than those with fewer CD8 + T cells (82.1% and 78.6%) (Fig. 1H).Meanwhile, TNBC patients with increased TAM in ltration displayed signi cantly poorer 5-year OS and DFS than those with decreased TAM in ltration (OS: 85.2% vs 96.6%; DFS: 80.3% vs 94.9%) (Fig. 1I).Furthermore, to assess the combined effects of TAZ, T cells and TAMs on prognosis, we strati ed the TNBC patients into three groups (Group I: patients with low TAZ expression, an increased number of T cells, and fewer TAMs; Group III: patients with high TAZ expression, fewer T cells, and an increased number of TAMs; and Group II: the remaining patients).The 5-year OS in groups I, II, and III were 100.0%,94.7%, and 65.0%, respectively (Fig. 1J).Meanwhile, the 5-year DFS in group I was 100.0%, which was signi cantly higher than that of group II (89.3%) and group III (65.0%) (Fig. 1J).
Therefore, the high expression of TAZ was positively correlated with fewer T cells while the increase in TAM in ltration was associated with poor prognosis in TNBC patients.

Hippo pathway promoted TNBC growth and metastasis by driving the immunosuppressive tumor microenvironment
To explore the role of Hippo pathway in the immunosuppressive microenvironment of TNBC, we stably silenced TAZ expression in mouse breast cancer cells and established a tumor orthotopic transplantation model with these cells.We found that the tumor sizes and weights in 4T1-shTAZ and E0771-shTAZ xenograft tumors were signi cantly lower than those of the control groups (Fig. 2A-B and S2A-B).The number and the largest diameter of metastatic nodules signi cantly decreased in the shTAZ mice than in the control group (Fig. 2C-D and S2C).Additionally, IHC staining and ow cytometry analysis showed an increased number of CD8 + T cells and fewer F4/80 + CD206 + TAMs in the 4T1-shTAZ and E0771-shTAZ xenograft tumors (Fig. 2E-H and S3), whereas TAZ silencing did not reduce the number of other tumorin ltrating immune cells (Figure S2D-E).Collectively, these results con rmed that Hippo pathway indeed decreased intratumoral T cell in ltration but promoted TAM recruitment, thus forming an immunosuppressive microenvironment, and promoting TNBC growth and metastasis.

TAZ directly regulated IL-34 and PD-L1 expression
To gured out how Hippo pathway modulated immune cell in ltration in TNBC, GSEA analysis of TNBC patient datasets revealed that high expression of TAZ enriched expression of cytokines and chemokines production (Fig. 3A).To identify cytokines or chemokines that may be regulated by TAZ, the RT 2 pro ler PCR array was employed.The effect of TAZ on the expression of immune genes (including 57 cytokines/chemokines and 25 immune checkpoints) was presented as downregulation of cytokines/chemokines in 4T1-shTAZ-and E0771-shTAZ-derived xenograft tumors and SUM1315-shTAZ and BT549-shTAZ cells (Table 7-8 and Figure S4A).As shown in the Venn diagram (Fig. 3B), only IL-34 was shared by all of the above groups.Notably, TAZ expression was positively associated with the expression of IL34 in TCGA TNBC patients (Fig. 3C).RT-qPCR, WB, and IHC showed that TAZ silencing signi cantly decreased IL-34 expression in TNBC cells and xenograft tumors (Fig. 3D&E and S4B).ELISA testing also indicated that IL-34 secretion was reduced in TAZ-silenced TNBC cells (Fig. 3D).Furthermore, luciferase reporter and ChIP assays on 4T1 and E0771 cell lines were performed to determine if TAZ could regulate IL-34 transcription; indeed, TAZ activated IL-34 transcription by directly binding to speci c regions on the IL-34 promoter (Fig. 3F and S4C-E).In the RT-qPCR-based assay, in addition to IL-34, we found that PD-L1 was also a gene targeted by TAZ.
TAZ silencing signi cantly decreased PD-L1 expression in 4T1 and E0771 xenograft tumors (Fig. 3G and S5A), which was con rmed by the results of IHC (Fig. 3H and S5B).Furthermore, TAZ showed a high a nity to speci c regions on the PD-L1 promoter in the luciferase reporter and ChIP assays (Fig. 3I and S5C-E).Clinically, TAZ expression was positively associated with IL-34 and PD-L1 expression, while IL-34 was positively correlated with CD68 + TAMs in TNBC samples (Fig. 3J and S4F).Additionally, high expression of IL-34 or PD-L1 was associated with poor prognosis in TNBC patients (Fig. 3K and S4G, S5F&G).Thus, Hippo/YAP signaling could directly mediate PD-L1 and IL-34 expression in TNBC.
The TAZ/IL-34 axis promoted proliferation and migration of macrophages by activating the p38 and mTOR signaling pathways Given that IL-34 is a newly discovered cytokine, which speci cally binds to its receptor, CSF-1R, and regulates the development, differentiation, and function of monocytes and macrophages[26], we next examined the role of IL-34 in macrophage proliferation and chemotactic migration in vitro.Indeed, we found that IL-34 could lead to concentration-dependent proliferation and chemotactic migration of macrophages (Fig. 4A&B).The phospho-kinase array and WB indicated that IL-34 increased the phosphorylation levels of p38 and mTOR, which could be inhibited by the CSF-1R inhibitor pexidartinib, a p38 inhibitor, or an mTOR inhibitor (Fig. 4C&D).The enhanced macrophage proliferation and migration could also be abolished by these inhibitors, indicating that IL-34 could induce the activation of the p38 and mTOR signaling pathways in macrophages (Fig. 4E&F).In addition, the CM from 4T1 or E0771 cells could enhance the proliferative and migratory abilities of macrophages, which could be suppressed by TAZ or IL-34 silencing, whereas IL-34 overexpression rescued the TAZ silencing-induced inhibition of the proliferative and migratory effects of 4T1 or E0771 cells on macrophages (Fig. 4G-I).Therefore, YAP/TAZ signaling could promote proliferation and migration of macrophages by upregulating the IL-34-mediated activation of the p38 and mTOR signaling pathways.

The TAZ/IL-34 axis contributed to tumor progression by inducing TAM in ltration
Considering the role of YAP/TAZ signaling in TAM activation, we performed in vivo assays to study the effect of the TAZ/IL-34 axis on regulating TAM activation and TNBC progression.In 4T1 and E0771 cells, TAZ or IL-34 silencing mitigated the tumor growth and metastasis, while IL-34 overexpression rescued the TAZ silencing-induced inhibition of TNBC progression (Fig. 5A-D and S6A-D).
We performed IHC staining and ow cytometry to evaluate TAM in ltration in xenograft tumors, and the results indicated that TAZ or IL-34 silencing decreased the number of F4/80 + CD206 + TAMs, while IL-34 overexpression rescued the TAZ silencing-induced reduction of TAM in ltration (Fig. 5E-H).We also found that TAZ silencing downregulated the expression of macrophage M2 markers arginase 1 (Arg1), interleukin 10 (IL-10), and mannose receptor (CD206), which could be rescued by IL-34 overexpression (Figure S6E).Hence, the TAZ/IL-34 axis contributed to tumor progression by inducing TAM in ltration.
TAZ functioned as a feedback mechanism inducing TAM in ltration to promote TNBC progression To further study the crosstalk between cancer cells and TAMs, we generated a co-culture model (Fig. 6A).
As expected, we found that the CM from 4T1 or E0771 cells could promote the proliferative and migratory abilities of macrophages, which were abolished by TAZ or IL-34 silencing (Fig. 6B-C).Interestingly, after co-culture with TAMs, 4T1 and E0771 cells showed an upregulation in TAZ expression and IL-34 and PD-L1 expression (Fig. 6D).To clarify the mechanism of increased TAZ expression in 4T1 and E0771 cells by TAMs, a cytokine array was conducted and showed that the most predominantly secreted cytokines by TAMs were transforming growth factor beta 1 (TGF-β1), C-X-C motif chemokine ligand 12, C-C motif chemokine ligand 1 (CCL1), CCL5, and CCL3 (Fig. 6E).Furthermore, anti-TGF-β1 neutralizing antibody signi cantly abrogated the TAM-induced upregulation of TAZ expression, whereas other neutralizing antibodies had no impact on TAZ expression (Fig. 6F).Therefore, TAM-derived TGF-β1 could promote the expression of TAZ, thereby forming a positive feedback loop to induce IL-34-mediated TAM in ltration in TNBC.

CSF-1R blockade sensitized TNBC to anti-PD-L1-mediated immunotherapy by reversing the immunosuppressive tumor microenvironment
Regarding the ability of TAMs to suppress T cell antitumor immune response [27], we investigated whether targeting TAMs through its CSF-1R inhibitor and inhibiting T cell antitumor response by anti-PD-L1 blockade could synergistically inhibit tumor progression in TNBC.The mice with orthotopic 4T1 and E0771 cells were treated with an isotype control antibody, the CSF-1R inhibitor pexidartinib, and anti-PD-L1, or a combination of pexidartinib and anti-PD-L1 (Figure S7A&B).The results showed that the combination of pexidartinib and anti-PD-L1 signi cantly mitigated tumor growth (Fig. 7A-C and S8A-E) and inhibited lung metastasis (Fig. 7D-E and S8F-I), without any signi cant effect on mice body weight or toxicity in the liver and kidneys (Figure S7C-E).Thus, CSF-1R signaling blockade and anti-PD-L1 blockade could result in a synergistic antitumor response in TNBC.
To further clarify the mechanism underlying the synergistic antitumor effect of the pexidartinib and anti-PD-L1 combination, the number of tumor-in ltrating immune cells in xenograft tumors was analyzed by IHC staining and ow cytometry.As expected, we observed a signi cant reduction in TAMs in the pexidartinib plus anti-PD-L1 treatment group and an increased number of CD8 + T cells in the combined treatment group (Fig. 7F, 7H, S9A and S9C).Immuno uorescence staining further validated the increase in the number of intratumoral CD8 + T cells in the combined treatment group (Fig. 7G and S9B).
Furthermore, Ki-67 and CD31 expression were signi cantly mitigated in the pexidartinib and anti-PD-L1 group, while caspase-3 expression was upregulated in the combined treatment group (Figure S10A-B).
Hence, the combination of pexidartinib and anti-PD-L1 treatment inhibited the proliferation and angiogenesis of cancer cells and promoted their apoptosis, thus enhancing the antitumor effect in TNBC.
Various immune cells are recruited into the tumor bed depending on the in ammatory factors secreted by the cancer cells, and these immune cells are reported to exhibit pro-tumor phenotypes [32].Among these tumor-in ltrating immune cells, TAMs are particularly abundant in TNBC [33].Several studies have substantially proven that intratumoral TAMs facilitate cancer progression and metastasis by promoting cancer cell proliferation, migration, and angiogenesis [34,35].To further investigate the mechanism by which TAMs affect TNBC progression, we isolated macrophages from xenograft tumors; we found that TGF-β1 was the most predominant cytokine secreted by TAMs.TGF-β1 activates the polarization of several immune cells and mediates the immunosuppressive microenvironment [36].TGF-β1 also upregulates TAZ expression via an SMAD3-dependent or SMAD3-independent mechanism [37,38], but this phenomenon is not clearly understood in TNBC.In this study, we con rmed that TAM-derived TGF-β1 positively regulated TAZ expression in TNBC cells.Further in-depth studies are recommended to investigate whether SMAD3 affects the role of TGF-β1 in promoting TAZ expression.Collectively, our results demonstrate the presence of a TAZ/IL-34/TAM feedback loop in TNBC growth and metastasis.
The gene expression pro les in this study also revealed that the immune checkpoint PD-L1 was a gene targeted by TAZ.Consistent with the ndings of a previous report [39], we found that TAZ promoted PD-L1 expression in both TNBC cells and tissues.Upregulation of PD-L1 and its ligation to PD-1 on T cells can trigger inhibitory signals, mitigate T cell antitumor response, and bypass immunosurveillance [40]; these mechanisms may explain the unfavorable survival outcome of patients in our study.Although anti-PD-L1 blockade has been clinically applied in the treatment of TNBC, its response rate and long-term bene ts are limited due to the presence of pre-existing immunosuppressive factors, especially TAMs [41,42].So there comes the need of developing new strategies to reverse the pre-existing immunosuppressive microenvironment and enhance the e cacy of immune checkpoint inhibitors in TNBC immunotherapy.In this study, the CSF-1R inhibitor pexidartinib signi cantly inhibited tumor growth and metastasis in combination with anti-PD-L1 blockade in mouse models.Moreover, the combined therapy signi cantly decreased TAM in ltration and activated CD8 + T cell response, thus reversing the immunosuppressive microenvironment in TNBC.CSF-1R inhibitors are considered novel targets in the treatment of solid tumors other than TNBC [43][44][45].To our knowledge, this is the rst study to explore the novel therapeutic bene ts of using the CSF-1R inhibitor pexidartinib in combination with anti-PD-L1 blockade in the treatment of TNBC.
In conclusion, this study revealed that Hippo pathway was associated with worse disease outcomes in TNBC and could increase the in ltration of tumor-associated macrophages through the TAZ/IL-34 axis, speci cally through its CSF-1R, leading to an immunosuppressive microenvironment and impairing the treatment e cacy of anti-PD-L1.Therefore, targeting the TAZ/IL-34 axis can be a novel immunotherapy strategy in the treatment of TNBC.

Conclusions
Altogether, we found that the Hippo pathway effector TAZ contribute to the bidirectional interactions between cancer cells and the TME in TNBC.We deciphered TAZ mediated the proliferation and migration of TAMs by regulating the IL-34/CSF-1R axis, and inhibited T cell in ltration by upregulating the expression of the immune checkpoint PD-L1, thus forming an immunosuppressive microenvironment in TNBC.Our ndings unveil the CSF-1R inhibitor pexidartinib in combination with anti-PD-L1 blockade can be used as a novel potential therapeutic approach for TNBC.

Figures
Figure 1 The

Supplementary Files
This is a list of supplementary les associated with this preprint.Click to download.
relationship between Hippo pathway and immune cell in ltration and its clinical signi cance in TNBC patients.(A) The landscape of immune cell in ltration of TNBC patients in the TCGA training cohort.High or low in ltration based on immune in ltration score estimated by ImmuCellAI.The activity of conserved YAP signature was estimated by GSVA algorithm.(B) Correlation of combined activity of conserved YAP signature with overall survival in TCGA TNBC patients.(C) Macrophage and CD8+ T cells in ltration between high and low activity of conserved YAP signature.(D) Representative images of IHC staining of TAZ, CD8, and CD68 in TNBC samples (up: high expression of TAZ and CD68 and low expression of CD8; bottom: low expression of TAZ and CD68 and high expression of CD8).Scale bar, 50μm.(E) Scatterplot of intratumoral CD8+ T cell counts according to TAZ IHC scores in TNBC samples.(F) Scatterplot of intratumoral CD68+ TAM counts according to TAZ IHC scores in TNBC samples.(G) Prognostic values of TAZ expression for OS and DFS.(H) Prognostic values of intratumoral CD8+ T cell counts for OS and DFS.(I) Prognostic values of intratumoral CD68+ TAM counts for OS and DFS.(J) Prognostic values of the combination of TAZ expression, CD8+ T cells, and CD68+ TAMs for OS and DFS (Group I: patients with low TAZ expression, an increased number of T cells, and fewer TAMs; Group III: patients with high TAZ expression, fewer T cells, and an increased number of TAMs; Group II: the remaining patients).Student's t-test in(C), Pearson's correlation test in (E) and (F), and Kaplan-Meier analysis in (B, G-J).DFS: disease-free survival; IHC: immunohistochemistry; OS: overall survival; TAM: tumor-associated macrophage; TAZ: transcriptional coactivator with PDZ-binding motif; TNBC: triple-negative breast cancer.

Figure 3 Figure 4 The
Figure 3

Figure 5 Effect
Figure 5

Table 1
Baseline clinical characteristics of TNBC patients.

Table 2
Clinicopathological variables correlated with TAZ expression in TNBC patients.

Table 3
Primary antibodies for western blot, immunohistochemistry, immuno uorescence, and ow cytometry.

Table 6
Primer sequences for IL-34 and PD-L1 promoter used for ChIP.

Table 7
Changes in the expression levels of immune genes in xenograft tumors.

Table 8
Changes in the expression levels of cytokines/chemokines in TNBC cells.