The participants in this study were 23 female breast cancer patients (Stage I, n = 6; Stage II, n = 8; Stage III and IV, n = 9; ages: 18–65 years) and 10 age-/sex-matched healthy people ( Table S1 ). The tumor stage was assessed using the TNM classification of the International Union Against Cancer from 2002. Three patients with Stage II breast cancer who had mastectomy were given ipsilateral normal breast tissues. The ethics committees of the ESI Post-Graduate Institute of Medical Science and Research, Kolkata (Approval No: ESI-PGIMSR/MKT/IEC/13/2017 dated December 22, 2017), the Institute of Post-Graduate Medical Education and Research Oversight Committee (Approval No. IPGME/IEC/13/2017 dated December 22, 2017), and Human Ethics Committee, Bose Institute (Approval No: BIHEC/2017-18/7, dated May 28, 2017) approved the collection of post-operative breast tumor tissue samples and peripheral blood from breast cancer patients/normal. All patients enrolled in the trial gave their informed consent.

To obtain PBMC, peripheral blood was centrifuged at 800×g for 45 min over the lymphocyte separation medium (Histopaque). For T-cell polarization, magnetic-bead sorted naïve CD4⁺CD45RA⁺ T cells from cord blood were activated with anti-CD3 and anti-CD28 beads (Invitrogen) before being allowed to develop into Tregs in the presence of recombinant TGFβ (5 ng/ml) and IL2 (50 U/ml) and anti-IFNγ antibody (PeproTech). Flow cytometry was used to determine the purity of the enhanced cells, which was consistently >90%. During flow cytometric analysis, LymphoGate was employed to detect lymphocytes. Treg differentiation was confirmed by flow cytometry and qPCR. Cells were cultured in a complete RPMI-1640 medium supplemented with 10% FBS at 37°C in a humidified 5% CO₂ incubator. All experiments were performed with mycoplasma-free cells.

THP1 (CVCL 0006), a well-established monocyte cell line, was grown in RPMI-1640 with 10% serum. Monocytes were treated with PMA (5 ng/ml; Sigma) to generate the M-M0 subtype. THP1 can be polarized from Mϕ-M0 to Mϕ-M1 by adding LPS (100 ng/ml; Peprotech) and IFNγ (20 ng/ml; PeproTech) to it. Mϕ-M0 was also polarized to the Mϕ-M2 subtype by treating it with IL4 (20 ng/ml; PeproTech) for 48 h. The cell line has been authenticated using STR (or SNP) profiling within the last 3 years by the National Centre for Cell Science, Pune, India. All experiments were performed on mycoplasma-free cells.

To generate ex vivo TME, tumor tissues were collected from breast cancer patients undergoing surgical procedures, minced, and stirred in cell dissociation reagent containing collagenase Type IV in DMEM/F-12 for 2 h. The tumor cell suspension was filtered through nylon mesh and centrifuged for 10 min at 250×g. To make a single-cell suspension, the cell pellet was washed with serum-free RPMI 1640 and resuspended in complete RPMI 1640 media. CD24 and ESA positivity or CD4-/CD8- and CD25 negativity were used to assess the purity of those cells flow cytometrically. T cells from healthy volunteers were over-layered on top of primary breast tumor cells (to mimic the TME in vitro) and cultured for 72 h to generate Tregs ex vivo ( Figure S1 ) (38).

Fluorophore-conjugated CD4 (FITC/-PE), CD25 (PE/PE-CY7), CD127 (AF-647), CTLA4 (PE), FOXP3 (AF-647), and CCR4 (BV-421/PE) antibodies (Biolegend/BD Bioscience) were used to stain cells for phenotyping. The expression of Th1- and Th2-specific transcription factors T-bet (Santa Cruz) and GATA3 (Santa Cruz) was used to determine the percentage of Th1 and Th2 cells, respectively. Cells were stimulated for 5 h at 37°C with cell activation cocktail and brefeldin-A (Biolegend) before being stained with anti-IL10 (BV-421) using the Cytofix/CytoPerm Plus kit (BD Biosciences). The fluorescent-conjugated annexin-V was used to assess cellular apoptosis (BD Bioscience). For the purification of naïve and Treg cells, CD4⁺ cells, labeled for CD45RA, CD25, and CTLA4, were subjected to high-speed cell sorting (FACS-ARIA, BD Biosciences) to obtain CD25⁺CTLA4⁺ Treg and CD45RA⁺ naïve T cells. Data were acquired and analyzed by FACS using FACS-Suite and FlowJo software (BD Biosciences). Quadrants were constructed based on signals utilizing FMO and unstained controls to quantify stained cells in contour plots. For t-SNE analysis, the lymphocyte population was gated and the CD4⁺ population was subsequently marked. The lymphocyte population from blood acquired from a breast cancer patient in the cohort was a down-sample with 5,000 cells, and the dimensionality reduction was done with the population to create t-SNE parameters.

Ex vivo-generated Tregs (2 × 10⁶ cells) were transduced with FOXP3-/CCR4-shRNA (Thermo-Fisher Scientific), FOXP3-/CCR4-cDNA (ADDGENE), or control vector by electroporation (Invitrogen, USA) in a single-pulsed method (voltage 260 V and capacitance 1050 μF). Transduced cells were cultured for 48 h before being exposed to downstream tests. We routinely achieved a transduction effectiveness of 85%–90% under these conditions without compromising cell viability.

Treg cells (2 × 10⁶ cells) were stimulated with PMA and ionomycin for 4 h at 37°C before RNA was isolated using TRIzol reagent (Thermo Fisher Scientific) and reverse transcribed using Verso-cDNA synthesis kit (Thermo-Fisher Scientific). FastStart Essential DNA Green Master (Roche) was used for qPCR analysis, which was carried out using a GeneAmp PCR-2720 (Applied Biosystems, CA, USA). Melting curves using LC96 SW1.1 software were used to assess amplification products using SYBR-green detection (Roche). A 2^–ΔΔCt approach was used to analyze the data. The expression level of the house-keeping gene, GAPDH, was used to normalize for variations in the amount of RNA. The respective primers list is given in Table S2 .

To cross-link the protein–DNA complexes, ex vivo-generated Tregs (2 × 10⁶ cells) were cross-linked in 1% p-formaldehyde for 15 min at 37°C. Soluble chromatin (7–20 μg) was treated with 4–5 μg of antibody overnight at 4°C after being sonicated and pre-cleared with protein-A agarose/salmon-sperm DNA. FOXP3, RNA POL-II (Santa Cruz), HAT1, acetylated-H3 (K14, K18, K23, and K27), anti-rabbit, and anti-mouse immunoglobulin-G antibodies (Cell Signaling Technology) were used for ChIP and Re-ChIP assays. In addition, a negative control IgG (Sigma-Aldrich) was employed. Following immunoprecipitation and washing, DNA from each immunoprecipitation was measured using distinct primer sets ( Table S2 ) for the FOXP3 and RNA POL-II binding sites on the CCR4 promoter by qPCR. After eliminating background immunoprecipitation assessed with nonspecific IgG or the input sample, the fold-change in occupancy of each protein at the CCR4-promoter region of Treg relative to the T naïve cell was estimated. DNA–protein complexes immunoprecipitated with anti-FOXP3 antibody in primary ChIP were eluted with 25 μl of 10 mM dithiothreitol for 30 min at 37°C and diluted 20 times with Re-ChIP buffer for the Re-ChIP tests (1% Triton X-100, 2 mM EDTA, 150 mM NaCl, and 20 mM Tris-HCl; pH 7.5). After that, the complexes were re-immunoprecipitated with anti-HAT1 antibody before qPCR examination of the precipitated chromatin DNA (39).

Breast tumor patient-derived Tregs and naïve T cells were washed in PBS with 1% BSA. Then 1 × 10⁴ cells were allowed to adhere to poly-L-lysine-coated slides. Cells were fixed with 4% p-formaldehyde and permeabilized with 0.2% saponin. After blocking with 3% BSA, cells were incubated with anti-FOXP3/-CCR4 (rabbit/mouse mAb), followed by fluorescence-tagged secondary antibodies (Invitrogen) and DAPI (BD Biosciences). Cells were visualized in Leica confocal microscope, with DPX mounting medium, at 60× magnification. Images were analyzed by ImageJ software (40).

Polycarbonate filters (pore size, 5 µm; Cell Culture Inserts, Corning) were used for Trans-well assay that helps direct the passage of the cells from the upper to the lower chamber. The upper chamber was plated with RPMI-1640 medium containing patient-derived Tregs and the lower chamber was plated with tumor cell supernatant containing recombinant CCL22 (PeproTech). After 4 h at 37°C, cells migrated at the lower chamber were collected, immunostained, and analyzed in a flow cytometer. To check the role of FOXP3 in Tregs migration, ex vivo-generated Tregs were first transfected with lentivirus containing desired clones and the transfected cells were placed in the upper chamber of trans-well plates.

Cells were homogenized in buffer (20 mM Hepes, pH 7.5, 10 mM KCl, 1.5 mM MgCl₂, 1 mM Na-EDTA, 1 mM EGTA, and 1 mM DTT) supplemented with protease and phosphatase cocktail inhibitor. A total of 60 μg of protein was resolved using 10% SDS-PAGE, transferred to nitrocellulose membrane, and probed with specific antibodies CCR4 (Santa Cruz) and FOXP3 (Santa Cruz). Secondary antibodies used (1:10,000) were conjugated with HRP (Sigma). After that, the immunoblots were developed by chemiluminescence method using luminol and coumaric acid. Equal protein loading was confirmed with anti-β-actin antibody (mAb) (Santa Cruz).

Packaging of lentiviruses was performed by transient transfection of HEK cells. One day before transfection, HEK cells were seeded in a T75 flask at 1 × 10⁵ cells/cm² in DMEM supplemented with 10% FCS, 1 mM pyruvate, and 40 mg/ml gentamicin. The calcium phosphate precipitation method was used to co-transfect cells using 7.5 μg of gag/pol packaging plasmid psPAX2, 7.5 μg of pGIPZ-Foxp3-shRNA/Foxp3-cDNA transfer vector, and 4 μg of envelope plasmid pMD2.G using the Profection mammalian transfection kit (Promega, USA). Transfections were carried out in 10 ml of DMEM without antibiotics, and the cells were grown for 16 h. After that, the medium was replaced with complete DMEM, and after 48 h, the supernatant was collected. Centrifugation at 1,500 rpm for 5 min at 4°C was used to remove cell debris, which was then passed through a 0.45-μm pore PES filter. For concentration, 20 ml of supernatant was centrifuged at 4,000 rpm for 20 min at 4°C in a 12% polyethylene glycol (PEG-8000) solution. In serum and antibiotic-free DMEM, pelted viruses were re-suspended and the aliquots were kept at −80°C for further use (41).

BALB/c and C57/BL6 mice weighing 20–25 g (in-house animal facility) were kept in a temperature-controlled environment with a light–dark cycle. All animal experiments were carried out following the National Institutes of Health’s Principles of Laboratory Animal Care (NIH publication No. 85-23, revised in 1985) as well as Indian laws on “Animal Protection” under the provision of the Bose Institute Animal Ethics Committee for the control and supervision of animal experiments (Reg. No. 95/99/CPCSEA; Approval No: IAEC/BI/47/2017). A total of 2 × 10⁶ isogenic mammary carcinoma 4T1 or B16/F10 melanoma cells (ATCC) were injected in the breast fat pad of BALB/c mice or subcutaneously in C57/BL6 mice, respectively. Mice were separated into five groups, each with three mice: (i) control, (ii) tumor alone (no transduction), (iii) control vector, (iv) Foxp3-shRNA, and (v) Foxp3-cDNA. The experiments were repeated twice. Lentivirus-bearing desired clones were injected into the tail vein after 7 days of tumor inoculation. Mice were tracked for the next 21 days. Every other day, the tumor volume was measured. The tumor volume was calculated using the formula V = (W2 × L)/2 (caliper measurements), where V represents the tumor volume, W represents the tumor width, and L represents the tumor length. After 21 days, the mice were sacrificed, and Tregs from PBMC and the tumor site location were examined using flow cytometry (42).

The correlation study between FOXP3 and CCR4 was executed using R2: Genomics Analysis and Visualization Platform. The same dataset was used to construct t-SNE plots for FOXP3 and CCR4. MatInspector software was used to analyze the CCR4 promoter. Multiple sequence alignments of CCR4 promoter between human and mouse was done by Clustal W. Unless otherwise stated, values are shown as the standard error of the mean (SEM). A 2-way ANOVA was used to compare different experimental groups, followed by a post-hoc Bonferroni modification of the multiple comparison t-test. A Student’s t-test was used to examine the significance of the differences between mean values when suitable. At p < 0.05, the results were considered significant. Graph Pad software was used to create a heatmap for the fold-change values. Prism software was used to analyze the data (Graph Pad software).

We obtained peripheral blood and post-operative breast tumor tissue from a cohort of 8 breast cancer patients undergoing surgery to establish the expression pattern of CCR4 in Tregs. We used a mix of phenotypic markers to further characterize human Tregs and their relationship to tumor progression characteristics. FOXP3 was employed as an internal marker, while CD4, CD25, and CD127 were used as surface markers (33). In comparison to age-/sex-matched healthy donors, patients with advanced breast cancer had higher proportions of CD4⁺CD25⁺CD127^-FOXP3⁺ Tregs in the peripheral circulation and tumor tissue, as shown in Figure 1A ( Figure 1A ; left panel). Based on the cell surface marker expression pattern, we chose CD4⁺CD25⁺CD127^- as a suitable set of surface markers to sort Tregs for further characterization. The tumor site had a higher percentage of CD4⁺CD25⁺CD127^-FOXP3⁺ Tregs than the peripheral circulation of breast cancer patients and age-/sex-matched healthy donors ( Figure 1A ; right panel).

CCR4 is a chemokine receptor that is selectively expressed by Tregs in various pathophysiological situations and plays an important role in lymphocyte organ-specific migration (35, 36). The distribution pattern of CCR4 in tumor-associated CD4⁺CD25⁺CD127^- Treg populations in healthy donor and tumor patients was determined using a comparative t-SNE method based on multi-color flow cytometry data ( Figure 1B ). The two populations of Tregs (CCR4⁺ and CCR4^-) are present in both healthy donors and tumor patients, according to CCR4 expression levels; however, the presence of the CCR4⁺ Treg population is higher than the CCR4^- Treg population in tumor patients, which indicates that Treg increases the expression of CCR4 to infiltrate in the TME. We intended to evaluate the status of CCR4 expression in other T-cell subsets (Th1 and Th2) because Tregs show a higher expression of CCR4 in tumor settings. When compared to healthy donors, CCR4 expression is significantly increased by Tregs during tumor condition; however, CCR4 expression is not significantly increased by other T-cell subsets ( Figure 1C ).

We evaluated the expression level of CCR4 in both the peripheral circulation (source) and the tumor site (destination) since it is the tumor-specific migratory marker of Tregs. Figure 1D shows that the CCR4⁺ Treg population was found in higher frequency in Tregs obtained from the tumor site (30%) than in Tregs obtained from the tumor patient’s peripheral blood (10%), implying that when the Treg acquires CCR4 expression in the peripheral circulation, it migrates to the tumor site. The graphical representation ( Figure 1D , right panel) also shows that CCR4⁺ Tregs account for the majority of Tregs in tumor tissue, whereas the percentage in the peripheral circulation of both the healthy donor and the tumor patient is relatively low, indicating that only CCR4⁺ Tregs can infiltrate the tumor site and CCR4^- Tregs cannot ( Figure 1D ). In comparison to CCR4^- Tregs, CCR4⁺ Tregs from both the peripheral circulation and tumor tissue had greater Ki67 expression. CCR4⁺ Tregs from both the peripheral blood and tumor tissue had greater Ki67 expression than CCR4^- Tregs, indicating that CCR4⁺ Tregs are more proliferative ( Figure 1E ). The tumor immune escape is facilitated by the eventual proliferation of these CCR4⁺ Treg cells in tumor tissue.

CCR4 is a chemokine receptor that responds to CCL22 and CCL17 ligands (27, 28). The interaction between CCL22/CCL17 and CCR4 triggers a signaling cascade in Tregs that results in changes in cell shape and increased motility. In our experiments, we discovered that only M2 macrophages (tumor-associated macrophages) present in the TME express higher levels of CCL22/CCL17 than other macrophage subsets ( Figure 2A ). Figure S2 shows the validation of macrophage subtype characterization. It was also clear that core breast tumor tissue showed higher levels of CCL22/CCL17 transcripts than neighboring non-tumor/normal tissue ( Figure 2A ). We used correlation coefficient plots to show the relationship between CCR4 and CCL22/CCL17 expression in core breast tumor tissue, which revealed a strong positive connection between CCR4 and CCL22/CCL17 in breast tumor tissue ( Figure 2A , lower panel). The ability of CCR4⁺ Tregs to migrate towards the CCL22 chemokine gradient was assessed using a trans-well migration assay in which Tregs from breast cancer patient blood were placed at the top of the chamber and a gradient of recombinant CCL22 was adjusted in the lower chamber containing cell culture media. It was observed that as the concentration of CCL22 increased, the percentage of migrating CCR4⁺ Tregs increased ( Figure 2B ). In the lower panel, a graphical representation of the same has been provided ( Figure 2B ). CCL22 and CCL17 levels in breast cancer tissue are associated with strong Treg infiltration and consequent immunosuppression; hence, it could be used as a prognostic marker. Several studies have also found that Treg cells dampen the immune system by producing the immunosuppressive cytokine IL10 (12). As shown in Figure 2C , the alteration of CCR4 expression in Treg has no negative impact on its immunosuppressive function since the level of intracellular IL10 is not affected. Furthermore, the Annexin-V positivity test revealed that CCR4 expression does not confer a survival advantage to Tregs. All of this suggests that the chemokine receptor CCR4 is exclusively involved in the tumor-specific migration of Tregs without affecting its survival or suppressive activities.

We decided to investigate FOXP3’s activity as a transcriptional activator for CCR4 because it plays such an important role in Treg development and function (37). Being a transcription factor, FOXP3 binds to the promoter region of its target genes. Because FOXP3 nuclear localization is critical for Treg function, we decided to investigate its nuclear translocation. Our findings show that nuclear localization of FOXP3 is more common in tumor patients’ lymphocytes than in normal lymphocytes ( Figure 3A ), implying that it plays a dynamic role as a transcription factor. Our t-SNE study of TCGA data from patients with breast invasive cancer reveals that high CCR4 expression in tumor tissue is linked to high FOXP3 expression ( Figures 3B, C ). With this in mind, we aimed to confirm the positive association between CCR4 and FOXP3 in other cohorts of breast cancer patients. To demonstrate the association between CCR4 and FOXP3 expression in core breast tumor tissue, we created a correlation coefficient plot, which further indicates that CCR4 expression is highly correlated with FOXP3 expression in breast tissue taken from different cohorts ( Figure 3D ).

We ablated or overexpressed FOXP3 in Treg cells and grew them on a breast tumor organoid bed to replicate the ex vivo TME to better understand the role of FOXP3 in CCR4 transcriptional activation. Flow cytometry was used to confirm Treg production ex vivo ( Figure S1 ). CCR4-mRNA expression is very low in FOXP3-ablated Tregs and very high in FOXP3-overexpressed Tregs, as seen in Figure 3E . After FOXP3 depletion/overexpression, CCR4 expression at the protein level as well as at the cell surface showed a consistent pattern with mRNA levels ( Figure 3F ). As a transcription factor, FOXP3 binds to the regulatory area of its target gene’s promoter, prompting us to investigate FOXP3’s binding site at the CCR4 promoter. For the same reason, we used in silico analysis to examine the CCR4 promoter and observed two FOXP3-binding sites at the −668 bp to −651 bp (distal site) and −124 bp to −108 bp (proximal site) regions of the promoter. Figure 3G shows the FOXP3 (DNA-binding domain) consensus promoter binding sequence (obtained from Jasper). This sequence is found in the regulatory region of all FOXP3 target genes; the CCR4 promoter contains the same DNA-binding region, prompting us to investigate FOXP3’s binding to the CCR4 promoter. The in silico finding was validated by ChIP analysis. FOXP3 binding at the distal site was stronger than at the proximal position ( Figure 3G ), showing that the distal region is the most critical FOXP3-binding site in the CCR4 promoter. We also examined RNA POL-II binding on the CCR4 promoter to confirm that the promoter region was open for FOXP3 binding, and the ChIP signal for RNA POL-II was highest in this region ( Figure 3G ).

We altered FOXP3 in ex vivo-generated Tregs to better understand the processes through which FOXP3 modulates CCR4 transcription. We found that ablation of FOXP3 decreased the promoter binding, but overexpression of FOXP3 boosted it ( Figure 4A ). Because of the permissive and repressive deacetylation state of histone-3 in naïve T cells, the CCR4 promoter stays inaccessible. FOXP3 overexpression increased HAT1 binding to the CCR4 promoter’s distal regions, modifying the chromatin structure and making it accessible to FOXP3 ( Figures 4B, C ). Interestingly, both FOXP3 and HAT1 binding is required for permissive (K23 and K27) or repressive (K14 and K18) acetylation of histone-3 in Tregs to increase CCR4 expression. Figure 4D shows a schematic depiction of the FOXP3-binding region on the CCR4 promoter at the distal region coupled with HAT1. FOXP3 expression in Tregs was ablated or overexpressed by genetic manipulation, and the ability of the altered Tregs to migrate was monitored using a trans-well migration assay in the presence of recombinant CCL22 to confirm our hypothesis that FOXP3 regulates Treg infiltration in the tumor site. The number of migrating cells in the lower chamber was much higher in FOXP3-overexpressed cells than in FOXP3-ablated cells, according to the data in Figure 4E . Concurrently, CCR4 expression was found to be higher in the migrating Treg cells ( Figure 4F ). In conclusion, the FOXP3/HAT1 complex induces acetylation of the CCR4 promoter in tumor Tregs, which regulates Treg infiltration in the TME.

We wanted to see if this phenomenon was true in the in vivo animal model because both breast cancer patient-derived Tregs and ex vivo-generated Treg cells showed a dependency of FOXP3 in the transcriptional regulation of CCR4, the key regulator of Treg infiltration in the TME. We employed 4T1-bearing breast carcinoma BALB/c mice and melanoma (B16/F10)-bearing C57/BL6 mice for this study. To genetically manipulate FOXP3 expression in mice, the mice were injected with lentivirus harboring Foxp3-shRNA or Foxp3-cDNA clones. The mice were monitored for 21 days after treatment. In both the breast carcinoma ( Figure 5A ) and melanoma ( Figure 6A ) models, FOXP3 ablation significantly reduced tumor volume and mass. The decrease in tumor size is accompanied by a decrease in CCR4⁺ Treg generation both in the peripheral circulation ( Figures 5B , 6B ) and at the tumor site ( Figures 5C , 6C ), demonstrating that FOXP3 transcriptionally activates CCR4 in both in vitro and in vivo settings. FOXP3 overexpression, on the other hand, enhanced tumor size ( Figures 5A , 6A ) by increasing the percentage of CCR4⁺ Tregs in the peripheral circulation ( Figures 5B , 6B ) as well as at the tumor site ( Figures 5C , 6C ).