Integrins α₅β₁ and α_vβ₃ Differentially Participate in the Recruitment and Reprogramming of Tumor-associated Macrophages in the In Vitro and In Vivo Models of Breast Tumor

Tumor-associated macrophages (TAMs) are a prominent component of the tumor microenvironment (TME), accounting for up to 50% of tumor mass (1). These heterogeneous groups of macrophages are functionally and phenotypically diverse and account for the immunosuppressive nature of the microenvironment, with increased TAM counts associated with aggravation of tumor growth, undesirable prognosis, and therapy resistance (1, 2). Signals from the tumor and its microenvironment shape immune cell behavior in the tumor’s stroma. The TME recruits myeloid cells, such as monocytes, granulocytes, and macrophages, which often exhibit an immunosuppressive, proangiogenic, and profibrotic characteristic (3–5).

Several tumor-derived chemoattractants, such as chemokines, ILs, and cytokines, are essential in recruiting myeloid cells into tumors (6, 7). Following recruitment, monocytes differentiate into macrophages and switch their functional phenotypes between classical inflammatory (M1) to an immunosuppressive phenotype (M2) by sensing their environment (8, 9). TAMs are shown to resemble the M2 phenotype, and they are delivered to facilitate tumor progression via matrix remodeling, promoting angiogenesis and contributing to immunosuppression in the tumor environment (10, 11). TAMs are protumorigenic macrophages that produce immunosuppressive cytokines, chemokines, and growth factors such as arginase, IL-10, and TGF-β, inhibiting antitumor immunity and encouraging tumor progression (12, 13).

Integrins are transmembrane heterodimeric glycoproteins that mediate cell migration in coordination with molecules such as GTPases and chemokine receptors (14). Signals from the binding of extracellular matrices (ECMs) with integrins also contribute to various cellular processes, such as adhesion, migration, differentiation, proliferation, and immunological functions (15, 16). Myeloid cells interact with specific ligands within the extracellular matrices of the TME to localize and navigate within the tumor site (17, 18). Integrins can influence the activation and function of myeloid cells in the TME and modulate the balance of proinflammatory and anti-inflammatory responses, affecting the tumor’s overall immune landscape. Our previous study has also demonstrated that integrin α₄ is associated with monocyte mobilization from bone marrow to the site of inflammation under systemic inflammatory conditions and that integrins α₅ and α_v could mediate polarization into an M1 and M2 phenotype (19). Thus, understanding the pivotal role of integrins in infiltrating myeloid cells into tumors is critical for developing immunomodulatory cancer therapeutic strategies that target the integrins or their downstream signaling pathways.

In this study, using M-CSF or MDA-MB-231 tumor supernatant (TS)–derived in vitro–generated TAMs from human peripheral blood monocytes and INF-γ– and LPS–derived M1 polarization or IL-4– and IL-13–derived M2 polarization or 4T1 TS-derived polarization in RAW 264.7 cells and a 4T1-induced murine breast tumor model, we investigated the role of surface integrin in tissue-specific recruitment and functional reprogramming of TAMs during tumor progression. Our study showed that integrin α₅β₁ drives the recruitment of monocytes from bone marrow and blood to peritoneum and tumor and correlated with the M1 phenotype. In contrast, TAMs expressing high levels of CD206 and MERTK (MER proto-oncogene, tyrosine kinase) exhibited increased α_vβ₃ expression and contributed to M2-type polarization. Our data derived from gene expression studies in sorted TAMs expressing specific integrin, functional blocking studies, and bioinformatics data analysis demonstrated an essential role of integrins α₅β₁ and α_vβ₃ in TAM reprogramming in mammary tumors. Our data show that adhesion receptors such as integrins play an essential role in the recruitment, polarization, and reprogramming of TAMs from inflammatory to immunosuppressive phenotypes in the tumor microenvironment and could be potential targets for TAM-targeting cancer therapeutics.

Human breast adenocarcinoma (MDA-MB-231, MCF-7), pancreas ductal carcinoma (PANC-1, MIA PaCa-2), mouse mammary carcinoma (4T1), and mouse macrophage RAW 264.7 cell lines were obtained from the National Centre for Cell Science (Pune, India). MDA-MB-231 and MIA PaCa-2 cells were cultured in DMEM (HiMedia, Kennett Square, PA) with 10% FBS, whereas MCF-7, PANC-1, 4T1, and RAW 264.7 cells were cultured in RPMI 1640 (Life Technologies) with 10% FBS and 1% antibiotic/antimycotic solution, all at 37°C with 5% CO₂. Conditioned medium (TS) was collected by incubating cells in serum-free DMEM or RPMI 1640 for 24 h after reaching 90% confluency. Supernatants were collected and centrifuged at 12,000 × g for 10 min, and aliquots were stored at −80°C for further use.

All human blood monocyte experiments were conducted with blood drawn from healthy donors following protocols endorsed by the Institute Human Ethics Committee, Indian Institute of Technology (Roorkee, India). PBMCs were isolated using Histopaque 1077 (Sigma-Aldrich, St. Louis, MO), followed by careful isolation of monocytes using the plastic adherence method. Briefly, 5 × 10⁵ PBMCs were seeded for 1 h in 400 µl of RPMI 1640 devoid of FBS in a 48-well flat-bottom plate. To differentiate monocytes into macrophages, they were cultured for 7 d in RPMI 1640 medium supplemented with 5% FBS and either TSs or M-CSF (Life Technologies, Thermo Fisher Scientific, Waltham, MA) as described by Chakraborty et al. (20). Cells were harvested at various intervals (days 1, 3, 5, and 7) and prepared for flow cytometry and quantitative PCR (qPCR). In a separate experiment, before adding the TSs, cells were pretreated with 10 µg/ml anti-integrins α₅ and α_v and 3 μM focal adhesion kinase (FAK) inhibitor (PF-573288) (Sigma-Aldrich, St. Louis, MO) and cultured for days 3 and 7 after seeding.

In additional experiments, monocytes were isolated using the MojoSort human pan monocyte isolation kit (BioLegend, San Diego, CA) and cultured on plates coated with ECM proteins (fibronectin, fibrinogen, vitronectin, and vimentin) (5 μg/ml) (Merck & Co., Rahway, NJ). The plates were then blocked using 2% BSA. Monocytes were cultured for 7 d in RPMI 1640 medium with 5% FBS in the presence of either M-CSF or MDA-MB-231 TS. On day 7, cells were collected and processed for gene analysis. In a different experiment, only fibronectin and vitronectin (5 μg/ml) were coated, and before adding cells to the wells, monocytes were incubated with anti-integrins α₅β₁ and α_vβ₃ and the cells were then incubated in RPMI 1640 medium in the presence of MDA-MB-231 TS. After 7 d of culture, flow cytometry was performed.

RAW 264.7 cells were polarized as described previously with modifications (21, 22). Briefly, RAW 264.7 cells were cultured in the presence or absence of FAK inhibitor (PF-573288, 3 μM) followed by stimulation with IFN-γ (20 ng/ml) combined with LPS (20 ng/ml) (Escherichia coli O26:B6; Sigma-Aldrich, St. Louis, MO) or IL-4 (20 ng/ml) combined with IL-13 (20 ng/ml) or culture supernatant of a 4T1 cell line and cultured for 24 and 48 h. The recombinant Abs were purchased from PeproTech (Rocky Hill, NJ). The cells were collected and flow cytometry was performed.

Female BALB/c mice (10–12 wk old) were obtained from the Indian Institute of Science Education and Research (Mohali, India), with approval from the Institute Animal Ethics Committee at the Indian Institute of Technology (Roorkee, India). Tumor induction involved a s.c. injection of 1 × 10⁶ 4T1 cells in 100 μl of sterile 1× PBS near the fourth nipple. Tumor growth was assessed periodically by measuring size with vernier calipers and calculating volume using the following equation: V = (L × W²) × 0.50, where V indicates volume, L indicates length, and W indicates the perpendicular width. Mice were sacrificed on days 3, 7, 10, 15, 23, and 30, and various samples, including blood, bone marrow, peritoneal lavage, and lungs, were collected. Tumor samples were obtained on days 10, 15, 23, and 30 for subsequent analysis.

On days 3 and 7, in vitro–generated macrophages derived from human peripheral blood monocytes were detached using ice-cold PBS, followed by surface marker staining using anti-human CD11b-FITC/allophycocyanin, CD14-PerCP, MHC class II (MHCII)–FITC, CD206-PE, CD86-PE/Cy7, and CD163-allophycocyanin/Cy7 Abs. The differentiated RAW 264.7 cells were collected after 24 and 48 h and stained using purified anti-mouse FcR (CD16/CD32), CD11b-BV510 (BD Biosciences, San Diego, CA), F4/80-AF488/PerCP, MHCII-PerCP, CD206(MMR)-BV605, MERTK-allophycocyanin, CD49e-PE/Cy7, and CD51-PE. After surface staining, the cells were immediately fixed and permeabilized using an intracellular fixation and permeabilization buffer set (Thermo Fisher Scientific, Waltham, MA), and intracellular staining was done using phospho–SRC (proto-oncogene tyrosine-protein kinase c-Src)-AF488 (Thermo Fisher Scientific, Waltham, MA), anti-STAT1 phospho-PE, and anti-STAT6 phospho-PE/Cy7.

For the in vivo study, blood, bone marrow, peritoneal lavage, lungs, and tumors were isolated on the above-mentioned days. To prepare single-cell suspensions of lung and tumor, minced lung and tumor were enzymatically digested using collagenase IV (Sigma-Aldrich, St. Louis, MO). Then, the mixture was mechanically disrupted by passing it over a 70-μm cell strainer. Contaminated erythrocytes were eliminated using RBC lysis buffer (BioLegend, San Diego, CA). For surface marker staining, purified anti-mouse FcR (CD16/CD32), CD11b–BV510, Ly6G-BUV395 (BD Biosciences, San Diego, CA), Ly6C-allophycocyanin/Cy7, F4/80-AF488/PerCP, MHCII-PerCP, CD206(MMR)-BV605, CD64-BV786 (BD Biosciences, San Diego, CA), CCR2-FITC, CXCR4-allophycocyanin, MERTK-allophycocyanin, CD49e-PE/Cy7, and CD51-PE Abs were used.

In vitro and in vivo samples were gated using BD Horizon fixable viability stain 450 (FVS450; BD Biosciences, San Jose, CA). All flow cytometry Abs were purchased from BioLegend (San Diego, CA) unless otherwise noted. All samples were fixed with 1% paraformaldehyde (HiMedia, Kennett Square, PA) and collected on a FACSAria Fusion flow cytometer and FACSLyric flow cytometer (BD Biosciences, San Jose, CA).

A single-cell tumor suspension was stained with Abs in a 500-μl vol for 30 min on ice. Live cells were gated using BD Horizon FVS450. Surface marker staining included Abs against purified anti-mouse FcR (CD16/CD32), CD11b-BV510, Ly6G-BUV395, Ly6C-allophycocyanin/Cy7, F4/80-AF488, CD49e-PE/Cy7, and CD51-PE. After washing, cells were filtered through 70- and 40-μm cell strainers without fixation. Cell sorting was performed using the BD FACSAria Fusion sorter. Tumor tissue macrophages were gated based on integrins α₅ and α_v on FVS450⁻CD11b⁺Ly6G⁻F4/80⁺ cells, excluding cell doublets. Sorted cells (α₅⁺, α_v⁺, and α₅⁺α_v⁺) were collected into 5-ml FACS tubes with 2 ml of complete RPMI 1640 medium without FBS. Ly6C^hi monocytes from bone marrow served as the control group. The postsorting purity of macrophages was determined, and gene expression was assessed using qPCR.

The total RNA from differentiated macrophages/TAMs generated in vitro and tumor tissues were extracted using TRIzol reagent. The total RNA of cells sorted from tumor tissues was extracted using an RNeasy kit (Qiagen, Germantown, MD). RNA was reverse transcribed using a high-capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA). Real-time PCR was carried out according to the manufacturer’s instructions using PowerUp SYBR Green master mix (Applied Biosystems, Foster City, CA). A StepOnePlus real-time PCR system (Applied Biosystems, Forster City, CA) was used for the quantitative real-time PCR analysis. Samples were normalized to the endogenous control (human GAPDH or mouse HPRT), and relative gene expression was measured using the 2^−ΔΔCt method. The primers used for the study are listed in Supplemental Table I.

Peripheral blood samples from naive and tumor-induced mice on days 3, 7, 10, 15, 23, and 30 were collected via retro-orbital bleeding. Serum was obtained by allowing blood to clot for 15 min at room temperature, followed by centrifugation at 1800 × g for 5 min at 4°C. IL-6 and IL-10 levels in serum samples were measured using a sandwich ELISA protocol with Abs from PeproTech (Rockey Hill, NJ). The color reaction was initiated with a tetramethylbenzidine liquid substrate system (BioLegend, San Diego, CA) and stopped with 2 N H₂SO₄ (HiMedia, Kennett Square, PA) before measuring absorbance at 450 nm using a BioTek Epoch 2 plate reader (BioTek, Winooski, VT).

An RNA sequencing dataset (GSE195857) was downloaded from the GEO database. This dataset contains the transcriptome of TAMs, mammary gland macrophages (MGMs) from healthy mammary fat pads, and bone marrow–derived monocytes from healthy and tumor-bearing mice (BMDM-Ts/BMDM-Hs). Quantification of transcript expression was done by the Salmon algorithm (23), followed by quality control. The data were further processed on Bioconductor using the DESeq2 package (24) with test = Wald, fitType = parametric, and a p value cutoff of 0.05 to identify the differentially expressed genes (DEGs). Functional enrichment analysis on DEGs was performed using the goseq package (25). The genes obtained from the top pathway after functional enrichment were mapped to a protein–protein interaction network using the STRING v12.0 database (http://string-db.org/) (26).

Statistical analysis was performed using a one- or two-way ANOVA with a Bonferroni correction. The Holm–Sidak method was used to analyze tumor sizes with various t tests. The values are presented as mean ± SEM, with p values ≤0.05 considered statistically significant across all datasets. Data were analyzed using GraphPad Prism 9.5.1 for Windows (GraphPad Software, La Jolla, CA). All flow cytometry data were analyzed using FlowJo v10.8.1 (Tree Star, Ashland, OR).

To investigate the distribution pattern of monocytes in the TME, flow cytometric analysis was performed on different days (days 3–30) postinduction to determine their recruitment from the bone marrow to blood, tumor, peritoneum, and lungs after induction of 4T1-induced breast tumor. Cells were isolated from the tissues and stained with different myeloid cell surface markers (CD11b, Ly6G, and Ly6C) to differentiate monocytes and neutrophils. CD11b⁺Ly6G⁻Ly6C^hi cells were gated to indicate monocytes in various tissues (Fig. 1A). In the bone marrow, the frequency of monocytes increased from day 3 to day 15 after tumor induction followed by gradually decreasing as compared with naive animals (Fig. 1A, first row). Conversely, Ly6C^hi monocytes increased in blood after day 10, indicating migration from bone marrow to blood after day 7 (Fig. 1A, second row). The frequencies of Ly6C^hi cells in the peritoneum and lung increased steadily from day 3 to 30. However, there was a reduction in Ly6C^hi cell frequency in tumor tissues (Fig. 1A).

After injection of 4T1 cancer cells in the mice, the growth of mammary tumors was measured on days 10, 15, 23, and 30, showing a significant increase in volume from day 10 to 30 (Fig. 1B). Monocyte distribution in the tumor, peritoneum (the nearest site to tumor tissue), and lungs (a distant site of metastasis) was studied during the course of tumor progression. The study revealed two Ly6C monocyte populations gated as CD11b⁺Ly6G⁻Ly6C^hi and CD11b⁺Ly6G⁻Ly6C^low and CD11b⁺Ly6G⁻F4/80⁺ macrophages in these tissues (Fig. 1C). As shown in Fig. 1D, in the tumor, there was a decrease in the number of Ly6C^hi monocytes and more Ly6C^low monocytes after day 10, alongside a significant increase in the number of F4/80⁺ macrophages, suggesting differentiation of Ly6C^hi monocytes into macrophages (Fig. 1D). Interestingly, peritoneum displayed no substantial change in Ly6C^low monocyte numbers but showed a significant increase in the number of Ly6C^hi monocytes and a huge increase in F4/80⁺ macrophages after day 10, indicating monocyte differentiation (Fig. 1E). In the lungs, there was no substantial change in monocyte/macrophage numbers initially (days 0–7); however, from day 10 on, there was a rise in Ly6C^hi and Ly6C^low monocytes, alongside a significant increase in F4/80⁺ macrophages (Fig. 1F). This sudden increase in the macrophage population in the lungs suggests that monocytes/macrophages have migrated from their precursor sites and infiltrated the lungs in large numbers after tumor induction.

CCR2, a G protein–coupled receptor, is a critical functional receptor for CCL2 that primarily directs the chemotactic responses of monocytes to the site of inflammation (27). Similarly, the CXCR4/CXCL12 axis has been correlated with breast tumor growth, metastasis, and invasion upon binding with the ligands, triggering signaling that orchestrates the retention of cells in the bone marrow. Several studies have correlated high levels of CXCR4 expression in cancers with poor prognosis and resistance to chemotherapy (28, 29). Our analysis revealed the CCR2 and CXCR4 expression in monocytes across bone marrow, blood, tumor, peritoneum, and lungs. In bone marrow and blood, CCR2 expression increased from day 3 to 30 in Ly6C^hi monocytes compared with naive animals (Fig. 2A). However, CXCR4 expression decreased in bone marrow from day 3 to 10. It increased in blood, which suggests the migration of monocytes from bone marrow to the blood (Fig. 2B). The histogram overlay demonstrates the significant changes in CCR2 and CXCR4 levels in the bone marrow and blood of naive and day 7 post-4T1 injection mice, indicating early migration activity. This is essential for capturing the immune response in its early stages and the mobilization of monocytes. These findings underscore the regulatory roles of CCR2 and CXCR4 in monocyte migration from bone marrow and blood during tumorigenesis.

We investigated myeloid cell trafficking from bone marrow to tumor, peritoneum, and lungs, analyzing CCR2 and CXCR4 expression in Ly6C^hi/low and F4/80⁺ cells. In the tumor, Ly6C^hi monocytes exhibited higher CCR2 and CXCR4 expression compared with Ly6C^low and F4/80⁺ cells, whereas CXCR4 increased in F4/80⁺ macrophages in the tumor after day 15 (Fig. 2C, 2D). Contrary to the monocytes, F4/80⁺ macrophages showed higher CCR2 and CXCR4 expression in the peritoneum than did Ly6C^hi and Ly6C^low monocytes, which were upregulated from day 3 to 30 (Fig. 2E, 2F). In the lungs, CCR2 and CXCR4 expression in Ly6C^hi/low, F4/80⁺ cells resembled naive animals initially (days 0–7) but increased after day 10 of tumor induction, suggesting myeloid cell migration to the lungs after day 10 (Fig. 2G, 2H). The histogram overlay shows the difference between the Ly6C^hi/low and F4/80⁺ populations on day 15, as well as a significant increase in CCR2 and CXCR4 expression levels in the tumor, peritoneum, and lungs.

Chemokines such as CCL2 and CXCL12 participate in the recruitment of monocytes into tumors, with CCR2 and CXCR4 serving as their respective receptors (30). MIP-1α or CCL3 and CCL5 also recruit and activate immune cells, promoting inflammation of the tumor and the progression of the disease (31, 32). The transcriptional expression of these chemokines (Ccl2, Cxcl12, Ccl3, and Ccl5) and receptors (Ccr2 and Cxcr4) in breast tumors was examined from day 0 to day 30. Ccl2 expression significantly peaked from day 0 to day 10 compared with Cxcl12, Ccl3, and Ccl5 (Fig. 2I). Receptors Ccr2 and Cxcr4 showed upregulated expression toward the early time point on day 3, followed by a significant decrease as the tumor progressed (Fig. 2J). The gene expression of CCL2/CCR2 and CXCL12/CXCR4 chemokines decreased significantly as the tumor grew, suggesting mRNA translating into proteins, which correlates with our previous finding on increased levels of CCR2 and CXCR4 in the tumor after day 10.

Integrin receptors mediate the migration of immune cells and are also necessary for immune cell phagocytosis, migration, and cellular adhesion (33). When analyzing Ly6C^hi monocytes across bone marrow, blood, tumor, and lungs after tumor induction, we found that among the different integrins (α₂, α₃, α₄, α₅, α₆, α_v) analyzed, the expression of integrin α₅β₁ was upregulated in the bone marrow and blood as the tumor progressed. Moreover, in the tumor and lungs, α_vβ₃ was upregulated in tumor-infiltrating monocytes and macrophages (data not shown). Further analysis of integrin α₅β₁ and α_vβ₃ expression patterns in bone marrow and blood samples in a series of time points collected on days 3, 7, 10, 15, 23, and 30 after tumor induction revealed upregulated α₅β₁ expression in Ly6C^hi monocytes (Fig. 3A, 3B). The histogram overlay shows that in the bone marrow and blood, the expression of integrin α₅β₁ significantly increased after day 15 after tumor induction as compared with the naive mice, and there was a significant decrease in the expression of integrin α_vβ₃ in the bone marrow.

In our study, we observed differential expression of integrins α₅β₁ and α_vβ₃ on monocyte and macrophage subsets in the tumor, peritoneum, and lungs. In tumors, α₅β₁ expression was higher in Ly6C^hi and Ly6C^low cells compared with F4/80⁺ macrophages, and integrin α_vβ₃ was upregulated in all three cell populations after day 15 with a more robust upregulation in F4/80⁺ macrophages (Fig. 3C, 3D). Peritoneal Ly6C^hi and Ly6C^low monocytes showed similar integrin expression patterns, with significant upregulation of both integrins in F4/80⁺ macrophages (Fig. 3E, 3F). In the lungs, both integrins α₅β₁ and α_vβ₃ were upregulated in F4/80⁺ macrophages after day 10 compared with Ly6C^hi/low monocytes (Fig. 3G, 3H). The histogram overlay shows that there was a distinct difference in the expression of integrins α₅β₁ and α_vβ₃ between these populations, that is, Ly6C^hi, Ly6C^low, and F4/80⁺ cells, during day 23 postinjection as compared with the naive cells. Our study concludes that in the later stage of tumor development, α_vβ₃ integrin expression is higher in F4/80⁺ macrophages compared with tumor-infiltrating monocytes in the tumor. In contrast, F4/80⁺ macrophages also expressed higher α₅β₁ and α_vβ₃ in both the peritoneum and lungs.

Integrin profiling of TAMs differentiated from human peripheral blood monocytes using culture supernatants from breast cancer (MDA-MB-231, MCF-7) and pancreatic cancer cell lines (PANC-1, MIA PaCa-2) revealed that MDA-MB-231 TS induced an M2 phenotype with higher CD206 and CD163 and lower MHCII and CD86 expression levels compared with other TS- and M-CSF–treated cells (Fig. 3I–L). Integrin expression profiles in TAMs from different TSs were examined during several days (days 1, 3, 5, and 7 after seeding), showing upregulation of α₅ and α_v integrins in MDA-MB-231 TAMs compared with control cells cultured with media alone (just media) (Fig. 3M). In contrast, TAMs from MCF-7 TS exhibited initial downregulation and late upregulation of α₅ and α_v integrins compared with control (Fig. 3N), whereas there was a downregulation of integrins α₅ and α_v in PANC-1 and MIA PaCa-2 TS-differentiated macrophages (Fig. 3O, 3P). The results show that MDA-MB-231 TS-treated macrophages have higher levels of integrins α₅ and α_v compared with other culture supernatants, which were used for further in vitro studies. RAW 264.7 cells stimulated with IFN-γ/LPS for M1 macrophages, IL-4/IL-13 for M2 macrophages, or 4T1 TS also showed increased α₅ and α_v expression compared with controls (media only) (Fig. 3Q, 3R). These findings highlight that MDA-MB-231 TS and 4T1 TS induce significant upregulation of integrins α₅ and α_v in human peripheral blood monocyte-derived macrophages and RAW 264.7 cells, respectively, suggesting that these integrins are key markers for macrophage polarization in breast cancer.

Previous studies on TAM phenotypic markers in the 4T1-induced tumor model demonstrated a progressive shift from Ly6C^hi to Ly6C^low populations, which were further categorized into M1 and M2 TAMs based on MHCII expression (34). According to this study, F4/80⁺ cells were gated on CD11b⁺Ly6G⁻ populations (Fig. 4A), yielding six groups: Ly6C^hiMHCII^low (gate 1), Ly6C^intMHCII^low (gate 2), Ly6C^lowMHCII^low (M2 TAMs; gate 3), Ly6C^hiMHCII^hi (gate 4), Ly6C^intMHCII^hi (gate 5), and Ly6C^lowMHCII^hi (M1 TAMs; gate 6) as shown in Fig. 4B. The frequency and cell number of M2 TAMs significantly increased from day 10 to day 30 after tumor induction, whereas Ly6C^hiMHCII^low cells were decreased after day 10 (Fig. 4C). The number of Ly6C^intMHCII^low cells increased after day 10 and then decreased subsequently. This suggests the polarization of Ly6C^hi cells into M2 TAMs via an intermediate population of the Ly6C^intMHCII^low phenotype. Additionally, the number of M1 TAMs decreased after day 10, possibly indicating repolarization to M2 TAMs later (Fig. 4C). The phenotype of these subsets was also verified by analyzing the surface marker expression levels of CD206, MERTK, and CD64. M2 TAMs displayed higher CD206 and MERTK expression levels compared with M1 TAMs and Ly6C^hi cells, whereas Ly6C^hiMHCII^low (gate 1), Ly6C^intMHCII^low (gate 2), Ly6C^hiMHCII^hi (gate 4), and Ly6C^intMHCII^hi (gate 5) populations exhibited higher CD64 expression compared with both M1 and M2 TAMs (Fig. 4D–F).

Because integrins α₅ and α_v were differentially expressed in the F4/80⁺ cells in the tumor, the expression pattern and kinetics of these two integrins were examined at various time points on M1 and M2 TAMs and intermediate populations. As shown in Fig. 4Gand 4H, the shift of the histograms and the increase in the mean fluorescence intensity (MFI) values suggest that M1 TAMs, M2 TAMs, and Ly6C^intMHCII^low populations showed an upregulated expression of integrin α₅ after day 10. In contrast, Ly6C^hiMHCII^hi (gate 4) and Ly6C^intMHCII^hi (gate 5) populations showed elevated expression of integrin α₅ on days 10 and 15, then decreased significantly, but there were no significant changes in the Ly6C^hiMHCII^low cells. Similarly, integrin α_v expression significantly increased in the M2 TAM population compared with Ly6C^hi cells and M1 TAMs (Fig. 4I, 4J). The Ly6C^hiMHCII^hi and Ly6C^intMHCII^hi populations also showed elevated expression of α_v on day 10 and reduced expression after day 15 (Fig. 4I, 4J). This suggests that M2 TAMs expressed significantly higher levels of integrins α₅ and α_v and gradually increased in expression from day 10 to day 30 after tumor induction as compared with other subsets, whereas M1 TAMs expressed significantly higher integrin α₅ as compared with different subsets.

In the TME, macrophages undergo a phenotypic shift from proinflammatory (M1) to anti-inflammatory (M2), facilitating tumor growth. CD206, a mannose receptor, and MERTK signaling contribute significantly to establishing an immunosuppressive milieu crucial for tumor progression (35, 36). Therefore, CD206^hi/low and MERTK^hi/low cells in M1 and M2 TAMs in the tumor and lungs at days 15 and 30 were analyzed, and the expression of integrins α₅ and α_v in CD206 and MERTK expressed cells were checked. Distinct populations were gated based on the expression of CD206 and MERTK on Ly6C^lowMHCII^low (M1) and Ly6C^lowMHCII^hi (M2) TAMs (Fig. 5A, 5B, 5I, 5J). In the tumor and lungs, M2 TAMs expressed significantly higher CD206 and MERTK than M1 TAMs on days 15 and 30, which correlates with the cell number (Fig. 5C, 5D, 5K, 5L) and frequency of the cells (Fig. 5A, 5B, 5I, 5J). The cell number of the CD206^hi cells in the M2 TAMs was higher than CD206^low cells on both day 15 and day 30, and there was a significant increase in the number of CD206^hi and MERTK^hi cells in M2 TAMs compared with M1 TAMs in both tumors and lungs (Fig. 5C, 5D, 5K, 5L).

In the previous results, as the prominent expression of integrins α₅ and α_v in the different TAM subsets was observed, we hypothesized that these integrins might have been involved in the reprogramming of TAMs. Therefore, we analyzed the expression of integrins α₅ and α_v in CD206- and MERTK-expressing cells. As shown in Fig. 5E–H, integrin α₅ and α_v expression levels in the CD206^hi/low and MERTK^hi/low cells were notably elevated in M2 TAMs compared with M1 TAMs at both day 15 and 30, suggesting that these adhesion receptors might be involved in TAM reprogramming. Specifically, CD206^hi and MERTK^hi cells within M2 TAMs showed increased expression of α₅ and α_v compared with CD206^low and MERTK^low, respectively, on days 15 and 30. However, no significant difference in α₅ expression was observed in the M1 TAMs.

Similarly, in the lungs, M2 TAMs expressing CD206^hi have higher integrin α₅ expression than do CD206^low cells, and there were no significant differences in CD206^hi and CD206^low in M1 TAMs (Fig. 5M), whereas integrin α_v in CD206^hi cells expressed M1 and M2 TAMs was significantly upregulated as compared with the CD206^low cells both at day 15 and 30 (Fig. 5N). M2 TAMs expressing MERTK^hi have higher α₅ and α_v expression than do MERTK^low cells (Fig. 5O, 5P). Also, the expression levels of integrins α₅ and α_v in both CD206^hi/low and MERTK^hi/low cells were significantly increased in M2 TAMs as compared with M1 TAMs (Fig. 5M–P). With CD206 and MERTK being essential markers of TAM polarization, these findings suggest the possibility of the involvement of integrin α_v as well as integrin α₅ in the reprogramming of M2 TAMs.

In subsequent experiments, we investigated the functional role of integrins α₅ and α_v in TAM reprogramming at the transcriptional level. Macrophage subsets expressing these integrins were sorted from day 15 tumor tissues based on FVS450⁻CD11b⁺Ly6G⁻F4/80⁺ expression. Three macrophage populations were identified based on integrin expression: α₅⁺, α_v⁺, and α₅⁺α_v⁺ (Fig. 6A). The frequency and purity of different sorted populations are shown in Fig. 6B. Interestingly, α₅⁺ macrophages showed enhanced levels of inflammatory cytokines such as Tnfa and Il6 as compared with α_v⁺ and α₅⁺α_v⁺ cells transcriptionally (Fig. 6C, 6D). However, α_v⁺ macrophages showed an elevated level of M2 markers Cd163 and Cd206 at the transcriptional level compared with other macrophage populations (α₅⁺ and α₅⁺α_v⁺) (Fig. 6E, 6F). Similarly, α_v⁺ macrophages also showed increased levels of protumorigenic genes such as Il10, Arg1, Tgfb, and metastasis and angiogenic marker Mmp9 transcriptionally (Fig. 6G–J).

M1 macrophage polarization is driven by STAT1 and STAT2, which enhance inflammatory cytokine production and immune responses against tumors (11). Conversely, STAT3 and STAT6 influence M2 macrophage polarization, fostering an immunosuppressive environment that promotes tumor growth and metastasis (37, 38). Our results showed that there was a significant increase in Stat1 and Stat2 levels in the α₅⁺ macrophages, whereas α_v⁺ macrophages have significantly increased Stat3 and Stat6 expression in transcript level as compared with control Ly6C^hi cells (Fig. 6K–N). Additionally, integrins α_vβ₅ and β₃ depended on peroxisome proliferator-activated receptor-γ (PPARγ) signaling to regulate M2 polarization in breast tumor, further supporting a tumor-promoting microenvironment (39, 40). The Pparg gene expression was higher in α_v⁺ macrophages as compared with α₅⁺ macrophages (Fig. 6O). This interaction between α₅ and α_v integrins and tumor extracellular matrix modulates these signaling pathways, influencing TAM polarization and tumor progression. Our study indicates that α_v⁺ integrins may play a prominent role in the differentiation and polarization of M2 TAMs and help in tumor progression.

Inflammatory macrophages generate mediators such as IL-6 and IL-10, fostering a mutagenic environment. As tumors evolve, macrophages transition into M2-like TAMs, further promoting tumor growth while dampening immune responses (41). In our study, IL-6 and IL-10 levels were significantly higher in serum from tumor-bearing mice from day 3 to day 30 compared with controls (Fig. 6P). Elevated IL-6 suggests enhanced cancer cell survival, whereas increased IL-10 indicates a regulatory effect on the immune response, potentially suppressing antitumor reactions.

To check the function of α₅ and α_v integrins in macrophage polarization, human peripheral blood monocytes were cultured in MDA-MB-231 TS in the presence or absence of integrin α₅ and α_v blocking Abs. By day 7, M-CSF– and TS-treated monocytes differentiated into macrophages, displaying elongated morphology compared with controls, which was notably reduced upon integrin α₅ and α_v blockade (Fig. 7A). Quantitative analysis confirmed that integrin blocking significantly decreased cell spreading (Fig. 7B) and elongation (Fig. 7C), highlighting the importance of α₅ and α_v in TAM differentiation. Furthermore, gene expression analysis revealed that blocking integrins α₅ and α_v led to reduced expression of M2 markers (Cd206, Cd163) and protumorigenic genes (Arg1, Vegf) in the differentiated TAMs compared with TS control and M-CSF on day 7 (Fig. 7D–G). This indicates that integrins α₅ and α_v are crucial for promoting M2-type polarization and the expression of genes associated with tumor progression. Additionally, there was no significant impact on the expression of M1 markers (Cd80, Inos2) with integrin blocking, suggesting the role of α₅ and α_v integrin specificity toward M2 polarization (Fig. 7H, 7I).

FAK is a cytoplasmic tyrosine kinase crucial for intracellular signaling by integrins, influencing adhesion, migration, survival, growth, and differentiation (42, 43). To study the impact of FAK inhibition on macrophage polarization, human peripheral blood monocytes were treated with the FAK inhibitor PF-573228 and cultured with M-CSF and MDA-MB-231 TS for 7 d. Results showed a significant decrease in both M1 (MHC II, CD86) and M2 (CD206, CD163) markers in M-CSF– and TS-induced macrophages after FAK inhibition compared with cells without FAK inhibition (Fig. 7J–M). Similarly, in RAW 264.7 cells, IFN-γ and LPS stimulation induced M1 polarization evidenced by increased expression of MHCII (Fig. 7N), whereas increased expression of CD206 and MERTK was observed in IL-4– and IL-13–stimulated cells (Fig. 7O, 7P). FAK inhibition significantly reduced MHCII expression, impairing M1 polarization, and also reduced CD206 and MERTK expression, impacting M2 polarization. Additionally, treatment with 4T1-conditioned medium effectively polarized RAW 264.7 cells into M2 macrophages, as indicated by elevated CD206 and MERTK and lowered MHC II levels. However, the presence of the FAK inhibitor notably decreased these markers, demonstrating a disruption in M2 macrophage differentiation (Fig. 7O, 7P). These findings highlight FAK’s role in macrophage polarization and the effects of its inhibition.

During tumorigenesis, tumor-infiltrating myeloid cells can support cancer cell survival, proliferation, and migration, partly by influencing ECM remodeling (44). To that note, the expression levels of ECM proteins such as fibronectin, vitronectin, fibrinogen, collagen IV, and vimentin were checked in tumor tissues (days 10, 15, 23, and day 30) and mammary gland tissues (tissues from the site of injection on days 3 and 7) and compared them with mammary gland tissue from naive mice as a control. Results showed increased fibronectin, vitronectin, and collagen IV expression in tumors compared with naive control, fibrinogen, and vimentin as the tumor progressed (Fig. 8A). We further explored the impact of ECM proteins on TAM polarization using M-CSF and MDA-MB-231 TS. Fibronectin and vitronectin enhanced the expression of integrins α and α_v in TAMs, indicating their regulatory role in TAM polarization through integrins (Fig. 7B, 7C). Blocking integrins α₅ and α_v in monocytes cultured with MDA-MB-231 TS and ECM proteins reduced M2 markers (CD206 and CD163) but did not significantly affect M1 markers (MHCII and CD86), except for integrin α_v, which lowered CD86 and MHCII in vitronectin-coated wells (Fig. 8D–G). This suggests that fibronectin and vitronectin, through integrins, primarily influence M2 TAM polarization.

The involvement of integrins α₅ and α_v has emerged as a focal point of investigation due to their known roles in cell adhesion, migration, and TAM reprogramming. RNA sequencing analysis from the 4T1 murine breast tumor model (GEO dataset: GSE195857) compared TAMs to MGMs and BMDM-Hs to BMDM-Ts. This revealed 4148 DEGs for TAMs versus MGMs and 401 for BMDM-Ts versus BMDM-Hs, with significant pathways including the “immune system” (Reactome ID: R-MMU-168256). Notable integrins such as Itgam (CD11b) and Itgax (CD11c) were significantly upregulated in both macrophages and monocytes (Fig. 9A, 9B). Additionally, downregulation of integrin α₅ suggests reduced proinflammatory pathways, whereas upregulation of integrin α_v points to a shift toward a protumorigenic and immunosuppressive phenotype.

Furthermore, the correlation of DE genes in the immune system pathway with the integrins α₅ and α_v was checked with the STRING database. In the BMDM-Ts versus BMDM-Hs comparison, integrins α₅ and α_v were found to strongly interact with VCAM-1 and NCAM-1 (neural cell adhesion molecule-1), which further associate with MMP9 and STAT3 (Fig. 9C, 9D). VCAM-1 and NCAM-1 facilitate cell adhesion and motility, crucial for inflammatory responses. MMP-9 promotes ECM remodeling, aiding cell invasion and metastasis. Activation of STAT3 in monocytes fosters their differentiation into immunosuppressive TAMs, which also facilitate the secretion of various protumorigenic factors, including cytokines (Il-10ra, Il-13ra1, and Il-4ra), promoting tumor progression and immune evasion. In the TAMs versus MGMs category (Fig. 9E, 9F), many DEGs such as FAK, also known as Ptk2 (protein tyrosine kinase 2) proto-oncogene protein tyrosine kinases (Src, Crk, Lyn), vasodilator-stimulated phosphoprotein (Vasp), and several integrins subunits (Itgal, Itgb5, Itga4, Itgb7) show strong association with both α₅ and α_v integrins. FAK and SRC kinases regulate cell adhesion, migration, and survival, whereas VASP enhances motility and invasion by modulating actin dynamics, and integrin subunits mediate cell–ECM interactions. This network highlights the central roles of α₅ and α_v integrins in cancer metastasis, emphasizing their potential as therapeutic targets in cancer treatment.

To explore how FAK inhibition affects macrophage polarization signaling, human peripheral blood monocytes were treated with the FAK inhibitor PF-573228. This treatment significantly reduced the phosphorylation of SRC, STAT1, and STAT6 in macrophages differentiated with M-CSF and MDA-MB-231 TS compared with untreated controls (Fig. 9G–I). In RAW 264.7 cells, FAK inhibition led to decreased expression of integrins αv and α5 in both M1 macrophages (induced by IFN-γ and LPS) and M2 macrophages (induced by IL-4 and IL-13), as well as in cells treated with 4T1 TS (data not shown). Further analysis showed that FAK inhibition reduced SRC, STAT1, and STAT6 phosphorylation in RAW 264.7 cells under various treatments (Fig. 9J–L), highlighting the importance of FAK-mediated integrin signaling in macrophage polarization. The reduced integrin expression impairs the activation of key signaling pathways necessary for macrophage polarization.

The balance between the M1 and M2 polarization states of TAM can influence the fate of tumors (6, 45). Integrins play a crucial role in orchestrating the infiltration of myeloid cells into the TME and are known to influence the behavior of TAMs and their communication with cancer cells (46, 47). Despite the abundance of literature on inflammatory molecules involved in tumorigenesis, the series of events that occur before the appearance of palpable tumors, driving immune cell recruitment and polarization of TAMs in the dynamic TME, have still remained unknown. Thus, our study performed an in-depth analysis of the molecular landscape governing monocyte and macrophage trafficking from the bone marrow to the systemic circulation and tumor tissue, including recruitment and polarization of TAMs in possible metastatic visceral organs, lungs, and a site close to the tumor development, that is, peritoneum. Combined with reprogramming and polarization, immune cell recruitment mechanisms will provide essential insights into the TME and potential targets for therapeutic approaches.

A previous comprehensive study on the macrophage phenotypes in breast and lung carcinoma suggested that Ly6C^hi subsets were the most prominent population, serving as a precursor for all distinct TAM subsets. According to that study, these inflammatory monocytes differentiate rapidly into TAMs (34). Using the 4T1 murine model of breast tumors, our timeline study indicated an increase in the frequencies of Ly6C^hi monocytes in the bone marrow, blood, and peritoneum from day 3 to day 30 after tumor induction. Their frequencies declined in the tumor tissues from day 15 to day 30 after differentiation into TAMs, where the Ly6C^hi cells changed into Ly6C^low cells and polarized into different subpopulations of M1 (Ly6C^lowMHCII^hi) and M2 (Ly6C^lowMHCII^low) TAMs. Our data suggest that the CCL2/CCR2 signaling axis facilitates monocyte infiltration from bone marrow to tumor tissues.

Integrins facilitate myeloid cell adhesion to the ECM and endothelial cells, aiding TAM recruitment within the tumor stroma and enhancing infiltration into the tumor parenchyma (48, 49). Our results show that integrin α₅β₁ was highly expressed on the Ly6C^hi monocytes within the bone marrow/blood during early tumor progression. These cells migrated to the tumors and surrounding tissues, peaking at day 15 before their decrease in bone marrow. This suggests that integrin α₅ mediates the early events of monocyte recruitment to tumors. These monocytes further differentiated into F480⁺ macrophages and expressed higher levels of integrin α_vβ₃. We found that both integrins α₅β₁ and α_vβ₃ increased in their expression with the progression of the tumors, thereby underscoring their function in monocyte differentiation and TAM polarization. We further observed a gradual increase in M2-type TAMs as tumors advanced, with Ly6C^hi monocytes continuously infiltrating into the tumors. In line with this observation, previous studies have shown that the Ly6C^lowMHCII^low subset of TAMs was the primary population associated with tumor growth, invasion, and metastasis in mammary adenocarcinoma (50).

Furthermore, using multicolor flow cytometry studies, we identified six TAM subpopulations based on Ly6C and MHCII expression, each showing unique integrin α₅ and α_v patterns. Integrin α₅ was upregulated from day 10 to 30 after tumor induction in the M1, M2, and intermediate populations (Ly6C^hiMHCII^hi and Ly6C^intMHCII^hi). It was highly expressed in the M1 TAMs when compared with the M2 TAMs. In contrast to M1 TAMs, integrin α_v was highly expressed in M2 and upregulated integrin α₅ at a later time point. In contrast, in the intermediate populations, the expression of integrins α₅ and α_v increased until day 15 but decreased thereafter. These findings suggest that integrins α₅ and α_v both may be involved in M2 TAM polarization in the tumor microenvironment. Our findings are consistent with a prior study by Shu et al. (39) that illustrates how integrin β₃ via PPARγ is essential for the polarization of M2 TAMs in a 4T1-induced breast tumor model and that blockade of integrin β₃ suppressed macrophage M2 polarization. However, in this study, the authors analyzed the entire F4/80-expressing cells and not the cellular subsets. A more detailed analysis of M1 and M2 TAMs, based on the surface expression of anti-inflammatory markers, such as CD206 and MERTK, in tumors and lungs, revealed that the expression of integrin α_v was higher in both CD206^hi and MERTK^hi M2 TAMs compared with M1 TAMs in both tumor and lung tissues.

In our studies, we have further validated the role of integrins in TAM polarization by performing detailed pro- and antitumorigenic gene profiling in TAMs expressing specific integrins that were sorted from 4T1-induced murine breast tumors. Our data showed that CD11b⁺F4/80⁺α_v⁺ TAMs showed higher M2 markers (Cd206, Cd163) and protumorigenic genes (Tgfb, Arg1), whereas CD11b⁺F4/80⁺α₅⁺ cells demonstrated an increase in inflammatory cytokine genes (Tnfa, Il6). This correlated with higher expression of stat1 and stat2 in α₅⁺ cells and Stat3, Stat6, and Pparg signaling molecules in α_v⁺ cells at transcript levels.

In our functional in vitro assays, blocking of integrins α_v and α₅ disrupted TAM differentiation and polarization, as shown from the lower expression levels of M2 polarization markers such as CD206 and CD163, which again supported the role of integrin α_v and α₅ in different stages of TAM polarization. In support of these findings, we had previously demonstrated that a C-terminal fragment of adhesion protein Fibulin7 (Fbln7-C) binds to monocyte integrin α₅β₁ and negatively regulates monocyte and macrophage migration, differentiation, and polarization via promoting STAT and ERK1/2, signaling. Administration of Fbln7-C also reduced the growth of 4T1-induced murine tumors and delayed the programming of TAMs in murine tumors (20). Similarly, in murine polymicrobial sepsis and LPS-induced endotoxemia models, integrin α₅ and integrin α_v demonstrated a robust correlation with the inflammatory and anti-inflammatory phenotype of macrophages, respectively (19). In this study, while F4/80⁺CD206^hiα_v^hi macrophages showed higher expression of TGFBR1, the F4/80⁺α₅^hi macrophages displayed increased p-SRC expression levels at 12 h after induction (19).

Interaction of surface integrins with ECM proteins present in the TME can play an essential role in cell adhesion and migration, proliferation, survival, and metastasis (16, 51). Integrins recognize multiple matrix ligands, such as fibronectin, vitronectin, collagen, and laminin, as well as cell surface receptors, such as ICAM-1 or VCAM-1 (51, 52). In our study, the expression of fibronectin and vitronectin in tumor tissues was upregulated at the transcript level and blocked integrin–ECM interactions, especially with fibronectin and vitronectin, which inhibited M2 TAM polarization. This may imply that interactions of integrins with the specific ECM proteins in the TME can play an essential role in TAM polarization and reprogramming.

The bioinformatics analysis of RNA sequencing data of BMDCs and TAMs acquired from a GEO dataset (GSE195857) was performed. Protein–protein interaction analysis via STRING revealed that SRC family kinase and FAK are key signaling molecules associated with integrins α₅ and α_v. Studies have shown that SRC and FAK can regulate TAM adhesion, migration, and polarization through downstream PI3K/AKT, Ras/MAPK, or STAT3 pathways (49, 53). FAK phosphorylates several downstream signaling molecules, which influence cellular responses such as migration, survival, and differentiation (54). Our results from the experiments with FAK blocking demonstrate that inhibition of FAK can significantly disrupt the polarization macrophages under tumor microenvironment-like conditions. This was confirmed in TAMs generated in vitro using RAW 264.7 cell lines and human PBMC-derived macrophages. A critical role of FAK in macrophage polarization in the above models was demonstrated by decreased MHCII expression in M1 macrophages and reduced CD206, CD163, and MERTK in M2 macrophages and an impaired expression of p-SRC, STAT1, and STAT6 under FAK blocking condition. Targeting these pathways could modulate TAM behavior and immune response in the TME.

In summary, our data demonstrate the role of integrins α₅ and α_v in regulating macrophage infiltration, differentiation, and polarization in breast cancer models. Understanding these mechanisms could aid in comprehending immunoediting, cancer progression, and developing TAM-based therapies.

The authors have no financial conflicts of interest.

We gratefully acknowledge Dr. Prerna Sharma’s valuable contributions in critically reviewing the manuscript.