Platelets have a crucial role as regulators of inflammation and are involved in various stages in BC development and dissemination (13). Tumor-activated platelets further contribute to cancer progression by promoting critical processes such as angiogenesis and metastasis. Platelets modulate innate immunity (antigen presentation by dendritic cells, monocyte recruitment and differentiation or neutrophil extracellular trap formation) and also promote thrombosis and metastasis (for example with the mechanism of lysophosphatidic acid-dependent (LPA) metastasization or formation of platelet clots) (Figure 1) (14). Moreover, the adaptive immune responses can be modulated also by platelets inducing the differentiation of T-helper 17 cells (13).

Platelet-derived growth factors (PDGFs) contribute to sustaining proliferative signals. Among them we recognize PDGF, transforming growth factor –beta (TGF-beta) and platelet -derived endothelial cell growth factor (PD-ECGF) that are often produced by BC cells and enhance their progression and aggressiveness (15). Platelet-derived extracellular vesicles (PEVs) are also considered potential mediators in the activation of signaling connected to migration in metastatic BC cell lines (16).

Furthermore, the link between hemostasis and BC assumes that platelets have a central role in disease progression (17). In peripheral blood, tumor cell interaction with adherent platelets arrest tumor cells thanks to adhesion proteins and crosslinking plasma protein ligands that support platelets to adhere to the vessel wall. Tumor cells that fail to attach are rapidly cleared from the circulation and undergo apoptosis. To facilitate adhesion to platelets, some cancer cells can upregulate aberrant surface proteins. The binding to platelets helps metastatic cells to arrest within the microvessels of their target organs, where then they extravasate, start to proliferate at the attachment site, or remain dormant for extended periods of time (18). PDGFs support the proliferation and extravasation of invading metastatic cells in the metastatic niche (Figure 1) (19).

Furthermore, BC cells secrete high levels of interleukin-8 (IL-8) in response to platelets that may activate their AKT pathway promoting an invasive capacity. Patients with BC receiving aspirin had lower circulating IL-8, and their platelets did not increase tumor cell invasion compared with patients not receiving aspirin (20, 21).

In BC, elevated platelet-related markers may be associated with poor prognosis. The meta-analysis of 17,079 individuals conducted by Guo et al. confirms that an high PLR is associated with poor overall survival (OS) as well as high risk of recurrence for BC patients (22). However the metanalisis includes few studies about aBC patients, therefore the specific relationship between PLR and aBC need to be better explored.

Neutrophils have a central role in inflammatory response; patients with various cancer types, including BC, often exhibit increased numbers of circulating neutrophils (9).

Neutrophils with an immature phenotype have been observed in the blood stream of cancer patients. Increased levels of tumor-induced granulocyte-colony stimulating factor (G-CSF) and granulocyte–macrophage-colony stimulating factor (GM-CSF) enhance hematopoiesis towards the production of myeloid cells, granulocyte–monocyte progenitors (GMPs) and neutrophil progenitors (23).

Neutrophils seem to be involved in BC progression promoting metastasis-initiating cells that drive cancer spread (24, 25). They can secrete immunosuppressive mediators and angiogenic factors such as reactive oxygen species, vascular endothelial growth factor (VEGF) and matrix metalloproteinase 9 (MMP-9) contributing to a pro-tumor microenvironment (26). Neutrophil-secreted factors alter the heterogeneity of cancer cells, favoring breast metastasis-initiating cells (27). In a BC model, neutrophils induced by tumor cells showed to suppress CD8+ T lymphocytes promoting metastasis through immunosuppression (28). Furthermore, it has been observed that neutrophils may support the metastatic potential of circulating tumor cells (CTCs) in patients with metastatic disease (29).

In BC, the formation of neutrophils extracellular traps (NET), web-like structures formed by DNA and intracellular contents expelled by these cells, has been linked to increased invasiveness and risk of venous thromboembolism (30). The tumor releases pro-inflammatory factors, pro-NETotic factors and extracellular vesicles into the circulation that can activate platelets and the endothelium causing NET release. NETs can capture CTCs, promote the formation of metastases and also the extravasation in the damaged endothelium and generate a highly inflammatory microenvironment for the pre-metastatic niche (Figure 1) (30). NETosis seems to be more frequently produced by morphologically circulating immature neutrophils that express a pro-metastatic behavior, as observed in an in vitro model of BC liver metastasis (31). IL-8 is able to cause neutrophils NETs release and at the same time has an important chemoattractant effect for these cells in the BC microenvironment (32).

NLR is the most widely evaluated inflammatory index. Its elevation is associated with poor prognosis in several cancers and showed to be an independent factor of outcome prediction (9, 33). The prognostic value of the NLR index has been studied in BC (34). A systematic review of fifteen studies analyzing a total of 8563 patients highlighted that a high NLR is associated with a poor OS and DFS in patients with BC especially in triple negative disease and HER-2 positive (HER2+) BC population rather than hormone receptor-positive (HR+) BC patients (35). In a retrospective study, that had the aim to determine the prognostic implications of NLR in the peripheral blood of patients with malignant bone metastasis collected from a prospective cohort, the ratio was significantly associated with tumor type (P<0.0001, included BC) (36).

The combination of NLR/PLR can be considered a more stable marker to changes as compared to single ratios, which may be influenced by concomitant drugs or conditions (e.g. infections or corticosteroids). Combined indexes may reflect also the immune balance and the patients’ immunogenic phenotype as a worse independent prognostic indicator from common prognostic factors such as grading, Ki-67, and molecular subtypes (37). However, conflicting data persist regarding the utility of NLR in predicting prognosis in patients with metastatic disease.

Similarly to NLR and PLR, lymphocyte-to-monocyte ratio (LMR) reflects the imbalance between adaptive and innate immune system in patients with advanced neoplasia and with an inadequate anti-tumor activity (38–40). Lower LMR has been associated with poor survival in BC (41–44). Few studies showed how lymphopenia can be a predictor of poor outcome in aBC patients with increased risk of disease progression and worse long-term survival, assuming a link to a weak anti-tumor response and lower tumor-infiltrating lymphocytes (TILs) (45, 46).

TILs are an emerging tissutal predictive biomarker for BC and their phenotype influences the TME. Infiltration of type 2 (CD4+ T-helper cells or Th2), including Forkhead box P3 (FOXP3) CD4+ regulatory T-cells, inhibits cytotoxic T-cells (CTL) function, supports proliferation and promotes an adaptive anti-inflammatory immune response that is responsible for tumor growth. Especially TNBCs may present a lymphocytic infiltrate >50% and are consequently termed “lymphocyte predominant BCs” (47).

The circulating lymphocyte count and lymphocytes characteristics, especially T-cell receptor diversity, have been investigated, either alone or in combination, as prognostic factors at diagnosis in aBC patients (48). It was observed that the severe restriction of TCR diversity (≤ 33%) was independently associated with shorter OS (48). In addition, the quantitative alteration of lymphocytes in the peripheral blood, and mainly the CD4+ lymphopenia, resulted to be strongly associated with aBC progression (46). In a study from Trédan et al., the cohort of patients with aBC treated in first line showed a median OS of 1.2 months for severe CD4+ lymphopaenic patients, 14.7 months for patients with mild CD4+ lymphopaenia and 24.9 months for non-CD4+ lymphopaenic patients (log rank p-value < 10−4) (46). Importantly, the relative majority of immunosuppressive cytokines (IL-6 and IL-10) and immunosuppressive circulating lymphocytes, like CD8+CD28- suppressor T lymphocytes, in peripheral blood of aBC patients have been associated with a shortened PFS (49).

The clinical benefit of targeting tumor-cell platelets interaction in aBC is still under question. Many studies support the idea of utilizing targeted platelet therapies to inhibit the platelet’s role in the malignancy. Platelets exposed to tamoxifen or ticagrelor release significantly lower amounts of pro-angiogenic VEGF and have less interaction with BC cells (98, 99). However the concomitant use of anti-platelet therapy in cancer patients has a rationale but carries many risks as the declining platelet function and counts as a consequence of disease progression or myelosuppressive effects of treatments.

As previously reported, TANs are particularly involved in tumor progression and studies on new drugs are evaluating therapeutic strategies on several fronts: inhibition of neutrophils recruitment in tumors, depletion of neutrophils in TME, targeting tumor-promoting TAN polarization (100). Targeting neutrophils as a treatment option has been investigated in many preclinical models with discouraging results due to the short life span of these cells (nearly 24 h in blood) (101). Low toxicities strategies to inhibit protumor neutrophils are warranted and expected as promising approaches. Treatments that target the mechanism of interaction between tumor cells and neutrophils are more encouraging.

INF-β and TGF-β are cytokines with a role in switching neutrophils polarization from N1 to N2. In a BC mouse model the blockade of TGF-β increased the percentage of N1 and the activity of CD8+ T cells (102). A phase I trial enrolling patients with solid tumors (including BC patients) has the objective to evaluate the efficacy of a selective and orally active TGF-β receptor 1 inhibitor (NCT03685591). This new TGF-β receptor 1 inhibitor combined with palbociclib in a xenograft BC model led to a significant increase in OS, suggesting the potential for such combination (103).

Neutrophils are considered the major productors of pro-angiogenic factors and the presence of a rich neutrophils infiltrate in TME has been associated with resistance to anti-VEGF therapies (104). Tumors enriched in neutrophils are also more likely resistant to ICIs. Consequently, there are many ongoing studies (phase I/II) evaluating the association between ICIs and new compounds against neutrophils in solid tumors, but results are still awaited (Table 3) (104). Chemokines and interleukins involved in the TAN recruitment (like CXCL1, 2, 5, 6, 8, IL-6, IL17) and their signaling are possible new targets for inhibitory drugs associated with ICIs enhancing their activity (105). The inhibition of enzymes involved in the protumor phenotype as nicotinamide phosphoribosyl transferase (NAMPT) or CXCR2 signaling resulted in an effective reduction of tumor growth and polarization to N1 (106).

Targeting MDSCs may also become a potential strategy to enhance antitumor activity of current treatments (Table 3). Entinostat, a selective HDAC1/3 inhibitor, can decrease the populations of MDSCs and FOXP3+ Tregs in murine models of mammary carcinoma (107). Combination of entinostat with nivolumab and ipilimumab is currently under evaluation in a phase I trial in patients with invasive and metastatic BC (NCT02453620) (Table 1). Entinostat showed promising preclinical and clinical data in HR+ endocrine-resistant BC. G-MDSCs and Mo-MDSCs manifested a reduction (14.67 vs +20.56%; p = 0.03 and -62.3 vs +1.97%; p = 0.002 respectively) in a post hoc analysis of samples from entinostat treated patients in ENCORE301 trial (108). CD40 was also significantly downregulated in the majority of MDSC subsets (109). Other drugs like IPI-549 (eganelisib) transported by liposomes, can inhibit PI3Kγ in MDSCs, resulting in downregulation of arginase 1 (Arg-1) that conduces to MDSCs apoptosis and reduction of their immunosuppressive activity to CD8+ T cells. This strategy synergizes with ICIs and inhibits tumor growth via facilitating the dendritic cell maturation and tumor infiltration of CD8+ T cells while decreasing the tumor infiltration of immunosuppressive regulatory T cells, MDSCs, and M2-like TAMs in solid tumors (81). Some trials targeting PI3Kγ involving BC patients are currently ongoing (Table 3).

Sitravatinib is an oral spectrum-selective tyrosine kinase inhibitor that targets the TYRO3/AXL/MERTK pathways and split the VEGFR2/KIT family receptor tyrosine kinases (RTKs). Inhibition of this pathway may promote the depletion of MDSCs in the TME and at the same time repolarize TAMs towards the M1 phenotype (110). The NCT04123704 trial is evaluating sitravatinib in aBC.

The myeloid lineage in solid tumors can also be targeted by MTL-CEBPA, a novel immunotherapy constituted by a small activating RNA (saRNA) that upregulates C/EBPα, a master regulator of myeloid cell differentiation with anticancer properties. Furthermore, this saRNA restores CEBPA gene transcription, and increases both CEBPA mRNA levels and protein expression at tumor cell level activating the expression of suppressor genes that are downregulated in certain types of cancer. MTL-CEBPA has been evaluated in a phase I trial including BC patients (111). The NCT04105335 trial is now recruiting patients with solid tumors and evaluating MTL-CEBPA in combination with pembrolizumab.

MDSCs are also depleted by other new drugs like ORIN1001 that targets and binds to the RNase domain of the Inositol-requiring enzyme 1 (IRE1) involved in stress adaptation mechanisms in tumoral cells and TME (112).

Macrophage-targeted treatment strategies instead are showing more promising results and are currently being evaluated in many clinical trials. These strategies include: inhibition of macrophage and macrophage precursors recruitment, depletion of TAMs, repolarization of TAMs to an antitumor phenotype, inhibition of tumorigenic factors and mechanisms promoted by TAM and enhancement of macrophage-mediated tumor cell killing or phagocytosis (Figure 2).

The disruption of macrophage recruitment is currently being exploited targeting the CSF1-CSFR and CCL2-CCR2 pathways with specific antibodies (Table 4). The CCL2 blockade showed to sequester monocytes in the bone marrow, instead the inhibition of CSF1 signaling can reduce monocyte development (64, 113). Trials with novel CSF1 inhibitors showed contrasting results. The NCT01596751 trial evaluated the tolerability of PLX3397 (pexidartinib), an anti CSF1, associated with eribulin on 67 aBC patients (phase I part) and then the effect on PFS in a TNBC cohort (phase II part), but results are still awaited. Pexidartinib showed tumor response associated with paclitaxel in the BC patient group of the NCT01525602 trial (114). Conversely, lacnotuzumab, another anti CSF1, when combined to carboplatin-gemcitabine did not show a greater antitumor activity with a worse tolerability profile (115). Emactuzumab, in phase I trials including advanced BC patients, showed a specific reduction of immunosuppressive TAMs, but did not result in clinically relevant antitumor activity (116, 117). In contrast, in the NCT02265536 trial a meaningful stable disease >9 months in two patients with aBC was obtained with LY3022855 (118). However, the interruption of these treatments seems to induce a rebound effect, with abnormal elevated circulating monocytes or accelerated metastases (119).

Other treatment options that can deplete TAMs are potentially constituted by antibodies targeting antigens expressed by TAMs such as the scavenger receptor A, CD52 and folate receptor β (120, 121). However, these targets have not been studied in breast cancer models.

Bisphosphonates are also under evaluation for their capacity to induce apoptosis in monocytic cells (122). They significantly reduce complications of breast cancer bone metastasis by inhibiting resident macrophages or osteoclasts, and recent clinical trials indicate additional anti-metastatic effects outside the bone microenvironment (123). In vitro, bisphosphonates cause increased macrophage death whereas in vivo inhibit the production of pro-angiogenic factors, such as MMP-9, other evidence suggests a shifting in TAMs to a pro-tumoricidal phenotype (122).

TAMs reprogramming to a M1 phenotype can be achieved by stimulating the costimulatory receptor CD40, complement receptor 3 (CR3), administrating Toll-like receptor 7 (TLR7) and 8 agonists, inhibiting IL10 or delivering IL-12 (124). For example imiquimod, a TLR agonist, can induce the production of proinflammatory cytokines by macrophages, therefore restoring the ability to attack BC cells (125). The topical application of imiquimod showed to reduce skin metastasis in aBC patients in association to nab-paclitaxel, but responses were fleeting (126). The subcutaneous administration of a TLR agonist has been experimented in a phase II trial involving also heavily pretreated aBC patients, showing modest results and a considerable risk of cardiac toxicity (127).

Other strategies that stimulate TAMs include agonistic anti-CD40 or inhibitory anti-CD47 antibodies. The co-stimulatory receptor CD40 is expressed on macrophages and usually binds the CD154 on T cells. The agonist action of specific antibodies can reverse immune suppression and drive antitumor T cell responses (128). A first-in-human study completed in 2017 showed that the injection of a CD40 agonist antibody into superficial lesions was well tolerated and associated with pharmacodynamic responses (129). Selicrelumab, a fully human CD40 agonist, is being experimented in some phase I trials including BC patients, both alone and in association to other drugs as vanacizumab, a bispecific antibody directed to Angiopoietin 2 (Ang2) and VEGF-A (NCT02665416). Other studies are evaluating the effect of this agonist especially in TNBC, for example in association to a FMS-like tyrosine kinase 3 ligand (FLT3) inhibitor or classical chemo/immunotherapy treatments (Table 4). New biotechnologies targeting CD40 are also in study, such as oncolytic adenoviral vectors or designed ankyrin repeat proteins (DARPins), which are genetically engineered antibody mimetic proteins typically exhibiting a highly specific and high-affinity target protein binding. MP0317 is a DARPin intravenously administered drug targeting fibroblast activation protein (FAP) and CD40 that is currently being evaluated in a phase I trial including also aBC patients (NCT05098405).

Regarding CD47, blocking the interaction between it and the signal-regulatory protein alpha (SIRP-alfa), a “don’t eat me” signal can re-activate the phagocytic activity of TAMs (130). Many tumors overexpress CD47, enabling immune escape from the innate immune system such as macrophages binding SIRPα and compromising the antigen presentation and T cell infiltration (131). Targeting CD47 can also enhance the anti-tumor effect of other therapeutic strategies. The combination of anti-CD47 with trastuzumab, significantly suppressed the growth of antibody-dependent cellular cytotoxicity (ADCC)-tolerant HER2+ BC via Fc-dependent antibody-dependent cellular phagocytosis (ADCP) (132). The ASPEN-01 open-label, multicentre, phase I dose-escalation and dose-expansion study, evaluated the association of evorpacept (an anti-CD47) plus either intravenous pembrolizumab or trastuzumab. The safety findings support the use of evorpacept and preliminary data on the antitumor activity suggest future investigation. However, this trial included only one patient with aBC (133). Many other phase I and II trials are ongoing, evaluating the CD47 blocking effect on solid tumors also in aBC patients (Table 4). The tolerability of anti-CD47 has been successfully evaluated in a trial including five aBC patients (134).

The CD47-SIRPα axis may also be targeted using the SIRPα factor. TTI-621 (SIRPαFc) is a soluble recombinant fusion protein that acts by binding human CD47 evaluated in the. NCT02663518 trial in various solid tumors.

The macrophage-1 antigen, also called CR3, is a complement receptor consisting of CD11b (integrin αM) and CD18 (integrin β2). CR3 is a pattern recognition receptor, capable of recognizing and binding to many molecules found on the surfaces of foreign cells enabling phagocytosis (135). The clinical trial NCT02981303 have evaluated the capacity of a new molecule constituted by a pathogen-associated molecular pattern (PAMP) to enhance innate immune cell killing and the maturation of antigen presenting cells when combined to ICIs in TNBC, meeting both safety and efficacy requirements (136). CR3 is the principal β2 integrin known to contribute to PAMPs recognition (137).

There is also evidence that targeting the Ang2-TIE2 may inhibit the functions of TIE2-expressing macrophages, a TAM subset endowed with proangiogenic activity in mouse tumor models (138). Ang2 is a ligand of the TIE2 receptor and modulates endothelial cell biology facilitating angiogenesis. Ang2 inhibition, by monoclonal antibodies, peptibodies, or CovX-Bodies, may determine antitumor responses and also inhibit the functions of TAMs (139). In BC the expression of Ang2 is correlated to more aggressiveness. The intravasation occurs in sites where a TIE2-expressing macrophage and an endothelial cell are in direct contact. Ablation of the activity of these macrophages blocks intravasation after Ang2-TIE2 axis inhibition (140).

Trebananib, a peptibody that inhibits the binding of angiopoietin 1 and 2 to TIE2 showed potential anticancer effect in a phase Ib and phase II studies, with manageable AEs (141, 142). Vanucizumab is another novel bispecific antibody inhibiting VEGF-A and Ang2 that demonstrated safety and anti-tumor activity in a phase I study of 42 patients with advanced solid tumors (143). Other two trials have evaluated the tolerability of nesvacumab (an antiAng2 antibody) in advanced neoplasms (NCT01688960, NCT01271972) showing a preliminary antitumor activity (144). Rebastinib, instead, is a TIE2 inhibitor that blocks the assembly of macrophages and endothelial cells involved in metastasization at the peripheral site (tumor microenvironment of metastasis) (145). The NCT02824575 trial hypothesizes that rebastinib combined with antitubulin therapy could improve clinical outcomes in BC by preventing intravasation. The nonsteroidal anti-inflammatory drug celecoxib showed an interesting activity in BC increasing the presence of M1 like macrophages, but the real effect is in doubt (146, 147). In this perspective, the NCT00075673 trial has evaluated the weekly administration of oral vinorelbine in combination with celecoxib in aBC. Results are still awaited.

In view of the central role of innate and adaptive immune systems in cancer development, immune cells are not only considered potential therapeutic targets, but also innovative vehicles for treatments. The genetic engineering to deliver, correct or enhance immune cells demonstrated to be successful. Chimeric antigen receptor (CAR) T-cell treatment has provided notable results in hematological tumors (148). Unfortunately, the same evidence has not been demonstrated in solid tumor, where T-cells encounter substantial difficulties in penetrating and surviving in the TME (149, 150). The extracellular matrix (ECM) is one of the major parts of TME, and it is a physical barrier to various kinds of anticancer therapies. MMPs can degrade almost all ECM components, and macrophages are an important source of MMPs (151). Amongst the cell types used in engineered cell immunotherapies, macrophages have recently emerged as prominent candidates for the treatment of solid tumors, including BC (Figure 3) ( 152). In a preclinical study, macrophages engineered with specific CARs (CAR-M), activated after the detection of the HER2 antigen on tumor cell surface (153). The activation of these engineered macrophages triggered by the internal signaling of CD147 determines the production of MMPs. The infusion of CAR-147 macrophages reduces the tumor collagen deposition (153). The initial in vitro tests failed to show strong antitumor activity of the CAR-147, however infusion of CAR-147 cells into the aggressive HER2-4T1 bearing mouse model showed significant tumor growth inhibition (153).

One first in human phase I trial is currently active and recruiting patients with HER2+ advanced solid tumors experimenting with engineered CAR-M (NCT04660929). Another active protocol has the objective to collect tumor samples to develop patients’ derived organoids from HER2-, HER2-low and HER2+ BCs to test the antitumor activity of newly developed CAR-M (NCT05007379).

In vitro studies showed that CAR-M infused in tumor models increased intratumoral T-cell infiltration, NK cell infiltration, dendritic cell infiltration/activation, and TILs activation and at the same time can reduce tumor growth (152).

The combination of CAR-M treatment with other anti-tumor therapies such as CAR-T cells, ICIs and chemotherapy may synergize and provide an optimal tumor control. Nevertheless, toxicity remains an important concern and further optimization of CAR products is required (150).

Finally, preliminary outcomes about the use of mesenchimal stem cells (MSCs) in BC are interesting. Genetically modified MSCs with the insertion of tumor suppressor genes, proapoptotic genes, immune involved genes can inhibit cancer cell growth. Moreover modified MSCs delivering anticancer agents into tumor tissue have been studied in several cancer types, results in BC are awaited (154).