During carcinogenesis, concomitant with the mitogenic-promoting role through ERs interactions of ER+ breast cancer cells, immunoediting occurs. Immune editing is usually reported to be a process by which tumour antigenicity changes due to the selective pressure of effector immune cells; although immunoediting may precede immune-evasion, they can be independent each other (20–22). The principal immune-evasive mechanisms include a down-regulation of antigen presentation, lack of immune effector and/or an increase in immune suppressor cells, as well as an up-regulation of checkpoint molecules (23). It is well known that tumour cells can be recognised by T cells if tumour-associated antigens (TAAs) are joined with human leukocyte antigen (HLA) molecules expressed on the surface of tumour cells or antigen-presenting cells. In about 30–40% of higher-grade breast cancers, classical HLA-A, HLA-B and HLA-C molecules, which are necessary to activate CD8+ T cells, are down-regulated; instead, the non-classical HLA-E, HLA-F and HLA-G molecules favouring the immune escape are upregulated. In one study, HLA class I down-regulation occurred in 64% of 141 ER+ breast cancer patients (24). In another study, HLA-F expression was found in 50% of 56 ER+ breast cancers (25). In another research, breast cancer biopsies from 52 patients with invasive ductal carcinoma and unselected for molecular sub-types suggested that the down-regulation of HLA-Ia, HLA-E and HLA-DR and the up-regulation of HLA-G and HLA-DQ by different processes were likely responsible for immune evasion and breast cancer aggressiveness (26). In an experimental study, MCF7 breast cancer cell lines carrying miR-18a hyper-expression and cultured with the THP-1 cell line, a human monocytic cell line derived from acute monocytic leukaemia, showed lower antigen presentation capability, higher differentiation of pro-tumorigenic M2 macrophages and increased invasiveness. In these cells, TAP -1, a crucial protein for antigen presentation, was again expressed following the inhibition of the Wnt pathway. Accordingly, in miR-18a high ER+ tumours, even though there was a dense lymphocyte infiltrate, a higher CD4+/CD8+ ratio and the M2 macrophage marker CD206 contemporaneously with the invasive marker MMP9 also occurred (27). A decreased rate of TILs has been reported in ER+ breast cancer. This may be due to the ER expression, which has been reported to both promote a Th2 immune environment and reduce MHC class II molecules in breast cancer cells (28). Besides HLA-A, antigen peptide transporter 1 and 2 protein expression (TAP1/TAP2), which are members of the ATP-binding cassette transporter family, are also down-regulated in high-grade breast cancer. This was found in 53 patients with breast carcinomas unselected for molecular subtypes (29) and in about 40% of metastatic lesions (30). Molecules involved in antigen presentation and interferon (IFN) response gene mutations account for another mechanism of immune evasion. For example, resistance to checkpoint blockade (31, 32) and other immune therapies can be due to mutations in beta2-microglobulin (B2 M), an element of MHC class I, and JAK1/2 kinases, downstream of IFN receptors. In particular, JAK2 and STAT3 were found to be significantly mutated in the metastasis/relapse screen, and all mutations in a cohort of 163 patients arose in ER+ breast cancers (33). CD16+ lymphocytes inversely correlated with the proportion of regulatory CD4+CD25+CD127- cells and the Ki-67 rate in tumour cells. Accordingly, a higher level of Ki-67 occurred concomitantly with decreased effector lymphocytes (CD8+ and CD16+) and a high percentage of regulatory CD8+CD11b-CD28- T cells (34). In addition, oncogenic pathway alterations are likely to play a crucial role in T cell cutting out or the inhibition of T cell activity (21). In particular, in breast cancer, PI3K pathway driver mutations involve 49% of luminal A and 7% of basal-like molecular subtypes. These driver alterations could contribute to immune-evasion or an immunosuppressive microenvironment (35), as well as to heterogeneity with respect to TILs in the breast cancer subtypes. Tumour cells can constitutively express indoleamine-pyrrole 2,3 dioxygenase (IDO) or up-regulate it in response to interferon-gamma and oestrogen receptor signalling, and IDO over-expression results in diminished local anti-tumour response (36–38). IDO has been established as a normal mechanism of peripheral tolerance and immuno-suppression, and high IDO levels have been found in ER+ breast cancer compared to ER- tumours. In ER+ tumours, it was higher in more advanced stages (39, 40). In an experimental mouse model of luminal breast cancer, CCL5 also named RANTES (regulated on activation, normal T cell expressed and secreted) expression directly correlated with tumour progression. Moreover, with a deeper analysis, it was uncovered that high tumour CCL5 levels induced the polarisation of CD4+ T cells toward a pro-tumorigenic Th2 phenotype (17). Taken together, these data suggest that in hormone receptor positive breast cancer, the interaction between tumour cells and the immune microenvironment occurs differently than in other breast cancer subtypes, involving the intercommunication of endocrine factors with pro-inflammatory status and immune cells modulated by the TME (17).

Cellular components of the TME usually include tumour-associated macrophages (TAMs), mesenchymal stem cells (MSCs) and/or cancer-associated fibroblasts (CAFs), endothelial cells, immune cells such as T and B cells, natural killer (NK) cells and myeloid-derived suppressor cells (MDSCs) (41).

Chronic inflammation is a well-recognised ancillary mechanism of cancer progression that joins common breast cancer risk factors such as menopause and obesity. The inflammatory TME milieu promotes a more aggressive, endocrine-resistant ER+ breast cancer phenotype (16, 17). IL-6, TGF-beta, TNF-alpha, NF-kB, COX-2 and PGE-2 are pro-inflammatory cytokines present in the TME of ER+ breast cancer that are capable to activate pro-tumoral pathways mediating proliferation, immune evasion and metastasis (16, 84). The principal pro-tumorigenic actions of IL-6 and TGF-beta cytokines have been just mentioned and reported (see Table 1A ). TNF-alpha, which is a ubiquitous TME cytokine secreted by tumour cells and macrophages, may promote metastasis of breast and other cancers. In an experimental study, TNF-alpha enhanced the invasive capacity of MCF-7 cells through an increased expression of metastasis-related genes (85). TNF-alpha may also up-regulate aromatase expression in stromal cells and may play a principal role in governing oestrogen biosynthesis in adipose tissue (86). TNF-alpha strongly induces E2 increased levels in TME by maintaining the surrounding fibroblast in an undifferentiated state. In turn, oestrogens themselves up-regulate transcription and secreted levels of TNF-alpha in ER+ breast cancer cells, therefore creating an autocrine-positive feedback loop (87). TNF-alpha also mediates cancer-related inflammation through the activation of NF-kB (17). The role of NF-kB signalling in tumour initiation and inflammation is well established. Constitutive activation of NF-kB has been shown in various types of cancer; NF-kB leads to a release of adhesion molecules and immune-regulatory cytokines (such as TNF-alpha, IL-1, IL-6 and IL-8), which converge leukocytes to sites of inflammation (88). Significant correlations between ER-alpha, TNF-alpha and NF-kB expression have been shown in breast cancer tissues (89). A study showed that pre-metastatic niche formation in the lungs was promoted by PGE2 produced by fibroblasts through dysfunctional dendritic cells (DCs) and suppressive monocytes. This process was propagated by tumour inflammation, mainly by interleukin-1beta. Ablation of the Ptgs2 gene (encoding COX-2) in fibroblasts reversed the immune-suppressive phenotypes of lung myeloid cells in multiple breast cancer models (90). Moreover, it was reported that obesity-related breast inflammation is linked with high aromatase activity, likely through the promotion of CYP 19 transcription by an increase in cyclooxygenase 2 (COX2) derived prostaglandin-E2 (PGE2) (17). Overall, ER and aromatase expression in stromal and immune TME point out an important immune-modulatory role of ER signalling in cancer biology.

It is well known that oestrogens favour the development and progression of some various cancers, including breast cancer (91). For an effective anti-cancer immune response, a series of stepwise events are involved. These steps have been defined as the cancer-immunity cycle (92, 93). In cancer patients, the cancer-immunity cycle is impaired, and many mechanisms facilitate immune evasion. Among them, T cells homing and infiltration of the tumour may be compromised, and/or more commonly, the effector immune cells might be suppressed by factors produced in the TME (94). It has been shown that in many cancer patients, immune system deficiency may be due to the presence of negative regulators of T cell responses (checkpoints) in lymph nodes and tumour tissues (immunostat function) (95). In particular, factors in the TME, including PD-L1 and PD-1 molecules, can regulate the response of activated anti-tumour T cells, acting as an immune rheostat or ‘immunostat’ (96, 97). This highlights that the cancer immune response depends on many carefully governed events; therefore, it may be addressed, at best, only if considered not as a simple but as a very complex process. Most functions of immune escape are E2/ERs induced by non-cancerous cells in the TME. It has been established that non-cancerous cells in the TME are involved in tumour progression, and the interaction among cancer cells, immune and stromal cells and extracellular mediators within the TME may influence the immune response to tumour and immunotherapy (21, 24, 98, 99). In particular, oestrogens may favour an immunosuppressive TME by enhancing Th2 responses, tumour-promoting cytokines (IL-4, IL-6, IL-17A and TNF) and M2 TAM infiltrations, unlike M1 TAM infiltration and Th1 responses, which are commonly joined with Th1 cytokines (IL-12 and IFN) ( Figure 1A ). Oestrogens induce the expansion of Tregs and MDSCs (57, 100, 101) as well as inhibitory activity of the antigen-presenting cells (APCs) (102). Particularly, it has been reported that when APCs are exposed to Tregs, their antigen-presenting function is down-regulated and the expression of immunosuppressive molecules and cytokines secretion are increased (103). In vitro, after E2 treatment, PD-1 ligand (PD-L1) up-regulation on ER+ endometrial and breast cancer cells through activation of the PI3K pathway has been observed (104).

Within tumours, T cell activation commonly up-regulates PD-1 expression, and PD1+ T cells are usually confined to tertiary lymphoid structures (TLS), which take part of the tumour stroma and are made by B and T cells (105). Cytotoxic T cell exhaustion and consequent immune escape derive from interactions between PD-L1+ tumour cells and PD-1+ T cells (106). Additionally, CAFs may favour a pro-tumour environment by paracrine secretion of E2 and IL-6. Oestrogens, through non-cancerous cells comprising the TME mediated actions and promoting angiogenesis and cytokines release in the TME, further promote tumour growth, combined with immune evasion. In conclusion, the evidence that E2 largely regulates the immune TME suggests an investigation into the reversal of tumour immune evasion by addressing TME as another beneficial target of anti-oestrogen therapy.

The principal mechanisms of tumour progression and immune suppression promoted by chronic inflammation and increased oestrogens in TME are summarised in Table 2 and Figure 1 . In particular, in the absence of anti-oestrogens, the E2 pathway prevails, and an immunosuppressive TME is favoured. Th2 responses and M2 TAM infiltration with the associated suppressive cytokines (IL-1, IL-4, IL-6, IL-8, IL-10, IL-17A) and suppressive mediators (TGF-beta, TNF-alpha, NFkB, COX2, PGE2) overcome Th1 responses and M1 TAM infiltration. E2 further contributes to immune suppression through the proliferation of Tregs, expansion of MDSCs, induced inhibitory activity of APCs, increased tumour cell PD-L1 expression and the inhibition of CD8+ T cell and NK cell-induced apoptosis ( Figure 1A ). When the immunosuppressive E2 actions are counteracted by anti-oestrogens (also see Table 1B ), the immune response increases while the concomitant immune suppression decreases; thus, in ER+ breast cancer in clinical benefit during anti-oestrogens, the common immunosuppressive TME can be reverted ( Figure 1B ). Finally, the addition of the INF-beta-IL-2 sequence boosts the effector immune response, increasing the immune stimulation that shifts the balance by further decreasing the immune suppression ( Figure 1C ).

Over the past three decades selective oestrogen receptor modulators (SERMs) or down-regulators (SERDs) and AIs, that inhibit E2 signalling, have strongly decreased breast cancer mortality (130). Welboren et al. (131), confirming other previous findings by Frasor et al. (132, 133), reported that tamoxifen antagonised most of the E2-up-regulated genes at the same time possessing agonistic behaviour on E2-downregulated genes (134). Moreover, many other experimental studies have reported that mechanisms promoting at the cancer cell level, tumour growth often join with the inhibition of the immune response (135). Based on this, it can be inferred that anti-oestrogen therapy has the potential not only to inhibit tumour growth and progression through the anti-proliferative action mediated by ERs via the ‘genomic’ and ‘non-genomic’ pathways (136), but also to favour the immune response in E2-sensitive tumours by the reversion of an immunosuppressive TME (137) ( Figure 1B ).

In one of our first reports on the proposed hormone immunotherapy in ER+ metastatic breast cancer patients (116), the clinical outcome was correlated with immunological data. It was found that ‘in patients with clinical benefit, eosinophils, total lymphocytes, CD4+, CD8+ and CD16 + 56+ cells significantly increased after interleukin-2 administration (from p<0.012 to p<0.000). In the patients with progressive disease only a slight increase in eosinophils occurred (p=0.038)’ (116). An update of laboratory results during hormone immunotherapy was the subject of two further papers. In the former (115), it was reported that ‘in clinical benefit interleukin-2 administration was followed by a significant increase in total lymphocytes, CD4+, CD8+, CD16 + 56+ (NK) cells, IL-6, IL-12, and C-reactive protein (CRP) (from p<0.04 to p<0.000) but no change in IL-10 and TGF-beta1 was observed. During progressive disease, no change was observed in the former parameters, concomitant with a significant increase in IL-10 (p<0.020) and a significant decrease in TGF-beta1 (p<0.023)’. In the latter (117), with additional data, it was found that ‘during clinical benefit as opposed to progression a significant increase in the total number of lymphocytes, CD4+, CD8+, CD16 + 56+ (NK) cells, CRP and IL-12 was confirmed but not IL-6. At the progression, both basally and after interleukin-2 stimulation the mean values of CD4+CD25+ were more than two-fold higher than during clinical benefit with a decrease of CD4+ plus CD8+ (T effector) CD4+CD25+ (Treg) ratio’. The G0-G1 state has been found as a condition favouring a successful immune manipulation and has been widely discussed by us in recent reviews (134, 138). Furthermore, the primary mechanisms of immune suppression currently acknowledged have been presented in details in Tables 1A , B , 2 . These immune suppression mechanisms, mediated by oestrogens and ERs in ER+ breast cancer cells as well as other cells in the TME, align with our earlier hypothesised model. This is schematically represented here and is likely down-regulated in patients experiencing clinical benefits during anti-oestrogen treatment ( Figures 1B , 2B, C , paracrine loops). This reversion of immune suppression by anti-oestrogens is documented in our study by lower levels of CD4+CD25+ T cells as well as higher IL-12 values during clinical benefit than at the progression. This likely allowed the beta IFN-IL-2 sequence to significantly boost the effector immune cells of the innate and adaptive immune response ( Figures 1C , 2 ) as demonstrated by the highly significant increase in total lymphocytes, CD8+, CD4+ T cells and NK cells reported during the clinical benefit of the studied breast cancer patients (115, 117). More in particular, oestrogens coming from peripheral circulation and secreted in TME by CAFs and TAMs promote tumour growth, invasion, immune suppression and angiogenesis through genomic and non-genomic pathways and the other reported mechanisms in ER+ breast cancer cells; CAFs, TAMs, immune cells and APCs further contribute with production and secretion of inflammatory cytokines (IL-1, IL-4, IL-6, IL-8, IL17A, TGF-beta, TNF-alpha, NFkB, COX2 and PGE2), suppressive mediators (ROS, IDO, NO, IL-10), as well as other described mechanisms ( Figure 2A ). In breast cancer patients in clinical benefit during anti-oestrogens, the induced G0-G1 state likely down-regulates the ER-alpha mediated actions and the tumour burden remains stable or decreases ( Figure 2B ). The administration of IFN-beta-IL-2 sequence in association with anti-oestrogens boosts the immune response by stimulating proliferation and activation of the effector immune cells; as a consequence, hormone resistance is delayed, and again, tumour burden remains stable or decreases ( Figure 2C ).

As expected, the benefit in terms of median PFS and OS time positively correlated with the rate of ER+ breast cancer cells and the type of response to hormone therapy; in fact, it was significantly better in those with higher ER+ rate and more in responding patients than in those in stable disease during hormone-immunotherapy (117, 138). It has been reported that immunity in cancer TME, also termed ‘immunostat’, is tissue-specific; and its regulation as tumour growth is multi-factorial in origin (139). However, in metastatic setting, most target therapies, at best, affect one or a few pathological molecular pathways. This can account for the relatively early development of resistance and the involvement of non-cancerous cells for the concomitant side effects. From this point of view, anti-oestrogens in metastatic endocrine-dependent breast cancer represent a unique model of a single drug addressing multiple pathological immunological and tumour-growth- promoting targets. This likely explains the relatively prolonged clinical benefit without relevant side effects. Our findings suggest that this peculiarity additionally can significantly be improved by a concomitant immune stimulation with IFN-beta-IL2 sequence.

Recently, the CCND1-CDK4/6-RB molecular pathway has been investigated as a useful target in increasing the clinical benefit of ER+, HER2- breast cancer patients on first-line hormonal therapy; this constitutive pathway controls and governs whether a cell move on or arrests at the G1-S phase (140, 141). In these ER+, HER2- metastatic breast cancer patients, CDK 4/6 inhibitors, namely ribociclib (121), palbociclib (122, 123) and more recently abemaciclib (124), due to significant increase of PFS in randomised clinical trials compared to antioestrogens alone, after FDA approval, have entered into clinical practice. Table 3 summarises the results reported in these main clinical trials. Median PFS ranged from 25.3 months with ribociclib (121, 125) to 28.2 months with abemaciclib (124). Median OS has not yet been reached in the clinical trial with abemaciclib, while in a more recent evaluation of ribociclib (121, 125) and palbociclib (122, 123) trials, no significant difference was found in the treated patients vs. controls for palbociclib (53.9 vs. 51.2 months), while it occurred for ribociclib (63.9 vs. 51.4 months; P = 0.008). However, in all these trials, grades 3–4 AEs have been reported in > 10% of the patients receiving CDK inhibitors. In our observational 2:1 controls-case study, median PFS and OS were 33 and 81 months, respectively, without any relevant AEs. Overall, these findings suggest that in ER+ metastatic breast cancer patients in clinical benefit, during the first-line salvage hormone therapy, a concomitant stimulation of the effector immune cells is a valid choice to hormone therapy alone. The recently reported reversion of the immune suppressive TME by anti-oestrogens (137) is consistent with our initial hypothesis. In this favourable condition, the increase of the immune response by the INF-beta-IL-2 sequence is a therapeutic intervention involving a physiological mechanism that likely accounts for no important concomitant AEs (18). Differently, the inhibition of the G1-S checkpoint by the CDK 4/6 inhibitors likely involves tumoral and non-tumoral cells with a common occurrence of relevant AEs. Moreover, notably, the cost per patient per year of the proposed immunotherapy with INF-beta-IL-2 sequence is about 8 to 18 times cheaper than that with the CDK 4/6 inhibitors. However, the last but not the least important issue to be taken into careful consideration is that the different rationale suggests a synergic activity of our proposed immunotherapy with antioestrogens and CDK 4/6 inhibitors.