Over the past recent years, it has become increasingly evident that tumor-infiltrating lymphocytes (TIL) can control the clinical progression of epithelial cancers [5]. In breast cancer, two large series in newly diagnosed or early-stage breast cancer reported a significant correlation between increased lymphocytic infiltration and better clinical outcomes. Denkert et al. [6] analyzed tumor-infiltrating lymphocytes (TILs: using hematoxylin/eosin sections and immunohistochemistry [IHC]) of pretherapeutic core biopsies from over 1,000 breast cancer patients. Mahmoud et al. [7] analyzed tumor-infiltrating CD8⁺ T cells (by IHC) in 1,334 breast cancer patients. In both studies, high lymphocytic infiltrates or CD8⁺ T cell counts were associated with improved patient outcome, independent of standard prognostic and predictive factors. In 2009 high risk, newly diagnosed breast cancer patients, TILs were also evaluated using the same method as Denkert et al. but this time on full-face H&E sections taken from surgical removal of the primary tumor [8]. In this study it was found that the prognostic benefit of TILs was largely restricted to the triple-negative breast cancer subtype (TNBC: negative for expression of estrogen receptor [ER], progesterone receptor [PR], and Human epidermal growth factor receptor 2 [HER2]). While TNBC patients usually have a worse, women with TNBC and high TILs had a numerically better disease-free outcome than the women with ER⁺/HER2⁻ negative tumors who usually have the best prognoses. These data suggest that some patients have preexisting antitumor immunity at diagnosis and this is associated with improved long-term clinical outcomes. The reasons why some breast cancers are associated with TILs and others are not are unclear.

Also in support of a role for tumor-infiltrating CD8⁺ T cells in controlling breast cancer progression, unsupervised gene expression profiling of breast cancer-associated stroma has revealed an immune gene signature enriched for CD8⁺ T cell genes associated with good prognosis [9]. Interestingly, Mahmoud et al. [7] reported a positive correlation between CD8⁺ T cells and high histological grade, younger patient age, and negative estrogen receptor (ER) status. Two other studies found a similar correlation between intratumoral CD8⁺ T cell counts and negative ER status [10, 11]. The underlying mechanisms for this observation are unclear (and discussed below).

In contrast to intratumoral CD8⁺ T cells, analysis of CD4⁺ T cells on clinical outcomes in breast cancer has resulted in inconsistent results. In general, the presence of CD4⁺ T cells in tumors has been associated with worse prognosis. For instance, IHC analysis of tissue microarrays derived from 179 treatment-naïve breast tumors revealed that high levels of CD4⁺ T cells and macrophages were correlated with reduced overall survival (OS). In contrast, high levels of CD8⁺ T cells combined with low levels of CD4⁺ T cells were correlated with improved OS [12].

The association between CD4⁺ T cell infiltrates and poor prognosis is generally attributed to CD4⁺ FOXP3⁺ T regulatory cells (Tregs). Tregs were first described as a population of T cells capable of suppressing immune responses in mice and were originally defined by the surface markers CD4 and CD25 [13]. Further investigation revealed that Tregs express and functionally depend on the transcription factor FOXP3 [14–16]. Given its essential role in Treg development, FOXP3 has become a popular marker of Tregs in cancer studies. Intriguingly, studies of the prognostic value of FOXP3⁺ T cells in cancer patients have led to discrepant findings. Tissue microarray analysis of FOXP3⁺ cells in 1,445 cases of primary invasive breast cancer revealed that the number of FOXP3⁺ cells was associated with a worse outcome, but was not an independent prognostic factor in multivariate analysis [17]. On the other hand, Liu et al. [18] reported that intratumoral FOXP3⁺ cells was an independent predictor of poor prognosis in 1,270 breast cancer cases. Other studies have assessed the impact of FOXP3⁺ cells on neoadjuvant breast cancer chemotherapy. One study reported that high infiltration of FOXP3⁺ cells after neoadjuvant chemotherapy was an unfavorable and independent predictor of relapse-free survival (RFS) and overall survival (OS) [19]. Likewise, Ladoire et al. [20] reported that a high CD8/FOXP3 ratio was an independent predictive factor of RFS and OS in HER2-overexpressing breast cancer following neoadjuvant chemotherapy. In contrast, Oda et al. [21] reported that a high level of FOXP3⁺ cells before neoadjuvant chemotherapy was an independent predictor of pathological complete responses. Others have also reported a positive correlation between high levels of FOXP3⁺ T cells and good prognosis [22, 23]. Taken together, the role of FOXP3⁺ cells in breast tumors still remains unclear and warrants further investigation.

Interestingly, a recent publication reported a favorable role for CD4⁺ T cells [24]. For the first time, CD4⁺ follicular helper T (Tfh) cells were found in tumor-infiltrating T cell populations, and their presence was associated with better prognosis. Accordingly, human breast tumors with high TILs were found to have higher frequency of CXCL13-producing Tfh cells. Notably, Tfh cells were found associated with tertiary lymphoid structures. Tertiary lymphoid structures (TLS) had been previously identified in lung and colorectal cancers, and their presence linked with better prognosis. It is now clear that Tfh cells are an important constituent of TLS in human breast tumors.

Intratumoral B cells have been associated with a favorable prognosis in breast cancer patients. An early gene expression study of 200 consecutive lymph node-negative cases reported a B cell metagene primarily formed by immunoglobulin heavy- and light-chain genes that was associated with metastasis-free survival in highly proliferating breast tumors [25]. Interestingly, immunoglobulin κ C (IGKC) as a single marker has also been shown to have similar predictive and prognostic value compared to the entire B cell metagene [26]. Building up of these findings, Mahmoud et al. [27] recently analyzed the density and localization of B cells in 1,470 breast cancer cases using IHC. Consistent with previous studies, increased B cell infiltration was associated with improved survival. Notably, B cell infiltration correlated with hormone receptor negativity and basal phenotype.

Macrophage infiltrates in breast cancer often correlate with a worse clinical outcome. Tumor-associated macrophages have been associated with suppressed antitumor CD8⁺ T cells, increased angiogenesis, and increased metastasis. The presence of CD68⁺ macrophages is generally inversely correlated with CD8⁺ T infiltrates in human breast tumors [12]. Interestingly, tumor-associated macrophages were found significantly increased after chemotherapy with paclitaxel and in patients who had received neoadjuvant chemotherapy compared to surgery alone. Using transgenic mouse models of breast cancer, Coussens and colleagues also demonstrated that tumor-associated macrophages can directly promote breast cancer metastasis, via IL-4 producing CD4⁺ T cells [28]. Other myeloid cells composed of incompletely differentiated cells, termed myeloid-derived suppressor cells (MDSC), have also been shown to enhance breast tumor growth in mice [29]. Human equivalents of MDSC have been identified [30]. Recently, Sceneay et al. [31] demonstrated that tumor hypoxia, via MDSC recruitment, can alter the lung microenvironment to make them more permissive for metastasis, thereby driving the formation of a pre-metastatic niche. Furthermore, increase in circulating MDSC in human cancer has been associated with stage 4 disease [32].

The correlation between lymphocytic infiltrates and clinical outcomes in breast cancer varies across molecular subtypes. In 2011, DeNardo et al. [12] performed a meta-analysis of CD68 and CD8 gene expression in 4,000 breast cancer cases and reported that a CD68high/CD8low immune gene signature correlated with reduced OS for basal or HER2⁺ breast cancer subtypes, but not for luminal breast cancers. Similarly, a metagene of STAT1 signaling – a surrogate of interferon response – was associated with better outcome in triple-negative breast cancer (TNBC) and HER2⁺ breast cancers, but not in luminal cases [33, 34]. Another independent group identified an immune response prognostic gene module in ER⁻, but not ER⁺ breast cancers [11].

Among ER⁻ breast cancers, accumulating evidence suggests that basal or TNBC breast cancers are particularly associated with lymphocytic infiltration. In 2012, Liu et al. [35] performed a large-scale IHC analysis of CD8 expression on 3,400 breast cancer cases representing different subtypes and reported that only core basal TNBC demonstrated a significant correlation between intratumoral CD8 staining and favorable prognosis.

A B cell metagene has also been associated with good outcome in TNBC. In a gene expression analysis of 579 TNBC, Rody et al. [36] revealed that a ratio of high B cell and low IL-8 metagenes was identified in 32 % of TNBC patients with good prognosis. Taken together, these studies suggest that clinical outcomes in ER⁻ breast cancers, especially TNBC, are particularly influenced by tumor immune responses.

Why immune infiltration appears more relevant for breast cancers which are ER negative or overexpress HER2 as opposed to other subtypes is unknown. We speculate that it could be due to the poorly differentiated nature and high genomic instability of these subtypes. In TNBC, the intrinsic nature of the tumor cells may help explain its propensity for inflammatory responses. For instance, signaling pathways able to downregulate ER and HER2 expression may be associated with increased pro-inflammatory activity. This is supported by the identification of lactoferrin as a repressor of hormone receptors and HER2 expression in breast cancer cells [37]. Lactoferrin concomitantly induces the production of cytokines and chemokines, including MIP-1, MIP-2, IL-10, and IL-4. Increased immune infiltration in TNBC may also be associated with increased genetic instability due to p53 loss of function and BRCA1/2 disruption, which are common features of this molecular subtype. In line with this model, a recent study in high-grade serous ovarian cancer revealed an important positive correlation between BRCA1 disruptions and high levels of TILs [38].

Among TAA, cancer-testis (CT) antigens represent potential antitumor antigens since they are not shared with normal somatic cells. Interestingly, analysis of CT antigens in human breast cancer revealed higher expression in ER^– vs. ER⁺ breast cancers and co-expression with basal cell markers [38]. Similar findings were reported by Curigliano et al. [39]. Consistent with these two studies, IHC analysis of eight CT antigens revealed significantly more frequent expression in ER⁻ vs. ER⁺ human breast cancers [40]. In a study of genes differently expressed in TNBC, Karn et al. [41] further reported that among genes showing no correlation with known markers of TNBC, the most prominent group of genes encoded for CT antigens. Taken together, these studies suggest that CT antigen expression is a common feature of TNBC.

In addition to being involved in the natural progression of cancer, immunity can affect the activity of various anticancer agents. Recent evidence suggests that some chemotherapeutic drugs, such as anthracyclines and oxaliplatin, rely on the induction of anticancer immune responses. In mouse models of cancer, chemotherapy with anthracyclines or oxaliplatin requires priming of IFN-γ-producing CD8⁺ T cells for optimal treatment response. In cancer patients, high levels of interferon (IFN)-gamma and CD8⁺ T cells are predictive of a good clinical response to anthracyclines. The immune-stimulating property of anthracyclines and oxaliplatin was shown to require preapoptotic translocation of calreticulin (CRT) on the tumor cell surface, post-apoptotic release of the chromatin-binding protein high-mobility group B1 (HMGB1), and extracellular release of adenosine triphosphate (ATP). CRT, HMGB1, and ATP act in concert to promote tumor antigen presentation by dendritic cells (DCs) via activation of CD91, TLR-4, and purinergic P2X7 receptors, respectively. It was recently demonstrated that chemotherapy-induced autophagy is essential for the release of ATP and subsequent anticancer immunity. Accordingly, autophagy-deficient tumor cells are unable to release ATP in response to anthracyclines or oxaliplatin and fail to elicit CD8⁺ anticancer T cells. This suggests that patients with autophagy-deficient tumor cells might benefit from therapeutic strategies designed to compensate this process in order to trigger immunogenic signaling.

Extracellular ATP appears as a central activator of chemotherapy-induced antitumor immunity. However, tumors can overexpress ecto-nucleotidases, which catabolize the hydrolysis of extracellular ATP into adenosine. Expression of these ecto-nucleotidases, such as CD39 and CD73, has two major consequences: decreasing the concentration of pro-inflammatory ATP and increasing the concentration of immunosuppressive adenosine. Notably, while the catabolism of ATP into adenosine monophosphate (AMP) is reversible, catalyzed by membrane-bound kinases, the conversion of AMP into adenosine by CD73 is irreversible. This places CD73 at a crucial checkpoint in the conversion of immune-activating ATP into immunosuppressive adenosine. Several groups have now demonstrated the importance of CD73 in the suppression of anticancer immunity. In breast cancer, CD73 is overexpressed in response to loss of ER expression [42]. Studies undertaken by the authors and confirmed by others revealed that targeted blockade of CD73 can effectively reduce tumorigenesis and metastasis of breast cancer in mice [43–47]. Furthermore, the authors recently demonstrated that CD73 expression is significantly associated with worse prognosis in TNBC patients and that the CD73-adenosinergic pathway promotes chemoresistance to anthracycline in mice [48]. Targeting CD73 therefore represents a novel approach with the potential to enhance the efficacy of chemotherapy for treatment of breast cancer.

Antitumor immunity also plays a major role in the efficacy of targeted therapies. Tumor-targeted monoclonal antibodies (mAbs), such as trastuzumab, rely in part on immune-mediated killing [48]. While innate immune responses, in particular via Antibody-dependent cell-mediated cytotoxicity (ADCC), appear to be important for trastuzumab activity [49], recent studies suggest that trastuzumab also stimulates adaptive antitumor immunity. Two studies in mice showed that trastuzumab-like therapy requires adaptive CD8-dependent immune responses to mediate optimal activity [50, 51]. These studies support a model whereby trastuzumab activates MyD88-dependent Toll-like receptors – most likely via the release of high-mobility group box 1 (HMGB1) following ADCC – stimulates the release of type I interferons (IFNs), and primes adaptive IFN-γ-producing CD8⁺ T cells. These studies raise the possibility that combination strategies may be used to capitalize on the adaptive tumor-specific immunity generated by anti-ErbB2 mAbs. Consistent with this notion, Stagg et al. [51] demonstrated that anti-PD-1 and anti-CD137 mAbs can each synergize with anti-ErbB2 mAb therapy. This data advocates that enhancement of T cell antitumor immunity should be evaluated in the clinical setting in combination with trastuzumab.

Data also exists suggesting that preexisting immunity could be important for trastuzumab efficacy. Evaluation of TILs in baseline primary tumor specimens before treatment found that those patients with high levels of TILs derived more benefit from trastuzumab added to their cytotoxic chemotherapy [52]. This data suggested that though trastuzumab has been thought to act primarily through inhibition of cell signaling, it may also serve to relieve tumor-mediated immunosuppression through yet undefined mechanisms. This data also supports that some breast cancer subtypes may be more amenable to immunotherapeutic approaches.