Abstract
Is breast cancer (BC) immunogenic? Many data suggest that it is. Many observations demonstrated the prognostic role of tumor-infiltrating lymphocytes (TILs) in triple negative (TN) and human epidermal growth factor receptor 2 (HER-2)-positive BC. TNBCs are poorly differentiated tumors with high genetic instability and very high heterogeneity. This heterogeneity enhances the ‘danger signals’ and select clone variants that could be more antigenic or, in other words, that could more strongly stimulate a host immune antitumor response. The response to chemotherapy is at least partly dependent on an immunological reaction against those tumor cells that are dying during the chemotherapy. One of the mechanisms whereby chemotherapy can stimulate the immune system to recognize and destroy malignant cells is commonly known as immunogenic cell death (ICD). ICD elicits an adaptive immune response. Which are the clinical implications of all ‘immunome’ data produced in the last years? First, validate prognostic or predictive role of TILs. Second, validate immune genomic signatures that may be predictive and prognostic in patients with TN disease. Third, incorporate an ‘immunoscore’ into traditional classification of BC, thus providing an essential prognostic and potentially predictive tool in the pathology report. Fourth, implement clinical trials for BC in the metastatic setting with drugs that target immune-cell–intrinsic checkpoints. Blockade of one of these checkpoints, cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4) or the programmed cell death 1 (PD-1) receptor may provide proof of concepts for the activity of an immune-modulation approach in the treatment of a BC.
Breast cancer immunogenicity
Should we use immunotherapy in breast cancer (BC)? The answer lies on another key question; is BC immunogenic? That means whether BC has the ability to provoke an immune response. Immunosurveillance is critical for BC. The proof that BC may elicit an immune response are: (1) the presence of infiltrating immune cells and lymphocytes (TILs) within the tumor microenvironment; (2) the prognostic value of an immunity-related gene signature; and (3) genetic instability leading to increased number of mutations, translated into more neoantigens. The characteristics mentioned above are more pronounced in triple negative BC (TNBC) and human epidermal growth factor receptor 2 (HER-2)-positive tumor, therefore they are considered the most immunogenic BC subtypes [Cimino-Mathews et al. 2015]. TNBC has been divided in six molecular subtypes, including an immunomodulatory subtype characterized by an immune gene expression pattern, high number of immune cells infiltrating the tumor stroma, higher programmed death-ligand 1 (PD-L1) expression and high mutation rate. These features make TNBC patients good candidates for immunotherapy [Lehmann, 2011; Emens, 2012]. What is the biological basis for immune activation observed in BC? The current view of BC progression is focused on a dysregulated network of cellular pathways and most important challenge in research is to identify the central molecular switches (so-called driver alterations). The immune system has a completely different view on biology, it mainly separates between immunogenic and nonimmunogenic alterations, regardless of the biological relevance of the involved pathways. The challenge would be to identify those alterations in tumor tissue that are able to trigger immune activation. We are learning from large-scale genomic analyses that genomic alterations are highly diverse; therefore, we would expect that the biological basis of immune activation might also be heterogeneous.
Strategies to modulate the immune system in breast cancer
Immune checkpoint inhibitors
The strength and duration of T-cell response is highly regulated. Immune checkpoint receptors on T-cell surface play a role in this setting giving positive or negative signals to T cells. Cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4), programmed cell death-1 and its ligands (PD-1, PD-L1/2) axis, lymphocyte activation gene-3 (LAG-3), T-cell immunoglobulin mucin-3 (TIM-3) are negative signals inhibiting T-cell immune response. Figure 1 summarizes all immunotherapy strategies currently under investigation in BC treatment. Other inhibitory receptors, not yet used as targets in clinical settings, are B7-H4 (B7 homologous H4), V-domain immunoglobulin (Ig) suppressor of T-cell activation (VISTA), B and T lymphocyte attenuator (BTLA), T-cell Ig and ITIM domain (TIGIT). B7 homolog 3 (B7-H3) could act as both negative and positive receptor [Shin and Ribas, 2015]. The goal of immune checkpoint inhibitors such as anti-CTLA-4, anti-PD-1/anti-PD-L1, is to ‘release the brakes’ enhancing T-cell activation by blocking a negative pathway. On the contrary, receptors like OX40 induce T-cell activity by enhancing positive signals [Linch et al. 2015]. Agonist drugs can therefore boost immune activity.
Immunotherapy strategies currently under investigation in BC treatment.
CTLA-4
After TCR engagement, CTLA-4 is upregulated to attenuate T-cell responses and prevent expansion of autoreactive T cells, primarily during the priming phase of within lymph nodes. The two anti-CTLA-4 monoclonal antibodies (mAbs) available, ipilimumab and tremelimumab, have been both tested in BC. Ipilimumab has been evaluated with or without cryoablation in 18 patients with early stage BC, causing T-cell activation in the bloodstream and a modest increase in the ratio of tumor CD8+/regulatory T (Treg) cells [Diab et al. 2014]. Tremelimumab combined with exemestane in 26 patients with advanced Estrogen receptor (ER)-positive advanced BC decreased TRegs and increased number of activated Inducible costimulator ligand (ICOS)+ T cells [Vonderheide et al. 2010]. There are two trials actually recruiting in advanced or metastatic breast cancer (MBC) (Table 1). The first one is a study of an inhibitor of STAT3 and cancer stem cells self-renewal in combination with ipilimumab [ClinicalTrials.gov identifier: NCT02467361]; the second one combines histone deacetylase inhibitor and immune checkpoint blockade with nivolumab and ipilimumab for metastatic solid tumors and locally advanced/metastatic HER-2-negative BC [ClinicalTrials.gov identifier: NCT02453620]. In the window opportunity trial in combination with cryoablation no toxicity has been observed. The most common immune-related adverse events (irAEs) described during treatment with ipilimumab in trials, including patients with melanoma, are dermatitis (pruritus, rash), enterocolitis, endocrinopathies (hypophysitis, thyroiditis), liver abnormalities (elevated serum liver tests, hepatitis), and uveitis, and will be reviewed in detail. Some irAEs involving the nervous system also have been described: central nervous system (CNS) or peripheral–sensory or motor deficits, as well as irAEs of the cardiovascular system, hematopoietic system, and others [Fecher et al. 2013].
Ongoing clinical trials with immune check-point inhibitors in BC treatment.
AC, Adriamicin and cyclophosphamide; AIPAC, Active Immunotherapy PAClitaxel; BC, breast cancer; CTLA-4, cytotoxic T-lymphocyte-associated antigen-4; HER-2, human epidermal growth factor receptor 2; HR+, Hormone receptor positive; FAK, focal adhesion kinase; ICLC, carboxymethylcellulose, polyinosinic-polycytidylic acid; KIR, Killer-cell immunoglobulin-like receptors; LAG-3, lymphocyte activation gene-3; MAb, monoclonal antibody; MBC, metastatic breast cancer; PARP, Poly (ADP-ribose) polymerase (PARP); PD-1, programmed cell death 1; PD-L1, programmed death-ligand 1; SBRT, stereotactic body radiation; SABR, stereotactic ablative radiotherapy; STAT3, signal transducer and activator of transcription 3; TNBC, triple negative breast cancer.
PD-1-PD-L1/2
The enthusiasm for anti-CTLA-4 mAb is obscured by the rising success of anti-PD-1 and anti-PD-L1 mAbs, demonstrating activity in several cancers with a better toxicity profile compared with anti-CTLA-4. PD-1/PD-L1 results in negative regulation of T cells primarily within the tumor microenvironment. Pembrolizumab (MK-3475), approved for melanoma and non-small cell lung carcinoma (NSCLC), has been tested in 27 heavily pretreated patients with metastatic TNBC expressing on tumor samples PD-L1 within the phase Ib study KEYNOTE-012. Tumor samples were screened for PD-L1 expression using a prototype immunohistochemistry (IHC) assay with the 22C3 antibody. Only patients with distinctive stromal or ⩾1% tumor nest cell PD-L1 staining were eligible. PD-L1 was assessed in archival tumor samples. Determining which patients derive benefit from PD-1/PD-L1-directed immunotherapy remains an important clinical question, particularly in light of the autoimmune toxicity of these agents. The use of PD-L1 (B7-H1) IHC as a predictive biomarker is confounded by multiple unresolved issues: variable detection antibodies, differing IHC cutoffs, tissue preparation, processing variability, primary versus metastatic biopsies, oncogenic versus induced PD-L1 expression, and staining of tumor, stroma versus immune cells. Emerging data suggest that patients whose tumors overexpress PD-L1 by IHC have improved clinical outcomes with anti-PD-1-directed therapy, but the presence of robust responses in some patients with low levels of expression of these markers complicates the issue of PD-L1 as an exclusionary predictive biomarker. An improved understanding of the host immune system and tumor microenvironment will better elucidate which patients derive benefit from these promising agents. There was 1 complete response (CR), 4 partial responses (PR) and 7 stabilized disease (SD), with 3 patients remained on pembrolizumab for at least 11 months. There was one treatment-related death due to disseminated intravascular coagulation [Nanda, 2014]. Atezolizumab (MPDL3280A) is a human IgG1 engineered anti-PD-L1 antibody, with a modified Fc region to avoid antibody-dependent cytotoxicity or complement-dependent cytotoxicity induction [Gibson, 2015]. A total of 54 TNBCs treated with MPDL3280A within a phase I study reached a 19% of objective response rate (ORR) and 3 responding patients were continuing to respond. The 24-week progression-free survival (PFS) rate was 27%. The duration of responses ranged from 0.1 to >41.6 weeks; median time to response was not reached [Emens et al. 2015]. The anti-PD-1 nivolumab (BMS-936558/MDX-1106) and the anti-PD-L1 durvalumab (MEDI4736) are currently under investigation in BC.
Lymphocyte activation gene-3
IMP321 is a recombinant soluble LAG-3Ig fusion protein studied in association with weekly paclitaxel in first line MBC. A total of 30 MBC patients received IMP321 in three cohorts (doses: 0.25, 1.25 and 6.25 mg). IMP321 induced both a sustained increase in the number and activation of antigen-presenting cells (APCs) such as monocytes and dendritic cells (DCs) and an increase in the percentage of natural killer (NK) cells and long-lived cytotoxic effector-memory CD8 T cells. Clinical benefit was observed for 90% of patients with only 3 progressors at 6 months. Also, the objective tumor response rate of 50% compared favorably to the 25% rate reported in the historical control group, the translational studies showed increased number of APCs, NKs and long lived effector memory CD8 T cells. No clinically significant local or systemic IMP321-related adverse events (AEs) were recorded, in line with a previous study in which IMP321 was used alone (i.e. without chemotherapy) [Brignone et al. 2010]. All trials ongoing with immune-checkpoint inhibitors in BC are reported in Table 1.
Cancer vaccines
The final effector of antitumor immune response lies mainly on cytotoxic T lymphocytes (CTLs) recognizing nonself antigens, ultimately leading to tumor cell killing. Tumor associated antigens (TAAs) can be recognized by the immune system triggering a specific CTL response. A cancer vaccine is an active specific immunotherapy strategy presenting TAA as short or long peptides, DNA, viral vectors, whole tumor cells, autologous DCs expressing TAAs (by pulsing DCs with peptide, by transfection or co-cultured with tumor cells). Vaccines can be associated with cytokines; DC vaccines can be modified to express costimulatory molecules or cytokines with TAAs in order to potentiate the immune activation. BC is the third most studied target of cancer vaccination, after melanoma and cervical cancer. Vaccines tested in clinical trials usually showed very low toxicity and the ability to induce an immunological response and generation of CTLs against the antigen used.
HER-2
A number of vaccines with different platforms, from peptides to DNA-based vaccines, have been tested in clinical trials against the HER-2, the most used antigen to generate BC vaccine. E75 (nelipepimut-S) peptide derived from the HER-2 extracellular domain, showed in the adjuvant setting with Granulocyte-macrophage colony-stimulating factor (GM-CSF) and booster inoculation a reduction of risk of recurrence, especially in node-positive and low (IHC 1+ or 2+) HER-2 levels disease tumor [Mittendorf et al. 2014a]. AE37 is a peptide derived from the HER-2 transmembrane domain instead. When combined with GM-CSF, AE37 elicits a strong immunological response and risk recurrence reduction in the adjuvant setting of high risk BC patients [Trappey et al. 2013]. Lapuleucel-T, a vaccine made by autologous Peripheral Blood Mononuclear Cells (PBMCs) loaded with a fusion protein made by HER-2 and GM-CSF, stimulated immune as well as clinical responses and extended disease stabilization [Park et al. 2007]. Intranodal vaccination has been also evaluated, with a DC-based vaccine targeting HER-2 evaluated in ductal carcinoma in situ (DCIS) [Sharma et al. 2012]. In another prospective trial, clinically disease-free node-positive and high-risk node-negative BC patients with tumors expressing any degree of HER-2 (IHC 1–3+) were enrolled. Patients were randomized to AE37 + GM-CSF versus GM-CSF alone and toxicity was monitored. Clinical recurrences were documented and disease-free survival (DFS) analyzed. The trial enrolled 298 patients; 153 received AE37 + GM-CSF and 145 received GM-CSF alone. The groups were well matched for clinicopathologic characteristics and toxicities have been minimal. At the time of the primary analysis, the recurrence rate in the vaccinated group was 12.4% versus 13.8% in the control group [relative risk reduction 12%, Hazard Ratio (HR) 0.885, 95% confidence interval (CI) 0.472–1.659, p = 0.70]. The Kaplan-Meier estimated 5-year DFS rate was 80.8% in vaccinated versus 79.5% in control patients. In a planned subset analyses of patients with IHC 1+/2+ HER-2-expressing tumors, 5-year DFS was 77.2% in vaccinated patients (n = 76) versus 65.7% in control patients (n = 78) (p = 0.21). In patients with TNBC (HER-2 IHC 1+/2+ and hormone receptor negative) DFS was 77.7% in vaccinated patients (n = 25) versus 49.0% in control patients (n = 25) (p = 0.12). The overall intention-to-treat analysis demonstrates no benefit to vaccination. [Mittendorf, 2016]. In the phase I dose-escalation study the safety and immunogenicity of an anticancer immunotherapeutic [recombinant HER-2 protein (dHER-2) combined with the immunostimulant AS15] was assessed in patients with early-stage HER-2-overexpressing BC [Limentani, 2016]. A total of 61 trastuzumab-naïve patients with stage II–III HER-2-positive BC received the dHER-2 immunotherapeutic after surgical resection and adjuvant therapy. They were allocated into four cohorts receiving different doses of dHER-2 (20, 100, 500 µg) combined with a fixed AS15 dose. Safety and immunogenicity (dHER-2-specific antibody responses) were assessed. After completing the immunization schedule (3 or 6 doses over 14 weeks) and a 6-month follow-up, the patients were followed for 5 years for late toxicity, long-term immunogenicity, and clinical status. The immunizations were well tolerated, and increasing doses of dHER-2 had no impact on the frequency or severity of AEs. Few late toxicities were reported, and after 5 years 45/54 patients (83.3%) were still alive, while 28/45 (62%) with known disease status were disease free. Regarding the immunogenicity of the compound, a positive association was found between the dHER-2 dose, the immunization schedule, and the prevalence of dHER-2-specific humoral responses. Among the patients receiving the most intense immunization schedule with the highest dHER-2 dose, 6/8 maintained their dHER-2-specific antibody response 5 years after immunization. The dHER-2 immunotherapeutic had an acceptable safety profile in early HER-2-positive BC patients. dHER-2-specific antibody responses were induced, with the rate of responders increasing with the dHER-2 dose and the number and frequency of immunizations. The phase I/II study evaluated the safety and clinical activity of dHER-2 and the immunostimulant AS15 in patients with HER-2-overexpressing MBC [Curigliano et al. 2016]. A total of 40 HER-2-positive MBC patients received up to 18 doses (12q2w, 6q3w) of dHER-2 immunotherapeutic, as first- or second-line therapy following response to trastuzumab-based treatment as maintenance. Toxicity was graded by the Common Terminology Criteria for Adverse Events (CTCAE) and clinical activity was evaluated by target lesion assessment according to the Response Evaluation Criteria in Solid Tumors (RECIST). Immunogenicity was assessed. The dHER-2 immunotherapeutic was well tolerated: grade 1/2 AEs were most common. No cardiac events were observed and one patient experienced an asymptomatic decrease of left ventricular ejection fraction below the normal range (47%). Both humoral and cellular immunogenicity to the dHER-2 antigen was observed. No patient discontinued the immunizations because of AEs but 35/40 withdrew prematurely, 34 because of disease progression (24/34 before or at the tumor assessment after dose 6). Overall, 1 patient achieved a complete response lasting 11 months and 1 patient had a partial response lasting 3.5 months. A total of 10 patients experienced SD ⩾ 26 weeks with 4/10 still in SD at the last tumor assessment after 47 weeks. Immunization of MBC patients with the dHER-2 immunotherapeutic was associated with minimal toxicity and no cardiac events. Clinical activity was observed with 2 objective responses and prolonged stable disease for 10/40 patients [Curigliano et al. 2016].
Mucine 1
Mucine 1 (MUC1), a glycoprotein involved in cell adhesion and signal transduction, is overexpressed in 70% of cancers and usually aberrantly glycosylated. Anti-glycosylated MUC-1 vaccine showed a trend toward improved PFS and OS when combined with hormonal therapy [Miles et al. 2011]. The study population consisted of 1028 women with MBC across 126 centers who had previously received chemotherapy and had either a complete or a partial response or no disease progression. All women received one-time intravenous cyclophosphamide (300 mg/m2) 3 days before subcutaneous injection of 100 μg sialyl-Tn (STn) conjugated to keyhole limpet haemocyanin (KLH) plus adjuvant (treatment group) or 100 μg KLH plus adjuvant (control group) at weeks 0, 2, 5, and 9. Subsequently, STn-KLH without adjuvant or KLH without adjuvant was then administered monthly for 4 months, and then quarterly until disease progression, without cyclophosphamide. STn-KLH vaccine was well tolerated; patients had mild-to-moderate injection-site reactions and reversible flu-like symptoms. Week-12 antibody testing revealed high specific IgG titers and a high rate of IgM-to-IgG seroconversion; the median IgG titers in STn-KLH recipients were 320 (anti-ovine submaxillary mucin) and 20,480 (anti-STn), with no detectable antimucin antibodies in the control group. The time to progression (TTP) was 3.4 months in the treatment group and 3.0 months in the control group. The median survival times were 23.1 months and 22.3 months, respectively [Miles et al. 2011]. An oxidized mannan-MUC1 vaccine showed, in the adjuvant setting, an encouraging reduction in recurrence rate (12.5% versus 60% in the placebo arm) after 15 years of follow up [Vassilaros et al. 2013].
Carcinoembryonic antigen
Carcinoembryonic antigen (CEA) is an adhesion molecule overexpressed in several cancers. The poxviral vaccine targeting both CEA and MUC1 is engineered to express T-cell costimulatory molecules and has been tested in advanced BC. Only a minority of patients, mainly with lower tumor burden and less pretreated, experienced durable responses [Mohebtash et al. 2011]. Overall, 26 patients were enrolled and given monthly vaccinations. Clinical and immune outcomes were evaluated. These patients were heavily pretreated, with 21 of 26 patients having 3 or more prior chemotherapy regimens. Side effects were largely limited to mild injection-site reactions. For the 12 BC patients enrolled, median TTP was 2.5 months (1–37+) and median overall survival was 13.7 months. A total of four patients had stable disease. Overall, one patient had a complete response by RECIST and remained on the study for >37 months. Other antigens used in BC vaccines are mammoglobin-A, telomerase, p53 and survivin [Cimino-Mathews et al. 2015].
Adoptive T-cell therapy
Adoptive cell transfer (ACT) means removing T cells from the patient to be genetically engineered to express chimeric antigen receptors (CARs), made by a single chain antibody structure extracellularly fused with an intracellular signaling domain including CD28 and CD137 costimulatory molecules, then reinfused into the patient. These potent CAR-T cells successfully treated lymphoblastic leukemia [Stagg and Allard, 2013]. However, CAR-T cells recognizing HER-2 caused a patient’s death probably due to cross reactivity [Morgan et al. 2010]. In an attempt to treat cancer patients with ERBB2-overexpressing tumors, authors developed a CAR based on the widely used humanized mAb trastuzumab (Herceptin). An optimized CAR vector containing CD28, 4-1BB, and CD3 zeta-signaling moieties was assembled in a gamma-retroviral vector and used to transduce autologous peripheral blood lymphocytes (PBLs) from a patient with colon cancer metastatic to the lungs and liver, refractory to multiple standard treatments. Following completion of nonmyeloablative conditioning, the patient received 10(10) cells intravenously. Within 15 minutes of cell infusion the patient experienced respiratory distress, and displayed a dramatic pulmonary infiltrate on chest X-ray. She was intubated and despite intensive medical intervention the patient died 5 days after treatment [Morgan et al. 2010]. A clinical trial with autologous cMET redirected CAR administered intratumorally is ongoing in BC [ClinicalTrials.gov identifier: NCT01837602].
Figure 1 summarized all potential immunotherapy strategies in BC.
Strategies to improve immunotherapy in breast cancer
Define the best setting
New drugs are usually first experimented in patients with metastatic disease, after standard options fail. However, the metastatic setting is less suitable for immunotherapy, due to a large immunosuppressive disease burden and highly evolved immunoescape mechanisms. A significant problem of BC is tumor recurrence after a curative treatment, also after several years especially in ER-positive disease. Immunotherapy, cancer vaccine in particular, could have several advantages in the adjuvant setting: (1) fewer isolated tumor cells have not established a large immunosuppressive environment; (2) due to low toxicity, cancer vaccines could be administered for a long period; (3) repeated booster vaccinations could keep long memory immune response [Soliman, 2013]. For example, the peptide vaccine E75 is in an advanced stage of clinical development, as a phase III trial is currently ongoing in the adjuvant setting in node positive disease (PRESENT study) [ClinicalTrials.gov identifier: NCT01479244]. Preinvasive BC could be another ideal setting for vaccine because of the small amount of slow growing disease. A DC vaccination against HER-2 injected intranodally was evaluated in 27 patients with DCIS. Overall, five patients had no evidence of disease at surgery, and there was an increased number of HER-2-specific Cytotoxic T lymphocyte interferon (IFN)-gamma secreting cells [Sharma et al. 2012]. A preventive vaccination strategy in women at high risk of developing BC could be potentially explored (specifically in patients with ductal carcinoma in situ, HER-2 positive and poorly differentiated).
Combined immunotherapies
Due to the complexity of anticancer immune surveillance and the dynamic interplay between cancer and immune system, any immunotherapy strategy engaging only one component of the immune system is clearly inadequate to mount an effective immune response and to overcome immune tolerance. Immunotherapy could act in synergy with other treatment strategies in order to hit the complex immune surveillance machinery at more than one level. A better understanding of the immunomodulatory properties of cytotoxic chemotherapy, targeted therapy and radiation therapy is therefore necessary to design combination strategies rationally. The combination of anti-CTLA-4 and anti-PD-1 in melanoma showed an impressive response rate and durable response compared with a single agent, at the price of an increase in toxicity rate [Larkin et al. 2015]. The same combination is currently under evaluation in different solid tumors including TNBC [ClinicalTrials.gov identifier: NCT01928394]. A combination of anti-PD-1 and anti-T-cell immunoglobulin and mucin-domain containing-3 (TIM-3) mAbs showed a very potent anticancer effect against experimental tumors, and a combination of anti-LAG-3 and anti-PD-1 therapies exhibited a synergistic activity, preventing exhaustion and anergy of T cells [Shin and Ribas, 2015]. Combining immunotherapies targeting co-inhibitory and co-stimulatory molecules means inhibiting the suppressive load on T cells while enhancing T-cell activity. For example, PD-1 blockade is likely to synergize with OX40 agonists in particular in highly PD-L1-expressing tumors: OX40 increases IFN-gamma production by T cells, which upregulates PD-L1 in many cancer cells, then OX40 promotes the expansion and survival of activated CD4 and CD8 cells expressing PD-1 [Linch et al. 2015]. Ibrutinib and idelalisib, B-cell receptor-signaling inhibitors used in B-cell lymphoid malignancy, also showed activity on T cells, promoting development toward T Helper 1 (Th1) phenotypes and suppressing Tregs, which makes a combination with immune checkpoint inhibitors very attractive [Dubovsky et al. 2013]. Trastuzumab, given concurrently with anti-HER-2 cancer vaccines, enhances HER-2 specific T-cell immunity anti-HER-2 antibodies [Soliman, 2013; Emens, 2012].
Combining immunotherapy and targeted therapy
Inhibition of phosphoinositide-3-kinase (PI3K) or AKT decreases PD-L1 expression and increases cancer cell killing by cytotoxic T cell (CTL) [Mittendorf et al. 2014b]. Drugs interfering with this molecular pathway could potentially enhance the activity of an associated immunotherapy treatment. However, there is still much more to learn about such combinations. For example, a trial with vemurafenib and ipilimumab in melanoma was stopped early due to significant hepatotoxicity, probably due to a mitogen-activated protein kinase (MAPK) pathway paradoxical activation [Ribas et al. 2013].
Combining immunotherapy and chemotherapy
Chemotherapy, historically considered immunosuppressive, on the contrary showed also positive immunomodulatory effects. Drugs like anthracycline, taxanes and cyclophosphamide can induce ‘immunogenic cell death’, characterized by TAA release from dying cancer cells as well as release of molecules promoting DC maturation. Subsequently DCs present TAAs to T cells inducing specific anti-tumor CTL activation [Kroemer et al. 2013]. Chemotherapy can enhance immune response by enhancing effector immune cells or by depleting immunosuppressive populations. In BC, taxanes can enhance NK and T-cell function [Carson et al. 2004], as well as increase TIL numbers in the neoadjuvant setting [Demaria et al. 2001]. Docetaxel increases levels of Th1-associated cytokines while decreasing negative inflammatory markers in the metastatic setting [Tsavaris et al. 2002]. Low doses of cyclophosphamide [Ghiringhelli et al. 2007] and paclitaxel [Zhang et al. 2008a] can induce selective depletion of Tregs, while docetaxel [Kodumudi et al. 2010] and gemcitabine [Nowak et al. 2002] can reduce the number of myeloid-derived suppressor cells (MDSCs). Paclitaxel, etoposide and 5-fluoruracil upregulate PDL-1 expression on BC cell lines, therefore promoting immunoresistance [Zhang et al. 2008b]. Interfering with the PD-1/PD-L1 pathway with anti-PD-1/PD-L1 immunotherapy could counteract this effect. Hormonal therapy can modulate the immune system as well, for example letrozole in the neoadjuvant setting reduces intratumoral FOX-P3 Tregs [Generali et al. 2009].
Combining immunotherapy and radiotherapy
Several studies demonstrated the immunomodulatory properties of radiotherapy (RT). RT causes immunogenic cell death (ICD), augments MHC-I expression both in normal and cancer cells, stimulates chemotaxis and recruitment of DCs and T cells to the tumor by inducing intracellular adhesion molecules, cytokines and chemokines, and inducing CTL priming. RT can destroy TILs within the tumor stroma, leading to temporary depletion of all immune cells, from CTL and NKs but also Tregs [Vatner et al. 2014]. Cancer cells surviving to RT undergo a mechanism of ‘immunogenic modulation’ making them more susceptible to immune-mediated killing [Gameiro et al. 2014]. The immune system mediated a rare phenomenon of RT called the ‘abscopal’ effect, characterized by a response in distant metastatic areas outside the radiation field after irradiation of other site of disease. The abscopal effect can therefore be facilitated combining RT and immunotherapy aiming at recruiting and activating DCs to improve the presentation to CTL of TAAs released in the irradiated site by ICD. This was the aim of the association of GM-CSF or intratumoral injection of autologous DCs within the irradiated tumor to RT. Toll-like receptor (TLR) agonists can activate intratumoral DCs expressing TLRs [Vatner et al. 2014]. A trial of RT to BC skin metastases combined with TLR7 agonist imiquimod and cyclophosphamide is ongoing [ClinicalTrials.gov identifier: NCT01421017]. Regarding immune checkpoints, preclinical studies in BC murine models showed synergistic activity of anti-CTLA-4 and RT combination due to the abscopal effect [Derer et al. 2015]. The anti-PD-1 antibody MK-3475 [ClinicalTrials.gov identifier: NCT02303366] and the OX40 agonist MEDI6469 [ClinicalTrials.gov identifier: NCT01862900] are currently under investigation with stereotactic RT to metastatic sites in BC. Investigating the optimal combination, correct dose and schedule is necessary to implement the promising association of RT and immunotherapy.
Timing
The correct timing of the various therapies used together is an aspect not adequately explored. The sequence can have an impact on the effect of each treatment, especially on the immune system. For example, anti-CTLA-4 induces the abscopal effect when it is concurrent to RT [Vatner et al. 2014]. Cancer vaccine induces first CTL stimulation and amplification, followed by Treg proliferation. Chemotherapy given late after cancer vaccines could potentially have stronger immunostimulating properties by depleting just the Treg compartment [dePillis et al. 2014]. A rational schedule could therefore contribute to better efficacy of combination treatment.
Identify biomarkers
Mutational load
Tumor mutational load has been recently associated with clinical efficacy from immune checkpoint inhibitors in NSCLC and melanoma [Snyder et al. 2014; Rizvi, 2015]. Any mutation could potentially produce neoantigens, therefore the level of mutagenesis of cancer cells may correlate with their immunogenicity. More neoantigens have a higher chance to be presented to T cells, widening the specific CTL repertoire against cancer. The recent study showing that mismatch-repair status predicted clinical benefit of anti-PD-1 pembrolizumab shows the importance of mutational load for immunotherapy [Le et al. 2015]. This study was conducted in a population of 41 patients with progressive metastatic carcinoma with or without mismatch-repair deficiency. Patients with mismatch-repair-proficient colorectal cancers, and patients with mismatch-repair-deficient cancers that were not colorectal. Indeed, mismatch-repair defects may cause a large number of somatic mutations in cancer cells increasing anti-PD-1 mAb effects. Median somatic mutation frequency of BC is lower than other tumor types, particularly compared with tumors with the highest level of mutations associated with extensive exposure to carcinogens (e.g. UV radiation in melanoma, and tobacco in lung cancer) [Lawrence et al. 2013]. Moreover, among BC subtypes, there is high variability in number and type of mutations (i.e. ER-positive versus ER-negative disease) [Stephens et al. 2012]. BRCA1/2 mutated BC display severe genomic instability, which could make mutated cancer more susceptible to immunotherapy. Therefore, the genomic landscape of BC could potentially have a role as a marker of response to immunotherapy.
PD-1/PD-L1 expression
Expression of PD-1 or PD-L1 in tumor cells or the tumor microenvironment has been variably assessed and has been used to select patients eligible for anti PD-1/PD-L1 treatment in clinical trials. In BC, PD-L1 upregulation has been associated with poor prognostic features such as large tumor size, high grade, high proliferation, ER-negative status and HER-2 positive status [Sabatier et al. 2015]. PD-L1 overexpression is also higher in inflammatory breast cancer (IBC) that in non-IBC [Bertucci, 2015]. TNBC has a greater expression of the PD-L1 gene compared with the non-TNBC subtype (expressed in 20% of TNBC) [Mittendorf et al. 2014b]. In basal-like tumor, PD-L1 was also associated with survival, better response to chemotherapy and increased immune-response [Sabatier et al. 2015]. Available data in immune checkpoint show responses among PD-1/PD-L1-negative tumors too and the real prognostic and predictive significance of these markers is still a matter of debate.
Tumor infiltrating lymphocytes
Immune cells are components of the tumor stroma. Intratumoral TILs have direct cell contact with carcinoma cells while stromal TILs are dispersed in the stroma. The latter seem to be superior and more reproducible than intratumoral TILs when used to quantify TILs in cancer. TILs are evident in hematoxylin/eosin stain, then IHC can further characterize the different class of infiltrated lymphocytes. Immune infiltrates are highly heterogeneous and a standardized methodology for TIL evaluation is needed [Salgado et al. 2015]. Tumors with high immune infiltrate (more than 50–60% of stromal TILs), defined as lymphocyte-predominant BC, have relatively good prognosis even if they belong to a bad subtype such as TNBC or HER-2+ [Salgado et al. 2015]. Accumulating preclinical and clinical evidence supports the use of TILs as a predictive and prognostic marker in BC, both in the neoadjuvant and adjuvant setting. The presence of TILs in the treatment of TNBC in the adjuvant setting is an independent prognostic factor for improved OS, decrease distant recurrence and increased metastases-free survival, while TILs seem to lose their prognostic role in luminal disease [Loi et al. 2013; Adams et al. 2014; Loi et al. 2014]. After neoadjuvant treatment such as anthracycline and paclitaxel chemotherapy or trastuzumab, increased TIL infiltration, increasing CD8+ cells and reducing Treg numbers could be associated with pathological CR and improving survival [Dieci et al. 2014; Denkert et al. 2010]. We do not know yet whether TILs are predictive of response to immunotherapy. We could speculate that BC with large or limited immune cell infiltrates likely may have different susceptibility to immunotherapy.
Conclusion
Modulating the immune system is a promising treatment strategy in BC, in particular in TNBC. However, there is still much to be done to implement immunotherapy in this disease. In our opinion, research should focus on: (a) identifying the best setting where modulation of the immune system has greater impact on natural history of the disease; (b) combining immunotherapy with new or standard treatment modalities, with schedule and timing rationally designed in order to increase immunotherapy; (c) identifying the most responding cancer subtype and identifying biomarkers to predict and to early assess activity of immunotherapy, in order to select patients with higher chances of response and to contain costs. It is expected that immunogenomic studies will elucidate the mechanisms that prevent or promote the development of a favorable antitumoral immunity. Such mechanisms could then be targeted to reprogram the microenvironment toward an immune permissive one, resulting in an enhanced efficacy of immunotherapeutic approaches. The role of cancer genomic instability in shaping immune response, and therefore influencing the response to immunotherapy, only now begins to be elucidated. Understanding how cancer genetics, host genetics and environmental factors collectively regulate the development of immune phenotypes of BC will have valuable clinical implications.
Footnotes
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Conflict of interest statement
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.