Introduction
Triple-negative breast cancer (TNBC) comprises 10 % to 20 % of breast cancers and is characterized by a lack of estrogen receptor (ER), progesterone receptor (PgR), and human epidermal growth factor receptor-2 (HER2) expression. TNBCs are usually composed of biologically aggressive and histologically high-grade tumors and tend to relapse within 3 years of diagnosis with the worst clinical outcome at this juncture [
1‐
3]. Previous clinical trials evaluating the efficacy of chemotherapy in the neoadjuvant setting revealed that TNBCs are more sensitive to chemotherapy and have higher rates of pathological complete response (pCR) than other breast cancer subtypes [
4,
5]. In contrast to patients with pCR and good clinical outcome, TNBCs with non-pCR have been reported to be associated with a markedly worse prognosis [
4]. Nevertheless, some patients with residual triple-negative tumors following neoadjuvant chemotherapy (NAC) are also known to achieve a relatively long-term survival without recurrence. Therefore, novel biological markers in residual tumors that can predict the survival of non-pCR patients and identify those who should receive new investigational therapeutic agents are required to further improve the patients with TNBC.
Tumor-infiltrating lymphocytes (TILs) have been reported to be associated with clinical outcome of the patients in a number of different malignant tumors, including breast cancer [
6‐
8]. TILs have been reported to differ among breast cancer subtypes and are considered a reliable marker of the efficacy of chemotherapy and trastuzumab in the TNBC and HER2-enriched subtypes [
9‐
12]. Dieci et al. recently reported that the presence of TILs in residual disease following NAC was associated with improved clinical outcome in TNBCs [
13]. The results of this study could certainly have clinical relevance for selecting the patients at risk of relapse in TNBCs, but subclassification of the lymphocytes constituting the tumor-associated immune system was not performed in their report. Subclassification of TILs is pivotal; for instance, cytotoxic T cells (CD8
+ T cells) have been reported to be associated with improved clinical outcomes in patients with breast cancer [
14‐
16], but other studies could not confirm this association [
17]. In addition, regulatory T cells defined as forkhead box protein 3 (FOXP3)
+ T cells play a pivotal role in suppressing anti-tumor immunity [
18,
19]. However, it is also true that the prognostic roles of FOXP3 remain controversial; for instance, breast cancers with FOXP3
+ TIL have been reported to be less sensitive to cytotoxic chemotherapy and have a worse prognosis [
17,
20], but others reported that those with FOXP3
+ TIL have a better prognosis [
21,
22].
The results of recent preclinical studies revealed that cytotoxic agents could possibly exert their antitumor activities by inducing an immune response against tumor cells [
23,
24]. In a small breast cancer series of 21 patients with residual disease following NAC, chemotherapy-induced lymphocytes in the tumor bed were observed in seven patients [
25]. The study of TNBCs by Dieci et al. above also reported that TILs were lower in pre- than post-chemotherapy specimens in 18 of 19 patients [
13]. Ladoire et al. analyzed the changes in subclassified TILs, focusing upon CD8 and FOXP3 status in a breast cancer series including 19 cases of ER-negative disease, and reported that the ratio of CD8/FOXP3 following chemotherapy may correlate with improved prognosis [
26]. However, more than half of the NAC regimens administered to patients in previous studies were anthracycline-based regimens, whereas the current standard is anthracycline- and taxane-based regimens. In patients with TNBC, the association of CD8
+ TIL and FOXP3
+ TIL status in residual tumors and changes induced by anthracycline- and taxane-based chemotherapy with prognosis has therefore remained unknown.
Here, in this study, we evaluated the prognostic significance of CD8+ TIL and FOXP3+ TIL in residual tumors following NAC comprising anthracyclines and taxanes and the changes of CD8+ TIL and FOXP3+ TIL before and after the therapy in a relatively larger TNBC cohort than those of previous studies. The aim of our study was to identify a reliable marker for more appropriate selection of high-risk patients eligible for more aggressive therapeutic agents, including novel investigational ones in development in patients with TNBC.
Discussion
Our study is the first to evaluate CD8+ TIL, FOXP3+ TIL, and the CD8/FOXP3 ratio in residual tumors and changes in these parameters after chemotherapy in more than 100 patients with TNBC. The results indicated that a high CD8+ TIL and the high CD8/FOXP3 ratio in residual tumors and increases in these parameters after NAC compared with pretreatment status in biopsy samples can accurately predict improved RFS and BCSS in TNBC patients with non-pCR following NAC.
A recent meta-analysis of studies exploring the prognostic value of TILs in patients with TNBC revealed that high TIL levels were significantly associated with better survival outcome in TNBCs, and the authors concluded that TIL status should be considered a strong prognostic factor in this subtype [
10]. Dieci et al. also reported the importance of TIL in residual disease after NAC in the largest cohort of 278 patients with TNBC, and their results have had a large impact in both preclinical and clinical practice [
13]. Focusing on the subsets of TIL, the tumor-related immune system has two paradoxically functional components: cytotoxic CD8
+ T cells and regulatory FOXP3
+ T cells. Cytotoxic T cells recognize foreign antigens on tumor cells through specific interactions with T-cell receptors, which lead to tumor cell death by inducing the release of proteins such as perforin and granzyme from activated T cells [
34‐
36]. By contrast, FOXP3
+ TILs, which diminish the immune response to self-antigens, have a critical role in suppressing anti-tumor immunity [
18,
19]. Therefore, the evaluation of these two major components of the tumor-related immune system is required for the assessment of chemotherapeutic efficacy in patients with breast cancer.
Based on the limited data on the detailed subclassification of TILs as CD8
+ TIL or FOXP3
+ TIL or not in TNBCs, studies evaluating the association of CD8
+ TIL or FOXP3
+ TIL status with clinical outcome have reported conflicting findings [
14‐
17,
20‐
22,
26,
27]. In the above-mentioned meta-analysis, the absence of CD8
+ TIL was reported to be associated with worse disease-free survival and overall survival (HR=0.24, 95 % CI 0.12–0.45,
P <0.0001, HR=0.58, 95 % CI 0.52–0.65,
P <0.0001, respectively), and the absence of FOXP3
+ TIL
+ also correlated with worse disease-free survival and overall survival (HR=0.44, 95 % CI 0.27–0.72,
P=0.001, HR=0.76, 95 % CI 0.60–0.96,
P=0.019, respectively) [
10]. This pooled estimation focusing on CD8
+ TIL and FOXP3 TIL
+ was considered valuable, but the results should be interpreted carefully because studies of the value of TIL in the neoadjuvant setting were excluded from this meta-analysis. Therefore, our present retrospective study of the status of CD8
+ TIL and FOXP3
+ TIL in residual tumors after NAC could provide results directly connected to present clinical practices, in which NAC is the major treatment mode for patients with TNBC. The positive correlation between high CD8
+ TIL in residual tumors after NAC and improved prognosis identified in this study is consistent with several previous studies investigating the value of CD8
+ TILs in the adjuvant setting for TNBCs [
14‐
16]. In addition, the importance of the CD8/FOXP3 ratio following chemotherapy in our TNBC cohorts supports the results of a unique small study that suggested that the ratio of CD8/FOXP3 after chemotherapy is correlated with prognosis [
26].
The additional prognostic value of stained TILs, CD8
+ TILs, FOXP3
+ TILs, and CD8/FOXP3 ratio should be discussed because there have been several studies which find the significance of unstained TILs in patients with TNBC [
8,
10‐
13]. In our cohort, the positive correlation was observed between the status of unstained TILs and CD8
+ TILs (R=0.432,
P <0.001). This is thought to be the statistical reason that unstained TILs were not significant in our multivariate analysis; on the contrary, CD8
+ TILs were significant. Based on the results, staining CD8 enabled us to identify the patients with poor prognosis more effectively than the unstained TILs which could be more available than stained TILs in clinical practices.
The results of preclinical studies indicate that the demise of immunogenic cells induced by cytotoxic agents could enable the cross-presentation of antigens, dendritic cell activation, and the induction of tumor-specific cytotoxic T cells labeled by CD8, a major component of the tumor-related immune system [
23,
24]. To further clarify the dynamic change in tumor infiltration by lymphocytes in patient samples, the changes in CD8
+ TIL and FOXP3
+ TIL caused by chemotherapy compared with baseline in biopsy specimens were determined for the first time in this study. We focused on the changes in CD8
+ TIL, FOXP3
+ TIL, and the CD8/FOXP3 ratio induced by anthracycline- and taxane-based NAC in a cohort of 78 patients with TNBC. Our study demonstrated that increases in CD8
+ TIL and the CD8/FOXP3 ratio were significantly associated with improved clinical outcomes. These results suggest that activation of the activated tumor-related immune system by cytotoxic agents might affect the micro-metastatic tumor cells in the bone marrow or blood vessels or other parts of the body, a hypothesis that certainly merits further investigation.
The present study indicates that the post-chemotherapy status of CD8
+ TIL and the CD8/FOXP3 ratio and the changes in these parameters caused by chemotherapy could be used to identify a subgroup of patients eligible for clinical trials of investigational drugs. Recently, a novel approach targeting programmed death 1 (PD-1) and the PD-1 ligand (PD-L1) pathway has been investigated in human patients with breast cancer [
37,
38]. PD-1 is a member of the CD28/cytotoxic T-lymphocyte antigen 4 (CTLA-4) family of co-stimulatory receptors and is a negative regulator of T-cell lymphocytes [
39,
40]. The PD-1 ligand PD-L1 is expressed on various tumor cells, including breast cancer cells [
38]. Early-phase clinical trials using anti-PD-1 and PD-L1 antibodies have demonstrated favorable objective responses in patients with non-small cell lung cancer, malignant melanoma, and renal cell carcinoma [
41‐
43]. For breast cancer, a phase II trial of an anti-PD-1 antibody will soon be initiated on the basis of the results of a phase I trial that demonstrated durable efficacy and acceptable safety in patients with heavily treated TNBC (the results of the study were presented at the 37th San Antonio Breast Cancer Symposium but have not yet been published). However, the association between the presence of CD8
+ TIL, FOXP3
+ TIL, PD-1, PD-L1, and the therapeutic efficacy of an anti-PD-1 antibody or PD-L1 antibody has not been examined. Further studies may be required for patient stratification for the development of these drugs.
As a potential limitation of our study, the status of TILs on biopsy specimens may not reflect that of the entire tumor before chemotherapy because of the tumor heterogeneity. By carefully estimating all pieces (four or more) of biopsy specimens, we minimized the limitation which was common among the studies with neoadjuvant settings. Further studies are needed to validate the results of the present study.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MM coordinated the study, analyzed data, and drafted the manuscript. KT and HH advised on data analyses and revised the manuscript. HT assisted with statistical analyses, contributed to the interpretation of data, and revised the manuscript. YT and SN performed immunohistochemistry, analyzed data, and drafted the manuscript. AS and GW provided the idea for the study, helped with data analysis, and revised the manuscript. HS, NO, and TI organized the study, directed data generation and analysis, and revised the manuscript. All authors read and approved the final manuscript.