Abstract
Background/Aim: Gangliosides (acidic glycosphingolipids) have crucial regulatory roles in normal physiological processes, as well as in pathological conditions, including tumor onset and progression. GD2 is highly expressed in triple-negative breast cancer (TNBC), particularly in cancer stem cells. However, little is known on the clinical impact of GD2 expression on the prognosis of TNBC. Consequently, we aimed to investigate the association between GD2 expression in TNBC and the prognosis of TNBC. Patients and Methods: We assessed GD2 expression in 76 patients with primary TNBC who had undergone surgery at our Institute between 2012 and 2015 using immunohistochemical analysis with a tissue microarray technique. We investigated the relationship between GD2 expression and clinicopathological factors in TNBC, recurrence-free survival (RFS), and overall survival (OS). Results: Increased GD2 expression was observed in 45% of TNBC patients. There was no significant association between GD2 expression and clinicopathological factors in TNBC. The 5-year RFS rate among patients with GD2-positive TNBCs was significantly worse than that among patients with GD2-negative TNBCs (75.4% and 94.9%; HR=4.931; 95%CI=1.024-23.752; p=0.027). The OS in patients with GD2-positive TNBCs tended to be inferior to that of patients with GD2-negative TNBCs (HR=5.357; 95%CI=0.599-47.939; p=0.092). Interestingly, in patients with GD2-positive TNBCs, a higher grade of tumor-infiltrating lymphocytes (TILs) displayed a significantly better impact on OS (TILs-high vs. TILs-low; p=0.04). Both univariate and multivariate analyses showed that GD2 expression negatively affected RFS (p=0.027, p=0.021, respectively). Conclusion: GD2 expression is an independent unfavorable prognostic factor for TNBC.
Globally, breast cancer is one of the most prevalent cancers among women (1). Although the mortality rate of early breast cancer is improving, metastatic and recurrent breast cancers continue to be a treatment challenge. Triple-negative breast cancer (TNBC), which lacks estrogen and progesterone receptors and does not overexpress human epidermal growth factor receptor 2 (HER2), has the worst prognosis among all breast cancer subtypes (2, 3).
Recent evidence revealed that TNBC tends to have higher immunogenicity due to genomic instability and a higher mutational load than other subtypes, leading to immunologically inflamed environments (4-6). Tumor-infiltrating lymphocytes (TILs) are more abundant in the tumor microenvironment of TNBC than in other subtypes, and patients with early TNBC and high levels of TILs typically have a favorable prognosis (7, 8). In addition, the expression of programmed death-ligand 1 (PD-L1) is possibly increased not only in tumor cells but also in immune cells within the tumor microenvironment compared to other subtypes (9, 10). Combination of immune checkpoint inhibitors with chemotherapy in patients with TNBC has become a new standard treatment (11, 12). However, owing to the paucity of targeted therapies for TNBC, patients are still commonly treated with conventional chemotherapy and radiotherapy without any particular specificity for TNBC, resulting in development of resistance. Therefore, development of new treatments for TNBC is urgently needed.
Gangliosides (acidic glycosphingolipids) have crucial regulatory roles in normal physiological processes, such as embryogenesis, as well as in pathological conditions, including tumor onset and progression (13). Physiologically, ganglioside GD2 is exclusively and modestly expressed in the neural tissues of the central and peripheral nervous systems, as well as in skin melanocytes (14). In contrast, GD2 is overexpressed in neuroectodermal tumors such as neuroblastoma and melanoma, soft tissue sarcoma, small lung cell cancer, and breast cancer (14-18).
Furthermore, GD2 has been reported to be abundant in CD44highCD24low cancer stem cells (CSCs), that are important for cancer chemoresistance (19). In breast cancer, it has been reported that GD2 expression was associated with aggressive features such as the triple negative phenotype, enhanced invasiveness, and a CSC-like phenotype (18).
Taken together, owing to the exclusive overexpression of GD2 in malignant tumor cells and its close correlation with CSCs, GD2 seems to have potential as a promising target for TNBC. However, little is known on the clinical impact of GD2 expression on the prognosis of TNBC. Consequently, in this study, we first attempted to clarify the impact of GD2 on the prognosis of patients with TNBC.
Patients and Methods
Patients. Seventy-six patients with triple-negative invasive BC, who underwent surgery for stage I to III tumors, between January 2012 and December 2015, at the Mie University Hospital (Tsu, Japan) were included in the study. Patients with ductal carcinoma in situ (DCIS), small tumors (<5 mm), tumors that had received neoadjuvant chemotherapy, or those with recurrent tumors were excluded. Table I shows the clinical and pathological characteristics of the patients enrolled in this study. This study was conducted in accordance with ethical principles and as per the tenets of the Declaration of Helsinki, and it was approved by the Institutional Review Board of Mie University Hospital (No. H2021-021). The requirement for written informed consent was waived because of the retrospective nature of the study.
Association between GD2 expression and clinicopathological factors of patients included in the study.
Histological evaluation. All the patients underwent surgical intervention for primary breast tumors. The specimens were formalin-fixed, paraffin-embedded (FFPE), and cut into 4-μm-thick sections for hematoxylin and eosin (HE) staining. Nuclear grades (NG) were assigned according to the histological classification of breast tumors in the General Rules for Clinical and Pathological Recording of Breast Cancer (Japanese Breast Cancer Society) using surgical specimens.
Immunohistochemistry (IHC) for GD2 and PD-L1 expression was performed using tissue microarrays (TMAs). As described in our previous report (10), TMAs were constructed from tumor blocks of surgical specimens, using 2.0-mm (diameter) tumor cores from selected blocks. These cores were assembled in a TMA format, and the paraffin-embedded TMA blocks were then sectioned at 4-μm thickness and subjected to IHC analysis.
IHC analysis. Estrogen receptor (ER) and progesterone receptor (PgR) positivity was evaluated against a cut-off of 1%, and HER2 overexpression was evaluated in accordance with the American Society of Clinical Oncology/College of American Pathologists 2013 guidelines. Fluorescence in situ hybridization to assess HER2 amplification was performed whenever equivocal results (2+) were obtained. Ki-67 expression was assessed by IHC using MIB-1 monoclonal antibody (mAb) (Table II). The HER2 protein expression status was determined by immunohistochemical analysis using an anti-HER2 mAb assay. IHC staining was performed using an automatic immunostainer (BenchMark XT; Ventana Medical Systems, Tucson, AZ, USA).
Antibodies and immunohistochemical assays.
Evaluation of PD-L1 expression. PD-L1 staining was performed using anti-PD-L1 mAb (SP142) (Table II). PD-L1 status was evaluated based on the proportion of the tumor area occupied by PD-L1-expressing tumor-infiltrating immune cells (%IC) at any intensity. According to the results of a clinical trial that used SP142 for TNBCs (11), PD-L1 expression was assessed against a cut-off of 1%.
Evaluation of GD2 expression. GD2 was stained using an anti-GD2 mAb (14G2a) (Table II). GD2 expression was evaluated based on the proportion of GD2-positive tumor cells at any intensity, and scored as negative (<1%), low (1-10%), intermediate (11-50%), or high (>50%) by two pathologists independently that were blinded to clinical information. For statistical analysis, these categories were divided into two groups: negative (<1%) and positive (≥1%).
Evaluation of TILs. H&E-stained slides were reviewed by two pathologists in accordance with the criteria of the International TILs Working Group 2014 (20). As shown in our previous report (10), lymphoplasmacytic infiltration was evaluated in the stromal area around the invasive tumor, and the average of several tumor areas was determined, excluding lymphocyte infiltration around the DCIS and normal lobules. TILs were analyzed as a continuous parameter and categorized into two groups using 10% as the cutoff value, where <10% stromal TILs were defined as the low TILs group and >10% as the high TILs group.
Outcomes. Clinicopathological data, including age, histology, size, NG, Ki-67 index, lymph node status, type of surgery, ER/PgR/HER2 status, and use of adjuvant treatments (chemotherapy and radiation) were obtained from routine reports. Patients received anthracycline-and/or taxane-based regimens and radiation therapy according to their own risks as adjuvant treatments.
Statistical analysis. Continuous data were analyzed using the Mann-Whitney U-test. The χ2-test was used to compare GD2-positive and GD2-negative groups. Kaplan-Meier analysis and log-rank tests were used to show differences in recurrence-free survival (RFS) and overall survival (OS) according to GD2 expression in combination with TIL infiltration. Cox regression proportional hazard models were used to estimate hazard ratios (HR) for RFS according to GD2 alone or in combination with TIL infiltration in both univariate and multivariate analyses. p<0.05 indicated statistically significant differences. Statistical analyses were performed using IBM SPSS Statistics 27.0 (SPSS Inc., Chicago, IL, USA).
Results
GD2 expression in TNBC. First, we assessed the expression of GD2 in TNBC on a 4-point scale (A-D). We found that 55% of the tumors (42/76) were GD2-negative, 30% (23/76) were GD2-low, 11% (8/76) were GD2-intermediate, and 4% (3/76) were GD2-high (Figure 1E). In this study, GD2-positive was defined as ≥1%, resulting in a GD2-positive rate of 45% (34/76) in patients with TNBC.
GD2 staining of samples of triple negative breast cancer. (A) High (>50%), (B) intermediate (10-50%), (C) low (1-10%), and (D) negative (<1%) expression of GD2 are shown. (E) Pie charts presenting results of GD2 staining for all examined cases.
GD2 expression in TNBC is not associated with known clinicopathological factors. Table I shows the clinical and pathological characteristics of the patients enrolled in this study. We attempted to characterize the clinicopathological features of patients with GD2-positive TNBC. However, there were no significant associations between GD2 expression and tumor histology, tumor size, nodal status, nuclear grade, Ki-67 index, pathological stage, PD-L1 expression, or the presence of TILs (Table I).
GD2 expression is significantly associated with worse recurrence-free survival of TNBC. Next, we assessed the effect of GD2 expression on the prognosis of TNBC. The median duration of follow-up was 73.5 months (range, 1 to 104). In 76 patients, the recurrence-free survival (RFS) of GD2-positive TNBCs was significantly worse than that of GD2-negative TNBCs (HR=4.931; 95%CI=1.024-23.752; p=0.027) (Figure 2A). The 5-year RFS rate of GD2-positive and GD2-negative TNBCs were 75.4% and 94.9%, respectively. GD2-positive TNBCs tended to have an inferior overall survival compared to GD2-negative TNBCs, but the difference was not significant (HR=5.357; 95%CI=0.599-47.939; p=0.092) (Figure 2B). The 5-year overall survival (OS) of GD2-positive and GD2-netative TNBCs were 87.4% and 97.4%, respectively. These results suggest that GD2-positive TNBC is associated with a poorer prognosis than GD2-negative TNBC.
Kaplan-Meier curves for recurrence-free survival (RFS) (A) and overall survival (OS) (B) in patients with GD2-positive (red line) and GD2-negative (blue line) TNBCs. TNBC: Triple negative breast cancer.
Association between GD2 expression combined with TILs and prognosis of TNBC. Numerous studies have reported that tumor-infiltrating lymphocytes (TILs) are likely to accumulate in TNBC due to the high immunogenicity of these tumors, and high amounts of TILs in tumor areas are associated with a favorable prognosis. Therefore, we investigated whether the poor prognosis caused by GD2 expression in TNBC is affected by the presence of TILs. We divided the included cases in the study into four groups based on GD2 expression (positive/negative) and the number of TILs (high/low). As shown in Figure 3A and 3B, TNBC with GD2-positive and low TILs levels had the poorest RFS and OS among the four groups, with statistically significant differences (RFS, p=0.031; OS, p=0.028). These results suggest that high amounts of TILs can attenuate the unfavorable effects of GD2 expression in TNBC.
Kaplan- Meier curves for RFS (A) and OS (B) in patients with GD2-positive and high-TILs (green line), GD2-positive and low-TILs (blue line), GD2-negative and high-TILs (red line), and GD2-negative and low TILs (brown line) TNBCs. TNBC: Triple negative breast cancer; RFS: recurrence-free survival; TIL: tumor-infiltrating lymphocyte; OS: overall survival.
GD2 is an independent prognostic factor for RFS of TNBC. Finally, we performed both univariate and multivariate analyses to investigate the prognostic factors in 76 patients with TNBC. Univariate regression model analysis showed that RFS was significantly correlated with tumor size (pT1 vs. pT2-4, p=0.005) and GD2 expression (negative vs. positive, p=0.027) (Table III). When these parameters were subjected to multivariate analysis, tumor size (odds ratio=13.366; 95% CI=1.661-107.533; p=0.015) and GD2 expression (odds ratio=6.407; 95% CI=1.324-30.966; p=0.021) were significant independent factors for RFS (Table III).
Univariate and multivariate analysis for recurrence-free survival.
Discussion
Herein we investigated the impact of GD2 on the prognosis of TNBC patients. In this study, GD2 was incrementally expressed in 45% of the patients with TNBC in our cohort. Furthermore, we found that the 5-year RFS of GD2-positive TNBC was significantly worse than that of GD2-negative TNBC (p=0.027) and that GD2 expression was an independent prognostic factor for RFS. Our data indicated that GD2-positivity is associated with an unfavorable prognosis, particularly, those patients with TNBC that were GD2-positive and lacking TILs displayed the poorest overall survival.
Previous studies using breast cancer cell lines showed that GD2 was more highly expressed in basal-like cell lines with TNBC features than in other subtypes (19). However, the data on FFPE samples of human breast cancer are not only limited but also controversial. In one report, GD2 was shown to be more frequently expressed in high-grade tumors, ER-negative tumors, and the elderly (18), and in another report, GD2 expression was observed more frequently in ER-positive tumors than in ER-negative tumors (21). GD2-positive cases of TNBC have been reported to be 67% (18) and 18% (21) in two previous studies and 45% in the present study, demonstrating considerable variation. These variations may be attributed to the different antibodies used for GD2-staining, the different cutoff values for IHC staining, and the evaluation criteria in each study. Since a standard GD2 evaluation method has not yet been established, we used the 14G2a clone of mouse antibody for our GD2 staining procedure. This antibody was produced in the 1980s and has been used in several clinical trials for neuroblastoma since the 2000s. We determined positivity as 1% of the cells are positive for the expression of the marker, as described in the Materials and Methods section. However, the appropriateness of the present criteria must be validated in a larger cohort. Especially, for clinical application of GD2-targeted immunotherapy, the establishment of GD2 evaluation methods required urgent attention.
GD2 has been reported to be associated with aggressive features and advanced stages of all subtypes (18, 22). Most TNBC tumors are highly aggressive, and none of the clinicopathological factors, including the presence of TILs and PD-L1 expression, correlated with GD2 expression. Nevertheless, in this study on patients with early-stage TNBC, we observed that patients with GD2-positive TNBCs showed significantly worse RFS than those with GD2-negative TNBCs, and GD2-positivity was found to be an independent unfavorable prognostic factor for RFS in TNBC. One possible reason is that GD2-positive cells may be resistant to conventional adjuvant chemotherapy and radiotherapy and persist durably, which strongly suggests the behavior of CSCs. Al-Hajj et al. reported that CD44highCD24low cells had breast cancer stem cell characteristics that were capable of initiating tumors and differentiating into other subtypes (23). Battula et al. reported that most of the CD44highCD24low cells in breast cancer cell lines expressed GD2 and had the characteristics of CSCs (19). Although we did not evaluate the expression of CD44/CD24 in tumor cells, our observations also suggest that GD2-positive cells may have cancer stem cell properties related to resistance to conventional chemoradiotherapy. We further examined RFS and OS only in patients who received postoperative adjuvant chemotherapy and observed that GD2-positive patients tended to have a worse prognosis, without statistical significance (data not shown). Unfortunately, the sample size was quite small, requiring further analyses with larger cohorts.
GD2 is a good antigen for immunotherapy of neuroblastoma. Anti-GD2 antibody has been clinically used for neuroblastoma and has shown prolonged survival when administered in combination with cytotoxic drugs (24). Additionally, although preclinical evaluation of dinutuximab (anti-GD2 antibody; ch14.18) against TNBC cells has been affirmatively reported (25), considering its efficacy and adverse events, this treatment strategy is unsatisfactory. Recently, advanced immunotherapy, such as GD2-targeting chimeric antigen receptor T (CAR-T) cell therapy, has been under development for refractory neuroblastoma (26).
One of the new findings in this study was that GD2-positive TNBCs with low amounts of infiltrating TILs showed worse prognosis than that of GD2-positive TNBCs with high TILs. This observation suggests that GD2-positive tumor cells exhibit aggressive behavior in the absence of tumor-reactive T cells, leading to the early recurrence of TNBC or short survival. Conversely, the infusion of GD2-positive tumor cell-reactive effector lymphocytes may be a promising treatment for GD2-positive TNBCs. However, it remains unclear whether the same strategy can be applied to refractory TNBCs. For examples, whereas GD2 expression in glioblastoma (GB) is preserved in both primary and metastatic tumors without antigen loss, which provides the advantage of GD2-targeted immunotherapy against GB (27), there are no data on GD2 expression in metastatic breast cancer. This point should be clarified in future studies to extend the clinical application of GD2-CAR-T cell therapy for the treatment of advanced-stage TNBC. In addition, although the recent report showed the repression of natural killer cell activity mediated by GD2-sialic acid immunoglobulin-like lectin (SIGLEC) interaction (28), until now, the machinery of GD2 mediated suppression of host anti-cancer immunity has not been fully elucidated. The mechanisms of potential targeted for GD2 therapy in TNBCs need to be better clarified.
This study had several limitations. One is a small sample size from a single Institute. Further analyses with larger cohorts are required to confirm that GD2 is a prognostic factor in TNBCs. The next limitation is the lack of standardized evaluation methods for GD2 using FFPE primary samples. Since GD2 is a glycolipid, not a protein, GD2 expression could be attenuated during the deparaffinization, leading to the underestimation of GD2 expression compared to immunohistochemistry using frozen primary tumor tissues. In addition, because our Institute actively treats patients with stages I to III (most cases are early breast cancer), the presented data are largely exclusive to patients with early-stage TNBC. Conversely, our observations demonstrated that GD2 expression has a negative impact on the prognosis, even in early-stage TNBC, and a novel treatment strategy should be applied in this early phase to improve the prognosis of patients with TNBC.
Conclusion
In summary, we are the first to report the clinical significance of GD2 expression in Japanese patients with TNBC, particularly those with early-phase TNBC. We showed that GD2 is an independent poor prognostic factor for TNBC. Moreover, we also indicated the potential advantages of cellular immunotherapy targeting GD2, such as GD2-CAR-T cell therapy, towards treatment of TNBC. In this context, we are currently preparing for a clinical trial using our novel GD2 CAR-T cell therapy.
Acknowledgements
We thank Dr. Hiroshi Shiku (Department of Personalized Cancer Immunotherapy Center for Comprehensive Cancer Immunotherapy, Mie University Graduate School of Medicine, Japan; d, September 4, 2022) for the conception of the study and for the fruitful discussions. We would also like to thank Ms. Masako Ichishi (Department of Pathology, Mie University Hospital, Japan) for immunohistochemical staining.
Footnotes
Author’s Contributions
KS contributed to the conception, design, and writing of the article. CH contributed to the analysis, interpretation, and writing of the manuscript. KN contributed to the analysis and interpretation. YK and HY contributed to the pathological analysis. M. Ishitobi, M. Ishihara, and TM contributed to data collection. TO and IT supervised this study. HF contributed to the conception, and supervised the study, and reviewed and edited the manuscript.
Funding
This study was supported in part by research grants from JSPS (KAKENHI, No. 21K08593), OKASAN-KATO FOUNDATION (Mie, Japan), and Ekiden for Life (Charity Load Marathon for Cancer Patients in Mie, Japan).
Conflicts of Interest
Kanako Saito received honoraria from Daiichi-Sankyo, Chugai, Kyowa Kirin, Eli Lilly, Novartis, Eisai, and Pfizer outside the submitted work. Mikiya Ishihara received honoraria from Chugai, Eisai, MSD, Ono, Daiichi-Sankyo, Otsuka, and Eli Lilly outside the submitted work. Hiroshi Fujiwara is a member of the department funded by T Cell Nouveau Inc. The other Authors have no conflicts of interest.
- Received June 14, 2023.
- Revision received July 16, 2023.
- Accepted July 19, 2023.
- Copyright © 2023 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).