Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 May 1.
Published in final edited form as: Surgery. 2020 Sep 19;169(5):1234–1239. doi: 10.1016/j.surg.2020.07.052

Prevalence and clinical relevance of tumor-associated tissue eosinophilia (TATE) in breast cancer

Konstantinos Chouliaras a,#, Yoshihisa Tokumaru a,#, Mariko Asaoka a, Masanori Oshi a, Kristopher M Attwood b, Kazuhiro Yoshida c, Takashi Ishikawa d, Kazuaki Takabe e,f,g,h,*
PMCID: PMC7969474  NIHMSID: NIHMS1630695  PMID: 32958266

Abstract

Background:

Tumor-associated tissue eosinophilia (TATE) has been associated with outcomes in a variety of solid tumors; however, its role in breast cancer is not well defined. We hypothesized that tumor-associated tissue eosinophilia is associated with a high mutation and neoantigen load, and we assessed its correlation with cancer outcomes.

Methods:

The Cancer Genome Atlas was analyzed for eosinophil signatures in breast cancer specimens. Descriptive analyses were performed, including the tumor-infiltrating cell composition using CIBERSORT, cytolytic activity score, and gene set enrichment analysis. Overall survival and disease-free survival were calculated using the Kaplan-Meier method.

Results:

Out of 1069 cases analyzed, 40 (3.7%) had tissue eosinophils (the tumor-associated tissue eosinophilia group). Tumor-associated tissue eosinophilia was noted in 32.5% luminal, 5% HER2-positive, and 15% TNBC subtypes. The single nucleotide variant–neoantigen load was significantly higher in the tumor-associated tissue eosinophilia group (P = .005), with a higher nonsilent mutation rate (P = .01). The tumor-associated tissue eosinophilia group had lower cytolytic activity (P = .02) but had enriched MYC-targeted (P = .002), E2F-targeted (P = .04), deoxyribonucleic acid repair (P = .03), and unfolded protein response gene sets (P = .05). Tumor-associated tissue eosinophilia was associated with a trend toward improved disease-free survival (P = .06) but presented no differences in overall survival (P = .56).

Conclusion:

Tumor-associated tissue eosinophilia was noted in 3.7% of breast cancers and was associated with a higher single nucleotide variant–neoantigen load and nonsilent mutation rate, similar to that of tumor-infiltrating lymphocytes in the triple-negative subtype. However, a lower cytolytic activity score and enriched cell proliferation–related gene sets implicate different roles for tumor-associated tissue eosinophilia than for tumor-infiltrating lymphocytes.

TOC summary

Tumor-associated tissue eosinophilia was associated with higher SNV-neoantigen load and nonsilent mutation rate, similar to tumor-infiltrating lymphocytes. The importance of this finding is in the further clarification of immune cells in breast cancer and its impact on cancer-specific outcomes.

Introduction

The interaction of different cancer cells with the immune system is in the forefront of oncology research. A better understanding of the tumor immune microenvironment is essential to unlocking novel treatments. Stromal tissue-infiltrating lymphocytes have been shown to correlate with improved disease-free survival and overall survival after neoadjuvant chemotherapy in triple-negative and HER2-positive breast cancer.1 Specifically, the presence of CD8+ T cells with features of tissue-resident memory has been associated with improved outcomes.2 The exact T-lymphocyte subset involved in this effect and the possible roles of other immune cells have not been elucidated.

Eosinophils are granulocytic leukocytes derived from bone marrow progenitor cells that demonstrate pleiotropic activities in several types of cancer. Eosinophils are antigen-presenting cells that also secrete cytokines that stimulate CD8+ T cells in cancer tissues.3 The role of tumor-associated tissue eosinophilia (TATE) is variable in different cancer types; for instance, TATE is associated with a poor prognosis in Hodgkin’s lymphoma but has the opposite effect in colon and prostate cancers.4 However, the role of eosinophils in breast cancer has not been well elucidated, partly because eosinophil infiltration is rarely detected by routine hematoxylin-eosin staining of breast cancer specimens.5 On the other hand, a pilot study reported 23 out of 26 positive samples using monoclonal mouse antihuman eosinophil peroxidase antibodies, indicating significant eosinophil activity in breast cancer.6 In a different study of 44 breast cancer/ ductal carcinoma in situ specimens, eosinophils were found in abundance adjacent to the core biopsy site more than 30 days after the biopsy.7 We theorized that these data suggest the possibility that the role of TATE in breast cancer is uncertain because of their not being identified by conventional detection methods.

Recently our group has been pursuing computational biological approaches to analyze the tumor immune microenvironment.8-15 We use publicly available large patient cohorts that have gene expression data linked to clinical data and apply computational algorithms to estimate the number of cells of interest by calculating the gene expression levels of cell surface markers in the bulk tumor. Because we use gene expression data, no bias is caused by the selection of antibodies, such as can occur in immunohistochemistry analysis. This approach also allows the detection of very small numbers of certain cell types, which is the case for eosinophils. Further, our method does not require newly collected fresh tissue, unlike flow cytometry, which is the gold standard for the analysis of immune cells.

Here, we hypothesized that TATE plays a significant role in tumor microenvironments associated with a high mutation and neoantigen load, leading to activation of eosinophils and their effector pathways.

Methods

Patient cohort

The Cancer Genome Atlas (TCGA) data set breast cancer cohort16 was used to obtain clinicopathologic and genomic data through cBioportal as previously described.9,10,12-14,17,18 A total of 1,069 patients were included in this study. Institutional review board approval was waived because TCGA is a publicly accessible, deidentified database.

Immune cell composition and cytolytic activity score

We used a computational algorithm, CIBERSORT,19 to estimate the cell compositions of the immune cells as previously described.8-15 There was no minimum threshold for the presence of TATE because it was used as a binary variable: detectable versus nondetectable. Additionally, the cytolytic activity (CYT) score was calculated using the expression values of granzyme A and perforin-1 as described in detail in previous publications.8-14,20

Gene set enrichment analysis with TCGA

Software provided by the Broad Institute was used for gene set enrichment analysis. Comparisons between the TATE and non-TATE groups were performed using the Hallmark gene sets.

Statistical analysis

Statistical analysis of the categorical clinicopathologic parameters was performed by Fisher exact test based on the number of subjects analyzed, and continuous variables were compared by analysis of variance or Student’s t test. Exploratory analyses of immune phenotypes were also performed. The observed between-group effect sizes were estimated using Cohen’s d, where d ≤ 0.2, 0.2 < d < 0.8, and d ≥ 0.8 correspond to small, medium, and large effects, respectively.21 Overall survival (OS) and disease-free survival (DFS) estimates were calculated using the Kaplan-Meier method. All statistical analyses were performed using R software version 3.6.1 and Bioconductor Version 3.10.

Results

Clinicopathologic characteristics of TATE-positive breast cancer patients

Out of 1,069 breast cancer cases identified in the breast cancer TCGA cohort, 40 (3.7%) were detected by CIBERSORT to have tissue eosinophils (the TATE group). The TATE versus non-TATE groups, respectively, included 29 versus 710 Caucasians (72.5% vs 69%), 7 versus 172 African Americans (17.5% vs 16.7%), 0 versus 61 Asians (5.9%), and 4 versus 86 patients of not otherwise specified ethnicity (10% vs 8.4%) (P = .35). The distributions of patients in the TATE versus non-TATE groups based on American Joint Committee on Cancer staging were, respectively, 11 versus 167 stage I (27.5% vs 16.2%), 21 versus 584 stage II (52.5% vs 56.8%), 8 versus 236 stage III (20% vs 22.9%), 0 versus 18 stage IV (17.5%), and 0 versus 24 unspecified TNM stage (23.3%) (P = .37). In the TATE versus non-TATE groups, respectively, there were 13 versus 411 estrogen receptor–positive/HER2-negative (32.5% vs 40%), 2 versus 38 estrogen receptor–positive/HER2-positive (5% vs 3.7%), and 6 versus 108 triple-negative tumors (15% vs 10.5%) (P = .263) and 13 versus 361 unknown receptor status (32.5% vs 35%) (P = .26).

Immune cell composition of TATE-positive breast cancer

Because eosinophils are one type of immune cells, we were interested to identify which other immune cells had infiltrated in TATE-positive breast cancer. The TATE group had prevalent monocytes (P < .001, Cohen’s d = 0.13) and T-follicular helper cells (P = .005, d < .001) compared with the non-TATE group (Fig 1). Naïve B cells (P =.045, d = 0.25), resting mast cells (P = .014, d = 0.04), and resting CD4 memory T cells (P = .03, d = 0.12) were less prevalent in the TATE versus non-TATE groups (Fig 2).

Fig 1.

Fig 1.

Open in a new tab

Boxplots of cell types that predominate in the TATE group and a nonsilent mutation rate comparison.

Fig 2.

Fig 2.

Open in a new tab

Boxplots of cell types that predominate in the non-TATE group and comparisons of transforming growth factor-β (TGF-β) responses and cytolytic activity (CYT) scores.

TATE-positive breast cancer has high levels of mutations and neoantigens

Cancers with a high mutation rate generate high levels of neo- antigens that lead to the attraction of TILs to tumors. Thus we hy- pothesized that TATE-positive breast cancer has a high mutation rate and neoantigens. The nonsilent mutation rate and single nucleotide variant (SNV) neoantigens were significantly higher in the TATE group (P = .01, d = 0.18 and P = .005, d = 0.09, respectively), whereas the transforming growth factor-β (TGF-β) response P = .005, d = 0.01) and the CYT score (P = .02, d = 0.14) were higher in the non-TATE group (Figs 1, 2).

Cell proliferation–related gene sets enriched in TATE-positive breast cancer

We previously found that breast cancer with a high mutation rate has an aggressive biology. Given that the TATE group had a higher mutation rate, we hypothesized that it is associated with an aggressive biology. Gene set enrichment analysis revealed enriched MYC (P = .002), E2F (P = .04), DNA repair (P = .03), and unfolded protein response target (P = .047) gene sets in the TATE group (Fig 3). The angiogenesis gene set (P = .03) was enriched in the non-TATE group (Fig 4).

Fig 3.

Fig 3.

Open in a new tab

Enrichment plot depicting hallmark gene sets enriched in the TATE group (MYC, E2F, DNA repair, and unfolded protein response targets).

Fig 4.

Fig 4.

Open in a new tab

Enrichment plot depicting the angiogenesis hallmark gene set that is enriched in the non-TATE group.

TATE-positive breast cancer has a trend for better disease-free survival

Given the favorable tumor immune microenvironment, we were interested to determine whether patients with TATE-positive breast cancer have longer survival times. The median DFS was 30.17 months (TATE) versus 25.32 months (non-TATE) (P = .06), and the median OS was 25.32 months (TATE) versus 27.50 months (non-TATE) (P = .56) (Figs 5, 6).

Fig 5.

Fig 5.

Open in a new tab

Kaplan-Meier curves of disease-free survival and overall survival comparing the TATE and non-TATE groups.

Fig 6.

Fig 6.

Open in a new tab

Kaplan-Meier curves of disease-free survival (DFS) and overall survival (OS) comparing the TATE and non-TATE groups by American Joint Committee on Cancer stage.

Discussion

In this analysis of the TCGA database, TATE was noted in 3.7% of breast cancer cases, predominantly in early-stage cancers without an obvious predominance of a specific subtype based on receptor status. This represents one of the largest studies showing the presence of eosinophil infiltration in breast cancer specimens, more fully addressing the controversy of TATE in breast cancer.5,6 This is likely reflective of the difficulty of identifying the presence of TATE in tissues routinely stained with hematoxylin and eosin. TATE detection has been found to be improved with the help of additional immunohistochemical stains. Our computational biological approach enabled us to overcome this challenge.

The role of TATE has not been clearly delineated in breast cancer; however, a retrospective analysis of 112 predominantly triple-negative breast cancers found a higher complete pathologic response and improved outcomes in patients with a higher peripheral eosinophil count.22 In a different study of tissue microarrays, the presence of mast cells correlated with low-grade tumors, whereas no eosinophils were identified.5 In our study, although not statistically significant, TATE was not identified in any stage III or stage IV cases, an observation that could reflect the role that eosinophils play in breast cancer either as regulator or effector cells.

A higher SNV neoantigen load and nonsilent mutation rate were correlated with the presence of TATE in this cohort, which is likely indicative of a higher immune response in these tumors. Different immunogenic profiles of different cancer types, including breast cancer, have been reported in recent publications.23,24 We thus postulate that eosinophils may be a marker of higher immunogenic potential associated with a higher neoantigen load and a nonsilent mutation rate.

In the gene set analysis, DNA repair gene sets were found to be enriched, which is an interesting finding in light of recently published data reporting the prevalence of eosinophils adjacent to core biopsy sites with an observed increase in tumor proliferation.7 It is possible that eosinophils are attracted to the biopsy site for a subset of cases and that they are actively altering the tumor microenvironment. Eosinophils have been implicated in angiogenesis partially because of the preformed vascular endothelial growth factor in their granules25; however, this effect has not been directly evaluated in breast cancer. In our analysis, angiogenesis gene sets were downregulated in the TATE group, casting doubt on the role of tissue eosinophils in breast cancer angiogenesis.

Although a significant role of eosinophils in pubertal breast development has been identified,26 their exact role in breast cancer has not been elucidated. Experimental data have indicated that breast development during puberty is dependent on the interactions between eosinophils and macrophages at the level of terminal end buds, with significant, likely inhibitory effects on mammary ductal branching.27 Macrophages and T-follicular helper cells were found to be prevalent in the TATE cancer group, indicating an interaction between these 2 cell types in the tumor microenvironment. Furthermore, the CYT score and TGF-β response were found to be higher in the non-TATE group, which would imply that they likely act as modulators rather than as phagocytes. The extent and significance of the interaction between tissue-infiltrating lymphocytes and tumor eosinophils could provide more insight into the role of TATE in breast cancer.

The TCGA database, although it is a unique data repository, does not contain long-term follow-up data and therefore does not allow for precise comparisons of cancer-related survival outcomes. More specifically, even though there was a nonstatistical improvement in DFS, the OS curves of the TATE group parallel those of the non-TATE group. Moreover, several patients in the TATE group were censored from the DFS curve within the first 5 years, likely because of short follow-up, thus preventing definitive conclusions.

Therefore our group recently published works that provide a rationale for our finding that the outcomes of the TATE group were not worse despite an aggressive tumor biology. We found that breast cancers with high mutation rates had enriched cell-proliferation gene sets, similar to our findings for the TATE group, but did not have worse survival because of the accumulation of anticancer immune cells.28 In another study, highly proliferative KRAS signaling–enriched triple-negative breast cancers were found to have a heightened inflammatory response, cytolytic activity, and infiltrating lymphocyte activity, which led to improved survival.29 Highly proliferative tumors, such as KRAS-enriched tumors, also respond better to chemotherapy, which, together with favorable tumor cell infiltration, contributes to better outcomes.

A significant number of cases in our data set were missing receptor status data, which prevents any clear comparisons of the presence of TATE among the different breast cancer subtypes. Furthermore, the small number of specimens with TATE prevents definitive conclusions because of the decreased statistical power, but our findings could serve as the basis for a large multicenter study to clarify the role of eosinophils in breast cancer.

The exact impact of the presence of eosinophils on breast cancer has not been elucidated in the literature, but our results indicate they have an important role in the breast cancer microenvironment. The exact trigger of eosinophil attraction remains unclear but may be correlated with a higher neoantigen load and nonsilent mutation rate in a small subset of cases. In this cohort, TATE was found to have an important regulatory role in interacting with macrophages and associated with decreased cytolytic action. In summary, eosinophils were found to be present in 3.7% of a population of 1,069 breast cancer cases in the TCGA database. Patient outcomes between the TATE and non-TATE groups were similar, but MYC, E2F, DNA repair, and unfolded protein response gene sets were found to be enriched in the TATE group.

Acknowledgments

Funding/Support

Supported by National Institutes of Health grant R01CA160688 (Dr. Takabe) and National Cancer Institute Roswell Park Cancer Institute Support Grant award number P30CA016056.

Footnotes

Conflict of interest/Disclosure

none.

Presented at the 15th Annual Academic Surgical Congress (ASC), Orlando, FL, February 5, 2020.

References

  • 1.Denkert C, von Minckwitz G, Darb-Esfahani S, et al. Tumour-infiltrating lymphocytes and prognosis in different subtypes of breast cancer: a pooled analysis of 3771 patients treated with neoadjuvant therapy. Lancet Oncol. 2018;19: 40–50. [DOI] [PubMed] [Google Scholar]
  • 2.Savas P, Virassamy B, Ye C, et al. Single-cell profiling of breast cancer T cells reveals a tissue-resident memory subset associated with improved prognosis. Nat Med. 2018;24:986–993. [DOI] [PubMed] [Google Scholar]
  • 3.Davis BP, Rothenberg ME. Eosinophils and cancer. Cancer Immunol Res. 2014;2: 1–8. [DOI] [PubMed] [Google Scholar]
  • 4.Prizment AE, Vierkant RA, Smyrk TC, et al. Tumor eosinophil infiltration and improved survival of colorectal cancer patients: Iowa Women’s Health Study. Mod Pathol. 2016;29:516–527. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Amini RM, Aaltonen K, Nevanlinna H, et al. Mast cells and eosinophils in invasive breast carcinoma. BMC Cancer. 2007;7:165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Samoszuk MK NV, Gluzman I, Pham JH. Occult deposition of eosinophil peroxidase in a subset of human breast carcinomas. Am J Pathol. 1996;148(3). [PMC free article] [PubMed] [Google Scholar]
  • 7.Szalayova G, Ogrodnik A, Spencer B, et al. Human breast cancer biopsies induce eosinophil recruitment and enhance adjacent cancer cell proliferation. Breast Cancer Res Treat. 2016;157:461–474. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Narayanan S, Kawaguchi T, Peng X, et al. Tumor infiltrating lymphocytes and macrophages improve survival in microsatellite unstable colorectal cancer. Sci Rep. 2019;9:13455. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Takeshita T, Yan L, Asaoka M, Rashid O, Takabe K. Late recurrence of breast cancer is associated with pro-cancerous immune microenvironment in the primary tumor. Sci Rep. 2019;9:16942. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Katsuta E, Maawy AA, Yan L, Takabe K. High expression of bone morphogenetic protein (BMP) 6 and BMP7 are associated with higher immune cell infiltration and better survival in estrogen receptor positive breast cancer. Oncol Rep. 2019;42:1413–1421. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Narayanan S, Kawaguchi T, Yan L, Peng X, Qi Q, Takabe K. Cytolytic activity score to assess anticancer immunity in colorectal cancer. Ann Surg Oncol. 2018;25:2323–2331. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Okano M, Oshi M, Butash AL, et al. Triple-negative breast cancer with high levels of annexin A1 expression is associated with mast cell infiltration, inflammation, and angiogenesis. Int J Mol Sci. 2019;20(17). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.McDonald KA, Kawaguchi T, Qi Q, et al. Tumor heterogeneity correlates with less immune response and worse survival in breast cancer patients. Ann Surg Oncol. 2019;26:2191–2199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Takeshita T, Asaoka M, Katsuta E, et al. High expression of polo-like kinase 1 is associated with TP53 inactivation, DNA repair deficiency, and worse prognosis in ER positive Her2 negative breast cancer. Am J Transl Res. 2019;11: 6507–6521. [PMC free article] [PubMed] [Google Scholar]
  • 15.Katsuta E, Qi Q, Peng X, Hochwald SN, Yan L, Takabe K. Pancreatic adenocarcinomas with mature blood vessels have better overall survival. Sci Rep. 2019;9:1310. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Koboldt D, Fulton R, McLellan M, et al. Comprehensive molecular portraits of human breast tumours. Nature. 2012;490:61–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Takahashi H, Katsuta E, Yan L, Dasgupta S, Takabe K. High expression of annexin A2 is associated with DNA repair, metabolic alteration, and worse survival in pancreatic ductal adenocarcinoma. Surgery. 2019;166:150–156. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Terakawa T, Katsuta E, Yan L, et al. High expression of SLCO2B1 is associated with prostate cancer recurrence after radical prostatectomy. Oncotarget. 2018;9:14207–14218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Newman AM, Liu CL, Green MR, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015;12:453–457. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Rooney MS, Shukla SA, Wu CJ, Getz G, Hacohen N. Molecular and genetic properties of tumors associated with local immune cytolytic activity. Cell. 2015;160:48–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Cohen J Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988. [Google Scholar]
  • 22.Onesti CE, Josse C, Poncin A, et al. Predictive and prognostic role of peripheral blood eosinophil count in triple-negative and hormone receptor-negative/HER2-positive breast cancer patients undergoing neoadjuvant treatment. Oncotarget. 2018;9:33719–33733. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Zhang X, Kim S, Hundal J, et al. Breast cancer neoantigens can induce CD8(+) T-cell responses and antitumor immunity. Cancer Immunol Res. 2017;5:516–523. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Tran E, Ahmadzadeh M, Lu YC, et al. Immunogenicity of somatic mutations in human gastrointestinal cancers. Science. 2015;350:1387–1390. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Puxeddu I, Alian A, Piliponsky AM, Ribatti D, Panet A, Levi-Schaffer F. Human peripheral blood eosinophils induce angiogenesis. Int J Biochem Cell Biol. 2005;37:628–636. [DOI] [PubMed] [Google Scholar]
  • 26.Coussens LM, Pollard JW. Leukocytes in mammary development and cancer. Cold Spring Harb Perspect Biol. 2011;3(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Gouon-Evans V, Rothenberg ME, Pollard JW. Postnatal mammary gland development requires macrophages and eosinophils. Development. 2000;127: 2269–2282. [DOI] [PubMed] [Google Scholar]
  • 28.Takahashi H, Asaoka M, Yan L, et al. Biologically aggressive phenotype and anticancer immunity counterbalance in breast cancer with high mutation rate. Sci Rep. 2020;10:1852. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Tokumaru Y, Oshi M, Katsuta E, et al. KRAS signaling enriched triple negative breast cancer is associated with favorable tumor immune microenvironment and better survival. Am J Cancer Res. 2020;10:897–907. [PMC free article] [PubMed] [Google Scholar]

RESOURCES