Though breast cancer was not though to be immunogenic cancer types in previous compared with other cancer types, the role of specific immune cells in clinical and prognostic of breast cancer still in controversial. In the present research, we used two different breast cancer cohort TCGA and METABRIC with confirmed clinicopathology and RNA-seq expression data to analysis the immune cells expression pattern and their relationship with overall survival. The density of 22 main TIICs in breast cancer were evaluated by CIBERSORT algorithm and we found that though some immune cells shows prognostic ability in TCGA and METABRIC cohorts, only macrophage M2 were testified as an independent prognostic factor by multivariate Cox analysis in both two cohort. Based on the independent prognostic factors including M2 macrophages we build a nomogram model to further improve the prognostic ability in breast cancer. By using the nomogram risk, we can best separate patients into high and low risk groups in all patients as well as in patients in different TNM stages. The nomogram low and high risk groups showed significant difference in some immune features and the predicted clinical response to immune checkpoint blockade were higher in high risk group patients.
The TME comprised by various cells and the predominant cells recruited to and activated in it are immune cells[25, 26]. Immune cells in TME are closely related to tumor processes such as growth, angiogenesis and metastasis[27–29]. The tumor progression may be effected by imbalanced host immune response, and inverse, the quantity and phenotype of immune cells can be also effected by tumor cells[30, 31]. Therefore, testing the immune status and found the possible way to change the infiltrated immune cells type in the TME may help us dope out strategies to improve the response rate of immunotherapy. Tumor infiltrating lymphocytes (TILs) account for most of immune cells in TME which included T cells, B cells and NK cells are highly heterogeneous in tumors and in some cases showed opposite functions and effects on survival. In TCGA cohort, those TILs showed no significant difference in overall survival, while in METABRIC cohort the T cells CD4 memory activated, T cells regulatory, NK cells resting and another TILs T cells gamma delta showed prognostic ability. The T cells, which reported to produce some regulatory cytokines such as TGF-β and IL‐10 and can be recruited to the cancer environment via CCL22/CCR4, were able to escape from immune surveillance, result in tumor immune tolerance and ultimately reduced tumor survival[33–35]. TILs play different functions in breast cancer and more research should be done in the future to help us take advantage of it.
Macrophages are also important in infiltration immune cells in cancer and it can be divided into M0, M1 and M2 subtypes. M0 subtype, which is inactivated and not have inflammatory or tumor-related function, can through different activation pathways polarized to two distinctly immunoregulatory imparity subtypes M1(pro-inflammatory) and M2(anti-inflammatory). M2 macrophages can secrete cytokines to inhibit inflammation and inhibit the proliferation and differentiation of T cells, it can also promote tumor proliferation and some cancer associated process such as angiogenesis[36–40]. TCGA cohort exhibit that the density fraction of macrophage M0 and M1 were higher in tumor tissues than normal tissues, while M2 were on the opposite. In breast cancer tissues the fraction of M2 were higher than M1, which means in breast cancer the macrophage M0 tend to differentiate towards M2 subtype. We found that only M2 macrophage neither M0 nor M1 macrophages is an independent risk factor to breast cancer patients. Many studies also report macrophages were associated with shorter survival in breast cancer patients and that may due to high M2 subtype polarization[11, 12, 41]. Based on the shift of macrophages, several approaches such as eliminate the macrophages, blockade the M2 subtype shift, reprogramming the macrophages shift to M1 subtype and inhibit the recruitment of monocyte into cancer may try to as possible anti-tumor therapies in cancer.
We build a nomogram model including the density level of macrophages M2 to better predict the OS in breast cancer. The patients were appropriate separated into high risk group and low risk group according to the nomogram score. We found that the low and high risk group were different in some immune signatures. The content of immune cell subtypes were different between low and high risk groups and as Thorsson demonstrate that the tumor with macrophage dominated especially with higher M2 macrophage content and low lymphocytes infiltrate had poor prognosis. DNA aneuploidy, as reported were related to cell proliferation and poor tumor differentiation. In our study, the high risk group patients had higher Aneuploidy score, proliferation and worse prognosis which consistent with the results from a large meta-analysis that aneuploid breast cancer patients had poor DFS and OS. The immunotherapy is an attempt in breast cancer patients when there is no other way out and the current immunotherapy response were relatively poor. It is very important to seek out the potential molecular mechanism of immunotherapy responsiveness and those who are suitable to receive immunotherapy. Tumor mutation burden (TMB) which was calculated as total count of variants divided by the whole length of exons is a novel biomarker for predicting immunotherapy effect. In small-cell lung cancer patients, high TMB means better treatment prognosis either in single or combined immunotherapy. In breast cancer TMB were associated with prognostic immune subclasses and those patients with high TMB and favorable immune-infiltrate subtypes had better prognosis. Neoantigens are a kind of peptides that are generated from somatic mutations and it’s immunotherapy response were still contradict in tumors. In order to release tumor-specific immune responses, the Neoantigens can result T cells recognizing cancer cells through T-cell receptors (TCR) interacted with major histocompatibility complex (MHC)[48, 49]. Our high risk group patients had higher Neoantigens but lower TCR score than low risk group patients so to improve specificity and sensitivity of neoantigen–MHC complexes may more helpful in those patients. The TIDE score that can indicate tumor immune escape and predict response to immune checkpoint blockade such as anti-PD1 and anti-CTLA4 is mainly used for melanoma and non-small cell lung cancer, but can still try to applied in other tumors. The high risk group patients had lower TIDE score means they experience less immune evasion and may more likely to benefit from immune checkpoint blockade therapy. From the above, we found that though the high risk group patients had worse prognosis they may benefit more from immunotherapy.
Twenty gene selected by multiple methods were supposed to be hub genes between high risk and low risk groups. Fourteen out of the twenty genes were immune genes and among the 14 immune hub genes, CXCR2 is the most studied one in cancer and immunology. Knockdown CXCR2 in mammary tumor cells reduced cell invasion, metastasis and enhanced the toxicity of chemotherapies. CXCR2 positive regulate neutrophil recruitment to cancer and can promote tumor metastasis through TNFα-activated mesenchymal stromal cells (MSCs). Macrophages also reported can increase expression of inflammatory chemokines and promoting tumor cell migration and invasion through activation of CXCR2. Therefore, targeting CXCR2 may represent a novel therapeutic strategy for breast cancers. Other genes, such as vasoactive intestinal polypeptide (VIP) could inhibit the cytotoxicity of NK cells and could inhibit its expressions of some immune-associated genes like NKG2D and NF- κB. Neurotensin (NTS) belong to Neuropeptide G protein-coupled receptors (GPCRs) are over-expressed in numerous cancer cells and were reported to stimulate tumor growth. Further study could try to focus on these hub genes to regulate tumor microenvironment and immune status, which may help to found some new therapeutic methods for breast cancer. transcription factors and immune genes network in low and high risk group showed significant difference. In high risk group, higher transcription factor TAT expression co-upregulated with hazard immune genes and co-downregulated with protective immune genes while the transcription factor PPARD were in adverse. Methods that inhibit the gene expression of TAT or improve the gene expression of PPARD may change the immune gene expression and improve the high risk group patients survival as well. Further GO analysis of these survival related immune genes in high risk group showed that they were enriched in cell population proliferation and immune-relevant biological pathways, such as immune response and response to stimulus, this can partly explain the high risk group had relatively poor survival but may have better immune response than low risk group patients.
Though we have used two different cohort and large patients to get and verity our results, there are still some limitations in our study. The first is data missing, as the immune cells infiltration data were calculated by CIBERSORT algorithm and some patients were excluded because of the p value of predictive accuracy higher than 0.05. The second is the population race imbalance. The TCGA is an America project and the METABRIC is a Canada-UK project, as most of the patients included in our study were white, the results should be further researched in people of other races. The last is the study failed to use experiment to prove the results and exploring the underlying mechanisms of the results. Therefore, future research needs to do to better testify our results.