Immune-Phenotyping and Transcriptomic Profiling of Peripheral Blood Mononuclear Cells From Patients With Breast Cancer: Identification of a 3 Gene Signature Which Predicts Relapse of Triple Negative Breast Cancer

Abstract

Background: Interactions between the immune system and tumors are highly reciprocal in nature, leading to speculation that tumor recurrence or therapeutic resistance could be influenced or predicted by immune events that manifest locally, but can be detected systemically. Methods: Multi-parameter flow cytometry was used to examine the percentage and phenotype of natural killer (NK) cells, myeloid-derived suppressor cells (MDSCs), monocyte subsets and regulatory T (Treg) cells in the peripheral blood of of 85 patients with breast cancer (50 of whom were assessed before and after one cycle of anthracycline-based chemotherapy), and 23 controls. Transcriptomic profiles of peripheral blood mononuclear cells (PBMCs) in 23 patients were generated using a NanoString gene profiling platform. Results: An increased percentage of immunosuppressive cells such as granulocytic MDSCs, intermediate CD14++CD16+ monocytes and CD127negCD25highFoxP3+ Treg cells was observed in patients with breast cancer, especially patients with stage 3 and 4 disease, regardless of ER status. Following neoadjuvant chemotherapy, B cell numbers decreased significantly, whereas monocyte numbers increased. Although chemotherapy had no effect on the percentage of Treg, MDSC and NK cells, the expression of inhibitory receptors CD85j, LIAR and NKG2A and activating receptors NKp30 and NKp44 on NK cells increased, concomitant with a decreased expression of NKp46 and DNAM-1 activating receptors. Transcriptomic profiling revealed a distinct group of 3 patients in the triple negative breast cancer (TNBC) cohort who expressed high levels of mRNA encoding genes predominantly involved in inflammation. The analysis of a large transcriptomic dataset derived from the tumors of patients with TNBC revealed that the expression of CD163, CXCR4, THBS1 predicted relapse-free survival. Conclusions: The peripheral blood immunome of patients with breast cancer is influenced by the presence and stage of cancer, but not by molecular subtypes. Furthermore, immune profiling coupled with transcriptomic analyses of peripheral blood cells may identify patients with TNBC that are at risk of relapse after chemotherapy.

Keywords: MDSC; NK cells; breast cancer; flow cytometry; immune gene signature; monocytes; nanostring; treg.

Figures

Figure 1

There are significantly more Treg cells in the peripheral blood of patients with breast cancer and these exhibit an overall more activated/suppressive phenotype: PBMCs from patients with breast cancer and individuals with no-known disease were rapidly defrosted, allowed to rest for 2 h at 37°C, washed, and incubated with an Fc blocking reagent before being stained with a cocktail of mAbs reactive with cell surface antigens. Intracellular staining was undertaken following treatment with Fix/Perm and incubation with a FoxP3 mAb. Data were acquired on a Beckman Coulter Gallios™ flow cytometer and analyzed using Beckman Coulter Kaluza™ software. A two-tailed Mann Whitney test was performed to determine any significant differences in the measured parameters between healthy individuals and patients with breast cancer. A Wilcoxon matched-pairs two tailed test was used to assess the significance between Pre/Post-chemotherapy. A Kruskal-Wallis test was used to assess the significance of any differences in the measured parameters between the different disease stages, and individuals with no-known disease (**P < 0.005). There was a significantly higher number of CD3+CD4+CD127negCD25+FoxP3+ (regulatory T (Treg) cells) in the PBMCs of patients with breast cancer (A). No differences were found between healthy controls and cancer patients for CD39neg/ICOSpos(B) or CD39pos/ICOSpos(C) or CD39pos/ICOSneg(E), however there were significantly fewer CD39neg/ICOSneg Treg cells in the peripheral blood of patients with breast cancer (D).

Figure 2

Greater percentage of inflammatory monocytes and granulocytic MDSCs (gMDSCs) in the peripheral blood of patients with breast cancer than in age-matched volunteers. PBMCs from patients with breast cancer and individuals with no-known disease were rapidly defrosted, allowed to rest for 2 h at 37°C, washed and then incubated with an Fc blocking reagent before being stained with a cocktail of mAbs reactive with cell surface antigens. Freshly isolated PBMCs from a proportion of patients were then stained for the expression of intracellular Arginase. Data were acquired using a Beckman Coulter Gallios™ flow cytometer and analyzed using Beckman Coulter Kaluza™ software. A two-tailed Mann Whitney test was performed to assess differences between patients with cancer and their corresponding controls, and a Wilcoxon matched-pairs two tailed test was used to assess the influence of chemotherapy. A Kruskal-Wallis test was used to assess the significance of any differences in the measured parameters between the different disease stages, and individuals with no-known disease (*P < 0.05; **P < 0.005; ***P < 0.0005). A significantly higher percentage of pro-inflammatory monocytes within the CD14+ population was present in the PBMCs from patients with breast cancer (A). Although no difference was found in the proportion of CD11b+CD33+ within DRneg cells, a significantly higher percentage of DRnegCD11b+CD33+ cells were CD15+ in patients with breast cancer (B). A significantly higher percentage of freshly isolated MDSCs from patients with breast cancer were Arginase+ cells, and the percentage of Arginase+ MDSCs was positively associated with disease stage (B). Similar proportions of CD56dimCD16+ and CD56brightCD16− NK cells were present in patients with breast cancer and healthy controls; however there were significant differences in the intensity of CD56 expression on both NK cell subsets (C).

Figure 3

Chemotherapy affects the absolute number of B cells and monocytes (A), but not the percentage of Treg cells, MDSCs and NK cells (B). Peripheral blood (5–20 mL) was collected into EDTA Vacutainers™ and processed within 2 h. An aliquot of whole blood was used to obtain T cell, B cell, monocyte and NK cell counts using BD Trucount™ beads (BD Biosciences) following the manufacturers protocol. For this, 100μL of blood was mixed directly in the BD Trucount™ bead tube with mAbs to CD3, CD8, CD4, CD19, CD56, CD45, and CD14. Tubes were incubated for 15 min at room temperature, protected from light, after which erythrocytes were lysed via a 15-min incubation at room temperature in BD Pharm Lyse™ lysing solution. Cells were analyzed by flow cytometry within 1 h. No significant differences were found in the number of the different immune cells between patients and healthy donors. B cells and monocytes were the only two cell types significantly affected by one course of chemotherapy (A). A paired t-test was used to compare the samples. (*P < 0.05; **P < 0.005; ***P < 0.0005) and (C) PBMCs from patients with breast cancer and individuals with no-known disease were rapidly defrosted, allowed to rest for 2 h at 37°C, washed, and incubated with an Fc blocking reagent before being stained with a cocktail of mAbs reactive with cell surface antigens. Intracellular staining was undertaken following treatment with Fix/Perm and incubation with a FoxP3 mAb. Data were acquired using Beckman Coulter Gallios™ flow cytometer and analyzed using Beckman Coulter Kaluza™ software. A two-tailed Mann Whitney test was performed to assess significant differences in the phenotypic profiles in patients with breast cancer and their corresponding controls, and a Wilcoxon matched-pairs two tailed test was used to assess the influence of chemotherapy. A Kruskal-Wallis test was used to assess the significance of any differences in the measured parameters between the different disease stages, and individuals with no-known disease (*P < 0.05; **P < 0.005; ***P < 0.0005). The percentage of Treg cells was not affected by one round of chemotherapy (data not shown). The abosulte number of monocytes (CD45+CD3negCD14+) was found to be significantly increased by chemotherapy (B) while that of B-cells (CD45+CD3negCD19+) was significantly decreased, whereas only the MFI of CD56dim NK cells was affected by chemotherapy.

Figure 4

Peripheral blood mononuclear cells (PBMCs) from a distinct subgroup of patients with TNBC differentially expressed immune-related genes. Total RNA was extracted from frozen PBMCs frozen PBMCs from 23 randomly selected patients, 14 with TNBC and 9 with luminal-A breast cancer (5 × 106 – 107 cells) using a Qiagen RNAesasy kit according to the manufacturer's instructions. Gene expression profiling was carried out using a NanoString nCounter™ FLEX platform (NanoString Technologies Inc.). The Pan-Cancer Immune Profiling Panel was used to assess immune function at the mRNA level according to the manufacturer's protocol. Raw data were acquired using the nCounter™ digital analyser with 555 fields of view (fov) being collected for each sample. Differential expression and further analysis was performed using Human PanCancer Immune Profiling Advanced Analysis (version 1.0.36). Principle component analysis (PCA) was used for dimensionality reduction and assessing sample grouping. Genes with a false discovery rate below 0.05 were considered as significant. No differences in the mRNA expression profiles in PBMCs from 9 patients with Luminal A disease and 11 of the patients with TNBC were observed. However, PBMCs from 3 out of the 14 patients with TNBC demonstrated a clearly distinct immune gene expression profile (A). (B,C) Show the 20 top most differentially regulated genes between these two groups [PBMCs from 11 patients with TNBC (TNBC1) and the 3 distinct TNBC (TNBC2)]. High expression of CD163 combined with high expression of CXCR4 and low expression of THBS1 was associated with better RFS (p = 0.00012). Interestingly, this combination of genes was not associated with any difference in RFS for patients with ER positive breast cancer (p = 0.7) (D).