Abnormal T regulatory cells (Tregs: FOXP3+, CTLA-4+), myeloid-derived suppressor cells (MDSCs: monocytic, granulocytic) and polarised T helper cell profiles (Th1, Th2, Th17) in women with large and locally advanced breast cancers undergoing neoadjuvant chemotherapy (NAC) and surgery: failure of abolition of abnormal treg profile with treatment and correlation of treg levels with pathological response to NAC

Abstract

Background: Host defences play a key role in tumour growth. Some of the benefits of chemotherapy may occur through modulation of these defences. The aim of this study was to define the status of regulatory cells in women with large and locally advanced breast cancers (LLABCs) undergoing neoadjuvant chemotherapy (NAC) and surgery.

Methods: Bloods were collected from patients (n=56) before, during and following NAC, and surgery. Controls (n=10) were healthy, age-matched females donors (HFDs). Blood mononuclear cells (BMCs) were isolated and T regulatory cells (Tregs) (n=31) determined. Absolute numbers (AbNs) of Tregs and myeloid-derived suppressor cells (MDSCs) were ascertained from whole blood (n=25). Reverse transcriptase polymerase chain reaction analysis determined Treg mRNA (n=16). In vitro production of Th1, Th2 and Th17 cytokines (n=30), was documented. Patients were classified as clinical responders by magnetic resonance mammography after two cycles of NAC and as pathological responders using established criteria, following surgery.

Results: Patients with LLABCs had significantly increased circulating Tregs (≥ 6 fold AbN and percentage (%)) and MDSCs (≥ 1.5 fold AbN (p=0.025)). Percentage of FOXP3+ Tregs in blood predicted the response of the LLABCs to subsequent NAC (p=0.04). Post NAC blood Tregs (%) were significantly reduced in patients where tumours showed a good pathological response to NAC (p=0.05). Blood MDSCs (granulocytic, monocytic) were significantly reduced in all patients, irrespective of the pathological tumour response to chemotherapy. NAC followed by surgery failed to restore blood Tregs to normal levels. MDSCs, however, were reduced to or below normal levels by NAC alone. Invitro Th1 profile (IL-1β, IL-2, INF-γ, TNF-α) was significantly reduced (p ≤ 0.009), whilst Th2 (IL-4, IL-5) was significantly enhanced (P ≤ 0.004). Th1 and Th2 (IL-5) were unaffected by NAC and surgery. IL-17A was significantly increased (p ≤ 0.023) but unaffected by chemotherapy and surgery.

Conclusion: Women with LLABCs have abnormal blood regulatory cell levels (Tregs and MDSCs) and cytokine profiles (Th1, Th2, Th17). NAC followed by surgery failed to abolish the abnormal Treg and Th profiles. There was a significant correlation between the circulatory levels of Tregs and the pathological response of the breast cancers to NAC.

Figure 1

(A) Correlation between % of regulatory FOXP3⁺ (CD4⁺ CD25⁺ CD127^- FOX P3⁺) T cells before and after NAC and subsequent histological response in the breast following NAC (n = 16), graded as complete (cPR, 5), very good (VG, 4), partial (3), poor and no pathological response (2 + 1) in the breast following NAC. Values shown are means ± standard deviations. Statistical analysis: Spearman’s correlation coefficient Rho = −0.620, p = 0.04 (2-tailed) (pre-chemotherapy) and Rho = −0.799, p = 0.05 (2-tailed) (post-chemotherapy) (B) Correlation between % of regulatory CTLA-4⁺ (CD4⁺ CD25⁺ CD152⁺) T cells before and after NAC and subsequent histological response in the breast following NAC (n = 16), graded as complete (cPR, 5), very good (VG, 4), partial (3), poor and no pathological response (2 + 1) in the breast following NAC. Values shown are means ± standard deviations. Statistical analysis: Spearman’s correlation coefficient Rho = 0.260, p = 0.78 (NS) (pre-chemotherapy) and Rho = −0.794, p = 0.01 (2-tailed) (post-chemotherapy).

Figure 2

Gene expression in LLABCs and HFDs, (A) FOXP3, (B) CTLA-4: Mean FOXP3 and CTLA-4 gene expression over β actin in the circulation in women with LLABCs (n = 16); baseline (T1), following eight cycles of NAC (T6) and post-surgery (T7). Patients allocated into groups according to histological responses elicited in the breast cancers after NAC: Group I-cPR; Group II-VGR; Group III-partial response; IV-poor or no response. Statistical analysis: (A) FOXP3 (Group I & II): T1 v HFD (p = 0.006); T1 v T6 (NS); T1 v T7 (NS); T6 v HFD (NS); T7 v HFD (p = 0.002). (A) FOXP3 (Group III & V): T1 v HFD (p = 0.050); T1 v T6 (NS); T1 v T7 (NS); T6 v HFD (NS); T7 v HFD (p = 0.001). (B) CTLA-4 (Group I & II): T1 v HFD (p = 0.011); T1vT6 (NS); T1vT7 (NS); T6 v HFD (NS); T7 v HFD (p = 0.002). (B) CTLA-4 (Group III & IV): T1 v HFD (p = 0.038); T1 v T6 (NS); T1 v T7 (NS); T6 v HFD (NS); T7 v HFD (p = 0.041). Group I & II (T1) v III & IV (T1) (NS). Clear columns are good pathological responders whilst shaded columns are poor pathological responders to NAC.

Figure 3

Representative phenotypic profiles of % CD4⁺ CD25⁺CD127^- FOXP3⁺ Tregs in good pathological responders (Groups I and II), 200,000 events acquired, and poor pathological responders (Groups III and IV), 200,000 events acquired. AbNs of monocytic (CD11b⁺ CD14⁺ HLA DR^- CD124⁺ CD33⁺) MDSCs, in good pathological responders (Groups I and 11) 400,000 events acquired, and AbNs of granulocytic (CD11b⁺ CD14^- CD66b⁺ CD124⁺ CD15⁺) MDSCs in good pathological responders (Groups I and II), 150,000 events acquired. Data shown are for women with LLABCs, (baseline, post-chemotherapy, post-surgery) and healthy donors.