We have demonstrated that women with LLABCs have significantly and substantially elevated circulating levels (%, AbNs) of Tregs (FOXP3+, CTLA-4+) and concurrent increased expression of Treg (FOXP3, CTLA-4) mRNA in T cells from blood. Also, circulating AbNs of MDSCs (monocytic, granulocytic) are significantly increased. This data comprehensively expands and defines more precisely the findings of the few studies published, to date, in early breast cancer [11, 35–38].
Recent publications have demonstrated elevated blood Tregs in aggressive primary breast tumours, where spread to regional nodes had occurred, and with metastatic disease, albeit Treg specific markers were not always used [38–41]. Not dissimilar findings have been documented in other tumour types [12–14, 16]. High circulating levels of MDSCs have also been documented with advanced disease .
NAC is being used to treat women with LLABCs [31, 32, 42]. Some of the beneficial effects are being attributed to modulation of anticancer defences, especially regulatory cells [23–25]. However, the impact of NAC on immunoregulatory cells in women with LLABCs is poorly defined. Aruga et al. (2009) demonstrated a correlation of FOXP3+ cells in tumours and subsequent response to NAC. They showed that low numbers of tumour infiltrating FOXP3+ cells was associated with a good pathological response but did not characterise blood Tregs . Ladoire et al. (2008) showed that a cPR in breast cancers with NAC was associated with the disappearance of tumour infiltrating FOXP3+ Tregs . Diaz-Montero et al. (2009) showed increased circulating MDSCs in patients with breast cancer and correlation with clinical stage, but documented enhanced MDSCs with NAC (ACT) .
Our study, to the best of our knowledge, has shown for the first time that the level (%) of circulating FOXP3+ Tregs predicted the likely response to NAC – high levels resulting in poor pathological responses in the cancers in the breast. By contrast, low levels of FOXP3+ Tregs were associated with a cPR, a recognised surrogate marker of good long-term outcome [31, 32, 42]. NAC (ACT, ACTCap) significantly reduced Tregs (% FOXP3+, AbN CTLA-4+), but only after 8 cycles and following a complete or substantial reduction in tumour mass. However, NAC and surgery failed to restore Tregs to the levels found in HFDs. Recently, Oda et al. (2012) described FOXP3+ infiltraes in breast cancer as independent predictors of pCR to NAC. The absence of FOXP3+ and CD8+ infiltrates post NAC was associated with the highest pCR rate (Oda, et al., 2012) . These observations are in keeping with our data and suggest that NAC-induced Treg depletion in the circulation is mirroed within the tumour microenvironment and is crucial for tumour cell death.
Tumour infiltration of breast cancers by Tregs is associated with long-term risk of tumour recurrence and reduced disease-free survival [11, 45–47]. However, a recent publication, asscessing FOXP3 expression in triple negative breast cancers, documented improved survival in patients with high levels of tumour infiltrating CD4+CD25+FOXP3+Tregs. The finding, which is at variance with published data, may reflect a unique and, as yet, poorly understood immune interaction in this uncommon pathological subset of breast cancer .
We intend to immunohistochemically characterise the Treg profile in the tumour microenvironment of the patients investigated in our study and establish its relevance to subsequent response to NAC.
mRNA levels in the Tregs mirrored the abnormal levels of blood Tregs in women with LLABCs. NAC and surgery non significantly reduced circulating FOXP3+ and CTLA-4+ Treg mRNA. Post-treatment Tregs mRNAs however were abnormally elevated, compared with levels documented in HFDs. A recent study has documented enhanced FOXP3 and CTLA-4 transcripts in BMCs in breast cancer, but did not assess effect of NAC or surgery .
Our study also documented that NAC reduced circulating levels of MDSCs (monocytic, granulocytic), after only 4 cycles. This was more pronounced after 8 cycles of NAC and the MDSCs were not reduced further by surgery, suggesting a direct cytotoxic effect on MDSCs, irrespective of tumour burden. These findings are at variance with the enhancement documented with ACT in a recent study . The different time course used may be a factor. In a study in patients with advanced non-small cell lung cancer granulocytic (CD11b+ CD14- CD15+) MDSCs were increased in blood. Following chemotherapy and/or surgery, the level of these MDSCs fell significantly supporting the findings of NAC in breast cancer in our study .
Different subsets of MDSCs in humans are able to induce both Th17 effector cells and immunoinhibitory Tregs when co-cultured with CD4+ T cells . MDSCs can also induce differentiation of FOXP3+ Tregs from monocyte-derived Th17 cells . Our study suggests that the beneficial effects of NAC may be due, in part, to a reduction of immunoregulatory cells by a direct effect on Tregs and/or through inhibition of generation by MDSCs. The presence of elevated residual Tregs, albeit low, may reflect persistent occult micrometastases and need for further therapy, designed to abolish the remaining foci of regulatory cells [10, 29, 33, 34]. Further careful studies need to be done to address this issue and ascertain possible beneficial clinical outcome from such an approach.
The Th1 profile (IL-1β, IL-2, INF-γ, TNF-α) from women with LLABCs was substantially and significantly reduced, compared with HFDs. The most pronounced reduction occurred in patients whose tumours responded poorly to NAC. By contrast, the Th2 profile (IL-4, IL-5) was substantially and significantly increased; IL-10 was non significantly increased, and IL-6 showed no changes. NAC with or without surgery had no effect on Th1 and Th2 profiles, except for IL-4, which resembled levels found in HFDs following chemotherapy. Polarisation of cytokine responses has been described in breast cancer . Our study characterised this more comprehensively, documenting the inability of NAC and surgery to alter the majority of these abnormal changes, even in the presence of significant reduction of tumour burden and levels of regulatory cells. This complex and ongoing immune suppression suggests a chronic or permanent dysfunction of host defences induced by the malignant state. Careful and long-term monitoring of this dysfunctional immune state may provide insight as to likely long-term clinical outcome.
Th17 cells play a crucial role in inflammation and autoimmunity, but their function in cancer is unclear [52, 53]. Animal studies have shown a close association between IL-17 production by tumour infiltrating lymphocytes and destruction of tumour cells by induction of Th1 lymphocytes and CTL antitumour responses [54, 55]. TGF-β and IL-6 induce Th17 differentiation, amplified by IL-1β and TNF-α. Increased Th17 cells have been documented in blood, lymph nodes and human tumours [16, 52]. Data suggests they may enhance anticancer defences in man [53, 56]. A recent study in NODscid mice bearing human ovarian cancer and transfer of autologous Th17 human cells indicated an anticancer collaboration between Th17 cells and CD8+ CTLs . High levels of tumour infiltrating Th17 cells in breast cancer has been shown to be associated with improved prognosis and reduction of metastases . In our study, production of IL-17A by Th17 lymphocytes (±FOXP3+) was significantly increased in good (2.8 fold) and poor (5 fold) clinical responders. NAC had no demonstrable effect on IL-17A production. High dose cyclophosphamide in animals and metronomic doses in humans can induce the differentiation of Th17 cells . The impact of high dose cyclophosphamide (as in our NAC protocol) in man has not been studied, but appears to lack any overt regulatory effect.