Impact of systemic therapy on circulating leukocyte populations in patients with metastatic breast cancer

Tumors affect the immune system, locally and systemically. The frequencies of specific circulating immune cell populations correlate with disease progression as well as prognosis of the patients. Although largely neglected, conventional antitumoral therapies often possess immunomodulatory properties and affect the levels of specific immune cell populations. Most information, however, derive from animal or in vitro studies. As this could impact prognosis as well as response to therapy, further studies of the effects of treatment on circulating immune cells in patients are warranted. In this pilot study, we evaluated a wide panel of circulating immune cells over time (up to six months) in ten patients with metastatic breast cancer receiving standard antitumoral regimens. Overall, endocrine therapy tends to enrich for natural killer (NK) and natural killer T (NKT) cells in the circulation, whereas both chemotherapy and endocrine therapy reduce the levels of circulating monocytic myeloid-derived suppressor cells (Mo-MDSCs). This indicates that the systemic immunosuppressive profile observed in patients tends to revert over the course of systemic therapy and holds promise for future combination treatment with standard antitumoral agents and immunotherapy.

and M1 macrophages) exert direct cytotoxic effect on tumor cells 12 . On the other hand, regulatory T cells (Tregs), Th2-skewed T helper cells (Th), anti-inflammatory neutrophils or macrophages (N2 neutrophils and M2 macrophages) and myeloid-derived suppressor cells (MDSCs) are involved in immunosuppression, tissue remodeling and angiogenesis, and are thus regarded pro-tumorigenic 12,13 . Today it is a well-documented fact that tumors actively modulate the immune system towards a more tumor-permissive phenotype, locally as well as systemically. Leukocytosis, lymphopenia and neutrophilia are common hematological findings that correlate with disease progression, and consequently poor prognosis, in several malignancies [14][15][16] . In breast cancer patients, similar systemic immune alterations are detected. Decreased levels of circulating lymphocytes (especially CD8 + CTLs) and myeloid dendritic cells (mDCs) or increased levels of monocytes, NK cells, Tregs and MDSCs correlate with stage and/or disease progression [17][18][19][20][21][22][23] .
Although conventional antitumoral treatments target the malignant cells, some approaches also have immunomodulatory properties [24][25][26] . Substantial efforts are made to investigate the efficacy of immunotherapy in breast cancer and recently a clinical trial showed prolonged progression-free survival in patients with advanced TNBC treated with the PD-L1 antibody atezolizumab in combination with nab-paclitaxel 10 . As the levels of various immune cell populations could impact prognosis and response to therapy, and the fact that immunotherapies most often is given in combination with conventional therapy, further studies of the effects of standard treatment on circulating immune cells in patients are warranted.
In this pilot study, we prospectively monitored patients with MBC to evaluate whether specific circulating leukocyte populations, including T lymphocytes, NK cells and monocytic-MDSCs (Mo-MDSCs) in peripheral blood are affected during standard systemic therapies.

Results
Patient characteristics. A total of ten female patients with MBC were included in this study (Table 1).
Median age was 58 years (range 44-77 years). Three patients had de novo metastatic disease, whereas seven patients were diagnosed with distant recurrence. Four patients had more than three metastatic loci and five patients had visceral metastasis. Eight patients had ER +/HER2-tumors, one had ER +/HER2+ disease and one patient had TNBC. Among the eight patients with ER +/HER2disease, five received endocrine therapy (ET; two patients received tamoxifen and three patients aromatase inhibitors) and three patients received chemotherapy. Chemotherapy regimens used were FEC (5-fluorouracil , epirubicin, cyclophosphamide) in two patients and docetaxel in one patient with ER +/HER2-MBC. The patient with ER +/HER2+ disease received trastuzumab in combination with capecitabine and the patient with TNBC was treated with capecitabine. One patient was diagnosed with early progression at first evaluation (after 3 months of endocrine therapy) whereas the mean progression-free survival (PFS) was 23 months (range 2.8-56.7 months). See Table 1 for specification of treatment regimens and clinical information.

Modest alterations in circulating T cell populations in patients receiving systemic therapy.
In order to investigate whether systemic treatment affect circulating immune cell populations, we collected venous blood at baseline (BL; i.e. before onset of treatment) and after one and three months of treatment. Freshly isolated peripheral blood mononuclear cells (PBMCs) were analyzed using flow cytometry and variations in specific immune cell populations as compared to BL assessed. Representative dot plots and gating strategies of all cell populations are depicted in Supplementary Fig. 1.
Compared to healthy donors, patients with MBC generally display slightly reduced levels of T lymphocytes at baseline (Fig. 1, left panels). Overall, endocrine therapy (white symbols in figures) tends to induce a transient increase in both total T lymphocytes (CD3 + , Fig. 1a), cytotoxic-(CD8 + T c /CTLs, Fig. 1b) and T helper (CD4 + Th, Fig. 1c) cells after one month of treatment. Chemotherapy as well as radiotherapy is often accompanied by lymphopenia 27 . In the patient treated with docetaxel, there was little or no effect on the levels of T lymphocytes (Fig. 1, right   www.nature.com/scientificreports www.nature.com/scientificreports/ figures) displayed a transient reduction in all T lymphocyte populations investigated (Fig. 1, right panels). Of note, these patients received not only capecitabine (Xeloda; a fluoropyrimidine that is converted to 5-FU once ingested), but were also treated with radiation after the baseline samples were taken. Interestingly, the patient that received 5-FU in combination with epirubicin and cyclophosphamide (FEC) displayed a continuous increase in all T lymphocyte populations in circulation ( Fig. 1, right panels, black circle). The effects are likely due to a general influence on all T lymphocyte sub-populations as the percentages of CD8 + T c /CTLs as well as CD4 + Th cells remain fairly constant, or even decrease, within the total lymphocyte population (all CD3 + cells) in the majority of patients including the patient that received FEC ( Supplementary Fig. 2a,b). Similarly, the majority of patients displayed unaffected or modestly increased levels of regulatory T cells (Tregs) during treatment ( Supplementary  Fig. 2c). The patient with TNBC did, however, display a more pronounced increase ( Supplementary Fig. 2c, right panels black box).
In order to assess more long-term effects, we followed five patients up to six months after treatment onset ( Supplementary Fig. 3). CTLs and Th cells continued to decline within the total lymphocyte population in one patient treated with FEC, whereas a continuous increase in the same populations were observed for two patients treated with endocrine therapy (Supplementary Fig. 3b,c). The levels of regulatory T cells remained above baseline and even increased further for four of five patients ( Supplementary Fig. 3d). Furthermore, as two patients received radiation therapy, we assessed the levels of circulating immune cells before and after radiation, but before initiating systemic therapy. Only very modest variations were observed for all lymphocyte populations analyzed ( Supplementary Fig. 4a-e).
Altogether this indicates that T lymphocytes are only modestly affected during systemic treatment and could potentially be harnessed by future immunotherapeutic regimens.
NK-and NKT cells are low at baseline but tend to increase during endocrine therapy. We next focused on circulating CD56 + CD3 − NK and CD56 + CD3 + NKT cells. Both populations are decreased in MBC patients at baseline as compared to healthy controls ( Fig. 2a,b, left panels), although not statistically significant. The levels of NK-and NKT cells markedly increase over the course of the treatment in the majority of patients ( Fig. 2a,b, right panels). One patient treated with chemotherapy (docetaxel) and one patient treated with endocrine therapy (tamoxifen) displayed lower levels at three months as compared to baseline (Fig. 2a,b, right panels). Interestingly, the levels of NK-as well as NKT cells remained high in comparison to baseline up to six months in patients treated with endocrine therapy, whereas they tended to normalize to baseline-levels in patients treated with chemotherapy (Fig. 2c,d).
As NK-and NKT cells are capable of recognizing and killing transformed cells without preceding antigen presentation, the observation that endocrine therapy may enrich for these populations will be of interest to look into in the future.

Mo-MDSCs are enriched in metastatic breast cancer patients and decrease during systemic therapy.
We were recently first to show that Mo-MDSCs (CD14 + HLA-DR low/− Co-receptor low/− cells) are enriched in the peripheral blood of breast cancer patients 23 . When analyzing the levels of Mo-MDSCs in MBC patients receiving systemic therapy, the levels were significantly higher also after one and three months of treatment, as compared to healthy controls (Fig. 3a). Compared to baseline, a transient increase could be observed at one month of treatment for the majority of patients (Fig. 3a, right panel). However, at three months the levels decreased dramatically for six out of ten patients. Similar trends were observed regardless treatment regimen and when looking at Mo-MDSCs within the total monocyte population (Fig. 3b) as well as all CD14 + monocytes (Fig. 3c). Interestingly, the patient with TNBC still displayed increasing levels of Mo-MDSCs at three months (Fig. 3a, right panel, black box). This may in part be related to the increased Mo-MDSC and monocyte levels observed after radiation therapy ( Supplementary Fig. 4f-h).
When following five patients up to six months after treatment onset, Mo-MDSC as well as total monocyte levels tend to normalize to the level of healthy controls (Fig. 4, left panels). A substantial, albeit not significant, drop in the percentage of Mo-MDSCs was seen in patients treated with endocrine therapy (Fig. 4a, left panel). Furthermore, as compared to baseline levels, Mo-MDSCs in PBMCs as well as within the total CD14 + monocyte pool either continued or started to decline at six months (Fig. 4a,b, right panels). Similar trends are observed for monocytes (Fig. 4c, right panels).
Overall, Mo-MDSCs decrease in the peripheral blood during systemic therapy, though more consistent with endocrine therapy than with chemotherapy. This indicates that endocrine therapy may revert the enrichment of Mo-MDSCs in the peripheral blood of breast cancer patients. How this may be related to prognosis will be of interest to further look into in the future.

Discussion
In the era of immunotherapy, a profound understanding of how specific immune cells are affected during conventional therapy is necessary. Although largely neglected, studies have indicated that some standard antitumoral agents also rely on their capacity to modulate the immune system for proper therapeutic efficacy 24 . In this pilot study, we evaluated a wide panel of circulating immune cells over time in patients with metastatic breast cancer receiving conventional antitumoral regimens. Although the patients monitored in this study reflect the general population well (with regards to hormone receptor status, age, as well as treatment) a larger study population is required to further validate the overall trends observed in this paper.
Chemotherapy is often accompanied by a transient lymphopenia. A transient reduction in T lymphocytes were observed in two out of four patients receiving chemotherapy. These patients were treated with capecitabine (converted to 5-FU once ingested) as well as radiotherapy. FEC and docetaxel, on the other hand, were rather associated with an increase in T lymphocytes or had no effect, respectively. Similar trends could also be observed (2019) 9:13451 | https://doi.org/10.1038/s41598-019-49943-y www.nature.com/scientificreports www.nature.com/scientificreports/ for NK-and NKT cells with the main exception that the levels were restored and even increased at three months in the patients treated with capecitabine. These differences in chemotherapy effects on lymphocyte populations may theoretically also be influenced by the different dose regimens, i.e. capecitabine is an orally given 5-FU  www.nature.com/scientificreports www.nature.com/scientificreports/ for the patient as this may activate homeostatic mechanisms that result in the proliferation of effector T cells such as tumor-reactive CD8 + CTLs 24,28 . Indeed, the levels of T lymphocytes were restored to baseline levels or above for the majority of the patients after three to six months of treatment.
Some agents such as cyclophosphamide and 5-FU have been suggested to selectively and transiently decrease the levels of immunosuppressive cells such as Tregs and MDSCs 24,28 . Indeed, Tregs were decreased or unaffected in three of four patients treated with chemotherapy. The dramatic increase seen in the fourth patient may be due to individual differences, to the TNBC phenotype of the tumor or to tumor progression. The same patient also displayed continuous increased levels of Mo-MDSCs, whereas the Mo-MDSC levels were markedly reduced for three of the other four patients. This is in line with previous findings showing that docetaxel treatment of primary human monocytes in vitro reduced the levels of Mo-MDSC-like cells while promoting the generation of pro-inflammatory M1 macrophages 26 . Circulating MDSCs, on the other hand, have previously been suggested to increase in breast cancer patients treated with doxorubicin and cyclophosphamide 22,29 . In both studies, granulocytic-MDSCs (G-MDSCs) were studied in patients with early stage breast cancer. In contrast to our results, Wesolowski et al. could not detect any variations in Mo-MDSCs 29 . This is likely due to differences in the patient groups being monitored; metastatic and early stage breast cancer, respectively, which is in line with our previous results 23 .
In our material, only two patients received cyclophosphamide (FEC). Four out of five treated with chemotherapy were, however, given 5-FU in some form (FEC or capecitabine). In mice bearing EL4 thymoma, 5-FU selectively deplete MDSCs thus restoring IFNγ production by CD8 + T cells 30 . Similar results were observed in 4T1-Neu-tumor bearing mice treated with docetaxel 31 . In patients, little is known about the impact of 5-FU on MDSCs. 5-FU in combination with folinic acid and oxaliplatin (FOLFOX) decreased the levels of G-MDSCs whereas 5-FU with folinic acid and CPT11 (FOLFIRI) tend to increase the MDSC levels in patients with colorectal cancer 32 . Thus, further clarification of the impact of 5-FU on MDSCs is required considering different dose regimens and combination treatments.
Information about how endocrine therapy affects circulating immune cells in cancer patients is scarce. In mice, tamoxifen was proposed to induce a shift from cellular Th1 to humoral Th2 immunity, while suppressing alloantigen-but not mitogen-induced T-cell proliferation in vitro [33][34][35] . Here, we observed a modest, but transient increase in T lymphocytes. This was especially pronounced for CD8 + CTLs after one month of treatment. Interestingly, patients treated with endocrine therapy also had an enrichment of NK and NKT cells in the peripheral blood. NK and NKT cells are well-known players in immunosurveillance and tumor rejection, and could potentially be exploited in future immunotherapies. Finally, a substantial reduction in the levels of Mo-MDSCs was observed in patients treated with endocrine therapy. This finding fits well with the observation that estrogens may drive MDSC accumulation 36 . To our knowledge, this is the first study to imply that MDSCs are affected by endocrine therapy. However, it is not possible to discriminate between direct effects of tamoxifen on MDSC accumulation and indirect effects mediated via tumor and host mechanisms respectively in this study. As the levels of circulating Mo-MDSCs correlate with disease progression, the underlying mechanisms and clinical implications to this observation will be of great interest to study in the future. The findings in this study are limited by the relatively small sample size but can be viewed as mainly hypothesis generating. Thus, larger studies regarding the impact of endocrine therapy on MDSCs, NK and NKT cells are warranted. We are currently analyzing different immune cell populations in a larger cohort and evaluating possible correlations to important clinicopathological factors and outcome.
To summarize, overall chemotherapy as well as endocrine therapy tend to have modest effects on lymphocyte populations, although endocrine therapy seems to induce an enrichment of NK-and NKT cells. In contrast, the Mo-MDSC levels decrease markedly in the majority of patients regardless of treatment type, although in particular in patients treated with endocrine therapy. Altogether these results indicate that the systemic immunosuppressive profile in patients with metastatic breast cancer tends to regress over the course of treatment, be it chemotherapy or endocrine therapy. Thus, combination treatment with standard antitumoral agents and immunotherapy will be interesting to follow.

Materials and Methods
Patient cohort and blood sampling. Peripheral blood immune cell populations were analyzed in ten patients, newly diagnosed with MBC, included in a prospective observational trial (ClinicalTrials.gov NCT01322893), during the course of systemic therapy. For details on this prospective observational trial, see previous publications 37 . Ethical permission has been obtained from the Research Ethics Committee at Lund University (Dnr 2012/689 for healthy blood donors, Dnr 2010/135 and Dnr 2011/748 for patients with breast cancer) and the study has been performed in accordance with the ethical standards in the Declaration of Helsinki. All participating patients and donors gave written informed consent.
Blood samples were taken at baseline (BL; before start of systemic therapy) and at one, three and six months during systemic therapy. Data from BL samples have previously been published as part of a larger dataset in refs 23,37 . See Table 1 for detailed patient information. Ten healthy blood donors were used as controls (90% female, mean age ± SEM; 39 ± 4 years). Peripheral blood from MBC patients and healthy controls was drawn in EDTA tubes and analyzed within 24 h as previously described 23 . Briefly, the peripheral blood was diluted 1:2 in PBS supplemented with 5 mM EDTA and 2.5% w/v sucrose and separated over Ficoll-Paque Plus (GE Healthcare, Uppsala, Sweden). The peripheral blood mononuclear cells (PBMCs) were collected. Flow cytometry. Freshly isolated PBMCs were washed once in FACS buffer (PBS supplemented with 3% fetal bovine serum and 0.1% sodium azide) and stained for flow cytometry. Antibodies used; CD14 clone M5E2, HLA-DR clone G46-6, CD86 clone IT2.2, CD163 clone GHI/61, CD3 clone HIT3a, CD4 clone RPA-T4, CD8 HIT8a, CD25 clone 2A3, CD127-biotin clone HIL-7R-M21, CD56 clone B159, all from BD Biosciences. All