Wound Fluid from Breast Cancer Patients Undergoing Intraoperative Radiotherapy Exhibits an Altered Cytokine Profile and Impairs Mesenchymal Stromal Cell Function

Abstract

Intraoperative radiotherapy (IORT) displays an increasingly used treatment option for early breast cancer. It exhibits non-inferiority concerning the risk of recurrence compared to conventional external irradiation (EBRT) in suitable patients with early breast cancer. Since most relapses occur in direct proximity of the former tumor site, the reduction of the risk of local recurrence effected by radiotherapy might partially be due to an alteration of the irradiated tumor bed's micromilieu. Our aim was to investigate if IORT affects the local micromilieu, especially immune cells with concomitant cytokine profile, and if it has an impact on growth conditions for breast cancer cells as well as mammary mesenchymal stromal cells (MSC), the latter considered as a model of the tumor bed stroma.42 breast cancer patients with breast-conserving surgery were included, of whom 21 received IORT (IORT group) and 21 underwent surgery without IORT (control group). Drainage wound fluid (WF) was collected from both groups 24 h after surgery for flow cytometric analysis of immune cell subset counts and potential apoptosis and for multiplex cytokine analyses (cytokine array and ELISA). It served further as a supplement in cultures of MDA-MB 231 breast cancer cells and mammary MSC for functional analyses, including proliferation, wound healing and migration. Furthermore, the cytokine profile within conditioned media from WF-treated MSC cultures was assessed. Flow cytometric analysis showed no group-related changes of cell count, activation state and apoptosis rates of myeloid, lymphoid leucocytes and regulatory T cells in the WF. Multiplex cytokine analysis of the WF revealed group-related differences in the expression levels of several cytokines, e.g., oncostatin-M, leptin and IL-1β. The application of WF in MDA-MB 231 cultures did not show a group-related difference in proliferation, wound healing and chemotactic migration. However, WF from IORT-treated patients significantly inhibited mammary MSC proliferation, wound healing and migration compared to WF from the control group. The conditioned media collected from WF-treated MSC-cultures also exhibited altered concentrations of VEGF, RANTES and GROα. IORT causes significant changes in the cytokine profile and MSC growth behavior. These changes in the tumor bed could potentially contribute to the beneficial oncological outcome entailed by this technique. The consideration whether this alteration also affects MSC interaction with other stroma components presents a promising gateway for future investigations.

Keywords: GROα; IL-1β; RANTES; VEGF; breast cancer; intraoperative radiotherapy; leptin; mesenchymal stromal cells; oncostatin-M; tumor microenvironment.

Figure 1

Absolute cell counts of myeloid (A), lymphoid (B), Treg (C) and apoptotic (D) cells. Indicated values: mean ± standard deviation, statistical analysis by double sided t-test. Myeloid analysis: N(WF) = 31 (IORT 19, control 12); N(PBMC) = 17 (IORT 11, control 6) Lymphoid analysis: N(WF) = 33 (IORT 21, control 12); N(PBMC) = 22 (IORT 15, control 7) Treg analysis: N(WF) = 24 (IORT 14, control 10); N(PBMC) = 16 (IORT 11, control 5) Annexin staining: N(WF) = 30 (IORT 20, control 10); N(PBMC) = 21 (IORT 14, control 7). (For PBMC, some patients did not consent to an additional blood sampling and for some patients, cell numbers were insufficient for analysis.).

Figure 2

Signal intensities and calculated fold changes of the 30 cytokines with most differing levels between IORT and control group in the human cytokine array. (A): relative signal intensity, (B): fold change (>0: higher levels in IORT-WF; <0: higher levels in control-WF). N = 8 IORT vs. N = 7 control. Indicated values: mean ± standard deviation, statistical analysis by double sided t-test, * p < 0.05.

Figure 3

Cytokine concentrations tested in the individual WF samples by ELISA for OSM (A), leptin (B), IL-1β (C), GROα (D), HGF (E), VEGF (F), uPA (G) and RANTES (H). N = 18 IORT (yellow) and 18 control (blue), each n = 3. Indicated values: mean ± standard deviation, statistical analysis by double sided t-test, * p < 0.05.

Figure 4

Functional analysis of MDA-MB 231 incubated with 0.5%/1% pooled WF concerning proliferation (A), wound healing (B) and chemotactic migration (C). All assays were reproduced in 3 independent experiments with each 8–12 technical replicates in each condition. Indicated values: mean ± standard deviation in percent over time. Statistical analysis by MANOVA.

Figure 5

Functional analysis of MSC incubated with 0.5%/1% pooled WF concerning proliferation with 0.5 % WF (A), proliferation with 1.0 % WF (B), wound healing in 1.0 % WF (C) and chemotactic migration in 0.5 % WF (D). All assays were reproduced in 3 independent experiments with each 8–12 technical replicates for each condition. Indicated values: mean ± standard deviation in percent over time. Statistical analysis by MANOVA, * p < 0.05, ** p < 0.01, *** p < 0.0001.

Figure 6

Cytokine concentrations in MSC conditioned media for GROα (A), VEGF (B), RANTES (C), HGF (D), uPA (E), I-309 (F), PDGF-BB (G), OSM (H) and leptin (I). All experiments were reproduced in three independent experiments, with each 8–12 technical replicates in each condition (IORT = yellow, control = blue). Indicated values: mean ± standard deviation, statistical analysis by double sided t test, * p < 0.05, ** p < 0.01).