Clonogenicity-based radioresistance determines the expression of immune suppressive immune checkpoint molecules after hypofractionated irradiation of MDA-MB-231 triple-negative breast cancer cells

Abstract

Only a subset of patients with triple-negative breast cancer (TNBC) benefits from a combination of radio- (RT) and immunotherapy. Therefore, we aimed to examine the impact of radioresistance and brain metastasizing potential on the immunological phenotype of TNBC cells following hypofractionated RT by analyzing cell death, immune checkpoint molecule (ICM) expression and activation of human monocyte-derived dendritic cells (DCs). MDA-MB-231 triple-negative breast cancer tumor cells were used as model system. Apoptosis was the dominant cell death form of brain metastasizing tumor cells, while Hsp70 release was generally significantly increased following RT and went along with necrosis induction. The ICMs PD-L1, PD-L2, HVEM, ICOS-L, CD137-L and OX40-L were found on the tumor cell surfaces and were significantly upregulated by RT with 5 x 5.2 Gy. Strikingly, the expression of immune suppressive ICMs was significantly higher on radioresistant clones compared to their respective non-radioresistant ones. Although hypofractionated RT led to significant cell death induction and release of Hsp70 in all tumor cell lines, human monocyte-derived DCs were not activated after co-incubation with RT-treated tumor cells. We conclude that radioresistance is a potent driver of immune suppressive ICM expression on the surface of TNBC MDA-MB-231 cells. This mechanism is generally known to predominantly influence the effector phase, rather than the priming phase, of anti-tumor immune responses.

Keywords: breast cancer; dendritic cells; immune checkpoint molecules; radioresistance; radiotherapy; tumor cell death.

Figure 1

Generation of radioresistant breast cancer clones is done by repeated irradiation of MDA-MB-231 breast cancer cells. Radioresistant (sub)clones of MDA-MB-231 cells were generated by repeatedly irradiating MDA-MB-231 wildtype (MDA-MB-231) and brain metastasizing MDA-MB-231 (MDA-MB-231 BR) tumor cells (A). This resulted in more radioresistant clones (MDA-MB-231 RR and MDA-MB-231 BR RR), as verified by clonogenic survival assay (B). Data are from three independent experiments. **p < 0.01; ***p < 0.001 (Student’s t-test).

Figure 2

Cell death induction after irradiation of the four MDA-MB-231 cell lines is dependent on tissue origin rather than on radioresistance. (A) After seeding on day 0, the four MDA-MB-231 cell lines, the WT and the WT-derived brain metastasis clone (BR) as well as the radioresistant (RR) clones derived from those cells (WT RR, BR RR), were treated with 5 × 5.2 Gy. On day 6, 7 and 8, cell death forms were analyzed with Annexin V/Propidium iodide (AxVPi) staining via multicolor flow cytometry. The gating strategy is shown in (B). After pre-gating on the singlets and consequently excluding the debris, the remaining cells were identified as viable, apoptotic, or necrotic as presented. The percentage of apoptosis (C–E) and necrosis (F–H) of the different cell lines 24 (C–F), 48 (D, G) and 72 hours (E–H) after irradiation is shown as median with interquartile range. The data are from nine independent experiments. For statistical analysis, each treated clone was compared to the WT via Mann-Whitney U test (*p < 0.05, **p < 0.01, ***p < 0.001).

Figure 3

Radioresistance (RR) drives the expression of immune suppressive checkpoint molecules on the surface of the four presented MDA-MB-231 cell lines 48 hours after hypofractionated irradiation. The gating strategy is presented in (A) After pre-gating on the singlets, the debris was excluded. Then the viable cells were detected via the Zombie NIR viable/dead stain. Immune checkpoint molecule (ICM) expression is presented in the graphs as ΔMFI (mean fluorescence intensity). It was calculated by subtracting the MFI of the Zombie-only-stained samples (AF ctrl) from the respective Zombie-and-antibody-stained samples of various ICMs expressed on the cell surface of the four cell lines. Exemplarily primary data are shown for PD-L1 and PD-L2 detection. The WT and the WT-derived brain metastasis clone (BR) as well as the radioresistant (RR) clones derived from those cells (WT RR, BR RR) were treated with 5 × 5.2 Gy. (B) The expression of immune suppressive (PD-L1: (C), PD-L2: (D), HVEM: (E) and immune stimulatory (ICOS-L: (F), CD137-L: (G), OX40-L: (H) ICMs is presented as median with interquartile range. Data are from seven independent experiments. For statistical analysis, a Mann-Whitney U test was performed to compare untreated and treated cells within one cell line. The same test was used to compare an irradiated radioresistant cell clone with its respective non-radioresistant one. A Kruskal-Wallis test with multiple comparisons was used to examine statistical differences between the ICM expression of the different clones compared to the WT within the respective untreated (#) and treated (*) group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ^#p < 0.05, ^##p < 0.01, ^###p < 0.001, ^####p < 0.0001.

Figure 4

Hsp70 release was significantly increased from irradiated compared to non-irradiated MDA-MB-231 cells. The graph shows the concentration of Hsp70 per 10⁵ cells (ng/ml) in the cell culture supernatant of WT and the WT-derived brain metastasis clone (BR) as well as the radioresistant (RR) clones derived from those cells (WT RR, BR RR), either untreated (blue bars) or after irradiation with 5 × 5.2 Gy (brown bars). Data is presented as median with interquartile range. Data are from six independent experiments. For statistical analysis, a Mann-Whitney U test was performed to compare untreated and treated (5 × 5.2 Gy) cells within one cell line (*p < 0.05, **p < 0.01). Furthermore, a Kruskal-Wallis test with multiple comparisons was used to compare Hsp70 concentrations between the treated WT and its clones.

Figure 5

Neither untreated nor treated (5 × 5.2 Gy) MDA-MB-231 cells increased the expression of activation markers on dendritic cells (DCs) 48 hours after co-incubation. (A) Human monocyte-derived DCs were differentiated from peripheral blood mononuclear cells (PBMCs) for 5 days before they were co-incubated with untreated and treated wild type (WT) MDA-MB-231 cells or radioresistant (RR) clones. 48 hours later, the expression of common DC activation markers was examined using multicolor flow cytometry. The gating strategy is presented in (B) After pre-gating on the singlets, the viable cells were detected. Then, gating on CD11c positive cells identified DCs. CD70 (D), CD83 (E), CD80 (F), CD86 (G) expression on the cell surface of DCs is presented in the graphs as ΔMFI. It was calculated by subtracting the Zombie-only-stained samples (AF ctrl) from the respective Zombie-and-antibody-stained samples, here shown exemplarily for CD70 (C). The data is presented as median with interquartile range. Data are from seven independent experiments. For statistical analysis, a Mann-Whitney-U test was used to compare activation marker expression on DCs with and without (w/o) maturation cocktail (MC). Further, a Kruskal-Wallis test was performed to compare DCs w/o MC with DCs which had been co-cultured with either untreated or treated cancer cells, respectively. (*p < 0.05, **p < 0.01, ***p < 0.001, ^#p < 0.05).