M2 Macrophages Infiltrating Epithelial Ovarian Cancer Express MDR1: A Feature That May Account for the Poor Prognosis

Abstract

Multi drug resistance protein 1 (MDR1) expression on tumor cells has been widely investigated in context of drug resistance. However, the role of MDR1 on the immune cell infiltrate of solid tumors remains unknown. The aim of this study was to analyze the prognostic significance of a MDR1+ immune cell infiltrate in epithelial ovarian cancer (EOC) and to identify the MDR1+ leucocyte subpopulation. MDR1 expression was analyzed by immunohistochemistry in 156 EOC samples. In addition to MDR1+ cancer cells, we detected a MDR1+ leucocyte infiltrate (high infiltrate >4 leucocytes per field of view). Correlations and survival analyses were calculated. To identify immune cell subpopulations immunofluorescence double staining was performed. The MDR1+ leucocyte infiltrate was associated with human epidermal growth factor receptor 2 (HER2) (cc = 0.258, p = 0.005) and tumor-associated mucin 1 (TA-MUC1) (cc = 0.202, p = 0.022) expression on cancer cells. A high MDR1+ leucocyte infiltrate was associated with impaired survival, especially in patients whose carcinoma showed either serous histology (median OS 28.80 vs. 50.64 months, p = 0.027, n = 91) or TA-MUC1 expression (median OS 30.60 vs. 63.36 months, p = 0.015, n = 110). Similar findings for PFS suggest an influence of MDR1+ immune cells on the development of chemoresistance. A Cox regression analysis confirmed the independency of a high MDR1+ leucocyte infiltrate as prognostic factor. M2 macrophages were identified as main part of the MDR1+ leucocyte infiltrate expressing MDR1 as well as the M2 marker CD163 and the pan-macrophage marker CD68. Infiltration of MDR1+ leucocytes, mostly M2 macrophages, is associated with poor prognosis of EOC patients. Further understanding of the interaction of M2 macrophages, MDR1 and TA-MUC1 appears to be a key aspect to overcome chemoresistance in ovarian cancer.

Keywords: M2 Macrophages; MDR1; TA-MUC1; ovarian cancer; prognosis.

Figure 1

MDR1 immunostaining of ovarian cancer cells. Membranous expression of MDR1 on ovarian cancer cells differs between the subtypes (A, p = 0.004) with mucinous (B, IRS = 6) and clear cell (C, IRS = 4) showing a higher expression than serous (D, IRS = 3) and endometrioid (E, IRS = 2). (B–E) are shown in 40× magnification (scale bar = 50 µm), 25× magnification is provided in the Supplementary Figure S3.

Figure 2

A MDR1+ leucocyte infiltrate was detected by immunohistochemistry in all subtypes: serous (A), clear cell (B), endometrioid (C) and mucinous carcinoma (D). (A–D) are shown in 40× magnification (scale bar = 50 µm), 25× magnification is provided in the supplementary Figure S4. The highest relative frequency of cases with MDR1+ leucocyte infiltration was found for serous histology (E, p = 0.042) followed by mucinous, clear cell and endometrioid.

Figure 3

A high MDR1+ leucocyte infiltrate is associated with poor prognosis especially in patients whose carcinoma shows serous histology and TA-MUC1 expression. The Kaplan-Meier estimates show that high MDR1+ leucocyte infiltration (>4/field of view, 25× lens) leads to decreased PFS (A, p = 0.059, n = 126) and OS (A, p = 0.057, n = 126), although not statistically significant. Late separation of the curves suggests long time effects mediated by the MDR1+ leucocyte infiltrate. In serous subtype these effects lead to significantly impaired PFS (p = 0.029, n = 91) and OS (p = 0.027, n = 91). (E–H) show combined survival analysis of a high MDR1+ leucocyte infiltrate and TA-MUC1. PFS (E, p = 0.029, n = 110) and OS (F, p = 0.015, n = 110) of patients with high MDR1+ leucocyte infiltration is significantly decreased in TA-MUC1+ cases (IRS > 0), which are even worse when the carcinoma shows also serous histology (G, PFS, p = 0.007, n = 81; H, OS p = 0.007, n = 81). Censoring events have been marked in the graphs (+).

Figure 4

Macrophages were identified as main part of the immune cell infiltrate. The immune cell infiltrate was quantified by counting positive cells per field of view (20× lens; n = 12) in immunofluorescence double staining. Most infiltrating cells were CD68 positive macrophages, followed by CD3 positive T-cells. Just a few CD56 positive NK-cells were detected.

Figure 5

Characterization of the immune cell subpopulation by immunofluorescence double staining. M2 macrophages were identified as main part of the MDR1+ leucocyte infiltrate, expressing besides MDR1 the M2 marker CD163 (D) as well as the pan-macrophage marker CD68 (B). The stained tissue slices of serous ovarian cancer tissue were analyzed in 40× and 63× (inserts) magnification. For most CD45 positive immune cells (green) a co-localization with MDR1 (red) was observed (A); co-expression of MDR1 (red) and CD68 (green) (B); co-expression of CD163 (red) and CD68 (green) (C); co-expression of MDR1 (red) and CD163 (green) (D); no co-expression of TLR2 (red) and CD68 (green) was detected (E). Cell nuclei were marked by DAPI (blue) staining.