RANK-ligand (RANKL) expression in young breast cancer patients and during pregnancy

Abstract

Introduction: RANKL is important in mammary gland development during pregnancy and mediates the initiation and progression of progesterone-induced breast cancer. No clinical data are available on the effect of pregnancy on RANK/RANKL expression in young breast cancer patients.

Methods: We used our previously published dataset of 65 pregnant and 130 matched young breast cancer patients with full clinical, pathological, and survival information. 85% of patients had available transcriptomic data as well. RANK/RANKL expression by immunohistochemistry using H-score on the primary tumor and adjacent normal tissue was performed. We examined the difference in expression of RANK/RANKL between pregnant and non-pregnant patients and their association with clinicopathological features and prognosis. We also evaluated genes and pathways associated with RANK/RANKL expression on primary tumors.

Results: RANKL but not RANK expression was more prevalent in the pregnant group, both on the tumor and adjacent normal tissue, independent of other clinicopathological factors (both P <0.001). 18.7% of pregnant and 5.3% of non-pregnant patients had tumors showing ≥10% of cells with 3+ RANKL expression. RANKL expression was significantly higher in progesterone receptor-positive, and luminal A-like tumors, with negative correlation with Ki-67 (all P <0.001). On the contrary, RANK expression was higher in triple negative tumors (P <0.001). Using false discovery rate <0.05, 151 and 1,207 genes were significantly correlated with tumor-expressed RANKL and RANK expression by immunohistochemistry, respectively. High RANKL expression within primary tumor was associated with pathways related to mammary gland development, bone resorption, T-cell proliferation and regulation of chemotaxis, while RANK expression was associated with immune response and proliferation pathways. At a median follow-up of 65 months, neither RANK nor RANKL expression within tumor was associated with disease free survival in pregnant or non-pregnant group.

Conclusions: Pregnancy increases RANKL expression both in normal breast and primary tumors. These results could guide further development of RANKL-targeted therapy.

Figure 1

Differences in RANK and RANKL expression between pregnant and nonpregnant breast cancer patients. (a) Expression of RANKL by immunohistochemistry using the H-score (y axis) in all patients, pregnant patients and nonpregnant patients both in primary tumor (blue) and adjacent normal breast epithelial cells (green). Mean RANKL H-score was higher in normal breast epithelial cells compared with tumor cells; 50.19 versus 16.13, P <0.001 in all patients; 32.88 versus 8.06, P <0.001 in nonpregnant patients; and 87.29 versus 32.53 in pregnant patients, P <0.001. (b) Expression of RANK by immunohistochemistry using the H-score (y axis) in all patients, pregnant patients and nonpregnant patients both in primary tumor (blue) and adjacent normal breast epithelial cells (green). Mean RANK H-score was higher in normal breast epithelial cells compared with tumor cells; 24.65 versus 12.48, P = 0.003 in all patients; 21.72 versus 13.95, P = 0.07 in nonpregnant patients; and 31.52 versus 9.41 in pregnant patients, P = 0.016. RANK, receptor activator for nuclear factor κB; RANKL, receptor activator for nuclear factor κB ligand.

Figure 2

Examples of RANK and RANKL immunohistochemical staining on primary breast tumors and normal breast tissue. (a) RANKL expression on the primary tumor in a patient diagnosed with breast cancer during pregnancy (H-score: 270, 80% of cells showing RANKL expression score 3+). (b) RANKL expression on adjacent normal epithelial cells in a patient diagnosed with breast cancer during pregnancy (H-score: 265, 80% of cells showing RANKL expression score 3+). (c) RANK expression on the primary tumor in a young breast cancer patient not diagnosed during pregnancy (H-score: 140, 20% of cells showing RANK expression score 3+). (d) RANK expression on adjacent normal epithelial cells in a patient diagnosed with breast cancer during pregnancy (H-score: 210, 40% of cells showing RANK expression score 3+). RANK, receptor activator for nuclear factor κB; RANKL, receptor activator for nuclear factor κB ligand.

Figure 3

Expression of RANK and RANKL according to breast cancer subtype and correlation with Ki67. (a) Expression of RANKL by immunohistochemistry using the H-score (y axis) according to breast cancer subtypes in all patients, nonpregnant patients and pregnant patients. RANKL expression was higher in luminal A tumors (ER+, HER2–, Ki67 < 20%) compared with all other subtypes, particularly in pregnant patients (P <0.0001), than in nonpregnant patients (P = 0.32). (b) Negative correlation between Ki67 score by immunohistochemistry (y axis) and RANKL H-score (x axis) (P <0.0001, Pearson correlation = –0.29). (c) Expression of RANK by immunohistochemistry using the H-score (y axis) according to breast cancer subtypes in all patients, nonpregnant patients and pregnant patients. RANK expression was higher in triple-negative tumors compared with all other subtypes in all patients (P <0.0001), nonpregnant patients (P < 0.0001) and pregnant patients (P = 0.05). (d) Positive correlation between Ki67 score by immunohistochemistry (y axis) and RANK H-score (x axis) (P <0.0001, Pearson correlation = 0.37). ER, estrogen receptor; HER2, human epidermal growth factor receptor 2; RANK, receptor activator for nuclear factor κB; RANKL, receptor activator for nuclear factor κB ligand.

Figure 4

Correlation between RANK and RANKL H-score and mRNA levels. (a) Positive correlation between RANKL mRNA expression (y axis) and RANKL H-score by immunohistochemistry (x axis) (P <0.0001, Pearson correlation = 0.89). (b) Positive correlation between RANK mRNA expression (y axis) and RANK H-score by immunohistochemistry (x axis) (P = 0.01, Pearson correlation = 0.19). RANK, receptor activator for nuclear factor κB; RANKL, receptor activator for nuclear factor κB ligand.