Osteoprotegerin (OPG), The Endogenous Inhibitor of Receptor Activator of NF-κB Ligand (RANKL), is Dysregulated in BRCA Mutation Carriers

Abstract

Breast cancer development in BRCA1/2 mutation carriers is a net consequence of cell-autonomous and cell nonautonomous factors which may serve as excellent targets for cancer prevention. In light of our previous data we sought to investigate the consequences of the BRCA-mutation carrier state on RANKL/osteoprotegerin (OPG) signalling. We analysed serum levels of RANKL, OPG, RANKL/OPG complex, oestradiol (E2), and progesterone (P) during menstrual cycle progression in 391 BRCA1/2-mutation carriers and 782 noncarriers. These studies were complemented by analyses of RANKL and OPG in the serum and mammary tissues of female cynomolgus macaques (n = 88) and serum RANKL and OPG in postmenopausal women (n = 150). BRCA-mutation carriers had lower mean values of free serum OPG in particular in BRCA1-mutation carriers (p = 0.018) compared with controls. Among BRCA1/2 mutation carriers, lower OPG levels were associated with germline mutation locations known to confer an increased breast cancer risk (p = 0.003). P is associated with low OPG levels in serum and tissue, particularly in BRCA-mutation carriers (rho = - 0.216; p = 0.002). Serum OPG levels were inversely correlated (rho = - 0.545, p < 0.001) with mammary epithelial proliferation measured by Ki67 expression and increased (p = 0.01) in postmenopause. The P-RANKL/OPG system is dysregulated in BRCA-mutation carriers. These and previously published data provide a strong rationale for further investigation of antiprogestogens or an anti-RANKL antibody such as denosumab for breast cancer prevention.

Keywords: BRCA1 and BRCA2 mutations; Breast cancer; Cancer prevention; Carcinogenesis; OPG; RANKL.

Supplemental Fig. 1

Case control differences in hormone values. Mean difference in (log) progesterone, oestradiol and RANKL/OPG complex in serum between women with and without a mutation in the BRCA1 and/or BRCA2 gene as a function of the menstrual cycle, with 95% confidence bands. LRT = likelihood ratio test, df = degrees of freedom, RANKL = receptor activator of NF-κB ligand, and OPG = osteoprotegerin.

Supplemental Fig. 2

Sum of absolute differences in hormone values. a) Sum of absolute mean differences in log serum progesterone, oestradiol, OPG, and RANKL between women with and without a mutation in the BRCA1 and/or BRCA2 gene as a function of the menstrual cycle with bootstrapped 95% confidence bands, and individual component serum differences. Black line represents the appropriate null value for comparison, based on a multivariate half-normal distribution. b) The ratio of the (geometric) mean differential expression relative to the null. RANKL = receptor activator of NF-κB ligand and OPG = osteoprotegerin.

Supplemental Fig. 3

Correlation of serum progesterone and RANKL (a,c) or RANKL/OPG complex (b,d). Correlation between luteal phase serum progesterone and free serum RANKL or RANKL/OPG complex in human (a,b) BRCA wildtype and (c,d) BRCA1/2 mutation carriers. CI = confidence interval, RANKL = receptor activator of NF-κB ligand, and OPG = osteoprotegerin.

Supplemental Fig. 4

Correlation of serum progesterone and OPG. Correlation between luteal phase serum progesterone and free serum OPG in human (a) BRCA wild-type and (b) BRCA1/2 mutation carriers. CI = confidence interval and OPG = osteoprotegerin.

Supplemental Fig. 5

Correlation of mammary gland progesterone receptor (PgR) expression in (a) alveolar and (b) ductal cells in cynomolgus macaques. PgR = progesterone receptor, IHC = immunohistochemical, OPG = osteoprotegerin, CEE = conjugated equine oestrogens, and MPA = medroxyprogesterone acetate

Supplemental Table 1

Summary of Tests of Difference between groups for mean functions of time. Likelihood ratio tests used to test between the models that include interactions of time functions and group plus individual group term and time functions versus models that only include the time functions. *LRT test of difference between groups based on 5 degrees of freedom (df) for spline functions or 3 df for trigonometric functions; ^†LRT test of difference between groups from a multivariate model with {progesterone, estradiol, RANKL, OPG} based on 20 df for spline functions and 12 df for trigonometric functions.

Fig. 1

Serum analysis of hormones. Analysis of (log) (a) progesterone, (b) oestradiol, (c)RANKL, (d) OPG, and (e) RANKL/OPG complex in serum from women with and without a mutation in the BRCA1 or BRCA2 gene as a function of the menstrual cycle. The ranges of concentrations were: sRANKL (2.9 pg/ml–2135 pg/ml), sOPG (10.8 pg/ml–1414 pg/ml) and sRANKL/OPG complex (17.4 pg/ml–225,101 pg/ml). RANKL = receptor activator of NF-κB ligand, OPG = osteoprotegerin, and CI = confidence interval.

Fig. 2

Case–control differences in hormone values. Mean differences in (log) (a–c) RANKL and (d–f) OPG in serum between women with and without a mutation in the BRCA1 and/or BRCA2 gene as a function of the menstrual cycle, with 95% confidence bands. RANKL = receptor activator of NF-κB ligand, LRT = likelihood ratio test, df = degrees of freedom, and OPG = osteoprotegerin.

Fig. 3

Association between the nucleotide position of the BRCA1/2 germline mutation and serum OPG levels. The estimated hazard ratio (HR) indicating the risk for breast cancer depending on the nucleotide position of the BRCA1/2 mutation was regressed on the log of free serum OPG (pg/ml), adjusted for age and menstrual cycle day at sample donation (a). Serum samples from human volunteers have been split according to the HR (< 0.9; < 0.9–1.1; > 1.1) given by the volunteers nucleotide position of the BRCA mutation and serum OPG levels blotted for each group. OPG = osteoprotegerin.

Fig. 4

Hormonal effects on RANKL and OPG in tissue and serum. Expression of mammary gland tissue (a) RANKL and (c) OPG mRNA and matched serum (b) RANKL and (d) OPG concentrations in ovariectomised adult female cynomolgus macaques treated for two years with placebo (control; n = 31), CEE at 0.042 mg/kg (n = 28) or CEE + MPA at 0.167 mg/kg (CEE + MPA; n = 29). Tests of group mean differences adjusted by Dunnett method for multiple comparisons with a control. The ranges of concentrations were: sRANKL (2.8 pg/ml − 191 pg/ml), sOPG 12.8 pg/ml-293 pg/ml). RANKL = receptor activator of NF-κB ligand, OPG = osteoprotegerin, CEE = conjugated equine oestrogens, and MPA = medroxyprogesterone acetate.

Fig. 5

Serum OPG and mammary gland proliferation. Correlation between serum OPG and mammary gland Ki67 (a) mRNA and IHC labelling in (b) alveolar epithelial cells and (c) ductal epithelial cells in ovariectomised adult female cynomolgus macaques treated for two years with placebo (control), CEE at 0.042 mg/kg CEE or CEE + MPA at 0.167 mg/kg (CEE + MPA). IHC = immunohistochemical, OPG = osteoprotegerin, CEE = conjugated equine oestrogens, and MPA = medroxyprogesterone acetate.

Fig. 6

Mammary gland proliferation stratified by RANKL and RANK expression in the mammary gland and OPG levels in the serum. Cynomolgus macaques were divided according to expression of (a) RANKL and (b) RANK in their mammary gland and then substratified according to the serum OPG level (high/low, above and below the median). OPG = osteoprotegerin, RANKL = receptor activator of NF-κB ligand, IHC = immunohistochemical, and RANK = receptor activator of NF-κB.

Fig. 7

Serum levels of OPG and RANKL. Serum levels of (a) OPG and (b) RANKL in premenopausal (n = 832) and postmenopausal (n = 150) women. The ranges of concentrations in postmenopausal women were: sRANKL (2 pg/ml–361 pg/ml) and sOPG (42 pg/ml–2237 pg/ml). OPG = osteoprotegerin and RANKL = receptor activator of NF-κB ligand.