Increased expression of 25-hydroxyvitamin D-1alpha-hydroxylase in dysgerminomas: a novel form of humoral hypercalcemia of malignancy

Abstract

Humoral hypercalcemia of malignancy (HHM) is a common paraneoplastic disorder usually associated with increased synthesis of parathyroid hormone-related peptide (PTHrP). Unlike non-cancer forms of hypercalcemia, HHM does not routinely involve increased circulating levels of the active form of vitamin D, 1,25-dihydroxyvitamin D3 (1,25(OH)2D3). Dysgerminomas are a notable exception to this rule, previous reports having described hypercalcemia with elevated serum 1,25(OH)2D3. To investigate the etiology of this form of HHM we have characterized expression and activity of the enzyme that catalyzes synthesis of 1,25(OH)2D3, 25-hydroxyvitamin D-1alpha-hydroxylase (1alpha-hydroxylase), in a collection of 12 dysgerminomas. RT-PCR analyses indicated that mRNA for 1alpha-hydroxylase was increased 222-fold in dysgerminomas compared to non-tumor ovarian tissue. Parallel enzyme assays in tissue homogenates showed that dysgerminomas produced fivefold higher levels of 1,25(OH)2D3 compared to normal ovarian tissue. Immunolocalization studies indicated that 1alpha-hydroxylase was expressed by both tumor cells and by macrophages within the inflammatory cell infiltrate associated with dysgerminomas. The immunological nature of the increased 1,25(OH)2D3 production observed in dysgerminomas was further emphasized by correlation between expression of 1alpha-hydroxylase and the endotoxin recognition factors CD14 and toll-like receptor 4 (TLR4). These data suggest that inflammatory mechanisms associated with dysgerminomas are the underlying cause of the increased expression and activity of 1alpha-hydroxylase associated with these tumors. We further postulate that this autocrine/paracrine action of 1alpha-hydroxylase may lead to increased circulating levels of 1,25(OH)2D3 and a form of HHM which is distinct from that seen with PTHrP-secreting tumors.

Figure 1

Immunolocalization of 1α-hydroxylase in dysgerminoma tissue and a non-neoplastic ovary. Paraffin-embedded biopsies from the index case dysgerminoma (A) and a non-dysgerminomatous benign cystic ovary (B) were used to immunolocalize 1α-hydroxylase (A and B; magnification, ×200). Brown staining for 1α-hydroxylase was observed in macrophages and tumor cells within the dysgerminoma, while non-neoplastic tissue showed only weak staining in cystal epithelial cells.

Figure 2

Expression of genes associated with vitamin D metabolism and signaling in dysgerminomas. Data show the expression of mRNAs assessed by real-time RT-PCR. In each case mRNA expression is represented by: i] ΔCt values (± SD) which represent the number of PCR amplification cycles (Ct) at which logarithmic PCR plots cross a calculated threshold line. ΔCt values = Ct of the target gene − Ct of the housekeeping gene; ii] fold-increase in mRNA expression relative to the mean of non-neoplastic ovarian samples expressed as an arbitrary value of 1. The target genes analyzed in this fashion were 1α-hydroxylase (1α-OHase), 24-hydroxylase (24-OHase), and vitamin D receptor (VDR). In each case, the mean fold-change in gene expression for the dysgerminoma index case is shown as a horizontal line. Data shown are the mean ± SD. ***, P < 0.001 compared with non-neoplastic ovarian tissue, based on ΔCt values.

Figure 3

1α-hydroxylase activity is increased in dysgerminomas. Substrate [³H]-25OHD₃ (10 nmol/L) was added to cell homogenates in the presence of NADPH (0.2 mol/L) as cofactor and incubated for 5 hours. Conversion of the inactive substrate to active [³H]-1,25(OH)₂D₃ was determined by TLC and reported as mean fmole/hour/mg protein of [³H]-25OHD₃ converted to active [³H]-1,25(OH)₂D₃ in either dysgerminomas (n = 7) or non-neoplastic ovaries (n = 4) ± SD. All statistics were performed on raw data using analysis of variance (***, P < 0.001 compared with non-neoplastic ovaries).

Figure 4

Expression of putative target genes for 1,25(OH)₂D₃ in dysgerminomas. Data show the expression of mRNAs assessed by real-time RT-PCR. In each case, mRNA expression is represented by: i] ΔCt values (± SD); ii] fold-increase in mRNA expression relative to the mean of non-neoplastic ovarian samples as an arbitrary value of 1. The target genes analyzed in this fashion were CD45, PTHrP, p21, and HOXA10. In each case, the mean expression level for the dysgerminoma index case is shown as a horizontal line. ***, P < 0.001 compared with normal ovarian tissue, based on ΔCt values.

Figure 5

Expression of 1α-hydroxylase (1α-OHase) in dysgerminomas and normal ovaries correlates with CD14 and toll-like receptor 4 (TLR4). Levels of mRNA for 1α-hydroxylase were compared to CD14 (A) and TLR4 (B) mRNA levels (relative to 18S rRNA) and data shown as scatter plot for mean ΔCt values for each dysgerminoma (n = 12) and non-neoplastic ovary (n = 8). Correlation statistics (R value and statistical significance p) were performed on raw ΔCt triplicate means using Pearson product moment correlation.