Chicken Ovalbumin Upstream Promoter Transcription Factor II in the Human Adrenal Cortex and Its Disorders¹

Chicken ovalbumin upstream promoter transcription factor II (COUP-TFII) is an orphan member of the steroid/thyroid hormone receptor superfamily. COUP-TFII has been demonstrated to negatively regulate the transcriptional activity of adrenal 4-binding protein, a steroidogenic cell-specific transcription factor that activates the transcription of various steroidogenic P450 genes. We therefore examined immunolocalization of COUP-TFII in the human adrenal cortex and its disorders, including functioning and nonfunctioning cortical tumors, to study its possible correlation with adrenocortical steroidogenesis. In nonpathological adrenal cortex, COUP-TFII immunoreactivity was marked in the nuclei of adrenocortical cells in definitive and fetal zones from 16 gestational weeks to 2 months after birth. Immunoreactivity for COUP-TFII was marked in the zona glomerulosa and weak in the zonae fasciculata and reticularis from 7 months to 8 yr of age, but thereafter markedly decreased in these zones (P < 0.05, between age 7 months to 8 yr and 24–62 yr of age, respectively). In adrenocortical tumors, COUP-TFII immunoreactivity was marked in the nuclei of tumor cells of aldosteroma (H score, 134 ± 15.9; P < 0.001 vs. Cushing’s adenoma and P < 0.05 vs. nonfunctioning adenoma and carcinoma), modest in nonfunctioning adenoma (82.7 ± 19.8) and adrenocortical carcinoma (79.6 ± 56.3), and low in Cushing’s adenoma (38.2 ± 24.5). Results from immunoblotting performed in seven cases of adenomas were consistent with those of immunohistochemistry. In the attached nonneoplastic adrenal cortex of the adenomas, immunoreactivity for COUP-TFII was markedly increased compared to that in nonpathological adrenal cortex in adults and was especially marked in the zona glomerulosa in the attached adrenal of aldosteroma (P < 0.001) and the zona fasciculata in that of Cushing’s adenoma (P < 0.05). COUP-TFII immunoreactivity was universally detected in stromal cells of the adrenal glands. These results suggest that COUP-TFII plays an important role in the regulation of steroidogenesis in human adrenal cortex and its disorders.

IT IS WELL known that adrenocortical neoplasms produce various corticosteroids excessively (1) and are associated with the abnormal expression of specific steroidogenic enzymes (2–5). Therefore, it is very important to examine the regulation of human adrenocortical steroidogenesis to obtain a better understanding of adrenocortical disorders.

Chicken ovalbumin upstream promoter transcription factors (COUP-TFs) belong to the steroid/thyroid hormone receptor superfamily (6–10). COUP-TFs are classified as orphan receptors because their ligands have not yet been identified. Various in vitro studies have demonstrated that COUP-TFs negatively regulate the transcriptional activity of various steroid/thyroid hormone receptors (11–14), including adrenal 4-binding protein (Ad4BP) or steroidogenic factor-1 (15, 16). Ad4BP is a steroidogenic cell-specific transcription factor that activates transcription of various steroidogenic P450 genes (17). In addition, Ad4BP expression has been demonstrated in the human adrenocortex and its disorders (18). Therefore, it is possible that COUP-TFs may play an important role as regulators of steroidogenesis. Two distinct types of COUP-TF have been identified in human: Ear-3/COUP-TF (COUP-TFI) (6, 7) and ARP-1/COUP-TF (COUP-TFII) (9, 10). Recently, Shibata et al. (19) reported overexpression of COUP-TFI messenger ribonucleic acid in nonfunctioning adrenocortical adenomas and suggested that COUP-TFI is one of the key regulators influencing steroid biosynthesis by adrenocortical adenomas. However, the expression of COUP-TFII has not been reported in the human adrenal in detail, and the biological roles of this transcription factor remain unclear. Therefore, in this study we immunohistochemically examined the expression of COUP-TFII in nonpathological and pathological specimens of the human adrenocortex and correlated these findings with those of Ad4BP.

Materials and Methods

Human adrenal

Fifty-one human adrenal specimens were examined in this study. Twenty-four specimens of nonpathological adrenal glands were obtained from autopsy files (16–35 weeks gestation, and 1 day to 62 yr of age) from Tohoku University Hospital (Sendai, Japan). Twenty-seven cases of adrenocortical tumors (six aldosteromas, six Cushing’s adenomas, six nonfunctioning adenomas with no clinical hormonal abnormalities, and nine adrenocortical carcinomas) were retrieved from the surgical pathology files of Tohoku University Hospital. Adrenocortical carcinomas were histologically diagnosed based on the criteria of Weiss (20). The specimens were fixed in 10% formalin for 24–48 h at room temperature and embedded in paraffin wax.

Primary antibodies

The antibody for COUP-TFII was provided by Dr. Sotirios K. Karathanasis (Cardiovascular Therapeutics, Parke-Davis Pharmaceutical Research Division, Warner-Lambert Co., Detroit, MI). The generation and characterization of the primary polyclonal antibody for COUP-TFII were described previously (21). Briefly, a bacterially expressed and affinity-purified peptide spanning the N-terminal and DNA-binding domains (amino acid residues 1–170) of COUP-TFII was used to raise antibodies in rabbits. The polyclonal antibody for Ad4BP was provided by Dr. K. Morohashi (National Institute for Basic Biology, Okazaki, Japan). The generation and characterization of the Ad4BP antibody have been described previously (17), and the application of this antibody in an immunohistochemistry study has been previously reported (18, 22).

Immunohistochemistry

Immunohistochemical analysis was performed using a streptavidin-biotin amplification method (Histofine Kit, Nichirei, Tokyo, Japan). After deparaffinization, slides were heated in an autoclave at 120 C for 5 min in citric acid buffer (2 mmol/L citric acid and 9 mmol/L trisodium citrate dehydrate, pH 6.0). The dilutions of primary antibodies used were as follows: COUP-TFII, 1:1500; and Ad4BP, 1:700. Antigen-antibody complex was visualized with 3,3′-diaminobenzidine solution [1 mmol/L 3,3′-diaminobenzidine, 50 mmol/L Tris-HCl buffer (pH 7.6), and 0.006% H₂O₂], and counterstained with methyl green. The immunohistochemical procedure was performed as carefully as possible under the same condition among the slides to evaluate the relative immunoreactivity (23). The immunoreactivity absorption test for COUP-TFII and Ad4BP consisted of incubating the antibody-antigen mixture containing equal volumes of optimally diluted antiserum to COUP-TFII or Ad4BP and its corresponding COUP-TFII or Ad4BP peptide solution, respectively, for 18 h at 4 C. After centrifugation, the resultant supernatants were used as preabsorbed antibodies. Negative controls for absorption were conducted, in parallel to the absorption test, by mixing COUP-TFII antiserum together with Ad4BP peptide solution or Ad4BP antiserum with COUP-TFII peptide solution. Normal rabbit IgG was used in place of the primary antibodies as a negative control.

Scoring of immunoreactivity

After completely reviewing immunohistochemical sections, relative immunoreactivity for COUP-TFII and Ad4BP in each zone of adrenocortex was classified into the following groups by blind ranking of each slide by three of the authors (T.S., J.T., and H.S.) independently: 2 = strongly positive, 1 = weakly positive, and 0 = negative. Disconcordant results among the observers were reevaluated together, using a multiheaded light microscope. The adrenals were classified into the following age groups in this study: 16–35 gestational weeks (n = 4), 1 day to 2 months (n = 4), 7 months to 8 yr (n = 5), 10–18 yr (n = 5), and 24–62 yr (n = 8), and statistical significance was evaluated among the age groups within each zone. Relative immunoreactivity of tumor cells was evaluated by an H scoring system, as described by McCarty et al. (24) with some modifications. Briefly, more than 500 tumor cells were counted in each case, and H scores were subsequently generated by adding together 2 × % strongly stained nuclei, and 1 × % weakly stained nuclei, giving a possible range of 0–200. Statistical significance was evaluated using a Bonferroni test, and P < 0.05 was considered as significant.

Immunoblotting

Immunoblot analysis was performed in seven specimens of adrenocortical adenoma (two aldosteroma, two Cushing’s adenoma, and three nonfunctioning adenoma). The tissues were homogenized in triple detergent lysis buffer containing 50 mmol/L Tris-HCl (pH 8.0), 150 mmol/L NaCl, 0.02% sodium azide, 0.1% SDS, 100 μg/L phenylmethylsulfonylfluoride, 1 μg/mL aprotinin, 1 μg/mL Nonidet 40, and 0.5% sodium deoxycholate at 4 C, followed by sonication. After centrifugation, 50 μg of the supernatant proteins (whole cell extracts) were subjected to SDS-PAGE (8% acrylamide gel). After SDS-PAGE, proteins were transferred to polyvinylidene difluoride (Immobilon P, Millipore Corp., Bedford, MA) in 25 mmol/L Tris, 250 mmol/L glycine, and 0.1% SDS for 3 h at 200 mA constant current. The blots were blocked in 5% nonfat dry milk, PBS, and 0.1% Tween-20 (Bio-Rad Laboratories, Inc., Hercules, CA) at 4 C overnight and then incubated with diluted antibody for COUP-TFII (1:500) or Ad4BP (1:1000) for 4 h at room temperature. After incubation with horseradish peroxidase-linked donkey antirabbit Ig (Amersham International, Aylesbury, UK; dilution: 1:1000) for 1 h at room temperature, antibody/protein complexes on the blots were detected using ECL plus Western blotting detection reagents (Amersham International). The immunointensity of the specific bands was measured by the LAS-1000 imaging system (Fuji Photo Film Co., Ltd., Tokyo, Japan).

Results

COUP-TFII- and Ad4BP-immunoreactive cells were observed in all adrenal glands examined. The immunoreactivity for COUP-TFII and Ad4BP was abolished by COUP-TFII and Ad4BP preabsorbed with the antigen, respectively, but remained unchanged in cells treated with COUP-TFII antiserum preabsorbed with Ad4BP peptide solution or Ad4BP antiserum preabsorbed with COUP-TFII peptide solution. No specific immunoreactivity was detected in normal rabbit IgG-treated slides.

Nonpathological adrenal cortex

The results of COUP-TFII immunoreactivity in cortical cells are summarized in Table 1A. COUP-TFII immunoreactivity was marked in the nuclei of cortical cells in both the definitive and fetal zones of the fetus (1.75 ± 0.50 and 1.50 ± 0.58, respectively; Fig. 1) and 1 day to 2 months after birth. At 7 months to 8 yr of age, immunoreactivity for COUP-TFII was strongly detected in the cortical cells of the zona glomerulosa (1.40 ± 0.55), whereas it was weak in the cortical cells of the zonae fasciculata and reticularis (0.80 ± 0.45, respectively). Relative immunoreactivity for COUP-TFII was markedly decreased in these zones in both adolescent and adult age groups: 10–18 and 24–62 yr (0.38± 0.52 in zona glomerulosa, 0.13 ± 0.35 in zonae fasciculata and reticularis; P < 0.05 vs. 7 months to 8 yr, respectively; Fig. 2). Marked immunoreactivity of COUP-TFII was consistently detected in the nuclei of stromal cells in all cases examined. The great majority of vascular endothelial cells was also positive for COUP-TFII immunoreactivity regardless of the vascular type, including artery, vein, and capillary.

Immunoreactivity for Ad4BP was marked in the nuclei of cortical cells in all zones from the fetus to the adult (Table 1B). Relative immunoreactivity for Ad4BP in each zone was not significantly different among the age groups examined.

Adrenocortical tumor

The results of COUP-TFII and Ad4BP immunoreactivity in adrenocortical tumors are summarized in Table 2. COUP-TFII immunoreactivity was markedly detected in tumor cells of aldosteroma (H score, 134 ± 15.9; P < 0.001 vs. Cushing’s adenoma and P < 0.05 vs. nonfunctioning adenoma and carcinoma; Fig. 3A), modest in tumor cells of nonfunctioning adenoma (H score, 82.7 ± 19.8) and adrenocortical carcinoma (H score, 79.6 ± 56.3), and low in tumor cells of Cushing’s adenoma (H score, 38.2 ± 24.5; Fig. 3B). COUP-TFII was strongly immunopositive in stromal cells, including vascular endothelium, in all cases.

Ad4BP immunoreactivity was markedly detected in tumor cells in all adrenocortical tumors examined, and the H score was not significantly different among the types of neoplasms examined.

Attached nonneoplastic adrenal cortex of adenomas

The results of COUP-TFII immunoreactivity in attached nonneoplastic adrenal cortex of adenomas are summarized in Table 3A. Relative immunoreactivity for COUP-TFII was increased in all three zones of the attached nonneoplastic adrenocortex compared to immunoreactivity in those of nonpathological adrenocortex in adult (age 24–62 yr), especially in the zona glomerulosa of the attached nonneoplastic adrenocortex in aldosteroma (1.50 ± 0.55; P < 0.001 vs. nonpathological adrenal ages 24–62 yr) and in the zona fasciculata of the attached nonneoplastic adrenocortex in Cushing’s adenoma (0.83 ± 0.75; P < 0.05 vs. nonpathological adrenal ages 24–62 yr; Fig. 4).

Ad4BP immunoreactivity was marked in the cortical cells in all three zones of the attached nonneoplastic adrenocortex (Table 3B), but was not significantly different among the tumor types examined.

Immunoblotting analyses of COUP-TFII and Ad4BP in adrenocortical adenomas

COUP-TFII and Ad4BP were detected as bands of 47 and 53 kDa, respectively, which correspond to those reported previously (18, 21), in seven cases of adenomas examined (Fig. 5). The results of immunointensity of COUP-TFII and Ad4BP in adrenocortical adenomas are shown in Table 4. The results of immunoblotting were consistent with those of immunohistochemistry (Table 2).

Discussion

In our study COUP-TFII immunoreactivity was significantly decreased in cortical cells of all zones with increasing age. In addition, COUP-TFII immunoreactivity was increased in the attached nonneoplastic adrenocortex of adenomas, especially in the zona glomerulosa of the attached nonneoplastic adrenocortex in aldosteroma, and in the zona fasciculata in Cushing’s adenoma. In the human adrenal, the production of corticosteroids, including cortisol and adrenal androgens, is known to be progressively increased after 5–8 yr of age (25–27). Nonfunctioning cortical adenomas can produce biologically active corticosteroids as do aldosteroma and Cushing’s adenoma, and the expression of steroidogenic enzymes is suppressed in the attached nonneoplastic adrenocortex of these adenomas (4, 5). Therefore, the results from our study show that the expression of COUP-TFII-immunoreactive protein tends to be inversely correlated to steroidogenesis activity in adrenocortical parenchymal cells and suggests that COUP-TFII may be involved as a repressor of steroidogenic enzyme expression in the human adrenocortex.

In the present study COUP-TFII immunoreactivity was marked in aldosteroma, modest in nonfunctioning adenoma and adrenocortical carcinoma, and low in Cushing’s adenoma. These results also suggest an inverse correlation between COUP-TFII immunoreactivity and P450c17 (CYP17) expression, because the expression of CYP17 is very low in aldosteroma and high in Cushing’s adenoma (3, 19, 28). Ad4BP has been demonstrated to stimulate transcription of the bovine CYP17 promotor, whereas COUP-TFs inhibit transcription of this gene by competing for binding to overlapping recognition elements (16). Therefore, COUP-TFII is postulated to inhibit the transcription of CYP17 gene in aldosteroma. However, COUP-TFII immunoreactivity was weak in zona glomerulosa of nonpathological adult adrenal in our study, in which P450c17 is not expressed (28). Therefore, overexpression of COUP-TFII in aldosteroma may be partially associated with the process of tumorigenesis. Recently, Shibata et al. (19) reported that COUP-TFI messenger ribonucleic acid expression is markedly elevated in nonfunctioning adenoma and low in aldosteroma and Cushing’s adenoma. This expression pattern is different from that of COUP-TFII protein expression in our study. Therefore, COUP-TFI and COUP-TFII may differently regulate the steroidogenesis of the human adrenocortex and its disorders.

Marked immunoreactivity for Ad4BP was consistently detected in cortical cells in all cases examined, including nonpathological adrenal cortex from fetus to adult, adrenocortical neoplasms, and the attached nonneoplastic adrenocortex of adenomas. These findings are consistent with previous reports regarding the development of the adrenal in the mouse (29) and adult human adrenal and its neoplasms (18, 19). Therefore, Ad4BP expression is considered to be essential for the maintenance of the biological characteristics of adrenocortical cells (18). COUP-TFs have been reported to inhibit trans-activation by Ad4BP in previous in vitro investigations (15, 16), and therefore, the relative amounts of Ad4BP and COUP-TFs may be important in regulation of the gene expression of steroidogenic enzymes.

In this study, COUP-TFII immunoreactivity was markedly detected in Ad4BP-negative stromal cells and the great majority of vascular endothelial cells. In the mouse, Pereira et al. (30) reported that COUP-TFII was expressed in mesenchymal cells of many organs, and Tsai and Tsai (14) postulated that COUP-TFII plays a significant role during tissue development and differentiation. The pathophysiological function of COUP-TFII in the vascular system remains unknown, but Pereira et al. (31) recently reported that a targeted deletion in the COUP-TFII gene results in embryonic lethality with defects in angiogenesis. Our present results are consistent with these reports, and expression of COUP-TFII in mesenchymal cells is therefore considered to play an important role in the development and homeostasis of human adrenal glands, possibly through parenchymal-stromal interactions. However, this hypothesis requires further investigation for clarification.

In summary, we examined the expression of COUP-TFII in human adrenocortex and its disorders. COUP-TFII expression is inversely correlated to steroidogenesis in the adrenocortex and is marked in aldosteroma, suggesting that COUP-TFII is one of key regulators of adrenocortical steroidogenesis.

We appreciate the assistance of Ms. Miki Yoshizawa, Department of Molecular Biology, Tohoku University School of Medicine (Sendai, Japan), and Ms. Chika Kaneko, Department of Pathology, Tohoku University School of Medicine, for their skillful technical assistance.