Previous studies have indicated that hCG has a weak intrinsic thyroid-stimulating activity. Differences in the molecular composition and biological activity of hCG in patients with trophoblastic diseases and pregnant women occur, but are not well defined. Therefore, we have studied the effect of serum samples and purified hCG preparations from patients with trophoblastic diseases on T3 release from human and porcine thyroid slices in vitro. We examined 30 serum samples from 13 patients with nonseminomatous testicular germ cell tumors, 3 from women with choriocarcinoma, and 5 from patients with hydatidiform moles. In all but 1 serum sample from the tumor patients, but in none of 11 serum samples of pregnant women, T3-releasing activity was found. Two patients with testicular cancer and 1 patient with molar pregnancy experienced episodes of frank hyperthyroidism. Isoelectric focusing on polyacrylamide gels of tumor sera (n = 15) revealed substantial amounts of acidic isoelectric variants, pI 3.3-3.9, which were only barely detectable in pregnancy sera. The percentage of acidic hCG variants with pI 3.3-4.0 to total hCG with pI 3.3-5.2, as determined by hCG (+hCG beta) RIA of the eluted fractions of polyacrylamide gel isoelectric focussing, varied from 12-45% in sera of tumor patients and from 0-4% in pregnant sera. We purified the acidic variants of hCG with pI 3.6-3.8 (hCGav) from the urine of our patients. The beta-subunit of purified hCGav had a slightly higher mol wt (35,750) than that of hCG CR 119 (34,190) on polyacrylamide gel electrophoresis. The hCGav showed a dose-dependant stimulation of T3 release and cAMP generation from human thyroid slices, whereas the other hCG fractions on isoelectric focussing had no thyrotropic effect in similar dose levels. The TSH-like activity of hCGav could be roughly estimated as 10 mIU TSH/IU hCGav. Anti-hCG (+hCG beta) antiserum, but not anti-hTSH antiserum, neutralized the biological activity of hCGav. These findings strongly suggest that acidic hCG variants act as functional stimulators of the human thyroid in vitro. Since these molecular variants of hCG can exist in patients with trophoblastic diseases in significant amounts, they could be responsible for some cases of hyperthyroidism in trophoblastic diseases.

Cyclooxygenases-1 and -2 are both expressed in neuronal cells in vivo. In the neuroblastoma cell lines NG108 and N2a, however, only cyclooxygenase-1 was detectable. Differentiation of the cells with retinoic acid increased cyclooxygenase-1 mRNA and protein expression within 24 and 48 h, respectively. A further increase was observed when the cells were concomitantly treated with the glucocorticoid dexamethasone (a 2±3-fold increase compared with retinoic acid alone). In the absence of retinoic acid, dexamethasone only slightly up-regulated cyclooxygenase-1 expression. The inhibitor of protein synthesis cycloheximide abrogated the effect of dexamethasone, indicating the involvement of newly synthesised proteins. Retinoic acid increased the transcription of cyclooxygenase-1 mRNA, determined with a luciferasecoupled promoter construct. Dexamethasone only slightly augmented cyclooxygenase-1-promoter activity but increased cyclooxygenase-1 mRNA stability. Other corticosteroids, hydrocortisone and aldosterone, also up-regulated cyclooxygenase-1 whereas neurosteroids or oestrogen were ineffective. Up-regulation was mediated primarily by the glucocorticoid receptor, because the receptor antagonist RU486 strongly reduced the effects of all corticosteroids. This indicated that in NG108 cells, the mineralocorticoid aldosterone may bind to the glucocorticoid receptor. Treatment of NG108 or N2a cells with corticosteroids did not alter the morphological phenotype obtained during differentiation. We thus show that corticosteroids, which down-regulate cyclooxygenase expression in most cell types, up-regulate cyclooxygenase-1 during neuronal differentiation.

These results indicate differential regulation of the two types of TNF receptors at the mRNA and protein level during monocytic differentiation. Glucocorticoids interfered with mRNA expression of TNF-R80 and protein expression of TNF-R60, but the rather limited effect leaves the question of its functional relevance open. In contrast to other cytokine systems, TNF receptors do not appear to be major targets of glucocorticoid action.

No abstract