Stromal-derived factor 1 (SDF-1) and CXCR4 comprise a unique chemokine/chemokine receptor pair, exhibiting important functions in morphogenesis and growth regulation as well as attractant properties on T lymphocytes. No data are available on SDF-1 and CXCR4 in normal or pathological thyroid tissues. SDF-1, CXCR4, and CD18 messenger ribonucleic acid (mRNA) as a marker of leukocytic infiltration were quantified in tissues affected by thyroid adenoma (n = 11) and Graves' disease (GD; n = 16) using competitive RT-PCR. SDF-1 mRNA levels differed significantly between autonomous adenomas and the corresponding normal tissue, but not in GD between patients with low or high leukocyte infiltration, thyroid peroxidase, and TSH receptor autoantibodies, respectively. We found a strong correlation between CXCR4 and CD18 mRNA, which indicates CXCR4 expression by leukocytes. To define the cellular source of SDF-1 and CXCR4 in thyroid tissue, we examined various thyroid-derived cells. Fibroblasts are the most potent producers of SDF-1, although thyrocytes also secrete SDF-1 in vitro. Leukocytes showed very weak SDF-1 mRNA levels and no secretion of the chemokine. Immunohistology confirmed and extended these results; SDF-1 expression was found in fibroblasts, but not or very weakly in CD45(+) leukocytes and thyrocytes. Only leukocytes were CXCR4(+). As examined by flow cytometry, the number of CD3(+) T cells expressing CXCR4 is significantly higher in the thyroid than in peripheral blood. SDF-1 seems to be involved in thyroid tissue homeostasis in thyroid adenoma, but not in the maintenance of lymphocytic infiltration in GD.

Thyroid glands affected by Graves' disease (GD) show striking leukocytic infiltration, mainly by T-cells. The mechanisms by which the various leukocytes are maintained in the thyroid are unknown. Growth-regulated oncogene-alpha (GRO-alpha) in interaction with its receptor CXCR2 is a chemoattractant for both T-cells and neutrophils and may be one of the chemokines involved in the cell maintenance. GRO-alpha and CD18 mRNA as a marker of leukocytic infiltration were quantified in thyroid tissue using competitive RT-PCR. We found very high GRO-alpha mRNA levels in all thyroid tissues. In GD patients (n=16), the GRO-alpha mRNA did not correlate with the CD18 mRNA level or thyroid peroxidase and TSH-receptor antibodies in patients' sera. In thyroid autonomy (n=10), the GRO-alpha mRNA levels were significantly lower in autonomous single adenomas compared with the corresponding normal tissue. In order to define the cellular source of GRO-alpha mRNA and protein, we examined various thyroid-derived cells. Thyrocytes, thyroid-derived leukocytes and fibroblasts showed basal GRO-alpha mRNA and protein expression, which was remarkably upregulated by different stimuli in vitro. The expression of GRO-alpha by thyroid carcinoma cell lines confirms that thyrocytes may actually produce GRO-alpha. As shown by flow cytometry and immunohistology, CD68+ monocytes/macrophages are the only cell population strongly expressing CXCR2 in the thyroid.

Thyroid glands affected by Graves' disease (GD) show striking lymphocytic infiltration, mainly by CD45RO(+) T cells. The mechanisms by which the various lymphocytic subsets are recruited and maintained in the thyroid are unknown. RANTES (regulated on activation, normal T cells expressed and secreted) in interaction with its receptors (CCR1, CCR3, CCR4 and CCR5) may be one of the favorite chemokines involved in the cell trafficking and maintenance. RANTES messenger RNA (mRNA) was quantified in the thyroid tissue of 16 patients with GD and 7 patients with thyroid autonomy (TA), using competitive RT-PCR. We found a clear correlation between the RANTES mRNA level and 1) the degree of T-cell infiltration (r = 0.68), and 2) the level of serum antibodies to thyroid peroxidase (r = 0.76) in GD but not in TA patients. There was no difference between the autonomous nodules and the quiescent surrounding tissue in TA patients. To define the cellular source of RANTES mRNA and protein, we examined various thyroid-derived cells. Lymphocytes showed a markedly higher basal RANTES mRNA and protein level (mean +/- SEM; pg/mL, n = 3; 140 +/- 30) than thyrocytes (12 +/- 5) and fibroblasts (9 +/- 2). Lymphocyte stimulation with PMA enhanced RANTES secretion significantly (4490 +/- 200). Fibroblasts responded to stimulation with interleukin 1 (530 +/- 220) and tumor necrosis factor alpha (2780 +/- 1790), whereas thyrocytes did not. However, some thyroid carcinoma cell lines showed very high basal and stimulated RANTES expression. Lymphocytes expressed the mRNA of all chemokine receptors that bind RANTES. The number of CCR3(+) and CCR5(+) T cells was significantly higher in thyroid-derived leukocytes than in those in the peripheral blood stream. We conclude that RANTES expression, mainly by lymphocytes, is perhaps involved in the maintenance of lymphocytic infiltration and, therefore, in the autoimmune responses in GD.

We evaluated the presence and number of eosinophils at varying stages in the human corpus luteum from 27 ovaries of women at reproductive age. Eosinophils preferentially accumulated in dilated microvessels of the thecal layer transforming into septa of the corpus luteum. The granulosa layer under luteinization, the thecal layer, and haemorrhages in the former antrum each contained low, moderate and high numbers of extravasated eosinophils respectively. Eosinophils decreased rapidly during the stages of secretion and regression. Semi-quantitative reverse transcriptase-polymerase chain reaction (RT-PCR) and enzyme-linked immunosorbent assay (ELISA) systems were used to investigate the expression and regulation of the eosinophil-attracting chemokines RANTES (regulated on activation, normal T cell expressed and secreted) and eotaxin in granulosa cells obtained from follicular aspirates from women undergoing IVF. Contaminating leukocytes were determined by CD18 mRNA quantification. Granulosa cells expressed RANTES (n = 3; 43 +/- 14 pg/ml, mean +/- SEM). 4ss-phorbol-12-myristate-13-acetate (PMA; 211 +/- 53) and tumour necrosis factor alpha (TNFalpha) (238 +/- 59), but not interleukin (IL)-1 up-regulated RANTES at significant levels. In general, higher basal and stimulated RANTES mRNA and protein were found in cultures with higher CD18 mRNA levels than in those with lower levels. We found only traces of eotaxin mRNA and no eotaxin secretion, even in stimulated granulosa cell cultures, independently of leukocyte levels. Taken together, this is the first study demonstrating the selective presence of eosinophils in human periovulatory structures. RANTES, but not eotaxin, may play an active process in the accumulation of these cells.