Relaxin-2 is a novel biomarker for differentiated thyroid carcinoma in humans

Abstract

Relaxin's role in differentiated thyroid cancer (DTC) has been suggested but its characterization in a large clinical sample remains limited. We performed immunohistochemistry for relaxin-2 (RLN2), CD68 (total macrophages), CD163 (M2 macrophages) on tissue microarrays from 181 subjects with non-distant metastatic DTC, and 185 subjects with benign thyroid tissue. Mean pixels/area for each marker was compared between tumor and adjacent tissue via paired-t test and between DTC and benign subjects via t-test assuming unequal variances. RNA qPCR was performed for expression of RLN2, RLN1, and RXFP1 in cell lines. Amongst 181 cases, the mean age was 46 years, 75 % were females. Tumoral tissue amongst the DTC cases demonstrated higher mean expression of RLN2 (53.04 vs. 9.79; p < 0.0001) compared to tumor-adjacent tissue. DTC tissue also demonstrated higher mean expression of CD68 (14.46 vs. 4.79; p < 0.0001), and CD163 (23.13 vs. -0.73; p < 0.0001) than benign thyroid. These markers did not differ between tumor-adjacent and benign thyroid tissue groups; and amongst cases, did not differ by demographic or clinicopathologic features. RLN1 and RXFP1 expression was detected in a minority of the cell lines, while RLN2 was expressed by 6/7 cell lines. In conclusion, widespread RLN2 expression in DTC tissue and most cell lines demonstrates that RLN2 acts in a paracrine manner, and that RLN1 and RXFP1 are probably not involved in thyroid cancer cell signaling. RLN2 is a biomarker for thyroid carcinogenesis, being associated with but not secreted by immunosuppressive macrophages. These findings will guide further investigations for therapeutic avenues against thyroid cancer.

Keywords: Carcinogenesis; Macrophages; Relaxin; Thyroid carcinoma; Tumor microenvironment.

Fig. 1.

A. Validation of relaxin antibody for immunohistochemistry. Serial sections of ovaries from nonpregnant or pregnant (day 16) mice were probed with anti-relaxin antibody or control rabbit IgG at the same protein concentration. Brown staining shows the detection of relaxin in the corpora lutea of pregnant ovaries, but not in nonpregnant ovaries or when using control IgG. Scale bar:100um. B: Western blot of purified RLN1, RLN2, RLN3, pro-RLN2, porcine (Pr) or mouse (Ms) Rln1, INSL3, or human insulin (INS) under non-reducing (left panel) or reducing (right panel) conditions. The bands migrating at the approximate size of relaxin, pro-relaxin, or relaxin B-chain are shown (arrows).

Fig. 2.

Composite analysis of fluorescent immunohistochemistry comparing intratumoral to tumor-adjacent tissue amongst n = 181 DTC subjects for relaxin (RLN2) (Panel A); and comparing intratumoral tissue from n = 181 DTC subjects to n = 185 benign thyroid tissue for RLN2 (Panel B), CD68 (parent macrophage marker) (Panel C) and CD163 (M2 macrophage marker) (Panel D). ****p < 0.0001.

Fig. 3.

Fluorescent immunohistochemistry for RLN2, CD68, and CD163. Representative TMA specimens from benign, papillary-DTC and follicular-DTC. Adjacent tissue specimens shown are from the same subjects as the corresponding DTC. Positive staining is shown for RLN2 (green), CD68 (red), CD163 (white), or nuclei (blue). Left column: low magnification (scale bar 50um). Right column: high magnification (scale bar 25 μm). The dotted rectangle shows the corresponding high magnification field.

Fig. 4.

Gene expression analysis of RLN2 (Panel A), RLN1 (Panel B) and RXFP1 (Panel C) in thyroid cancer cells. RNA was extracted from thyroid cancer cell lines and subject to quantitative PCR. Data are shown as mean relative gene expression, n = 3. *p < 0.05 by ANOVA.