Serum CD5L as potential biomarker of thyroid hormone status during pregnancy

Abstract

The thyroid hormone (TH) status is routinely assessed by thyrotropin (TSH) and thyroxine (T4). Both biomarkers are mainly regulated by TH receptor beta, whereas many peripheral organs employ the alpha receptor. Serum cluster of differentiation 5-like molecule (CD5L) is a liver-derived protein under control of both TH receptor isoforms. However, clinical data on its relation to TH status are sparse. An additional biomarker of TH status is needed in particular during pregnancy, where the routine biomarkers become dynamically disturbed. This study aimed to determine possible covariates regulating serum CD5L and to test its potential suitability as additional TH biomarker during pregnancy. A sandwich ELISA for serum CD5L was established using newly raised antibodies. Circadian effects and the impact of liver disease on serum CD5L concentrations were assessed. Serum samples from pregnant women with well-characterized TH and trace element status were analyzed, and CD5L concentrations were correlated with other indicators of TH status including TSH, fT4, fT3, copper, and selenium concentrations. The new quantitative assay for CD5L showed high accuracy. Serum CD5L was stable in dilution and refreezing experiments and did not show strong circadian variance or dependency on liver disease. In serum of pregnant women, CD5L correlated positively to fT3, but not to fT4 or TSH. Significant positive correlations of CD5L were observed with serum levels of the TH-responsive trace elements selenium and copper. The data support the potential suitability of serum CD5L as an additional marker of TH status, with potential value for pregnancy and thyroid disease.

Keywords: circadian rhythm; cirrhosis; copper; diagnostics; feedback; hypothyroidism; inflammation; selenium.

FIGURE 1

Recombinant expression of human cluster of differentiation 5‐like molecule (CD5L) and ELISA development. (A) Human CD5L was recombinantly expressed in insect cells and purified by Ni‐NTA agarose affinity. The size marker (M) is indicated, along with the stained proteins in the fractions resulting from elution using 40 mM (lanes 2, 3) or 250 mM (lanes 5, 6) imidazole. (B) After immunization, two CD5L‐mAbs were selected and used to establish a quantitative ELISA for CD5L as a manual sandwich test. Six assay runs were conducted and displayed highly reproducible dose–response curves with recombinant CD5L. (C) The CV were determined over a wide concentration range of recombinant human CD5L and indicated a relatively large functional assay sensitivity. (D) Using a standard human serum pool, the determination of CD5L proved robust to thawing, yielding comparable results from samples that underwent up to six separate rounds of defrosting. (E) Adapting the ELISA to an automated assay system yielded again highly reproducible results. (F) The functional assay sensitivity characterized by a CV of less than 10% extended over approximately two orders of magnitude. CV, coefficient of variation; M, molecular weight marker [kD]; LLOQ; lower limit of quantification; ULOQ, upper limit of quantification, anchors; points denoting slope stability in a four parameter logistic algorithm for quantitative analysis of sigmoid functions.

FIGURE 2

Circadian variation of serum cluster of differentiation 5‐like molecule (CD5L) concentrations in healthy subjects. Serum samples collected at different times of the day from 10 human adults were analyzed for CD5L concentrations. The results indicate some relatively stable inter‐individual differences, but no strong variations with regard to circadian rhythmicity.

FIGURE 3

Potential variation of serum cluster of differentiation 5‐like molecule (CD5L) concentrations regarding sex, age, or stage of liver disease. Serum samples from patients with liver cirrhosis were analyzed for CD5L. Serum concentrations did not differ with regards to (A) sex, or (B) age of the patients. Severity of liver dysfunction is expressed as number of points in the CHILD‐Pugh score, and requirement for liver transplant is characterized by the number of MELD points. Serum CD5L concentrations correlated neither to (C) the CHILD‐Pugh, nor to (D) the number of MELD points. N = 105 samples, group comparison by nonparametric two‐sided Wilcoxon test, correlation analysis by nonparametric Spearman rank‐order correlation.

FIGURE 4

Associations of serum cluster of differentiation 5‐like molecule (CD5L) with the routine biomarkers of the thyroid hormone (TH) axis in pregnant women shortly before delivery. Serum concentration of CD5L showed (A) no significant correlation to free T4 (T4), but (B) a positive correlation to fT3. (C) No significant association between serum CD5L and thyrotropin (TSH) concentration was observed. Correlation analysis by nonparametric Spearman rank‐order correlation, R and p‐values are indicated.

FIGURE 5

Associations of serum cluster of differentiation 5‐like molecule (CD5L) in pregnancy with serum biomarkers of Se and Cu status. The serum CD5L concentrations in pregnant women showed significant positive correlation with (A) total serum Se, (B) total serum Cu, and (C) the product of total serum Cu and total serum Se concentrations. In comparison, (D) serum CD5L concentrations were significantly associated with serum selenoprotein P (SELENOP), but not with (E) serum CP or (F) the product of serum SELENOP and CP in pregnancy. Comparisons by nonparametric rank‐order Spearmen correlation; the R and p‐values are indicated as inset in the figures.