Differences between the glycosylation patterns of a pituitary thyroid-stimulating hormone calibrator (pitTSH) and serum samples have been shown to be responsible for nonidentical epitope expression and for introducing discrepancies in TSH measurements. We studied the feasibility of developing new candidate reference materials by remodeling recombinant TSH (recTSH) to generate potential mimics of serum TSH. Methods: Terminal sialylation and/or inner fucosylation of recTSH were remodeled by a

more »

... ination of enzyme treatments followed (or not) by lentil lectin-Sepharose affinity chromatography. The resulting TSH preparations were screened for epitope similarity in 23 immunoassays mapping 3 antigenic clusters common to the pitTSH 2nd International Reference Preparation (IRP) and the recTSH 1st IRP and then challenged against a pool of 63 patients with increased serum TSH (>60 mIU/L). Results : pitTSH was poorly correlated with serum TSH, with a mean (SD) slope of 2.124 (0.001), in contrast to recTSH [slope, 1.178 (0.056)]. Comparison of variably sialylated preparations with recTSH gave slopes of 0.860 (0.057) for desialylated TSH, 1.064 (0.057) for ␣2,3/ 6-oversialylated recTSH, and 0.953 (0.033) for ␣2,6-resialylated recTSH, indicating that TSH forms enriched in sialic acid closely resemble serum TSH. Further testing against serum TSH showed satisfactory agreement with both TSH preparations containing ␣2,6-sialic acid [slopes, 1.064 (0.057) and 0.953 (0.033)], particularly in the absence of nonfucosylated forms [0.985 (0.044)]. Conclusions: Glyco-engineered recTSH preparations enriched in sialic acid and inner fucose are promising candidates for future reference materials. These preparations may have advantages over existing preparations used for standardizing TSH measurements.