The importance of chromatographic resolution when analyzing steroid biomarkers.

The analyses of endogenous substances as biomarkers presents challenges that are distinctly different from the analyses of drugs or other xenobiotic substances. This is particularly true for estrogens. When no matrix is available which does not contain some level of the biomarker of interest, specificity cannot be demonstrated. Therefore it cannot be known whether the analyte signal includes a response from another substance. This uncertainty is increased by the fact that biomarkers are often created as part of a complex biosynthetic process that also creates a large number of substances with very similar structures and sometimes the same mass. Because of this, the two most powerful selectivity tools in the analysis of drugs, mass selective detection and MS/MS, are often rendered ineffective. The only remaining selectivity tool is chromatography and as will be demonstrated these separations can be very challenging. Failure to achieve specificity is perhaps the leading cause for inaccuracy of biomarker data and inter-laboratory variability.

Recombinant cell ultrasensitive bioassay for measurement of estrogens in postmenopausal women.

A recent analysis of data from nine studies provided convincing evidence that plasma estradiol measurements predict the risk of breast cancer in normal postmenopausal women. However, the median values detected by the various assays used in this study varied by 5-fold. These and other published data in normal postmenopausal women suggest that assays measuring low plasma estradiol concentrations suffer from problems of sensitivity, specificity, and precision. Availability of a practical, low-cost, specific, precise, and ultrasensitive estrogen assay might allow enhanced prediction of the risk of breast cancer and provide an objective means of selecting postmenopausal women for breast cancer prevention. A recombinant cell ultrasensitive bioassay (RCUB) for estrogen was recently validated for use in prepubertal children. We postulated that the RCUB might also prove useful for measurement of postmenopausal levels and designed the present study to examine this possibility. Thirty normal postmenopausal volunteers provided blood samples for measurement of estrogen by RCUB and, for comparison, by RIA. The estrogenic activity measured by RCUB [mean +/- sd, 11.9 +/- 10.9 pmol/liter (SI units, 3.23 +/- 2.96 pg/ml] was significantly lower than estradiol levels measured by RIA [43.7 +/- 44.0 pmol/liter (11.9 +/- 12.0 pg/ml)] in our volunteer subjects (P < 0.00001). Nonetheless, plasma estradiol levels measured by bioassay were significantly correlated with the estrogenic activity measured by RIA (r = 0.84) and by gas chromatography/tandem mass spectrometry (r = 0.85). To obtain biological evidence of the validity of the RCUB, we related plasma estrogen levels to body weight and body mass index and found highly significant correlations (r = 0.54 and r = 0.53, respectively). Surprisingly, 28 of 30 postmenopausal women were found to have estrogen levels in the prepubertal range with the RCUB. The levels detected by RCUB were similar to those previously reported using an ultrasensitive but less practical yeast bioassay. These results provide validation for the RCUB in postmenopausal women and suggest that it might prove useful for selection of women for drug therapy to prevent breast cancer.