Influence of antisera, purification procedures, and tracers on validity of a testosterone radioimmunoassay

Testosterone analysis in prostate cancer patients

Testosterone is an essential steroid hormone associated with a wide variety of biological processes in humans. In prostate cancer, androgen signaling is an important driver of tumor cell growth. Depletion of gonadal testosterone, achieved by surgical or chemical castration, prevents androgenic signaling and temporally reduces, stops or reverses tumor growth before inevitable progression to castration-resistant prostate cancer occurs. Additional treatment strategies targeting androgenic signaling have become available, although these are without curative intent. While circulating testosterone is also associated with disease risk and potential clinical utility, the main use in the clinical lab is monitoring adequate castration and subsequent resistance to therapy. Adequate castrate testosterone concentrations are currently based on over 50 year-old double-isotope derivative assays that are disputed in automated immunoassay (IA) analysis. The debate has been further fueled with the introduction of mass spectrometry-based assays for testosterone, offering a substantial increase in sensitivity and specificity. In this review, we discuss testosterone regulation and androgen deprivation therapy in prostate cancer. We provide an overview of the developments in testosterone analysis for monitoring adequate castration and resistance to therapy. Current clinical practice and future clinical utility will be discussed. Finally, clinical and research recommendations will be presented.

Kisspeptin-54 stimulates gonadotropin release most potently during the preovulatory phase of the menstrual cycle in women

CONTEXT

Kisspeptin, the endogenous ligand of the G protein-coupled receptor 54, is a key regulator of the hypothalamo-pituitary-gonadal (HPG) axis. GPR54-null mice exhibit reproductive dysfunction, and exogenous kisspeptin potently stimulates the HPG axis in rodents, primates, and human males. The effects of kisspeptin administration to human females are unknown.

OBJECTIVE

Our objective was to investigate the effects of kisspeptin on LH release during the menstrual cycle in female volunteers.

DESIGN

Bolus sc kisspeptin-54 was administered to female volunteers, and plasma gonadotropins were measured.

SETTING

The study took place at a hospital clinical research facility.

VOLUNTEERS

Subjects were healthy female volunteers with regular menstrual cycles.

INTERVENTION

1) Volunteers received a sc bolus injection of kisspeptin-54 (0, 0.2, 0.4, 0.8, 1.6, 3.2, and 6.4 nmol/kg; n = 3-4 per dose) in the follicular phase; and 2) volunteers (n = 8) received a sc bolus injection of either kisspeptin-54 (0.4 nmol/kg) or saline in random order during each phase of the menstrual cycle.

MAIN OUTCOME MEASURES

Plasma gonadotropins were measured.

RESULTS

1) Kisspeptin-54 caused a dose-dependent increase in mean LH over time at doses from 0.2-6.4 nmol/kg. 2) Kisspeptin-54 increased plasma LH compared with saline injection in all phases of the cycle. The effect of kisspeptin was greatest in the preovulatory phase and least in the follicular phase of the cycle [mean increase in LH over baseline (IU/liter) +/- sem for follicular phase was 0.12 +/- 0.17; preovulatory phase, 20.64 +/- 2.91 (P < 0.001 vs. follicular phase); luteal phase, 2.17 +/- 0.79 (P < 0.01 vs. follicular phase)].

CONCLUSION

Elevation of plasma kisspeptin in human females potently stimulates LH release in the preovulatory phase and provides a novel mechanism for manipulation of the HPG axis in women.

Normal or early development of puberty despite gonadal damage in children treated for acute lymphoblastic leukemia

To determine the timing of pubertal development and the frequency of gonadal dysfunction in children who survive acute lymphoblastic leukemia, we assessed pubertal status and the plasma levels of sex steroids, gonadotropin, and inhibin in 45 children (20 girls and 25 boys) who had received combination chemotherapy along with 24 Gy of irradiation to the cranium (modified LSA2L2 protocol). We also reexamined testicular biopsy specimens, obtained at the time of the cessation of chemotherapy, for the presence of germ cells. Germ-cell damage, indicated by marked elevations in the plasma level of follicle-stimulating hormone (P less than 0.001 for the comparison with normal children), was evident in both sexes and was confirmed in the boys by the absence of germ cells in the testicular biopsy specimens and by the small size of the testes for pubic-hair stage. Only 44 percent of the pubertal girls had measurable plasma inhibin levels, as compared with more than 93 percent of normal pubertal girls. Although plasma sex-steroid levels were normal, the secretion of luteinizing hormone in response to stimulation with gonadotropin-releasing hormone was elevated in the pubertal children (P less than 0.01 for the comparison with normal controls)--a finding that suggests compensation for decreased gonadal function. Despite clear evidence of gonadal damage, girls had early menarche at a mean age (+/- SD) of 11.95 +/- 0.91 years, as compared with the Australian standard of 12.98 +/- 1.11 years (P less than 0.01). Thus, in girls, puberty was early despite primary gonadal damage. Thirteen of 23 boys reached puberty at a mean age of 12.36 +/- 0.73 years. We conclude that treatment for acute lymphoblastic leukemia may lead to primary gonadal damage in both sexes, regardless of the age at treatment, but that the secondary characteristics of puberty develop at a normal age or, in girls, relatively early.

Biases in the assays of steroids and their binding proteins

Many immunoassays exhibit a bias. In such assays, the results are inaccurate, i.e. they deviate from the true value. The biases are mostly due to interfering compounds originating from the biological material assayed and/or from reagents. Sometimes systematic errors in calculation are also involved. The magnitude of the bias should be determined for every assay method, in order to make possible an assessment of reliability of the method. Biases frequently occur also in the measurements of steroid binding proteins, such as receptors, sex hormone binding globulin, corticosteroid binding globulin, etc. These biases are mostly due to the assumption that the measurements are performed under saturation conditions. These biases can be avoided by conducting the measurements at several concentrations of the ligand and by an appropriate correction for non-specific (low affinity) binding. In the assays of "free" steroids, biases are frequently encountered because of the disturbance of equilibrium in the course of the measurement proper. These biases can be minimized by a careful choice of experimental conditions.

Reliability of progesterone measurements in urine

A radioimmunoassay (RIA) of progesterone in urine is described. After the addition of labelled progesterone, morning urine was extracted with n-hexane adn the residue was either directly subjected to RIA, or chromatographed on celite prior to RIA. The progesterone from celite chromatography was radiochemically pure. RIA after chromatography was therefore considered valid. The non-chromatographed procedure resulted in overestimations, the degree of which was inversely proportional to progesterone content. The results obtained by the two procedures were well correlated (r = 0.88 and 0.93, for 2 different groups of samples.

Validation of radioimmunoassay for estradiol-17 beta by isotope dilution-mass spectrometry and by a test of radiochemical purity

Isotope dilution-mass spectrometry (ID-MS) was used as a reference method to determine the concentration of estradiol-17 beta (E2) in five different plasma pools (concentrations ranging from 0.040 to 65 nmol/l). The same plasma pools were also subjected to radioimmunoassay (RIA) using five different antisera of largely varying specificity. With the best antiserum (E), a direct RIA apparently gave accurate results (i.e. results statistically indistinguishable from those obtained by ID-MS) at all levels except the lowest one (0.040 nmol/l). It was shown, however, that the apparent accuracy of this RIA to some extent could be due to a lowering effect of lipids in the serum masking a lack of specificity of the antibodies. With the least specific antiserum (A), accurate results were obtained only after chromatography. However, in the assay of the lowest concentration of E2 with this antiserum there was a significant overestimation, even after chromatography. The other three antisera (B, C, D) of average quality gave accurate results in assays of plasma diethyl ether extracts in various numbers of the plasma pools tested, depending on their intrinsic specificity. This specificity was not correlated with the cross-reaction reported for individual antisera. ID-MS is difficult to use in most laboratories. We demonstrate here that the validity of a RIA may in this case be assessed by a relatively simple method, the test of radiochemical purity (RP-test). This test is based on the measurements of specific activity (e.g. dpm/pg) in small consecutive fractions of the chromatographic zone which is usually employed for the RIA.

Problems of validity in steroid immunoassays

A valid steroid immunoassay is the one which correctly describes the contents in unknown samples. Such an assay has to be accurate (devoid of bias) and precise. Assay accuracy can be assessed by a comparison of the results with those obtained by a reference method, e.g. gas liquid chromatography-mass spectrometry. Alternatively, the accuracy assessment can be made in terms of the purity of the compound assayed. The test of parallelism or the test of radiochemical purity can be used for this purpose. The former test, however, is not fully reliable. The degree of precision is of great importance especially when validity of individual measurements is considered. A low reproducibility (a high between-assay variation) leads to results of uncertain validity. An increased standardization and simplification of assays may result in a substantial improvement of the precision, and thus validity.