Main

Numerous studies report on qualitative changes of LH in various physiologic and clinical conditions, as reflected by alterations in the ratio of biologic to immunologic activity of the hormone, i.e. its B/I ratio (for reviews, seeRefs. 1–3). However, the biochemical features of the LH molecules, responsible for such changes, have largely remained unidentified. With the advent of immunometric assays with improved sensitivity it has become apparent that many, although not all, of the previous findings on changes of the B/I ratio of LH have been due to a bias introduced by the conventional RIA methods, which overestimate low hormone concentrations(4–6). Therefore, the B/I ratios of LH at low hormone levels are erroneously low, and whenever there is an increase in LH concentrations, a greater fold-increase occurs in B-LH, resulting in artifactual elevation of the B/I ratio. We have earlier demonstrated the existence of this type of bias in the B/I ratios estimated during the pulsatile LH secretion, after gonadotropin-releasing hormone stimulation and during treatment with its agonists(5, 6). One of the earlier findings which needs reevaluation is the increase of the B/I ratio of LH during pubertal maturation(7–13). Because this finding is now widely cited in textbooks on endocrinology, it is important to test whether it can be verified by using the recent assay methods with improved sensitivity and specificity. In addition, we report here that in agreement with Debertin and Pomerantz(13), forskolin in a carefully selected concentration increases the sensitivity of mouse Leydig cells to LH stimulation, and it can be used to increase the sensitivity of thein vitro bioassay of human serum LH.

METHODS

Subjects. Fourteen healthy boys entering puberty were studied. Their ages at the beginning of the study ranged from 11 y, 5 mo, to 12 y(median, 11.8 y). The boys were examined every 3 mo for the stage of puberty[according to Marshall and Tanner(14)] for a total time of 18 mo. A peripheral blood sample was taken for hormone measurements at each visit. Seven of the boys were in pubertal stages G1 and 7 in stage G2 at the time of the first visit, and at the time of the last visit, the boys had advanced two stages, half of them being in stages G3 and G4. Duration of each pubertal stage was correlated and compared with the longitudinal standards af pubertal development(14). All subjects had pubertal progression within normal limits. Some data on the physical and endocrine maturation of the subjects during the course of the study are presented inTable 1. Testicular volume of 2 cm3 was considered the limit of stages G1 and G2. The length and width of each testis was measured for mean testis volume(15). The study was approved by the Ethical Committee of The Children's Hospital, University of Helsinki.

Table 1 Clinical and hormonal data of the 14 boys, substantiating pubertal progression
Full size table

Hormone measurements. I-LH was measured using an IFMA (LH spec, Delfia, Wallac OY, Turku, Finland). The assay has a sensitivity of 0.04 IU/L, and intraassay CV of 5.4% at normal LH level (1-5 IU/L). All samples were analyzed in the same assay run.

B-LH was measured by the mouse interstitial cell in vitro bioassay, essentially as described by us before(16, 17), with the following modifications. LH-free serum from healthy men treated with testosterone enanthate (200 mg/wk, intramuscularly)(18) for contraception was used to equalize the serum content of all standard and unknown bioassay samples. The LH-free serum had unmeasurable (<0.04 IU/L) level of LH as assayed by IFMA. The sensitivity of the assay was improved, as demonstrated recently by Debertin and Pomerantz(13), by inclusion of 1.5 μmol/L forskolin in the incubation medium. This increased the sensitivity of the assay about 10-fold, when the dose response curves of incubations in the presence and absence of forskolin were compared (see Fig. 1). The interassay CV of the improved bioassay was 20.2% and that of intraassay 9.1% at normal LH levels (1-5 IU/L). The sensitivity of the individual assays ranged from 0.005 to 0.01 mIU/tube (0.05 to 0.1 IU/L), which is 5-10-fold higher than the sensitivity of the LH in vitro bioassay in the absence of forskolin(16). Upon setting up the sensitized in vitro bioassay, multiple dilutions of unknown serum samples were assayed, and they displayed parallelism of testosterone responses with those of the LH standards. Therefore, when the study material was analyzed, duplicate serum samples at two dilutions ranging from 1:2 to 1:10 were assayed. Hence, the amount of serum per incubation tube ranged from 5 to 25%. The dilutions were selected according to the expected concentration (on the basis of IFMA). If parallelism of the unknown serum dilutions did not occur with the standards, the sample was reassayed.

Figure 1
figure 1

Standard curves of the in vitro bioassay of LH in the presence and absence of 1.5 μmol/L of forskolin (FK) and different concentrations of LH-free human serum. The LH standard used was the WHO 2nd IRP 80/552.

Full size image

The same standard, WHO 2nd International standard 80/552 was used in both assays, and the LH concentrations were expressed as international units per liter (IU/L).

The serum levels (in Table 1) of FSH were measured by an immunofluorometric assay kit (Delfia, Wallac OY), and those of testosterone using a RIA after chromatographic separation(19).

Statistical analyses. The data are expressed as mean ± SEM. The statistical analysis was performed using one-way analysis of variance, followed by Duncan's new multiple range test to identify the groups differing. A p value less than 0.05 was considered statistically significant.

RESULTS AND DISCUSSION

The effect of forskolin on sensitivity of the in vitro bioassay is demonstrated in Fig. 1. Inclusion of a 1.5 μmol/L concentration of forskolin in the incubations increased the sensitivity of the assay about 10-fold, in the presence of 10% serum to 0.005 mIU/assay tube (or 0.05 IU/L). It is also seen that increasing the percentage of serum in the bioassay tubes decreased the sensitivity. This is due to as yet unidentified inhibitory factors in the peripheral serum(16). We find it therefore important to include the same amount of serum in each assay tube, including the standards. This is achieved by equalizing the serum content of all standard and unknown samples with LH-free human male serum(16). In this way the dose response curves of standards and serum samples at different dilutions are parallel. The net effect of forskolin on the sensitivity is thus about 10-fold when standard curves containing the same amounts of serum are compared. The increased sensitivity by forskolin increased the variability of the assay. The intra- and interassay CV values (9 and 20%, respectively) are still tolerable, although 2-fold higher than in assays performed without forskolin(16). The increased sensitivity of this assay is an asset which is not offered by the other recent developments in in vitro bioassays of LH, using either Leydig tumor cells(20) or cell lines expressing recombinant LH receptor(21).

We also paid special attention to evaluating the linearity of the B-LH and I-LH responses (Fig. 2). When the B-LH and I-LH levels of all samples analyzed were correlated, the coefficient was moderate(r = 0.72). However, when we took into account the fact that there was considerable interindividual variation in the B/I ratios of LH (see below), and grouped the subjects according to their average B/I ratio (<2, 2-3.4, and >3.5), good correlation coefficients (r = 0.90-0.95) were found between the two measurements (Fig. 2). The correlation was linear, and neither of the assays appeared to over- or underestimate the low hormone concentrations, i.e. the regression lines crossed the axes at or near the origin. Hence, the B/I ratio measurements are not influenced by bias in the absolute LH levels, which are lowest in the early stages of puberty. We have previously shown that the lack of such linear correlation throughout the concentration range of LH measured results in biased B/I ratio measurements(5, 22).

Figure 2
figure 2

Correlation between the B-LH and I-LH concentrations measured in the serum samples of 14 peripubertal boys (n = 98). All results are presented in panel A (r = 0.72). Inpanels B-D, the observations are grouped according to the mean B/I ratio of each subject. Those with B/I ratio < 2 are in panel B(r = 0.92), those between 2 and 3.5 in panel C(r = 0.95), and those above 3.5 in panel D (r = 0.90). All correlations were significant at the p level < 0.001.

Full size image

The mean levels of B-LH, I-LH, and the B/I ratios at the different stages of puberty are presented in Fig. 3. I-LH increased between pubertal stages I and IV from 0.42 ± 0.13 to 2.24 ± 0.34 IU/L (p < 0.01) and B-LH from 1.35 ± 0.49 to 5.04 ± 0.78 IU/L (p < 0.01), but no concomitant changes occurred in the mean B/I ratios of LH (2.58-2.84). The same was apparent when data on the individual subjects are inspected (Fig. 4). Although the B/I ratio of the individual samples varied between 0.6 and 5.9, the variation of this parameter within the samples of one subject was rather small. The CV percentages ranged between 9 and 40, with the mean for the whole group of 14 subjects being 22%. The large interindividual variation in the B/I ratios of LH has been suggested to be due to individual differences in the composition of LH isoforms with differing carbohydrate moieties and intrinsic bioactivities(1–3), but it is also possible that there are substances in the peripheral serum that modify the activity of LH molecules at the receptor site [for references, see Huhtaniemi(23)]. Although we measured the LH levels in single serum samples, we trust that they reflect reliably the overall B/I ratio of LH in the subjects. We have previously shown that, using our assay methods, the B/I ratio of LH does not change during the LH secretion pulses(6).

Figure 3
figure 3

The mean (± SEM) levels of B-LH (top panel), I-LH (middle panel), and the B/I ratios (bottom panel) of the 14 peripubertal boys studied according to their stage of puberty (I-IV). The number of subjects analyzed at each pubertal stage is presented in brackets at the bottom of bars of the bottom panel. When different letters are above the bars these results differ significantly from each other (p < at least 0.05).

Full size image
Figure 4
figure 4

The B/I ratios of LH in the individual serum samples of the 14 peripubertal boys with advancing chronologic age, indicating a large interindividual, but small intraindividual, variation of the ratio. The lines connect the samples of each individual. The numbers in the lines indicate the pubertal stage at the individual sampling times as follows: 1-1, stage I; 2-2, stage II; 3-3, stage III; 4-4, stage IV.

Full size image

The increase in I-LH between stages I and IV was 5-fold(Fig. 3). It was somewhat less that the over 10-fold increase between stages I and IV that was recently demonstrated by Garibaldiet al.(24) using a sensitive immunoradiometric assay. This is explained by the fact that our subjects were at the late phase of pubertal stage I, because truly prepubertal boys have LH levels below 0.3 IU/L when monitored by IFMA(25, 26). The concomitant increase in LH between stage I and IV is only 2-3-fold when measured using conventional RIA(8–10, 12, 24). We have considered this previously the main reason why the B/I ratios tend to increase when the absolute LH levels increase, as monitored by RIA. However, we also identified another difference. The B-LH increase was in our measurements 4-fold, whereas the three earlier studies on LH bioactivity during pubertal maturation demonstrated 8-12-fold increases(8–10, 12). Dissimilar assay sensitivities are the most likely reasons for the difference. Our assay had a sensitivity below 0.1 IU/L, whereas that of the previous studies was 0.5 IU/L(12) or over 1 IU/L(8–10). The low levels are therefore subject to greater error in the previous studies. Another clear difference is the greater increase of B-LH in the earlier studies beyond pubertal stage I. The different standards used, different approaches to eliminate the confounding effects of the serum matrix, different species of interstitial cells (mouse or rat), and their preparations may be additional confounding factors. Many of the LH bioactivity measurements in pubertal children have been carried out using the rat interstitial cell in vitro bioassay(7–9, 11, 12). However, we have previously shown that there is a good correlation between the B-LH levels measured with the rat and mouse bioassays(27).

We paid special attention to testing the linearity of the B- and I-LH responses in our assays (Fig. 2) and found no apparent changes in the B/I ratio of LH during pubertal maturation from pre- to mid- and late stages, where the steepest increase in the ratio has previously been reported(7–11). Therefore, in contrast to earlier reports, we conclude that the pubertal maturation is not contributed to by increased intrinsic bioactivity (i.e. B/I ratio) of LH. Our different finding was obtained using an immunoassay with improved sensitivity and specificity and an in vitro bioassay with improved sensitivity and good correlation with I-LH measurements. In our opinion, the previous contention that the quality of LH changes during pubertal maturation may need reevaluation. The current in vitro bioassays of LH are still suboptimal in sensitivity, precision, and accuracy, and further method development is needed before the overall value of gonadotropin bioactivity/immunoreactivity relationships can be judged. The good correlation between B-LH and I-LH levels in the present measurements, together with the lack of changes in B/I ratios of LH during puberty, indicates that the highly sensitive IFMA of LH gives a correct estimate of LH bioactivity as well, emphasizing the clinical usefulness of this assay for evaluation of the pubertal process.