Analytical construct of reversible desensitization of pituitary-testicular signaling: illustrative application in aging

Abstract

Luteinizing hormone (LH) administered in pharmacological amounts downregulates Leydig cell steroidogenesis. Whether reversible downregulation of physiological gonadotropin drive operates in vivo is unknown. Most of the analytical models of dose-response functions that have been constructed are biased by the assumption that no downregulation exists. The present study employs a new analytical platform to quantify potential (but not required) pulsatile cycles of LH-testosterone (T) dose-response stimulation, desensitization, and recovery (pulse-by-pulse hysteresis) in 26 healthy men sampled every 10 min for 24 h. A sensitivity-downregulation hysteresis construct predicted marked hysteresis with a median time delay to LH dose-response inflection within individual T pulses of 23 min and with median T pulse onset and recovery LH sensitivities of 1.1 and 0.10 slope unit, respectively (P < 0.001). A potency-downregulation model yielded median estimates of one-half maximally stimulatory LH concentrations (EC(50) values) of 0.66 and 7.5 IU/l for onset and recovery, respectively (P < 0.001). An efficacy-downregulation formulation of hysteresis forecasts median LH efficacies of 20 and 8.3 ng·dl(-1)·min(-1) for onset and offset of T secretory burst, respectively (P = 0.002). Segmentation of the LH-T data by age suggested greater sensitivity, higher EC(50) (increased LH potency), and markedly (2.7-fold) attenuated LH efficacy in older individuals. Each of the three hysteresis models yielded a marked (P < 0.005) reduction in estimated model residual error compared with no hysteresis. In summary, model-based analyses allowing for (but not requiring) reversible pituitary-gonadal effector-response downregulation are consistent with a hypothesis of recurrent, brief cycles of LH-dependent stimulation, desensitization, and recovery of pulsatile T secretion in vivo and an age-associated reduction of LH efficacy. Prospective studies would be required to prove this aging effect.

Fig. 1.

Schema of dose-response hysteresis concept. Hormone (effector) concentrations within a pulse (top left) drive target gland secretion rates (top right). Relationship between input (concentration) and output (secretion) is estimated as a dose-response logistic function (bottom). Onset (solid arrowheads) and recovery (open arrowheads) phases of dose-response interface are estimated simultaneously with a hysteresis inflection point (delay time; see appendix).

Fig. 2.

Salient differences in luteinizing hormone (LH) secretion by age using dual-waveform integrative deconvolution model. Data are box-and-whisker plots from 13 young (Y) and 13 older (O) men. Median and interquartile (25% and 75%) and interdecile (10% and 90%) confidence intervals with individual extreme values are shown. MPP, mass released per pulse; S, secretion.

Fig. 3.

Individual onset (initial) and recovery (delayed) dose-response parameters in 26 individuals. Results from sensitivity (A), potency (B), and efficacy (C) downregulation models are shown. T, testosterone.

Fig. 4.

Dose-response estimates for LH concentration-dependent drive of T secretion in 1 young (A) and 1 older (B) man. Top rows: deconvolution-calculated T secretion (solid lines) and dose-response hysteresis-predicted T secretion (dashed lines) rates (vertical bars on x-axis denote T pulse-onset times). Middle rows: mean dose-response [initial (solid curve) and delayed (dashed curve)] estimates with allowable random effects on efficacy (dotted lines) and allowable hysteretic shifts in potency (left), sensitivity (middle), and efficacy (right). Bottom rows: time-shifted reconvolved LH concentration profile (dashed curve with vertical bars on x-axis to denote LH-pulse locations) and unshifted LH concentration profile (solid line with ◊ for unshifted LH secretory-pulse locations).

Fig. 5.

Age enhances Leydig cell T secretory sensitivity assessed before [onset (left)] and after [recovery (middle)] analytically estimated hysteretic downregulation of LH-stimulated T secretion in 26 [13 older (O) and 13 young (Y)] healthy men. Age augments the difference between onset and recovery sensitivities of LH-T drive (right), denoting greater desensitization. Box-and-whisker plots (see Fig. 2) show results on a natural logarithmic scale.

Fig. 6.

Potentiation by age of submaximal LH drive of T secretion in healthy men. Data are LH EC₅₀ values estimated in the sensitivity-hysteresis (downregulation) model (Fig. 1). Lower values denote greater LH potency. Box-and-whisker plots show results on a natural logarithmic scale.

Fig. 7.

Attenuation by age of estimated LH efficacy (asymptotically maximal LH-stimulated T secretion rates) in the sensitivity-downregulation model. See Fig. 4 for data format.

Fig. 8.

Three-step deconvolution procedure-based equation system (appendixes a and b).

Fig. 9.

Three-step deconvolution analysis applied to 24-h LH concentration-time series. Solid lines, measured LH concentrations; dashed lines, deconvolution-predicted fit; x-axis ticks, estimates of pulse times; ◊, daytime-waveform interval. Top: stage I (difference equation model). Middle: stage II (intermediate equation model). Bottom: stage III (integral equation model).

Fig. 10.

Deconvolution plot similar to that of Fig. 9, but for T concentration data. Solid line, fitted LH concentrations; dashed line, time-delayed LH feedforward signal.

Fig. 11.

Detailed deconvolution of LH (left) and T (right) concentration profiles shown in Figs. 9 and 10. Top: measured concentrations, fits, and pulse times. Middle: estimated secretion rates, pulse times, and daytime interval. Dashed lines (top) are reconvolution curves (fits of data). Solid lines are measured hormone concentrations (top) and calculated secretion rates (middle). ◊ on x-axes denote day-night waveform demarcations [concentrations (top) and secretion rates (middle)]. Bottom: day- and nighttime secretory-burst waveforms (gamma probability densities).