GRF Dose Response Curves in the Presence and Absence of Somatostatin

Brazeau et al. (14) obtained dose-response curves for a range of doses of native hpGRF-44 in the presence of (1) zero SRIF, (2) 1 nM SRIF-14, or (3) 1 nM SRIF-28 (Fig. 7). It is assumed that the pituitary cells are initially desensitized to constitutive activation (equivalent to a level of GRF of r0), it can be seen, by applying Eq. (5), that the initial concentration of free receptors is f0 = kbfT / (k r0 + kb). Expression (9) can then be rewritten where Ko = kxr0 + kb. This was fitted by nonlinear least squares to these three sets of data allowing different SRIF concentration-dependent parameters, kr, kb, but all remaining parameters the same, resulting in the fitted curves also plotted on Fig. 7. Where the true value of a parameter is close to zero it sometimes happens that the least squares estimate for a specific data set is negative and this happened in this case for kb at zero dose of SRIF. In order for other parameters dependent on it to make sense it was necessary to constrain the estimate of kb to be positive. However, the fit changed very little as did the other parameters except kr for zero SRIF (which was inversely proportional to kb) for various

GRF dose-response curve

GRF dose-response curve

log(grf dose)

Fig. 7. Scaling the model parameters to real data. To use the model for quantitative rather than qualitative predictions it is necessary to set the parameters of the model appropriately, and then to show that the model thus adjusted provides a good predictor of data collected independently in comparable conditions. The symbols indicate measured GRF dose-response curves for GH release from isolated somatotrophs, using data taken from ref. 14. GH released (as percentage of control) is plotted against log10 (hpGRF-44 dose) when given alone (symbol X), with 1 nM SS-14 (O), and with 1 nM SS-28 (+). The curves show fits of the model to these data; best fits have been found using nonlinear least squares (see Eq. 13).

log(grf dose)

Fig. 7. Scaling the model parameters to real data. To use the model for quantitative rather than qualitative predictions it is necessary to set the parameters of the model appropriately, and then to show that the model thus adjusted provides a good predictor of data collected independently in comparable conditions. The symbols indicate measured GRF dose-response curves for GH release from isolated somatotrophs, using data taken from ref. 14. GH released (as percentage of control) is plotted against log10 (hpGRF-44 dose) when given alone (symbol X), with 1 nM SS-14 (O), and with 1 nM SS-28 (+). The curves show fits of the model to these data; best fits have been found using nonlinear least squares (see Eq. 13).

lower limits to kb close to zero. The estimates were as follows: zero SRIF, kb = 10 3, kr = 1.53 x 103; 1 nM SRIF-14, kb = 0.25, kr = 4.9; 1 nM SRIF-28, kb = 0.45, kr = 2.4. The common estimates of kj and r0 for the three sets were kj = 0.013, r0 = 13.6. Again the model predictions fitted well and the parameters fitted separately for each set show the orderings expected a priori.

Was this article helpful?

0 0

Post a comment