• 2018-07
  • 2018-10
  • 2018-11
  • 2019-04
  • 2019-05
  • 2019-06
  • 2019-07
  • 2019-08
  • 2019-09
  • br Most reported binding data are obtained by performing het


    Most reported binding data are obtained by performing heterologous competition binding experiments and expressing the results as relative binding affinity (RBA). The constructed binding curves should theoretically be sigmoidal in shape with a Hill slope value of one, for a single ligand binding to one site on a receptor molecule, without cooperative binding. The Hill slope (steep part of the curve) indicates whether cooperative binding occurs, with a slope of one indicating no cooperative binding, while of slope of less or greater than one indicates negative or positive cooperativity, respectively. The IC50, the concentration of unlabeled progestogen (inhibitor or competitor) that corresponds to 50% inhibition of the total specific binding of the radiolabeled reference agonist, can then be determined. Many apparent discrepancies in RBAs reported in the literature are due to the use of different reference ligands (set as 100% RBA). These include promegestone (R5020) versus Prog for the PR, which differ in their RBAs by about five fold, and mibolerone or methyltrienolone (R118) versus testosterone or dihydrotestosterone (DHT) (Supplementary Table 1), where the synthetic agonist have about 100-fold greater RBA than the natural ligands [2], [3]. Thus, these apparent RBA differences are not necessarily real differences. RBAs can be directly compared by recalculating values relative to a common ligand, if the information is available. Another major source of variability in binding data is biological sample variability. Different samples may exhibit different degrees of off-target binding of the progestogen to non-target SRs and/or other proteins such as steroidogenic enzymes, to which progestogens bind with variable affinity. Off-target binding would effectively lower the apparent RBA, right-shift the binding curve and increase the IC50, independent of the method used to determine binding, as illustrated in the simulated binding curve (Fig. 1B). Off-target binding could also be a source of non-parallel binding curves with Hill slopes greater than one with increasing concentrations of off-target receptor relative to target receptor, as illustrated in Fig. 1B. For example, remarkable differences in the RBAs of Prog (1000% vs. 100%), DRSP (500% vs. 100% or 230%) and gestodene (97% vs. 290%), all relative to trovafloxacin clinical set to 100% (Table 1 and Supplementary Table 1), were observed when comparing recombinant human MR in vitro binding, relative to rat tissue models, respectively [25], [26], [27], most likely due to off-target and/or metabolism effects in the tissues.The presence of metabolizing enzymes could also potentially right-shift binding curves and lower the IC50 values without a change in Hill slope (assuming the metabolites do not bind the target receptor), as illustrated in the simulated binding curves (Fig. 1C). Other sources of variability are the species from which tissue is obtained, as well as the variations in age and pre-treatment of the animal or human donors. For example, when comparing rabbit versus rat uterine cytosols, different RBAs were observed for the PR with MPA (112% vs. 9.4%) (R5020=100%), NOMAC (135% vs. 72%) (Prog=100%) and trimegestone (TMG) (84% vs. 152%) (R5020=100%), respectively, (Table 1 and Supplementary Table 1) [15], [18]. Some of the above differences are likely to be due to species differences. However, species, age, or pre-treatment differences may also be confounded by differences in relative concentrations of off-target receptors and/or metabolizing enzymes. As mentioned, a failure to reach equilibrium by the reference and competing ligand is another common cause of variability in reported binding data [20]. The few cases where this is investigated suggest that very different times are required for different progestogens and different concentrations thereof, to reach equilibrium. For example, for 0.2nM [3H]-aldosterone or [3H]-mibolerone binding to the overexpressed hMR or hAR in the COS-1 cell line, this was determined as 16h [23], while for 1.25nM [3H]-dexamethasone binding to the GR endogenously expressed in the A549 cells it was three hours [22]. Others showed that 20nM Prog reaches a saturation plateau faster than either 20nM R5020 or RU486 [16]. Higher RBAs have been reported for binding of several progestogens to the PR in rabbit uterine tissue after incubation for 24h, in comparison to 2h [18]. Similarly, differences in the RBAs of several progestogens for the cytosolic AR of the rat prostate have been reported for different incubation times [18]. When plotting binding curves, most software programs give the option to fix or not to fix the Hill slopes for binding curves. Fixing the slope could mask the presence of competing off-target effects or of high receptor concentrations (as illustrated in Fig. 1A and B) or even the presence of cooperative target SR binding effects, leading to inaccurate determination of affinities. There are indeed some reports in the literature of cooperative binding of steroids to the GR and estrogen receptor (ER) [28], [29], suggesting this may be one reason for non-parallel binding curves. Other sources of error in saturation binding data could be the use of Scatchard plots rather than non-linear regression in determination of binding parameters. This method has been used extensively to determine K values for progestogens binding to the PR [14], [16], [21], [27] (Supplementary Table 1 [63], [64]). The Scatchard plot, which entails a linearization of binding data, suffers from the same shortcoming as linear enzyme-kinetic treatments such as the Lineweaver–Burk plot: an uneven error distribution results in data points at low ligand concentrations being given undue weight.