Recommendation 1
1a. Define the bias type: We recommend using the more specific terms ‘pathway-biased ligand/signaling’ and ‘physiology-biased ligand/signaling’ rather than just ‘biased ligand/signaling’. This is because, although the person who generated the data may have only thought of one preferred meaning of bias, a substantial portion of the community will inevitably understand the meaning of bias differently.
1b. Include multiple and explain chose of reference ligands: To allow interpretation across bias types and studies, we recommend that experiments include multiple ligands that could serve as reference ligands e.g., a commonly used tool compound and clinical drug, the principal endogenous agonists and/or a pathway-balanced ligand, where available. Furthermore, we recommend to always explain the choice of reference ligands, e.g.: The reference ligand A was chosen, as it A) is the principal physiological agonist, B) has a relatively balanced signaling through pathways P1 and P2 (provide slope coefficient), or C) a clinical or tool compound for which enhanced pathway specificity is desired through the introduction of bias.” .
Reason: Bias cannot be reported without the use of a reference ligand, as a presumed bias may otherwise in fact be due to other differences in e.g., the baseline, window and sensitivity of assays (‘system bias’ and ‘observational bias’ in Figure 1). System bias affects all ligands equally (including the reference ligand), unless the ligand is labile (e.g. degraded by cellular components (peptidases, lipases, etc.) or is less stable in solution (oxidation, reduction, etc.). Therefore, a pairwise relative comparison of a reference and tested ligand will normally not be influenced by system bias.
Disclaimer: The use of an arbitrary reference ligand does not cancel the effect of system bias on the pairwise relative comparison ofpathways in pathway-biased ligand/signaling studies (it does so only for ligands). I.e., a reference ligand cannot support the claim that a ligand has an isolated/preferred signaling via one pathway unless the reference ligand is selected based on balanced signaling.
Case study: Depending upon the reference agonist used, a given tested agonist could be biased towards a given pathway or not. The dependence of bias on the choice of reference agonist is illustrated in the relative activation of G protein and β-arrestin pathways by the κ-opioid receptor with the agonist naphthoyl-β-naltrexamine (β-NNTA) (White et al., 2014). Thus, when salvinorin A is chosen as the reference agonist, β-NNTA is 8.9-fold biased toward G protein signaling, only 2-fold biased toward G protein if salvinorin B is the reference, and 4.3-fold biased toward β-arrestin recruitment (0.23 bias for G protein) if dynorphin-A is the reference. Such evidence clearly demonstrates that it is important to report the identity of the reference compound in bias studies and to restrict the conclusions to the comparison made.