Terminology summary
Ligand bias: Ligand-dependent preferential activation, by some agonists, of some transducer pathways over others originating from a single receptor subtype relative to a reference ligand (and tested in a given cellular system).
System bias: Bias due to differences in the cellular system, including so called “tissue bias”. System bias can be due to differential expression of intracellular transducers or effectors, or of proteins acting on the receptor as an allosteric modulator (e.g. RAMPs; kinases; PDZ proteins).
Functional selectivity: Functional selectivity is the observed response combining ligand- and system-bias (see also (Stallaert, Christopoulos & Bouvier, 2011)). It covers all instances where a ligand produces disproportional responses, relative to a reference ligand, in different pathways connected to the same receptor.
Observational bias: An artificial bias introduced by the experimental setup. Its final effect on the observed response is quantitatively equivalent to, and thus indistinguishable from, biologically induced system bias. Therefore, ligand bias assessment strategies that are meant to eliminate system effects from observed responses handle observational bias as a “system bias”. However, this effect can also be experimentally minimized by measuring signals at similar levels of the compared pathways by using equally sensitive assays, and by ensuring that at least one assay for each pathway has sufficient sensitivity.
Choice of reference ligand determines the meaning of bias, which can be a benchmark or pathway- or physiology-biased ligand/signaling
The GPCR field uses several fundamentally different meanings of biased ligand/signaling and these depend on the choice of reference ligand (Table 1). As these have different meanings, each of them serves a distinct purpose, and substantial portions of the research community will continue to use their favorite type of bias to fulfill the needs of their given studies. Therefore, these recommendations do not prohibit the use of any of these different types of biases, but strive to explain their principles, advantages, and limitations to aid any researcher in designing and interpreting different types of biased signaling studies. A fundamental recommendation of this paper, however, would be that authors are explicit in their descriptions of which version of bias they describe and which reference ligand they used.
Many biased ligand/signaling studies use a particular reference ligand because it is a much-used tool compound (e.g., isoprotenerol/isoprenaline for the β2-adrenoceptor) or a relevant clinical agent (e.g., an existing drug to be replaced in a new drug discovery project). In this case, the (only) meaning of the observed biased ligand/signaling is that it differs from that of the reference ligands, which in its turn can have a pathway- or physiology-bias. In contrast, ‘pathway-biased ligand/signaling’ utilizes the reference ligand with the most balanced signaling across pathways, allowing the research objective to identify ligands for or functionally dissect a specific pathway. Finally, ‘physiology-biased ligand/signaling’ utilizes the principal physiological agonist as the reference ligand and consequently can conclude whether the tested ligands display “natural” or “unnatural” signaling. The three types of biases are described in more detail in the following sections.