When everything is a master regulator, nothing is

Abstract

The term “master regulator” emerged in the 1960s and 1970s and referred to autoregulatory transcription factors that sat atop a developmental lineage. Since that time, usage of the term has increased and broadened to the point where it has lost clear meaning. Here we discuss the term “master regulator” with the goals of developing a consensus view of its definition and stimulating discussion on use of similar terms. We propose that the designation “master regulator” be reserved for transcription factors that are: 1) positioned at the top of a regulatory hierarchy specifying a cell lineage (and potentially specific cell states, such as hypoxia); and 2) sufficient to drive the transcriptional program characterizing that lineage or state. It is hoped that this piece will provide a precedent for use of additional terms applied to incompletely understood biological processes, resulting in experimentation that sheds light on such processes.

“You keep using that word. I do not think it means what you think it means.” – Inigo Montoya

Use of clear and precise language is beneficial when reporting and discussing science. Many terms that describe specific biological processes are understood to refer to the same thing by virtually all who use them (e.g., “cell cycle”). Sometimes terms are coined to describe complex biological functions that are not fully understood at the time of the term's origin. As understanding of the biology improves, the application of the term may evolve or expand, and clarity in the use of such terms can be lost, necessitating reevaluation of how the term ought to be used. Here we discuss the term “master regulator,” with the goal of providing such clarity and the hope that this piece will stimulate discussion on its use.

Studies performed in the 1960s and 1970s by Howard Holtzer's lab on cells of the myogenic, erythrogenic, and chondrogenic lineages led to the suggestion that complex programs of lineage-specific gene expression could be triggered by a very small number of regulatory factors (Holtzer et al., 1972; Tapscott, 2005). Careful experimentation demonstrated that the nucleoside analogue 5-bromodeoxyuridine (BUdR) could reversibly inhibit cell differentiation in a manner consistent with this compound directly targeting a very limited number of genes, perhaps even a single gene (Holtzer et al., 1972; Weintraub et al., 1973; Tapscott, 2005). Such genes were referred to as a “master switch” for the lineage. In a 1979 monograph, Susumu Ohno postulated the existence of a “master regulatory gene” to explain how a genetic sex-determination mechanism might work (Ohno, 1979). Ohno hypothesized about the need for a gene that “occupies the very top of a regulatory hierarchy” and that “should not be under the regulatory influence of any other gene.” Ohno further proposed that the master regulatory gene and its encoded protein engage in an autoregulatory, positive feedback loop initiated by maternal stores of the protein, consequently allowing it to sit atop the hierarchy and be both “the regulator and the regulated.” “Master regulatory gene” or “master control gene” terminology was subsequently applied to the MAT mating type loci in the budding yeast, Saccharomyces cerevisiae (Herskowitz, 1989) and to homeotic genes in the fruit fly, Drosophila melanogaster (e.g., Antp) (Rosales-Vega et al., 2024). These genes encode transcriptional regulators necessary for determining cell type and segmentation pattern/appendage identity, respectively. The term “master regulator” arguably received its greatest prominence with the 1987 identification of MyoD as a transcription factor capable of reprogramming multiple types of nonmuscle cells into myoblasts (Davis et al., 1987; Weintraub et al., 1989; Tapscott, 2005). It is worth noting that, consistent with Holtzer's original conception, BUdR blocks myogenic differentiation by inhibiting expression of MyoD (Tapscott et al., 1989).

Each of the factors described above shares the property of directly regulating a downstream transcriptional program that is sufficient for acquisition of cell identity. Furthermore, proposal of the concept of master regulatory factors promoted experimental pursuit of such factors, leading to ground-breaking discoveries. It is therefore reasonable to consider them as setting the benchmark for a definition of what is required for a factor to be classified as a master regulator. However, since the initial descriptions of these factors as such, use of the term master regulator has increased and broadened to the point where it has lost clear definition (Chan and Kyba, 2013). The term master regulator is now often used beyond transcriptional networks to refer to factors that are rate-limiting for a complex biological process or to factors that function broadly across multiple processes. As one example, Smoothened (Smo), a critical cell surface signal transducer in the Hedgehog pathway, has been designated as a “master regulator of adult liver repair” (Michelotti et al., 2013). In this study, Smo was conditionally genetically removed from hepatic stellate cell–derived myofibroblasts in adult mice, resulting in diminished production of liver progenitors and a consequent combination of beneficial and deleterious effects on liver regeneration after injury (Michelotti et al., 2013). However, Hedgehog signaling functions in an enormous number of biological processes during development and homeostasis (McMahon et al., 2003; Dessaud et al., 2008), and the downstream transcriptional programs it regulates are distinctive to the cell types receiving the signal. Some of the Hedgehog pathway's functions involve promoting cell identity (e.g., ventral neural tube neurons; Dessaud et al., 2008), but they are not transportable to other cell types in the way that forced expression of MyoD can reprogram multiple nonmuscle cell types to the muscle lineage. For example, forced activation of Smo in epidermal cells does not reprogram them to liver progenitors or hepatocytes (Youssef et al., 2012). It takes nothing away from the importance of Smo function during liver regeneration, or the excellence of the Michelotti et al. study, to say that Smo's activity does not meet the definition of master regulator. Similarly, the protein kinase MARK2 has been referred to as a “master regulator of the microtubule and actomyosin cytoskeletons and focal adhesions to mediate directional migration of cancer cells” (Pasapera et al., 2022). MARK2, a member of the AMPK-related kinase family performs multiple functions in, and is undoubtedly important to, this complex process, but many other factors are also involved in a broad and rate-limiting manner (Seetharaman and Etienne-Manneville, 2020). Furthermore, MARK2 and other members of its protein kinase family are activated via phosphorylation by another kinase, LKB1, which is itself referred to as a “master upstream kinase” (Lizcano et al., 2004). Again, we do not question the significance of the science in these papers, but rather point out that usage of the term “master regulator” has strayed from its origins.

In 2013, Chan and Kyba proposed that a master regulator can be defined as a factor that: 1) is expressed at the inception of a developmental lineage or cell type; 2) participates in the specification of that lineage by regulating multiple downstream genes either directly or by promoting expression of other transcription factors that in turn collaborate with them to drive lineage-specific gene expression; and 3) when misexpressed, has the ability to respecify the fate of cells ordinarily destined to form other lineages (Chan and Kyba, 2013). These properties describe a state of sufficiency for master regulator function and emphasize the centrality of reprogramming activity in this function. One might further argue that a master regulator should naturally also be required for determination of a specific developmental lineage, but the state of necessity applies to the gene and its homologues as a set, in the case where a master regulator has a redundant homologue. For example, despite its exceptional reprogramming ability, MyoD is not essential for determination of the skeletal muscle lineage (Rudnicki et al., 1992) because its function is shared with a structurally related transcription factor, Myf5 (Rudnicki et al., 1993). Mice with combined genetic removal of MyoD and Myf5 lack skeletal muscle (Rudnicki et al., 1993), revealing that each can compensate for the other and that they are together necessary for lineage determination (a third related transcription factor, MRF4, has modest determination function early in embryogenesis [Kassar-Duchossoy et al., 2004] but MyoD;Myf5 double mutant mice are born without detectable skeletal muscle). Additionally, MyoD is dependent on other transcription factors. It interacts physically and/or genetically with other transcription factor families that are rate-limiting for MyoD function in development and reprogramming and that therefore play critical roles in myogenesis. Among such factors are Mef2 and Six proteins (Molkentin et al., 1995; Santolini et al., 2016). However, these factors are expressed in many lineages and do not possess reprogramming ability on their own. Therefore, despite their importance in promoting MyoD activity, they do not qualify as master regulators themselves.

MyoD is unusual in that relatively few transcription factors are able, singly, to reprogram other cell types (Graf, 2011). An additional example is the transcription factor PPARγ, which is able to reprogram fibroblasts to the adipocyte lineage and is required in mice for formation of adipose tissue (Tontonez and Spiegelman, 2008). More commonly, members of more than one distinct transcription factor family cooperate to provide specific reprogramming abilities. Such factors can be viewed, together, as a master regulatory set. The best known example is that of Oct4, Sox2, and Nanog in maintenance of pluripotency in embryos in vivo and in their ability to reprogram somatic cells to induced pluripotent cells (Hochedlinger and Jaenisch, 2015). MyoD and Oct4/Sox2/Nanog (acting together) hold additional properties in common that help illuminate master regulator function. For example, MyoD activates and maintains its own expression, providing an autoregulatory loop similar to that proposed by Ohno (Thayer et al., 1989). Analogously, Oct4, Sox2, and Nanog act collaboratively to drive expression of the Oct4, Sox2, and Nanog genes (Jaenisch and Young, 2008). Therefore, both MyoD and Oct4/Sox2/Nanog display positive autoregulation, which presumably provides stability to gene expression levels and, thereby, stability to the myoblast or pluripotent stem cell fate, respectively. Furthermore, MyoD and Oct4/Sox2/Nanog directly regulate expression of lineage-specific genes, as well as of other transcription factors and signaling molecules that further participate in regulation of lineage-specific gene expression (i.e., MyoD and Oct4/Sox2/Nanog both generate feed-forward systems) (Tapscott, 2005; Balsalobre and Drouin, 2022).

It is clear that only transcription factors are able to satisfy the criteria required to meet the definition of the term master regulator as it was originally intended for use. Furthermore, the term most easily fits transcription factors that drive cell lineage identity. There are, however, transcription factors that dominantly drive broad gene expression programs that lie at the heart of other biological processes, and it has become the custom to refer to these as master regulators also. A good example here are the hypoxia-inducible factors (HIFs), which are activated by a state of hypoxia and viewed as master regulators of oxygen homeostasis (Wicks and Semenza, 2022). The HIF-1α, -2α, and -3α proteins are oxygen-sensitive factors that directly drive gene expression via hypoxia response elements in target genes as well as alter gene expression via activation of various transcriptional repressors, microRNAs, and additional regulatory factors (Wicks and Semenza, 2022). While not sitting atop a transcriptional program dictating cell identity, the HIFs control, and are required for, a complete transcriptional program that grades responses to oxygen tension. The mechanistic parallels make an argument for the use of the term master regulator when speaking of transcription factors like the HIF proteins. However, even within the domain of transcription factors, the term seems to be prone to creeping overuse. Besides the ability to specify a function, be that conferral of developmental identity or a coordinated response to a certain signal, are there additional guardrails that would help from depleting the term of meaning, as would happen if every transcriptional regulator was referred to as the master regulator of its particular subset of targets? Two additional requirements that seem reasonable to impose are the following: First, the master regulator of a process should exhibit the principle of sufficiency in the way that the master regulators of developmental lineages display sufficiency by way of lineage reprogramming when ectopically expressed. Ectopic expression of the master regulator, or of the activated form of the master regulator, should be sufficient alone to activate the process. Processes that occur only in response to several independent inputs mediated by independent transcription factors would then not be said to have a master regulator. Second, the master regulator should include under its aegis other regulators, such that it is not simply a transcription factor that regulates a set of target genes, but a transcription factor that additionally regulates other transcription factors and through them, their targets. Otherwise, why distinguish the regulator of a process with the title “master regulator” as opposed to simply “regulator”?

We have argued here that the literature would benefit by limiting use of the term master regulator to factors that hold the properties of those factors to which the term was originally applied. One may ask, what benefits would such a restriction actually bring? As stated at the beginning of this piece, we believe that clarity and precision in use of defined scientific terms is always helpful. More specifically, development of the concept of master regulatory factors led to creative experimentation that ultimately resulted in identification of such important factors. We propose, therefore, that the master regulator designation be reserved for transcription factors that sit atop a cell lineage (and perhaps a cell state, such as hypoxia) and are sufficient to drive the transcriptional program characterizing that lineage or state. It is our hope that this piece will stimulate discussion as to how and when to give transcription factors that occupy the apex of a complex, environmentally sensitive process, as exemplified by HIFs, the designation of master regulator. Similarly, we hope that gaining a consensus view on a clear definition of the term master regulator will serve as a precedent for use of new phrases that may be applied to incompletely understood biological phenomena, resulting in experimentation that sheds light on them.