Why all MODY variants in transcription factor genes are dominantly inherited

Roman Zug^*

• Johannes Gutenberg University Mainz, Mainz, Germany

Maturity-onset diabetes of the young (MODY) is a rare inherited form of diabetes caused by mutations in a single gene and characterized by an early onset, typically before the age of 25 years. MODY represents the most common form of monogenic diabetes and is due to impaired development and function of pancreatic β cells, resulting in deficient secretion of insulin. MODY is inherited predominantly in an autosomal dominant mode, which is remarkable because other forms of monogenic diabetes do not show this inheritance pattern (Bonnefond et al., 2023). Why are most MODY variants dominantly inherited? In an attempt to address this issue, Li et al. (2022) asked: "Could it be because of some biological property of the insulin-secreting pancreatic β cells that makes them susceptible to the deleterious effects of heterozygous but not homozygous variants?" Interestingly, however, rather than giving an answer to this question, Li et al. proposed that autosomal recessive forms of MODY "are at least as common as the dominant ones, but have not been discovered yet". In their view, the preponderance of autosomal dominant MODY does not have a biological reason, but rather reflects our inability to detect recessive variants. While this is theoretically possible, I here argue that there is indeed a plausible biological reason why MODY is mostly dominant. In the following sections I will outline the reasoning step by step, together with empirical evidence. Finally, I discuss why the hypothesis also helps to explain incomplete penetrance of MODY variants, and I disprove the idea that homozygous variants are less deleterious. Although the protein products of MODY genes serve a variety of molecular functions, by far the largest functional group comprises transcription factors (TFs) (Figure 1). These TFs act as master regulators of pancreatic development, and of β cell differentiation and function in particular (Arda et al., 2013;Conrad et al., 2014;Dassaye et al., 2016;Wortham & Sander, 2021). Therefore, MODY cases that are due to impaired pancreas and β cell development, caused by variants in master TF genes, should be considered developmental disorders (Zug, 2022).Out of 19 autosomal genes in which MODY-causing variants are known, 10 code for TFs (Figure 1) (Bonnefond et al., 2023). These TF genes are given in Table 1. Note that I do not include two other TF genes, KLF11 and PAX4 (nor the genes APPL1, BLK, and WFS1), because there is insufficient evidence that variants in these genes actually cause MODY (Laver et al., 2022;Sriram et al., 2025). Of the 10 known MODY-associated TF genes, the three most common ones alone are estimated to account for more than two thirds of all MODY cases: HNF1A (52%), HNF4A (10%), and HNF1B (6%) (Shields et al., 2010). All known MODY-causing variants in TF genes show dominant inheritance. Why is that? In order to answer this question, let us look at these variants in more detail. All MODYassociated variants in TF genes cause loss-of-function (LOF). This is not surprising, as most mutations cause LOF. What is surprising, though, is that LOF of a single allele is sufficient to cause a clinical phenotype. In other words, a 50% reduction in gene expression is not tolerated. This pronounced dosage sensitivity is called haploinsufficiency.Strikingly, all MODY-associated TF genes are haploinsufficient and hence intolerant to heterozygous LOF variants (Table 1). Haploinsufficiency is a manifestation of genetic dominance, as a phenotype is already visible in the heterozygous state (Zschocke et al., 2023). Haploinsufficiency represents a particularly strict form of gene essentiality, which can be defined as a considerable reduction in organismal fitness associated with a gene's LOF (Bartha et al., 2018). Accordingly, in haploinsufficient genes, there is strong negative selection even against heterozygous LOF variants, reducing the frequency of such variants in the population ('selective constraint') (Zeng et al., 2024). Haploinsufficiency is a hallmark of master regulator TF genes and can lead to a plethora of developmental disorders, MODY being one of them (Seidman & Seidman, 2002;Zug, 2022). Most often, MODY is caused by LOF variants in the coding regions of TF genes, but it can also be due to LOF variants in the cis-regulatory elements (CREs) of these genes, as has been shown, for example, for HNF1A (Gragnoli et al., 1997) and HNF4A (Hansen et al., 2002). Why are heterozygous LOF variants in TF genes (or their CREs) sufficient to cause MODY? Building upon earlier work (Wilkie, 1994;Veitia, 2002;Johnson et al., 2019), I have recently proposed the hypothesis that developmental disorders caused by TF haploinsufficiency result from defects in cell fate determination, which can be traced to disrupted bistability in the underlying gene regulatory network (Zug, 2022). Bistability is a crucial feature of dynamical systems that are able to make binary choices such as cell fate decisions. Bistability means that a system can be resting in two alternative stable states but not in intermediate states, resulting in switch-like threshold effects. A system is able to generate bistability if its components engage both in positive feedback and ultrasensitivity (Ferrell, 2002). Positive feedback prevents the system from resting in intermediate states.Ultrasensitivity means that an increase in the input signal first has little effect, but then produces higher and higher levels of output, as represented by a steep sigmoidal curve.Ultrasensitivity filters small stimuli out of the feedback loop, allowing the system to have a stable off-state (Ferrell, 2002), and often comes in the form of cooperativity (Zhang et al., 2013). Both positive feedback and cooperativity are pervasive features of cell fate determination. Cell fate is mainly controlled by the assembly of master regulators at celltype-specific enhancers, often called super-enhancers (Figure 2A). Genes associated with super-enhancers include those encoding the master regulators themselves, thus establishing autoregulatory positive feedback loops. Super-enhancers assemble a high density of master regulators, allowing for extensive cooperative TF-DNA binding (e.g., via dimerization). The requirement of both positive feedback and cooperativity for proper cell fate determination helps to explain the distinct dosage sensitivity (that is, haploinsufficiency) of master regulators: it is the high level of cooperativity (an instantiation of ultrasensitivity) that makes the positive feedback loops particularly sensitive to changes in TF concentration. Therefore, heterozygous LOF variants that disrupt positive feedback or cooperativity are sufficient to interfere with proper cell fate determination and eventually lead to developmental disorders such as MODY (Figure 2B) (Zug, 2022). Although bistability has been invoked before to explain MODY etiology (in the context of the HNF1A-HNF4A positive feedback loop : Ferrer, 2002;Kaci et al., 2024), these studies ignore cooperativity, failing to account for both requirements of bistability. Here I collect evidence in support of the hypothesis that heterozygous LOF in MODYassociated TF genes disrupts positive feedback or cooperativity and thus causes the disease. As shown in Table 1, all known MODY-associated TF genes (1) are haploinsufficient, (2) are master regulators of pancreatic development, and of β cell fate and function in particular, (3) engage in positive feedback and cooperativity, and (4) exhibit strong or extreme selection against heterozygous LOF, as estimated by a powerful Empirical Bayes approach (Zeng et al., 2024). Moreover, for HNF1A and HNF4A, disruption of positive feedback (Hansen et al., 2002) and of cooperativity (Hua et al., 2000;Singh et al., 2019) has been identified as the cause of MODY. Taken together, this evidence strongly supports the hypothesis that heterozygous LOF variants in master regulators of β cell fate are sufficient to disrupt TF positive feedback or cooperativity and thus cause MODY. This explains why MODY-associated variants in TF genes are dominantly inherited. Shaw-Smith et al., 2014Carrasco et al., 2012;Xuan et al., 2012Arda et al., 2013Charron et al., 1999;Xin et al., 2006 0.115 (extreme selection) Bonnefond et al., 2012Carrasco et al., 2012;Xuan et al., 2012Meng et al., 2018Charron et al., 1999;Xin et al., 2006;Chia et al., 2019 0.318 (extreme selection) Yamagata et al., 1996bQian et al., 2023;Ng et al., 2024Ferrer, 2002;Hansen et al., 2002Cujba et al., 2022 Yamagata et al., 1996aNg et al., 2019, 2024Ferrer, 2002;Hansen et al., 2002Jiang & Sladek, 1997;Lu et al., 2008 0.051 (strong selection) Iacovazzo et al., 2018Olbrot et al., 2002;Nishimura et al., 2015Raum et al., 2006Zhao et al., 2005 10 -3 to 10 -1 ), heterozygous LOF has a fitness effect on par with the strongest selection measured for common variants. Under extreme selection (shet > 10 -1 ), fitness effects are equivalent to a >10% chance of embryonic lethality (Zeng et al., 2024). The hypothesis outlined above also helps to better understand why not every individual carrying a pathogenic variant actually develops the disease, a phenomenon termed incomplete penetrance (Kingdom & Wright, 2022). Many MODY-associated TF genes show incomplete penetrance, e.g. HNF1A, HNF1B, HNF4A, NEUROD1, PDX1, and RFX6 (Mirshahi et al., 2022;Li et al., 2023;Sharp et al., 2025;Sriram et al., 2025). Even though incomplete penetrance can be caused by a range of factors, it has a strong genetic basis (Kingdom & Wright, 2022). Goldschmidt (1938) explained incomplete penetrance by assuming stochastic fluctuation in gene expression combined with some threshold effect.The stochastic nature of gene expression is now well established (Raj & van Oudenaarden, 2008). Goldschmidt's postulated threshold effect can be readily explained as well, at least with respect to master regulator TFs: as outlined above, these TFs engage in positive feedback and cooperativity, which allows for bistability and, thus, threshold effects. Heterozygous LOF variants bring gene expression levels close to the threshold, but only those variants for which gene expression happens to lie below the threshold will elicit a disease phenotype (Figure 3). This idea explains incomplete penetrance of variants in TF genes, including those associated with MODY (Zug, 2022). Lastly, I would like to address the assumption made by Li et al. (2022) that β cells are susceptible to heterozygous but not homozygous variants. I argue that this assumption is wrong, at least with respect to TF genes. The reason is that, in TF genes, variants in the homozygous state have generally more severe consequences than in the heterozygous state, a phenomenon termed semi-dominance (Zschocke et al., 2023). In many MODYassociated TF genes, homozygous LOF variants are generally thought to be embryonically lethal or result in early mortality, e.g. in GATA4 (Kuo et al., 1997), GATA6 (Morrisey et al., 1998), HNF1A (Harries et al., 2009), HNF1B (Barbacci et al., 1999), HNF4A (Chen et al., 1994), and ONECUT1 (Philippi et al., 2021). In other MODY-associated TF genes, homozygous variants exist but cause severe syndromic and usually neonatal diabetes, such as in MAFA (Iacovazzo et al., 2018), NEUROD1 (Rubio-Cabezas et al., 2010), PDX1 (Stoffers et al., 1997b) and RFX6 (Smith et al., 2010). Therefore, the question is not why MODY-associated TF genes are not susceptible to homozygous variants (they are very much so), but rather why homozygous variants are tolerated at all (at least in some genes), given the high intolerance of these genes even to heterozygous variants.