Hox genes: The original body builders

As biologists, we are constantly fascinated by the diversity and complexity of living organisms. To understand the origins of diversity in the animal kingdom we must understand animal development, and perhaps the group of genes that has most consistently captured the attention of developmental biologists since they were discovered are the Hox genes. These genes, which encode highly conserved homeobox-containing transcription factors, are present in a wide range of organisms, from fruit flies to humans. First discovered in the fruit fly Drosophila melanogaster, they were originally appreciated as playing a crucial role in determining the body plan of an organism [1]. And yet, research on Hox genes has gone far beyond animal development by informing at least two additional areas of biology. For one, they are critical drivers of animal evolution: changes to how and when they are deployed, and in the networks of downstream genes they regulate, have facilitated changes to animal body plans [2,3]. Hox gene research has also shed light on how families of related transcription factors with very similar DNA binding specificities can carry out distinct functions in vivo [4]. Arguably, no other set of genes have had such an important impact on such disparate and important areas of biology.

In this Special Issue, we present a wide range of articles that reflect all three of the fields on which Hox research has had a profound impact: animal development, animal evolution, and transcription factor mechanisms. For evolutionary insights, we have three fascinating articles. The first, by Mulhair and Holland [5, this issue], builds on the intriguing observation that most Hox genes are clustered in animal genomes and that their expression along the main body axis correlates with their position within these clusters. Mulhair and Holland’s contribution is a tour-de-force effort that uses publicly available genome sequences for no less than 243 insects, representing 13 orders, to analyze trends in the cluster-level organization of these genes. Large order-specific differences in Hox cluster size, organization, and the duplication, loss, and emergence of new homeobox genes (e.g. the explosion of zen orthologs in Lepidoptera) suggests that Hox genes have many species-specific functions and modes of regulation that are yet to be discovered. The article by Wanninger [6, this issue] addresses the evolutionary origins of Hox genes and the relationship of Hox gene number to animal complexity. Wanninger first provides an analysis of when Hox genes emerged and lost during evolution by depicting several different scenarios that can account for the currently available sequence data. One conclusion is that rather than relying only on gene expression to determine the evolution of morphological characters, it is better to include datasets of comparative morphology and gene-gene interactions. Third, Turetzek et al. [7, this issue] take a deep dive into the organization and expression of Hox genes in spiders. Spiders, with their distinct body plans relative to better studied arthropods such as fruit flies, provide the opportunity to ask if body plan modifications do indeed correlate with changes in Hox expression and gene number. The answer, based largely on the observation that spiders have two Hox clusters and multiple divergent Hox expression patterns, is almost certainly yes. Despite these differences, some Hox-dependent functions, for example, suppression of legs in the abdomen, are likely to be conserved between spiders and flies.

Hox genes were originally called homeotic genes when first described in Drosophila due to their dramatic ability to transform an entire appendage or segment from one identity to another [8]. These transformations of body parts were reminiscent of natural variations in animal body morphologies first described by William Bateson, who initially coined the term ‘homeosis’ [9]. For example, the Drosophila Hox gene Ultrabithorax (Ubx) normally dictates the fate of the third thoracic segment and, when its function is removed genetically, that segment is transformed to a second copy of the second thoracic segment, resulting in flies with two pairs of wings instead of one (as shown in the cover image of this issue) [10]. Remarkably, analogous transformations of segment identity have been observed in many other animals where loss-of-function genetics is feasible [11–13], suggesting that Hox segment identity functions are ancient and highly conserved. However, what was not initially obvious from these dramatic transformations is that Hox genes also play a pivotal role in specifying the identities of non-ectodermal tissues, including the nervous system and mesoderm. Three articles in this issue explore these Hox-dependent functions. In the article by Pinto et al. [14, this issue], an approach based largely on genome-wide studies is used to compare the targets, cofactors, and specificity mechanisms (see more below) used by Ubx in the Drosophila mesoderm and ectoderm. Two articles examine the role of Hox genes in the nervous system of two very different model systems: vertebrates and the worm, C. elegans. In the article by Smith and Kratsios [15, this issue], multiple Hox-dependent neuronal fate examples are described in the worm that lead to novel insights. For example, in addition to their role in neuron specification and development, Hox function is also required in terminally differentiated neurons to maintain their fates. At the opposite end of the complexity spectrum, Miller and Dasen [16, this issue] summarize the current state of our understanding for how Hox genes are themselves regulated in vertebrates, initially by broad gradients of retinoic acid and fibroblast growth factors, and later maintained by the Polycomb Group (PcG) of regulators. Notably, the PcG-mediated maintenance of Hox expression patterns in vertebrates fits well with the terminal selector functions highlighted by Smith and Kratsios in worms.

Hox proteins have also been illustrative in explaining how individual members of transcription factor families can execute highly distinct functions in vivo while having very similar DNA binding specificities in vitro, a phenomenon that has been called the transcription factor specificity paradox. One solution to this paradox is that Hox proteins bind with cofactors that increase differences in DNA binding preferences [4]. The article by Pinto et al. [14, this issue], suggest that cofactor-based mechanisms to achieve specificity may differ between tissue types, such as mesoderm and ectoderm, and argue that there are likely many more Hox cofactors than we currently know about. The article by Merabet and Carnesecchi [17, this issue] highlight another, not mutually exclusive, mechanism that is also likely address the Hox specificity paradox, namely, that differences in Hox expression levels can impact their function. Although it has been long appreciated that Hox function is impacted by gene dose (a proxy for expression level), this article makes the point that this is a much more widespread phenomenon than was generally appreciated and could contribute to the evolution of many Hox-regulated morphologies. In addition to providing an updated review of the cofactor model for Hox specificity, the article by Bobola and Sagerstrӧm [18, this issue] underscores the important point that the classic Hox cofactors – the TALE (three amino acid loop extension) homeodomain proteins such as Meis and Pbx – also cooperate with many non-Hox transcription factors. Although it was long known that TALE proteins have non-Hox functions [19–21], Bobola and Sagerstrӧm suggest that it may be appropriate to flip the traditional view: it may be more accurate to say that Hox proteins are cofactors for the TALE transcription factors. One of the arguments in favor of this view is that the TALE factors are bound to chromatin prior to Hox protein expression, suggesting that the TALE factors may be acting as pioneer transcription factors, which are able to bind to sites initially made inaccessible by nucleosomes. Interestingly, and along the same lines, the article by Paul et al. [22, this issue] points out that some Hox proteins can themselves act as pioneer factors, and that differences in the pioneering ability between Hox proteins may also impact binding specificity in vivo. Notably, and consistent with the view put forth by Bobola and Sagerstrӧm, the pioneering activity of some Hox proteins is dependent on TALE transcription factors. Last but not least, the article by Salomone et al. [23, this issue] also summarizes the cofactor models for Hox specificity, but emphasizes that there is often a trade-off between specificity and affinity: the most specific Hox-TALE binding sites tend to be low affinity. This raises the question of how low affinity binding sites can be sufficiently bound in nuclei, which typically do not have high concentrations of transcription factors. One likely solution is that the distribution of Hox proteins within nuclei – as probably the case for most transcription factors – is non-uniform and concentrated in local hubs, providing high local concentrations. For many non-Hox transcription factors, these hubs have been shown to form via interactions between intrinsically disordered regions (IDRs) that can form liquid-liquid phase separated condensates [24]. Interestingly, as pointed out by Salomone et al., intriguing evidence that this mechanism is relevant to Hox proteins comes from the characterization of Hox mutations that lead to both alterations in their IDRs and to human diseases, such as cancer and synpolydactyly.

As a group, these ten articles summarize multiple important and novel insights stemming from research into Hox genes that help inform a wide variety of questions currently debated by biologists. Although many questions remain and will require future research to answer, it is particularly noteworthy and satisfying that whatever insights have been obtained thus far – ranging from deep evolutionary questions to the mechanisms underlying human diseases – ultimately stem from the instinctive curiosity of Drosophila geneticists who first discovered these fascinating genes and who never could have imagined where their initial discoveries would lead.

Cover image caption:

As first discovered by E. B. Lewis, loss of function mutations in the Hox gene Ubx results in flies with a nearly complete duplication of the second thoracic segment, resulting in two pairs of wings (right), in contrast to wild type flies (left) with only one pair of wings. Photograph credit to Nicholas Gompel (http://gompel.org/).