Introduction

Aniridia is a human genetic disease that is manifested by alterations in the structure and function of the eye, including reduced iris size, absence of the fovea, and lens deformities (Glaser, et al. 1992). First documented as a genetic disease over 150 years ago, it has since become a model for autosomal dominant genetic disorders because of the high penetrance of its mutant alleles, the ease of diagnosis at birth, and a similar incidence in various populations (Glaser, et al. 1995). It was not until recently, however, that the aniridia gene (AN) was mapped (to chromosome band 11p13) and determined to be the gene PAX6  (Glaser, et al. 1992), a regulator of development of the eyes and central nervous system. There is a PAX6 dosage effect in aniridia ranging from mild loss of visual acuity and cataracts to severe nervous system defects  and anopthalmia (complete absence of the eyes) (Glaser, et al. 1994). 

Many of the studies that led to the identification of PAX6 as the human aniridia gene were conducted in mammals and insects and demonstrated that PAX6 is highly conserved among vertebrates and lower animals. PAX6 homologues have been found in mice (Small eye, or Sey), rats, zebrafish, quail, and the fly Drosophila (eyeless, or ey) with amino acid sequence identities of approximately 90% (Quiring, et al. 1994); PAX6 is 96% identical in its amino acid sequence to the pax(zf-a) in the zebrafish, but the two species diverged over 400 million years ago (Glaser, et al. 1992). This degree of conservation approaches that of histones, some of the most highly conserved proteins known. More important, however, is the finding that Drosophila and vertebrates have homologous PAX6 genes with similar functions and similar mutant phenotypes. In Origin of the Species, Darwin found it difficult to explain the evolution of structures as dissimilar as simple vertebrae eyes and compound insect eyes; he speculated that the structures might have developed separately through convergent evolution (Zuker, 1994). However, in both insects and vertebrates, PAX6 is expressed in the embryo just prior to and during formation of the eye in the region of its development. Based on that evidence, and the finding that misexpression of a PAX6 homologue in flies could induce ectopic eyes, it has been suggested that PAX6 is a master regulatory gene that induces eye development in a broad range of animals (Halder, et al. 1995). 

The PAX multigene family

The PAX6 protein is one of many transcription factors that induce embryonic differentiation along the major body axes. In response to concentration gradients of other regulatory proteins, these transcription factors bind to specific DNA sequences of other genes and regulate their expression, thus translating positional information into developmental patterns for distinct structures (Glaser, et al. 1994). 

PAX6 is a member of the PAX multigene family of transcription factors that help regulate embryonic differentiation. Like PAX6, many other PAX genes are expressed in the developing nervous system and are believed to help regulate neurogenesis. In mice, Pax1 is expressed in, among other areas, the developing vertebral column and thymus; Pax3 in the early neural tube; Pax6, upon closure of the neural tube, in the developing hindbrain and forebrain and several other areas of the developing nervous system; and Pax8 in the neural tube, hindbrain, and thyroid. PAX3 mutant phenotypes include deafness, depigmentation, and spina bifida (failure of neural tube closure) (Strachan, et al. 1994). 

Nine unlinked PAX genes have been identified by their homology to the paired (prd) segmentation gene expressed in Drosophila larvae. These genes encode proteins that all include a 128-amino acid sequence-specific DNA-binding domain, the Paired box, which can regulate the expression of other genes. Recent research on the binding of the Pax6 paired domain has revealed a possible structure of three a-helixes (Strachan, et al. 1994), a consensus DNA-binding sequence, and evidence for conformation changes in the protein upon binding (Epstein, et al. 1994). 

PAX genes (4,6,3,7),  also contain another common DNA-binding element, the homeobox. The homeobox, first discovered in Drosophila, encodes a 60-amino acid homeodomain that is thought to be part of more than 0.2% of the total number of vertebrate genes. The homeodomain, too, contains three a-helices, one of which is responsible for target sequence recognition (Glaser, et al. 1992). Recent research has also shown that the paired domain and homeodomain may interact cooperatively to recognize multiple DNA binding sites. (Jun and Desplan, 1996). Unlike the HOX   (Homeo boX) family of homeobox-containing genes, which regulate many aspects of embryonic morphogenesis, PAX genes are not clustered but are dispersed throughout the genome (Mark, et al. 1997). Particular attention has been devoted to the molecular biology of the PAX6 gene because of its high degree of conservation and its seeming ability to regulate development of both the compound and the simple eye structures. 

PAX6 Gene Structure

Human PAX6 is transcribed as a 2.7kb mRNA and encodes a 422-amino-acid protein that includes the paired box, the homeo box, and a third possible DNA-binding motif, the PST domain (Proline, Serine, and Threonine-rich sequence; Glaser, et al. 1995. See Figure 1). Interestingly, PAX6 contains an alternative mRNA splice-site in the paired domain which can result in a 42-nucleotide insertion; the insertion allows the carboxy terminal subregion of the paired domain to recognize a novel DNA sequence, allowing PAX6 to regulate an expanded or restricted set of genes depending on how the mRNA is spliced (Epstein, et al. 1994). PAX6 extends over 22kb and contains 14 exons and intron sequences in the homeobox itself. In addition, a CCAGCATGC translation start site in exon 4, a TAA stop codon in exon 13, a transcription start site and promoter region with TATA, CAAT, and GC regulatory elements, and three possible polyadenylation signals have all been characterized in several converging lines of research. (Glaser, et al. 1992). 

Mutations in various positions within PAX6 give rise to gene dosage effects that support the hypothesis that PAX6 regulates gene expression during development by means of concentration gradients with other transcription factors. In one family, truncation of PAX6 in the PST domain by a point mutation in exon 12 led to cataracts and decreased visual acuity in the father; truncation of PAX6 in the paired domain by a point mutation in exon 6 led to iris absence, cataracts, severe decreased visual acuity, and other ocular malformations in the mother; and a daughter compound heterozygote with a copy of each of the parent's mutated PAX6 genes died eight days after birth with severe central nervous system and craniofacial defects and anopthalmia (Glaser, et al. 1994). 

Drosophila as a PAX6 model system

Research on PAX6 has been facilitated by the discovery of a PAX6 homologue in Drosophila, the eyeless (ey) gene (Quiring, et al. 1994). Mutations in ey produce eye defects similar to those produced by Sey (mouse) and PAX6 mutations. Drosophila provides a convenient and useful model for PAX6 because its growth is rapid, its genetic and embryonic mechanisms have been well-characterized, and it is simple and inexpensive to maintain. Like PAX6, ey is expressed in the embryonic nerve cord, specific regions of the brain, and in eye precursors. 

Much PAX6 research focuses on the developmental pathways that lead to eye formation. Upstream and downstream regulatory genes have been investigated for ey in Drosophila. Hypothesized downstream targets of ey include eyes absent, sine oculis, and dachshund (Halder, et al. 1995). Though few upstream regulatory protein products of PAX6 or ey have been identified, a number of regulatory elements in the PAX6 and ey genes have been identified through gel-shift and footprinting assays and in transgeneic in vivo studies of mice expressing a lacZ reporter under the control of various Pax6 regulatory elements (R. Maas, personal communication). Other possible upstream regulators of Pax6 include activin A (Pituello, et al., 1995) and sonic hedgehog (another developmental control gene) (Ericson, et al., 1997). 

PAX6 as a master regulator

Because ey expression is not affected by mutations in other eye-determining genes, and because PAX6 is so highly conserved, PAX6 is hypothesized to be a "master regulator of eye development" (qtd. in Displan, 1997). This hypothesis was recently supported by the finding in Drosophila that misexpression of ey could induce ectopic eye formation on appendages (Halder, et al. 1995). The ectopic eyes were fully formed and included the full complement of cell types and structures, including photoreceptors. 

Later research, however, revealed that genes supposedly "downstream" of ey could also induce ectopic eyes. The products of Sine oculis (so), which encodes a homeobox-like domain, and eyes absent (eya), which encodes a novel nuclear protein, form a complex and can induce ectopic yes similar to those formed by ectopic ey expression (Pignoni, et al. 1997). More significantly, so and eya could together induce ey expression, which is not consistent with the idea of ey as the master regulatory of those genes. In addition, dachshund (dac), which encodes another novel nuclear protein and is induced by ey expression, and eya misexpression resulted in full ectopic eye production as well, while dac and eya alone could each weakly induce ectopic eye formation (Chen, et al. 1997). 

These gain-of-function experiments suggest that the protein products of ey and of its human homologue PAX6 operate not in a hierarchical linear pathway, but as a network with numerous feedback loops. A second possible hypothesis is that, because eye regulatory genes are activated several times during development, they are turned on in sequence at each stage. The repeated use of the same regulatory genes during eye development has been explained in conjunction with the high degree of conservation of the genes: as eye formation progressed during evolution from simple photoreceptors to the complex visual systems in insects and vertebrates, the same regulatory genes were co-opted for each new developmental pathway (Desplan, 1997). 

It is also possible that PAX6 and other eye development genes play a role in the development of other organs, such as the pancreas (St-Onge, et al., 1997). In addition, an eya homologue in humans, when mutated, has no apparent affect on eye formation. These results suggest that PAX6 may be involved in the larger process of organogenesis rather than only oculogenesis. 

Conclusions

The highly conserved PAX6 and its Drosophila homologue ey are key players in a highly complex developmental pathway leading to formation of both simple and compound eyes and possibly other organs as well. Research in several laboratories is directed towards characterizing the complex network of regulatory genes involved. Several PAX6 enhancer elements show promise as sites for upstream regulation of PAX6, and possibly even downstream products of eya, so, or dac may play a regulatory role. 

Several medical applications have arisen in recent years based on research on the PAX6 gene family. Because the gene has been sequenced, prenatal diagnosis of aniridia is now possible. In addition, some evidence suggests that PAX6 may be expressed by damaged eye tissue to induce limited regeneration; artificial upstream regulation of PAX6 may eventually be used to induce such regeneration. Finally, some cancers, including alveolar rhabdomyosarcoma, may be caused by PAX mutations (Strachan, et al. 1994). These findings suggest potential therapeutic applications for PAX6 research and may lead to a more complete understanding of its role in eye development.