Homeobox genes: going for growth

In this issue of Gut, Domon-Dell and colleagues¹ report the influence of butyrate on the intestinal homeobox gene Cdx2 [see page 525]. Butyrate is a byproduct of fibre fermentation by colonic bacteria and is known to be a substrate for colonocytes. Over the last few years there has been increasing interest in the function of genes involved in controlling the anatomical relationships between various organs. This article therefore touches on the role of homeobox genes and dietary factors in gut growth and differentiation.

The development of multicellular animals is dependent on expression of a hierarchy of genes that sequentially provide increasingly detailed positional information. Many of these genes contain a highly preserved “homeobox” sequence that codes for a DNA binding homeo domain. In insects, a group of such genes—specifying the individual characteristics of body segments and known as homeotic selector genes of the Antp-type (the defining gene is named Antennapaedia)—are clustered in a complex known as the HOM cluster. These are expressed topographically in the same order as they occupy in chromosomal DNA. They are conserved strongly during evolution so that their homologues appear as four paraologous gene complexes known as Hox complexes which specify positional information in mammalian embryos.

Other homeobox genes are scattered in the genome and of particular interest are homologues of a Drosophila homeobox gene called Caudal (Cad). This is known to be important in determining anatomical positioning as Drosophila larvae lacking Cad gene products show the phenotype of deletions of many of the posterior segments. The first mammalian Cad homologue (Cdx1) was isolated from an embryonic mouse cDNA library with two further members of the Caudal family (Cdx2 and Cdx4) having subsequently been cloned in mice. Cdx3 was originally identified in the pancreas of hamsters but has subsequently been shown to be identical to Cdx2.

Mutation of Cdx1 or 2 results in a skeletal homeotic shift, supporting an important role of the Cad genes in axial patterning. Homozygous deletion of Cdx2 is a lethal mutation, probably relating to its importance in the development of extra-embryonic tissue. The initial reports of the phenotype of Cdx2 +/− (that is, heterozygote) mice considered that it resulted in distal bowel tumorigenesis,² suggesting a role for Cdx2 as a tumour suppressor gene. More recently, additional studies suggest that this may have been a histological misinterpretation with the “tumours” being areas of gastric metaplasia.³ It therefore seems likely that Cdx2 directs endodermal tissue towards a caudal differentiation and that Cdx2 haplo insufficient areas develop as forestomach epithelium. Intercalary growth subsequently fills in the missing tissue types at the discontinuity between gastric and colonic epithelium resulting in the polypoidal lesions seen.

In addition to the key role that homeobox genes play during development, it is likely that at least some of these genes also play important functions on proliferation and differentiation during adult life. Results of manipulating Cdx-1 expression have given somewhat contradictory results. Transfection studies overexpressing Cdx-1 in the intestinal cell line, IEC6, has been reported to increase proliferation and differentiation⁴ but also transformational and tumorigenic activity.⁵ However, when the same approach was used on the human colonic cell lines HT29⁶ or Caco2 cells,⁷ it had virtually no effect on proliferation when transfected on its own but enhanced the growth inhibitory effect of Cdx-2 in cotransfected cells.⁶ An alternative approach is to reduce the endogenous levels of Cdx-1 using antisense RNA. This has been tried and shown to slow growth of Caco-2 cells,⁷ providing additional support for the idea that Cdx-1 plays a role in controlling proliferation. The data in support of Cdx-2 functioning as a tumour suppressor gene are somewhat stronger: Cdx2 overexpression reduces cell growth in IEC, Caco-2, and HT29 cells and Cdx2 expression is decreased in relation to the tumour grade in human colorectal cancers cells and in chemically induced tumours in the rat.⁸ In addition, the morphology of transfected IEC and Caco-2 cells is altered with increased differentiation and a rise in expression of small intestinal digestive enzymes such as sucrase-isomaltase.⁷ Taken together, it appears likely that Cdx2 functions as a regulator of intestinal cell differentiation in addition to its developmental role in embryogenesis.

Control of Cdx-1 and 2 expression is poorly understood although alteration in oncogenic Ras activation, which is an early and frequent event in colorectal cancers, is thought to decrease Cdx-2 expression, acting via the PKC-jun-fos pathway while increasing Cdx-1 expression by acting through the Raf-MEK1 pathway (for an excellent review, see Freud and colleagues⁹). In the paper reported in this issue of Gut, Domon-Dell et al have focused on the interactions between Cdx-2 expression and butyrate and found that its presence stimulates Cdx-2 expression. In their discussion, they provide further (unpublished) information regarding potential mechanisms for this effect, such as the possibility of an atypical butyrate response element by which it might act at a molecular level although the final answer remains unclear. Whichever mechanisms are involved, it seems likely that this is an example of direct nutrient regulation of intestinal cell function and its findings have applicability beyond the current study.

The usual impression of a “growth factor” is a peptide that binds to specific external receptors, inducing a secondary signalling cascade resulting in cell proliferation. When considering complex interactions, such as the control of growth and differentiation of gut cells, it is important to appreciate that the somewhat arbitrary labelling of a molecule to an individual function—for example, considering “epidermal growth factor” simply as a stimulant of proliferation—can be misleading as it is now clear that such factors have multiple effects. For example, although they are often considered separately, the distinction between cytokines and growth factors are sometimes blurred as the “cytokine” interleukin 8 has been shown to stimulate migration of human colonic epithelial cells,¹⁰ a process normally associated with “growth factors”, and peptides normally considered to be “growth factors” can influence immunological function. Similarly, molecules such as glutamine or butyrate which have been generally considered to be simple energy providers, and vitamins such as A and D, which were at one time thought to have limited biological functions, are now known to influence many other activities of the cell, such as development, differentiation, and proliferation (for example, see Nagpal and Chandraratna¹¹). The current study therefore provides further evidence for additional actions of relatively simple molecules on multiple functions within a cell, acting through pathways such as specific DNA response elements. The distinction between nutrition and pharmacology is therefore less clear than it first appears.