Mechanisms of transcriptional regulation by WT1 (Wilms’ tumour 1)

Abstract

The WT1 (Wilms’ tumour 1) gene encodes a zinc finger transcription factor and RNA-binding protein that direct the development of several organs and tissues. WT1 manifests both tumour suppressor and oncogenic activities, but the reasons behind these opposing functions are still not clear. As a transcriptional regulator, WT1 can either activate or repress numerous target genes resulting in disparate biological effects such as growth, differentiation and apoptosis. The complex nature of WT1 is exemplified by a plethora of isoforms, post-translational modifications and multiple binding partners. How WT1 achieves specificity to regulate a large number of target genes involved in diverse physiological processes is the focus of the present review. We discuss the wealth of the growing molecular information that defines our current understanding of the versatility and utility of WT1 as a master regulator of organ development, a tumour suppressor and an oncogene.

INTRODUCTION

The WT1 (Wilms’ tumour 1) protein was among the first tumour suppressor genes to be cloned. Research over the last two decades has revealed much about the functions of WT1 in development and disease, but also many surprises. Indeed, it has become clear that WT1 can act as either a tumour suppressor or an oncogene, it is a transcriptional regulator that can either activate or repress genes, and it is also a post-transcriptional regulator. Through these activities WT1 plays critical roles in development, adult tissue maintenance and also as an epigenetic regulator. These diverse activities are at least partly because of the complex nature of the WT1 gene, its products, post-translational modifications and interaction partners. The purpose of the present review is to summarize the growing molecular information that describes WT1 function. We begin with a brief overview of the structural features of WT1, its isoforms and modifications. We will briefly discuss the role of WT1 as a post-transcriptional regulator and its involvement in RNA metabolism. We then primarily focus on the role of WT1 as a dichotomous regulator of transcription, its target genes and known interaction partners.

WT1 and development

The WT1 protein is indispensable for normal development of the genitourinary system and also several other organs and tissues. WT1-null mice are embryonic lethal with complete agenesis of the kidneys, gonads, heart and diaphragm. WT-null mice die because of heart failure caused by thinning of the epicardium where WT1 appears to be required for proliferation of vascular progenitors [1,2]. WT1-null mice also lack the spleen and adrenal glands [2,3]. Later studies determined an important role for WT1 in the development of neuronal tissues. WT1-null mice do not develop olfactory epithelia or retina ganglia, and show decreases in the proliferation of neuronal progenitor cells [4,5]. More recently, WT1 has been shown to play a role in the development of the peripheral taste system [6], suggesting a significant function for WT1 in the formation of sensory tissues.

Many studies have determined the direct and necessary role of WT1 for normal kidney function. The kidney is formed by interactions between the metanephric mesenchyme and ureteric bud. WT1 is expressed in the metanephric mesenchyme, which condenses to form the epithelium of the nephron [7]. In the absence of WT1, these cells undergo apoptosis. In gain-of-function experiments, microinjection of a WT1-expressing plasmid into isolated embryonic kidneys caused normal nephron development [8]. A previous study used siRNA to ablate WT1 expression at specific times in organ explants of the developing kidney [9]. They demonstrated that WT1 is needed at several stages of kidney development: the condensation of the metanephric mesenchyme on to the ureteric bud, and during the MET (mesenchyme–epithelial transition).

How WT1 controls the development of the kidney is not fully clear. The identification of direct WT1 target genes is crucial to fully understand the role of WT1 in development. A recent study by Kreidberg and co-workers coupled ChIP to a mouse promoter microarray (ChIP-chip) and identified several essential kidney development genes including many previously unknown WT1 target genes involved in nephron progenitor differentiation in vivo [10]. Such studies clearly demonstrate the essential transcriptional role of WT1 in kidney development. It is interesting to note that although, in some instances, WT1 causes neuronal progenitor cells to proliferate, in some other tissues such as the kidney it switches cells from proliferation to differentiation. These different biological functions are tissue-context-dependent and might also be isoform specific. Recent work in uncovering WT1 target genes is providing some new insights into how WT1 functions in different tissues.

WT1 and adult tissue maintenance

In the adult kidney, WT1 expression is restricted to a highly specialized cell type: the podocytes that make up the filtration barrier between blood and urinary spaces [11]. Combining WT1-knockout and inducible yeast artificial chromosome transgenic mouse models, a study demonstrated that reduced expression of WT1 results in the disruption of glomerular health causing either crescentic glomerulonephritis or mesangial sclerosis [12]. Moreover, the two podocyte-specific genes Nphs1 (nephrosis 1) and Podxl (podocalyxin) are down-regulated in WT1-knockout mice [12], which underscores the crucial role of WT1 in podocyte maintenance.

Apart from podocytes, WT1 is also expressed in the Sertoli cells in the testes, granulosa cells in the ovaries, mesothelium and approximately 1 % of cells in the bone marrow of adult mice [11,13–16]. A previous study shed new light regarding the need for WT1 in adult tissues: the specific deletion of WT1 in adult mice caused rapid deterioration of multiple tissues in addition to kidney failure [17]. These included atrophy of the spleen and pancreas and failure of red blood cell formation. Furthermore, within 7–9 days of WT1 deletion, a widespread loss of bone and body fat volume was seen. That study further showed that WT1 loss causes a reduction in the expression of IGF1 (insulin-like growth factor 1), which is a known key regulator of homoeostasis and aging [17]. Collectively these studies show that WT1 plays a role(s) in kidney function in adulthood where it is required for podocyte function and kidney homoeostasis. Transgenic mice continue to provide new insights into WT1 function and the use of tissue-specific Cre lines can be extended to reveal further roles for WT1 in adult tissues. Moreover, a detailed analysis will probably reveal the functions of WT1 in more discrete tissues such as the sensory systems.

WT1 and cancer: tumour suppressor and oncogene

WT1 manifests both tumour suppressor and oncogenic activities, but the molecular mechanisms behind these two activities are still not well understood. The function of WT1 as a classic tumour suppressor is linked to the formation of Wilms’ tumours where WT1 was originally identified [18,19]. Wilms’ tumours are a malignancy of the kidneys that occurs in 1 in 10 000 children [18,20]. The tumours are also associated with hereditary genitourinary disorders such as DDS (Denys–Drash syndrome) and Frasier syndrome [21]. In Wilms’ tumours, a WT1 germline mutation can lead to predisposition to the disease. Also, WT1 is inactivated in many Wilms’ tumours causing increased cell survival and proliferation [22]. Several other genes are mutated in Wilms’ tumour such as WTX [Wilms’ tumour on the X; also known as AMER1 (APC membrane recruitment protein 1)], β-catenin and p53 [18,22]. In addition, LOH (loss of heterozygosity) or LOI (loss of imprinting) on chromosome region 11p15 is observed in up to 70 % of the tumours [23,24]. However, the mechanisms by which one or more genetic modifications cause Wilms’ tumour are unknown. A study by Huff and co-workers shed light on a potential mechanism of action of WT1 in Wilms’ tumours [25]. They produced the first mouse model that mimics a subset of human Wilms’ tumours by engineering mice to sustain somatic deletion of WT1 and constitutive Igf2 expression. That study determined that mice with these combinations developed tumours very early. Furthermore, mechanistically, a lack of WT1 blocked mesenchyme differentiation and thus kidney development leading to cell proliferation and Wilms’ tumours [25]. Thus the Wilms’ tumour mouse model presents a powerful tool in the analysis of WT1 in Wilms’ tumour that provides a better understanding of its tumour suppressor roles in tumorigenesis.

Whereas WT1 behaves as a tumour suppressor gene in Wilms’ tumour, a wealth of data on the expression of WT1 in adult tumours suggest a role as an oncogene. WT1 is expressed in tumours derived from tissues in which it is not normally expressed such as the colon, breast, desmoids and brain [18,26–29]. Furthermore, evidence using antisense oligonucleotides against WT1 demonstrated that ablation of WT1 protein led to cell-cycle arrest at the G₁ phase and increased apoptosis of breast cancer cells [30]. Interestingly, WT1 is transcriptionally up-regulated under conditions of low oxygen by HIF (hypoxia-inducible factor) as found in a tumour environment [31]. Furthermore, later studies revealed a role for WT1 in angiogenesis and vascularization, which is an important step for tumour growth. WT1 has been shown to be involved in endothelial cell proliferation and migration [32], and more recently WT1 has been shown to regulate the level of anti-angiogenic VEGF (vascular endothelial growth factor) isoforms [33]. The role of WT1 in leukaemia appears to be complex. WT1 is expressed in most acute leukaemias and it is currently used as a marker of poor prognosis and low survival rates [34]. WT1 is also expressed in immature myeloid leukaemias and CML (chronic myelogenous leukaemia) [35]. Collectively these studies strongly suggest an oncogenic role for WT1 in leukaemogenesis. However, studies also show WT1 mutations in a considerable proportion of both adult and child AML (acute myeloid leukaemia) supporting a tumour suppressor role of WT1. The majority of mutations involve insertions, deletions and mis-sense mutations such as those observed in patients with DDS [22].

How does WT1 act as both a tumour suppressor and oncogene? The oncogenic or tumour suppressive effect of WT1 might result from the cellular conditions at specific time of development and how this responds to modifications in or expression of WT1 [22]. This in turn might be explained by the varying context of cell type and cell status at the moment of WT1 alteration, presence of other downstream genetic modifications and/or microenvironment [22]. This places WT1 in the growing group of proteins such as c-Myc that can exhibit either pro- or anti-tumour activities. The generation of mouse models in which WT1 can be selectively overexpressed in specific tissues will probably shed light on the oncogenic activities of WT1. Regarding tumour treatment, peptide vaccines against WT1 are showing promising results in repressing tumour growth in patients with leukaemia, breast and lung cancer [36]. In addition, the treatment of tumour cell lines with cytotoxic drugs leads to proteolysis of WT1 by the serine protease HtrA2 (HtrA serine peptidase 2) [37]. This process could potentially be exploited to target adult tumours where WT1 acts as an oncogene. The emergence of WT1 as a target for cancer therapy necessitates a greater understanding of the molecular mechanisms of WT1 action, how the different forms of WT1 contribute to these activities and functional consequences of mutations within the WT1 gene. Mouse models to study oncogenic WT1 will help in this regard.

WT1 STRUCTURE

WT1 isoforms

The WT1 gene spans approximately 50 kb at chromosome locus 11p13 and was among the first tumour suppressors to be cloned [38–40]. Many functions have been ascribed to WT1. These include acting as a transcriptional activator and a repressor, a tumour suppressor and oncogene, a function(s) in RNA metabolism, and roles in development, tissue maintenance and disease.

This functional complexity of WT1 can, at least in part, be explained by the generation of multiple protein forms derived from the WT1 gene. At least 24 isoforms of WT1 are generated by a combination of alternative translation start sites, alternative RNA splicing and RNA editing. In mammals, exons 5 and 9 are subjected to alternative splicing, generating four different major splice isoforms [41,42]. These isoforms are also the best studied. Alternative splicing of exon 5 gives rise to a 17-amino-acid insertion (+ 17AA) between the proline/glutamine-rich N-terminus and the C-terminal zinc finger domain of WT1 (Figure 1). Alternative splicing at the end of exon 9 results in either inclusion or exclusion of three amino acids, lysine, threonine and serine (KTS), between the third and fourth zinc finger of WT1 (Figure 1). These two isoforms (+KTS and −KTS) are conserved in all vertebrates, from zebrafish to humans, and imbalanced expression of these variants leads to developmental abnormalities. Fish and non-mammalian vertebrates appear to express only these two variations [43]. This suggests that the different KTS isoforms perform different functions and comprise one of the most important variations. Indeed, WT1 can bind both DNA and RNA and this is influenced by the KTS insertion. Numerous studies have demonstrated that WT1 isoforms lacking the KTS insertion bind to DNA more strongly, and act as transcriptional regulators [44]. The gene product that contains the KTS insertion also acts as a transcriptional regulator, but in addition is associated with post-transcriptional processes. The physiological role of the +KTS and −KTS variants has been studied by mouse models that have a specific isoform deleted. Interestingly, the +KTS variants are essential for the male sex determination programme [45]. The gonads of mice lacking WT1 + KTS variants lack expression of male-specific genes such as Sox9 [SRY (sex-determining region Y) box 9] and Mis [Mullerian-inhibiting substance; also known as Amh (anti-Mullerian hormone)] and develop along the female pathway. It also has been shown that lack of WT1 + KTS isoforms specifically leads to defects in the development of the olfactory system [5].

Figure 1. Functional motifs of the Wilms’ tumour 1 protein WT1.

A linear schematic is shown with numbering indicating the amino acids. The alternative splice sites (17AA and KTS) are shown above along with the post-translational modifications (SUMOylation and phosphorylation). Regions of the protein involved in self-association, nuclear export and nuclear acid binding are shown above. The binding sites of some of WT1 cofactors, such as BASP1, HtrA2, WTX, STAT3, p53, Par-4, CBP and SRY, are indicated below the schematic diagram. Lightning bolts indicate HtrA2 cleavage sites in WT1. A, activation domain; R, repression domain; Zn, zinc finger.

Additional isoforms of WT1 also arise from alternative translation start points. They include the use of an upstream CTG start codon or an internal ATG start codon at the end of exon 1 [18]. Another WT1 isoform, AWT1, arises from the use of an alternative promoter that resides within exon 1 [46]. The authors of that study demonstrated that the AWT1 isoform is imprinted in normal kidney and expression is confined to the paternal allele. However, the properties and function of AWT1 remain to be investigated in vivo.

Determining the role(s) of the different WT1 gene products, gene regulation, and their effects in development and disease is likely to remain a major challenge for some time. Expression of single WT1 isoforms in cultured cells is of questionable physiological relevance, and RNAi-mediated depletion of specific WT1 isoforms is, in large part, not possible. Despite these caveats, such analysis can still yield important insights. Indeed, ectopic expression of individual isoforms of WT1 revealed the function of KTS in directing the interaction of WT1 with RNA processing factors. Although transgenic models have shed light on the role of the KTS insertion in WT1, unfortunately they have not been as fruitful with some of the other WT1 isoforms [43,44]. However, it is clear that our understanding of the roles of WT1 in both development and tissue homoeostasis is not yet complete and other mouse models of WT1 isoforms may still provide insights into their function. Indeed, this may be particularly important in studying the potential oncogenic roles for the different isoforms.

The transcriptional regulatory domains of WT1

The primary sequence of WT1 contains a proline- and glutamine-rich region in the N-terminal domain and four zinc fingers at the C-terminus of the C2-H2 ‘Krüppel’ type [48]. WT1 is a dichotomous transcriptional regulator that can either activate or repress transcription. WT1 has domains that function independently to activate or repress transcription when fused to the GAL4 DNA-binding domain. Specifically, the transcriptional repression domain is contained within WT1 residues 71–180 and the transcriptional activation domain between residues 180 and 250 (Figure 1) [47,48]. The N-terminal region of WT1, specifically two distinct domains residing at amino acids 1–45 and 157–253 mediate self-association [49]. A physiological role for WT1 self association into dimers or multidimers is not yet understood. However, self association of WT1 supports the model of a dominant-negative effect of WT1 mutants on WT1 wild-type protein. Indeed DDS is typified by WT1 proteins that are C-terminally truncated (or contain C-terminal point mutations) that render them defective in binding nucleic acids and yet act in a dominant-negative manner [50].

Refined mutagenesis of the WT1 repression domain identified a subregion (residues 71–101) termed the SD (suppression domain) [50]. The SD inhibits transcriptional activation by WT1 in in vitro transcription assays, but does not inhibit basal activator-independent transcription [20,50]. Moreover, the WT1 SD can inhibit a heterologous transcriptional activation domain in cis both in vivo and in vitro [50]. The WT1 SD functions by recruiting a transcriptional co-repressor, BASP1 (brain acid-soluble protein 1), that inhibits transcriptional activation by WT1 [51]. The SD of WT1 is 90 % conserved across species compared with approximately 50 % identity for the whole of WT1 [50]. To date, the SD has not been identified in any other transcription factors.

Post-translational modification of WT1

It has been shown that WT1 can be SUMOylated at two lysine residues within the N-terminus [52] (Lys⁷³ and Lys¹⁷⁷; Figure 1). WT1 had previously been shown to interact with the SUMO-1 (small ubiquitin-related modifier-1) E2-conjugating enzyme Ubc9 (ubiquitin-conjugating enzyme 9) [53]. The sites of WT1 SUMOylation lie within the transcriptional repression domain of WT1, but it is not clear if this modification affects the transcriptional regulatory functions of WT1. WT1 conjugates SUMO-1 and, although WT1 associates with nuclear speckles, the disruption of WT1 SUMOylation does not appear to affect its subnuclear localization [44,52]. Thus it is probable that SUMOylated WT1-binding partners might help to retain WT1 within the speckles. Indeed the WT1-binding partner BASP1 is also SUMOylated, but by SUMO-3 and not SUMO-1 [54]. This raises the potential for the differential regulation of WT1 and BASP1 by SUMOylation. However, it remains to be determined how SUMOylation of WT1 affects its function or association with its cofactors. A more detailed analysis of WT1 SUMOylation in developing tissues or cell line models of differentiation may shed further light on this post-translational modification.

WT1 can also be phosphorylated by PKA (protein kinase A) or PKC (protein kinase C) at Ser³⁶⁵ and Ser³⁹³ in the second and third zinc fingers respectively (Figure 1) [55,56]. Phosphorylation of WT1 blocks the ability of WT1 to bind to DNA. This, in turn, affects transcription and decreases transcriptional repression activity in vivo, but leaves the RNA-binding activity intact [55]. The authors also reported that the localization pattern of WT1 in the nucleus is altered by phosphorylation resulting in the cytoplasmic retention of WT1. These findings strongly suggest that phosphorylation negatively regulates the transcriptional role of WT1 by sequestration of WT1 protein in the cytoplasm and inhibition of the WT1 DNA-binding activity [55]. It will be interesting to test if phosphorylation of WT1 affects its post-transcriptional functions or localization to the polysomes (see below), perhaps by modulating interactions between WT1 and cytoplasmic proteins. Thus post-translational modifications represent ways in which the activity of WT1 can be modified. The existence of other WT1 post-translational modifications, such as acetylation or methylation, has yet to be determined. To date, resource studies utilizing global proteomic analysis of post-translational modifications have not used cell types that express WT1. With the number of such proteome-wide analyses growing it is probable that future works will incorporate WT1-expressing cells.

WT1 AS A DICHOTOMOUS REGULATOR OF TRANSCRIPTION

WT1 target genes: growth and development

WT1 functions as a transcriptional regulator that activates or represses a large variety of target genes. The identification of WT1 target genes is an ongoing and difficult process due to the isoform- and context-specific roles of WT1. Many potential target genes of WT1 have been identified; however, it is not fully understood which genes are the biological targets of WT1 and how WT1 orchestrates their regulation in development, tissue homoeostasis or disease. WT1 target genes can be thematically grouped. WT1 can induce expression of genes involved in organ development and the MET. Similarly, WT1 regulates growth factor genes and their receptors. Depending on the cellular context, WT1 can also regulate genes involved in apoptosis. Emerging target genes demonstrate a transcriptional role for WT1 in cytoskeleton organization and cell adhesion, actin biogenesis, cell signalling such as through MAPK (mitogen-activated protein kinase) and WNT, and epigenetic regulation. Table 1 presents a list of all the proposed WT1 target genes, the methods of validation and the organisms in which they were identified.

Table 1.

WT1 target genes

(a) Growth and development
Target gene	Organism(s)	Method of validation	Reference(s)
CXXC5(WID)	Human, mouse	ChIP-chip, ChIP, reporter assay, organ culture morpholino	[10,80]
BMP7	Mouse	ChIP-chip, ChIP, organ culture morpholino	[10]
Pax2	Mouse, human	ChIP-chip, ChIP, DNase I footprinting, reporter assay	[10,57]
Sall1	Mouse	ChIP-chip, ChIP, reporter assay, organ culture morpholino	[10]
Hey1	Mouse	ChIP-chip, ChIP, reporter assay, organ culture morpholino	[10]
Egr3	Mouse	ChIP-chip, ChIP	[10]
Fyn	Mouse	ChIP-chip, ChIP	[10]
Hoxd4	Mouse	ChIP-chip, ChIP	[10]
Sox11	Mouse	ChIP-chip, ChIP, reporter assay, organ culture morpholino	[10]
Pbx2	Mouse	ChIP-chip, ChIP, reporter assay, organ culture morpholino	[10]
Zfr	Mouse	ChIP-chip, ChIP	[10]
Hhat	Mouse	ChIP-chip, ChIP	[10]
Arnt2	Mouse	ChIP-chip, ChIP	[10]
Erbb2	Mouse	ChIP-chip, ChIP	[10]
Smad7	Mouse	ChIP-chip, ChIP	[10]
Pbx1	Mouse	ChIP-chip	[10]
Pixnb1	Mouse	ChIP-chip, ChIP, reporter assay, organ culture morpholino	[10]
BMP4	Mouse	ChIP-chip, ChIP, qRT-PCR	[6,10]
Smad4	Mouse	ChIP-chip, ChIP	[10]
Smad3	Mouse	ChIP-chip, ChIP	[10]
Vegfa	Mouse	ChIP-chip, ChIP	[10]
PTIP	Mouse	ChIP-chip, ChIP	[10]
Sulf2	Mouse	ChIP-chip, ChIP	[10]
Smad6	Mouse	ChIP-chip, ChIP	[10]
Smo	Mouse	ChIP-chip, ChIP	[10]
ER1	Human	Reporter assay, ChIP	[135]
Amhr2	Mouse	Microarray, EMSA, reporter assay, ChIP	[70]
IFI16	Human	Microarray, reporter assay, ChIP	[78]
VEGF	Human	EMSA, reporter assay, ChIP	[136,137]
EPO	Human	EMSA, reporter assay, ChIP	[138]
NTRK2 (TrkB)	Human	EMSA, reporter assay, ChIP	[74]
IGF1R	Human, Chinese hamster	EMSA, reporter assay, ChIP	[54,65,107,139–141]
Spry1 (sprouty 1)	Mouse	Reporter assay, ChIP, EMSA	[73]
Amh (MIS)	Mouse	ChIP, gel shift	[75]
POU4F2	Human	Reporter assay	[142]
Sf1	Mouse	DNase I footprinting, EMSA, reporter assay	[143]
CCNE	Human, mouse	Reporter assay, EMSA, ChIP	[131,144]
Ccng1	Mouse	Microarray	[145]
Igfbp4	Mouse	Expression microarray	[145]
VDR	Human	Reporter assay, EMSA, ChIP	[127,131,145,146]
IGF1	Human	Reporter assay	[140]
AREG	Human, mouse	DNase I footprint, reporter assay, EMSA, qRT-PCR, ChIP, microarray	[68,76,127,128,131]
EREG	Human, mouse	Microarray, qRT-PCR, reporter assay, ChIP	[76]
Amh (MIS)	Human	Reporter assay, EMSA	[117]
HBEGF	Human, mouse	Microarray, reporter assay, ChIP	[76]
IL11	Human, mouse	Microarray, reporter assay	[76]
CX3CL1	Human, mouse	Microarray, ChIP, reporter assay	[76]
IGFII	Mouse	DNase I footprinting, gel mobility-shift assay, reporter assay	[66,88,147–149]
INSR	Human	Reporter assay, DNase I footprinting, gel mobility-shift assay	[150–152]
MDK	Human	Gel mobility-shift assay, DNase I footprint, reporter assay	[153]
Sdc1	Mouse	Gel mobility-shift assay, DNase I footprint, reporter assay	[58]
NOV (NovH)	Human	RNase protection assay, primer extension, reporter assay	[67]
EGFR	Human	EMSA, reporter assay, DNase I footprinting	[64]
RARA	Human	Gel mobility-shift assay, reporter assay	[63]
LOC100766733 (Inhibin-α)	Chinese hamster	Reporter assay, RNase protection assay	[154]
TGFB	Monkey	EMSA, reporter assay	[62]
ETS1	Human	EMSA, reporter assay, ChIP	[32,127,131]
PDGFA	Human, mouse	Gel mobility-shift assay, reporter assay	[53,61,155,156]
RBBP7 (RbAp46)	Human	Suppression subtractive hybridization PCR	[157]
SLC6A6 (TauT)	Human	Reporter assay, EMSA	[158]
WT1	Human	Reporter assay	[159,160–162]
CORO1B	Mouse	ChIP, reporter assay	[163]
Snai2 (Slug)	Mouse	ChIP, reporter assay	[164]
Inha	Mouse	ChIP, reporter assay	[165]
(b) Differentiation
Target gene	Organism(s)	Method of validation	Reference(s)
Egr3	Mouse	ChIP-chip, ChIP	[10]
Nab2	Mouse	ChIP-chip, ChIP	[10]
Sox11	Mouse	ChIP-chip, ChIP, organ culture morpholino	[10]
Scx	Mouse	ChIP-chip, ChIP, organ culture morpholino	[10]
SNAI2 (SLUG)	Human, mouse	Microarray, ChIP, reporter assay	[76]
Sema6d	Mouse	ChIP-chip, ChIP	[10]
Rps6ka3 (Rsk2)	Mouse	ChIP-chip, organ culture morpholino	[10]
Snai2	Mouse	ChIP, reporter assay	[166]
CSF1	Human	Reporter assay	[167]
Scel	Mouse	ChIP	[168]
Sulf1	Mouse	ChIP	[168]
HOXA10	Human	EMSA, reporter assay	[169]
Nphs1	Mouse, human	Reporter assay, DNase footprinting, EMSA	[81,82]
TBXA2R	Human	EMSA, reporter assay, ChlP	[170]
EpoR	Mouse, human	EMSA, ChIP, reporter assay	[171]
SRY	Human	Reporter assay, gel mobility-shift assay	[71]
DAX1	Monkey, human	Footprinting analysis, reporter assay, mobility-shift assay	[172]
Podxl	Rat, mouse	EMSA, reporter assay, ChIP	[54,69]
p21	Human	Expression analysis	[173]
c-Myb	Human	Footprinting, EMSA, reporter assay	[174]
Gαi–2	Pig	Reporter Assay, gel mobility-shift assay	[175]
(c) Cytoskeleton organization and cell adhesion
Target gene	Organism(s)	Method of validation	Reference(s)
Col4a1	Mouse	Reporter assay	[176]
Cik4a2	Mouse	Reporter assay	[176]
Actn1	Mouse	ChIP-chip, ChIP	[10]
Zyx	Mouse	ChIP-chip, ChIP	[10]
Lsp1	Mouse	ChIP-chip, organ culture morpholino	[10]
Myo1B	Mouse	ChIP-chip	[10]
CDH1	Human, mouse	ChIP, reporter assay, EMSA	[166,177–179]
ITGA4	Human	ChIP, EMSA, reporter assay	[180]
NES	Human	Reporter assay	[181]
Adamts16	Mouse	EMSA, ChIP, reporter assay	[182]
(d) WNT signalling
Target gene	Organism(s)	Method of validation	Reference(s)
DKK2	Human	ChIP-ChIP, microarray	[79]
DACT1	Human	ChIP-ChIP, microarray	[79]
TBL1X	Human	ChIP-ChIP, microarray	[79]
CCND2	Human	ChIP-ChIP, microarray	[79]
LEF1	Human	ChIP-ChIP, microarray, ChIP, qRT-PCR	[6,10,79]
JUN	Human	ChIP-ChIP, microarray	[79]
NLK	Human	ChIP-ChIP, microarray	[79]
BTRC	Human	ChIP-ChIP, microarray	[79]
Wnt4	Mouse, human	Reporter assay, ChIP	[72,129]
(e) MAPK signalling
Target gene	Organism(s)	Method of validation	Reference(s)
DUSP16	Human	ChIP-ChIP, microarray	[79]
DDIT3	Human	ChIP-ChIP, microarray	[79]
MAPKAPK2	Human	ChIP-ChIP, microarray	[79]
DUSP6	Human	ChIP-ChIP, microarray, reporter assay, ChIP	[79,183]
RRAS2	Human	ChIP-ChIP, microarray	[10,79]
DUSP5	Human	ChIP-ChIP, microarray	[79]
ERK1	Human	ChIP-chip	[10]
(f) Apoptosis
Target gene	Organism(s)	Method of validation	Reference(s)
A1/BFL1	Human	Reporter assay, ChIP	[86]
BCL2A1	Human	Reporter assay, EMSA, RNase protection assay	[92]
BCL2	Human	Reporter assay, EMSA, RNase protection assay	[85,92]
c-MYC	Human, mouse	Reporter assay, microarray, EMSA, ChIP	[54,128,184–186]
n-MYC	Human	Reporter assay, EMSA	[187]
JunB	Human, mouse	ChIP, microarray	[37,76,131]
BAK	Human, mouse	ChIP, EMSA, reporter assay, RNase protection assay	[54,92]
(g) Epigenetic regulation
Target gene	Organism(s)	Method of validation	Reference(s)
DNMT3a	Human	ChIP, reporter assay	[93]
Rest	Mouse	ChIP-chip, ChIP	[10]
Jmjd1a	Mouse	ChIP-chip	[10]
Jmdj3	Mouse	ChIP-chip	[10]
Usp16	Mouse	ChIP-chip	[10]
Myst2 (HBo1)	Mouse	ChIP-chip	[10]
SRPK1	Human	ChIP, reporter assay	[33]
(h) Others
Target gene	Organism(s)	Method of validation	Reference(s)
STIM1	Human	ChIP	[188]
REN	Human	Gel shift, reporter assay, ChIP	[127,189]
Nos2	Rat	Reporter assay	[190]
MMP9	Human	ChIP	[191]
TERT	Human	EMSA, reporter assay	[192]
MDR1	Human	EMSA, reporter assay	[193]
TSP1	Human	Reporter assay	[194]
ODC	Human, mouse	Reporter assay, DNase 1 protection and methylation interference assays	[195]
SLC20A1	Human, mouse	Microarray, ChIP reporter assay	[76]
IL1RAP	Human, mouse	Microarray, ChIP reporter assay	[76]
CD95L	Human	ChIP, reporter assay	[196]

Many of the candidate target genes underscore the growth-regulatory role of WT1, which guides WT1 in development. For many of the target genes identified soon after the discovery of WT1, the transcriptional effects were determined with transient transfection of reporter assays and therefore, in many cases, it is still not clear if they are bona fide WT1 target genes in vivo. Examples include the developmental regulatory gene PAX2 (paired box 2) that is repressed by WT1 during normal kidney development [57] and the syndecan-1 protein, which also reflects the expression of WT1 during kidney development [58]. Syndecan-1 has been shown to have a role in MET in nephrogenesis [58]. Other examples include many growth factors and their receptors such as IGF1R (IGF1 receptor) [59], PDGF-A (platelet-derived growth factor α) [60,61] TGFβ (transforming growth factor β) [62], Rar-α (retinoic acid receptor-α) [63], EGFR [EGF (epidermal growth factor receptor)] [64], insulin receptor [65], IGF2 [66] and the NovH (nephroblastoma overexpressed gene) gene [67]. For some of these genes, further validation using techniques such as ChIP was later performed. The emergence of DNA microarrays provided a new tool to determine biologically relevant target genes of WT1. An early study used microarrays to show that WT1 transcriptionally activates the amphiregulin promoter [68]. Amphiregulin mirrors the pattern of WT1 protein expression in the developing kidney and can induce tubule branching in kidney organ explants. Using the same approach, Podxl was identified as a target gene [69]. Podocalyxin is expressed in the differentiated and mature podocyte cells of the kidney and is a major structural membrane protein of podocytes. DNA microarrays were also used to show that WT1 is an activator of Amhr2 (anti-Mullerian hormone type 2 receptor) [70]. That study generated new insights into the function of WT1 in early gonad development and sex determination. WT1 might mediate this function through the regulation of the sex-determining gene SRY. WT1 activates the SRY gene and initiates a regulatory gene cascade involved in the male sex determination and differentiation pathway [71].

WT1 target genes that explain its function in development and growth have also been validated by ChIP. Examples include WNT4 (wingless-type MMTV integration site family 4), a stimulator of renal development [72], SPRY1 (sprouty homologue 1) a regulator of receptor tyrosine kinase signalling and kidney development [73], TrkB [tropomyosin receptor kinase B; also known as NTRK2 (neurotrophic tyrosine kinase, receptor, type 2)] a growth factor receptor important for coronary vasculature development [74], ETS1 a stimulator of tumour vascularization via regulation of endothelial cell proliferation and migration [32], and MIS a positive regulator of Mullerian duct regression [75].

Since the detailed mechanisms by which WT1 functions in organogenesis in different tissues remains unclear, significant attention has been directed at identifying and further verifying target genes. New emerging approaches such as genome-wide expression profiling analysis in cells expressing inducible WT1 have identified novel target genes that may be important in organogenesis. One such study identified a number of direct WT1 target genes including the EGF family ligands epiregulin and HBEGF (heparin-binding EGF-like growth factor), the chemokine CX3CL1 [chemokine (C-X3-C motif) ligand 1], and the transcriptional factors SLUG [also known as SNAI2 (snail family zinc finger 2)] and JUNB (jun B proto-oncogene) [76]. The genes were validated further by ChIP and luciferase reporter assays. Additionally, the authors used an in vitro embryonic kidney culture system to demonstrate the physiological significance of the novel target genes. Overexpression of the WT1-regulated growth factors epiregulin, amphiregulin, CX3CL1 and the cytokine IL-11 (interleukin-11) significantly increased ureteric bud branching morphogenesis [76]. Although microarray analysis had been performed before [68], the authors improved the sharpness of the data profile by using tetracycline repressible expression profiles at multiple induction times which significantly reduced ‘noise’ and eliminated stochastic changes [76]. Expression profiling analysis in search for WT1 target genes has also been performed by comparing the induction of WT1 with the repression of WT1 [77]. To do so, the authors have created tetracycline-inducible WT1 (−KTS) isoform and WT1 (+ 17AA, + KTS) isoform in a developmentally relevant cell line HEK (human embryonic kidney)-293, which does not express endogenous WT1, and performed a microarray-based expression profile. They also generated a stable transfectant of the mouse mesonephric cell line M15 that endogenously expresses all WT1 isoforms, to express a full-length antisense WT1 construct that abolished the production of all four WT1 isoforms. That study demonstrated that by examining the complementary overlap gene expression changes caused by WT1 induction and repression, the most pronounced change was seen in genes involved in the mevalonate pathway of cholesterol biosynthesis [77]. Since, the effect of WT1 on its target genes is context-dependent, cell line-dependent and isoform-dependent, the study aimed to delineate physiological relevant targets of WT1 by the induction and repression of all WT1 isoforms. ChIP can be used to further validate that genes involved in cholesterol biosynthesis are direct targets of WT1.

Genome-wide analyses have also been performed to find WT1 target genes involved in Wilms’ tumours. A study compared the microarray profiles between WT1 induction in the Saos-2 osteosarcoma cell line and Wilms’ tumours and identified IFI16 (interferon-inducible 16), a transcriptional modulator, as a WT1 target gene [78]. IFI16 is overexpressed in WT1-replete tumours and loss of IFI16 and WT1 inhibits Wilms’ tumour growth. Both genes have different expression profiles in healthy kidneys, but are aberrantly co-expressed in Wilms’ tumours. That study reveals a pathogenic link between WT1 and IFI16 and identified IFI16 as a target gene that supports cell growth.

A previous study also identified WT1 target genes using Wilms’ tumour cell lines. They manipulated WT1 levels in a Wilms’ tumour-derived CCG99-11 cell line by conditional overexpression and shRNA-mediated knockdown [79]. Target genes regulated by WT1 were identified using gene expression profiling in combination with ChIP followed by microarray analysis (ChIP-chip). The microarray was performed in both GOF (gain of function) andLOF (loss of function) systems of WT1. To prioritize WT1 target genes, they compared genes differentially regulated by WT1 in both systems with genes identified in ChIP-chip arrays. Genes directly regulated by WT1 were functionally grouped into MAPK signalling, axon guidance and Wnt pathways, (see Table 1) [79]. As nine target genes were involved in Wnt signalling the authors concluded that one essential role of WT1 in development and/or disease is to regulate the Wnt signalling pathway. In addition, it is significant that the authors used both GOF and LOF of WT1 to identify direct targets. The loss of function of endogenous WT1 leads to loss of all WT1 isoforms and identifies those targets bound by all isoforms, whereas only the WT1 — KTS isoform was induced in the GOF [79]. For the latter, the study could not identify genes that are directly regulated by any of the other WT1 isoforms. Nevertheless, the comparison of genes differentially regulated by WT1 in either GOF or LOF genes with genes identified in ChIP-chip arrays led to identification of new direct targets. Specifically, WT1 regulates the Wnt pathway genes and this might partly explain its function in development.

Many of the efforts to identify WT1 target genes have been dependent on the use of immortalized cell lines and for some identified target genes a biological function in vivo has not been clarified. A previous study sought to identify bona fide WT1 target genes involved in renal development by using a genome-wide ChIP-chip approach using chromatin obtained from embryonic kidneys [9]. Many of the WT1 target genes identified were further validated by direct ChIP assays also using chromatin from embryonic kidneys [10] (see Table 1). Moreover, the target genes were further validated biologically in a novel modified WT1 morpholino LOF model in embryonic kidney explants. Therefore, to gain insight into the role of WT1 during kidney development, the authors used a novel modified antisense ‘vivo-morpholino’ delivery system to examine the effects of WT1 knockdown in kidney explants. Essential kidney development genes previously identified previously such as BMP7 (bone morphogenetic protein 7), PAX2 and SALL1 (spalt-like transcription factor 1) were found in that study. In addition, many novel WT1 target genes, such as LSP1 (lymphocyte-specific protein 1), PBX2 (pre-B-cell leukaemia homeobox 2), PLXDC2 (plexin domain-containing 2), SCX (scleraxis homologue), SOX11 and RPS6KA3 (ribosomal protein S6 kinase, 90 kDa, polypeptide 3), involved in nephron progenitor differentiation in vivo were identified. The authors conclude that this set of target genes clarifies the renal agenesis caused by the absence of WT1. Interestingly, WT1 target genes were also implicated in actin cytoskeleton organization, actin biogenesis, cell–cell signalling, cell adhesion, TGFβ/BMP signalling genes, Wnt signalling and MAPK pathway genes. In summation, by applying ChIP-chip to biological approaches, the study has identified many genes that are co-expressed with WT1 in nephron progenitor cells and explain the critical role of WT1 in nephron progenitor during kidney development in vivo [10].

WT1 target genes: differentiation

WT1 plays a central role in the control of differentiation. In the genome-wide analysis performed by Kreidberg and co-workers, the WT1 target genes identified included several involved in actin cytoskeleton organization and biogenesis, cell adhesion, and cell–cell signalling [10]. Specifically, MYO1B (myosin IB), LSP1 and ACTN1 (actinin α1) were found to be WT1 target genes expressed during kidney development and involved in cytoskeletal interactions. Previously, two transcriptome profiling reports during pre-implantation of mouse development found a requirement for the expression of genes involved in cytoskeletal interactions before major differentiation events [10]. Elevated expression of actin cytoskeleton genes and cell adhesion might be needed to facilitate the morphological changes that enable blastomeres to undergo differentiation [10]. Thus similar genes might be obligatory to cause the dramatic morphological changes associated with nephron progenitor differentiation around ureteric bud tips. In summation, WT1 function in differentiation can be explained, at least in part, by its role in activating genes involved in cytoskeletal interactions and cell–cell adhesion.

The ChIP-chip analysis in mouse embryonic kidney also identified WT1 novel target genes that might mediate WT1 function during nephron progenitor differentiation in vivo [10]. They include RPS6KA3, which has a known function in neurite differentiation along with SOX11, NAB2 (NGFI-A-binding protein 2) involved in Schwann cell differentiation, EGR3 (early growth response factor 3) involved in cellular growth and differentiation, SEMA6D (sema domain, transmembrane domain and cytoplasmic domain 6D) involved in myocardial patterning, SCX involved in heart valve and Sertoli cell differentiation, and CXXC5 (CXXC finger protein 5) which has a dual role in kidney development [80] and myelopoietic progenitor differentiation [10]. These target genes were identified as WT1 target genes in embryonic kidney tissues. Specifically, WT1 target genes, SOX11, SCX, RPS6KA3 and CXXC5, were not only co-expressed with WT1 in nephron progenitors in kidney tissues, but WT1 vivo-morpholino treatment of embryonic kidney explants resulted in reduced expression of these target genes [10]. WT1 morphants exhibited specific reductions in the expression of the target genes above in nephron progenitors or metanephric mesenchyme. Thus this genome-wide study identified novel kidney genes that explain WT1 function in nephron progenitors in vivo.

Other examples of candidate genes that mediate the function of WT1 in differentiation are Podxl and Nphs1 [69,81,82]. WT1 regulates the expression of the major glomerular podocyte membrane protein podocalyxin. Inducible expression of WT1 leads to induction of podocalyxin and supports a role for WT1 in activation of the glomerular differentiation programme [69]. Nphs1 is another WT1 target gene and crucial component of podocytes that is required for glomerular filtration [81,82].

WT1 target genes: apoptosis

Several genes that regulate apoptosis such as Bak (BCL2-antagonist/killer), Bcl2 (B-cell CLL/lymphoma 2), Bcl2A1, c-Myc and JunB are WT1 target genes [83,84]. High levels of Bc1-2 have been reported in sporadic Wilms’ tumours that express elevated wild-type + 17AA isoform [85]. Furthermore, induction of WT1 in murine myeloblast cells leads to up-regulation of another Bcl2 family member A1/BFL1 [86]. The anti-apoptotic gene A1/BFL1 is highly expressed in high-risk leukaemias along with WT1. WT1 also mediates an anti-apoptotic role through the transcriptional control of cell-surface receptors such as EGFR or IGF1R, that mediate survival signals [64,87,88]. Thus, in at least some cellular contexts, WT1 elicits oncogenic potential through inhibition of apoptosis.

The anti-apoptotic function of WT1 is also evident in development. WT1-null mice display an increase in apoptosis of metanephric mesenchyme, which leads to complete agenesis of the kidneys [1]. Furthermore, a study that used organ explants of the developing kidney to monitor the effects of WT1 ablation by siRNA, demonstrated that WT1 loss causes the metanephric mesenchyme to undergo apoptosis [9]. In addition, inhibition of WT1 expression by RNA interference in leukaemic cell lines also leads to apoptosis [89]. Underscoring an anti-apoptotic role for WT1, it has been found that expression of WT1 inhibits p53-mediated apoptosis through binding, stabilization and inactivation of p53 function [90].

Pro-apoptotic effects are observed when WT1 is overexpressed in cell lines such as U2OS, Hep2B and Saos-2 [64,91]. Osteosarcoma cell lines expressing WT1 under an inducible tetracycline-regulated promoter mediate apoptosis through reduced synthesis of the EGFR [64]. Also, WT1 can negatively regulate growth factor receptors such as EGFR and the insulin receptor, repressing survival signals and promoting death [91]. WT1 can also directly induce apoptosis through the regulation of the pro-apoptotic gene BAK [92]. That study demonstrated that expression of wild-type WT1 blocked cellular proliferation and DNA synthesis causing a rapid induction of apoptosis. The physiological conditions under which WT1 promotes apoptosis are not yet clear, but it is very probable that the tumour suppressor functions of WT1 include the induction of apoptosis pathways. The new mouse models of Wilms’ tumour might prove valuable in shedding further light on this aspect of WT1 function in vivo.

WT1 target genes: epigenetic regulation

Recent studies have identified the products of WT1 target genes as being involved in the epigenetic regulation of gene expression. A recent study by Malik and co-workers demonstrated that WT1 transcriptionally regulates DNMT3A (DNA methyltransferase 3A) and WT1 status influences genome-wide promoter methylation [93]. Overexpression of WT1 in Wilms’ tumour cell lines leads to an increase in DNA methylation at several promoters and thus silencing of gene expression. DNMT3A has diverse and complex functions in facilitating transcription of genes by non-promoter CpG island methylation and by demethylation of DNA. In addition, a study by Kreidberg and co-workers using a genome-wide ChIP approach identified several WT1 target genes encoding proteins that modify chromatin [10]. Examples include Jmjd1a [KDM3A (lysine-specific demethylase 3A)], Jmdj3, REST (RE1-silencing transcription factor), Usp16 (ubiquitinspecific peptidase 16; also known Ubp-M) and Myst2 (histone acetyltransferase 2) [94]. Further study of how WT1 regulates these genes, and the interplay with direct interactions between WT1 and chromatin and DNA modification complexes, will provide insights into the epigenetic effects of WT1.

ROLE OF WT1 AS A POST-TRANSCRIPTIONAL REGULATOR

WT1 interacts with splicing factors

Although the present review focuses on the transcriptional role of WT1, there is a wealth of evidence describing the role of WT1 in RNA metabolism. However, the physiological relevance and function of WT1 as a regulator of post-transcriptional processes is still not clear. It been proposed that WT1 −KTS isoforms function optimally as transcriptional factors, whereas the + KTS isoforms might act post-transcriptionally [95]. WT1 + KTS isoforms were found to localize with Sm antigens which are a major snRNP (small nuclear ribonucleoprotein) constituent, in nuclear speckles [96]. The presence of WT1 in speckles is a characteristic shared by splicing factors [96]. WT1 + KTS was found to interact with the splicing factor U2AF65 and was incorporated into spliceosomes. These early findings provided the first evidence of a potential new role for WT1 in pre-mRNA processing [97]. In addition, WT1 + KTS associates with the splicing factor WTAP (WT1-associated protein) and they both localize in the speckles [98]. WT1 + KTS also interacts with the RNA-binding protein and splicing factor RBM4 (RNA-binding motif protein 4). Moreover, WT1 + KTS is able to inhibit the effect of RBM4 on alternative splicing in vivo [99]. Thus, WT1 may modulate alternative splicing through its interactions with splicing factors.

WT1 binding to mRNA

Early studies by Ward and co-workers revealed that WT1 can bind to specific RNA sequences [100]. Both WT1 and its closely related EGR1 can bind to the same DNA sequences of the Igf2 gene. However, WT1, but not EGR1, can bind to Igf2 exonic RNA sequences through its zinc fingers. Specifically, WT1 zinc finger 1, which is not found in EGR1, is important for RNA binding [100]. That zinc finger 1 is critical for RNA binding was also confirmed in vivo in Xenopus oocytes [101]. RNA sequences bound to WT1 were also identified by SELEX (systematic evolution of ligands by exponential enrichment), an in vitro selection procedure that enriches RNA sequences bound by RNA-binding proteins [102]. Interestingly, WT1 associates with actively translating polysomes [103]. Specifically, the WT1 + KTS isoform, but not WT1 − KTS, increases the levels of an unspliced RNA retaining an intron and promotes the association of this RNA with polyribosomes. This study provides a link between the role for WT1 in RNA metabolism and translation. Further studies are needed to fully understand the latter.

Following on from the RNA-binding studies, a search began for candidate mRNA targets [95]. WT1 + KTS isoform was found to bind ACTN1 mRNA via its zinc finger domain [95]. ACTN1 is the first likely physiological mRNA candidate of WT1. Indeed, a previous study also identified actin as a novel interaction partner of WT1 [104]. That study further determined that perturbation of actin influences the cell dynamic properties leading to movement of WT1 off the polysome fractions and cancelled nucleocytoplasmic shuttling of WT1. Interestingly, actin perturbation also changed the WT1 DNA- and RNA-binding abilities [104]. These findings add to the role of WT1 in RNA metabolism in response to actin cytoskeletal changes. Although these studies support a role for WT1 in RNA metabolism, the physiological relevance of a post-transcriptional role of WT1 remains to be confirmed. Of note, a study determined that WT1 + KTS, but not WT1 − KTS, up-regulated Mash1 [also known as Ascl1 (achaete–scute complex homologue 1)] mRNA and protein post-transcriptionally [5]. The authors conclude that WT1 + KTS, which has a role in mRNA processing, acts upstream of Mash1 to promote the normal development of the olfactory system. Many other mRNAs might be regulated by WT1 + KTS in vivo and future work will hopefully shed more light on the post-transcriptional roles of WT1 in development and disease.

THE INTERACTION PARTNERS OF WT1 DNA-BINDING PROTEINS

Several proteins have been identified as WT1 interaction partners, which we have separated into five categories (Table 2). WT1-interacting partners in the first two categories are involved in transcriptional regulation, whereas members of the last three are involved in post-transcriptional regulation, proteolysis of WT1 and epigenetic regulation (Table 2). The interacting partners of WT1 involved in RNA metabolism and processing are discussed in the section above. As transcriptional regulators, members of the first two categories are the DNA-binding transcriptional factors and cofactors that modulate transcription. One of the early studies identified that WT1-mediated transcriptional activation was inhibited by a dominant-negative effect of WT1 mutants identified in patients with DDS that could not bind DNA [105]. In the same study, it was shown that WT1 was able to homodimerize by using the first 180 amino acids of WT1 [82,105]. Thus WT1 is able to self-associate and the association of mutant with wild-type WT1 may be responsible for the dominant-negative activity of mutant WT1 in disease.

Table 2.

WT1-interacting partners

(a) DNA-binding transcription factors
Protein	WT1-interaction domain	Reference(s)
STAT3	Residues 1–281	[113]
ERα	Not determined	[139]
SREBP	Zinc fingers	[77]
Pax2	Residues 1–242	[115]
SRY	Zinc fingers	[116]
p53	Zinc fingers (1–2)	[106,107,110]
BMZF2	Zinc fingers	[197]
p73	Zinc fingers	[108]
p63	Not determined	[108]
Hsp70	N-terminus	[114]
TFIIB	Activation domain (residues 180–250)	[50]
SF1	N-terminus	[117]
TBP	Residues 245–297	[122]
WT1	N-terminus	[49]
ZNF224	Zinc fingers (−KTS)	[198,199]
ZNF255	Zinc fingers	[197,198]
(b) Transcriptional co-regulators
Protein	WT1-interaction domain	Reference(s)
PINCH-1	Zinc fingers	[118]
WTX	Zinc fingers	[119]
WTIP	Not determined	[200]
IFI16	Zinc fingers	[78]
CITED2	Zinc fingers	[123]
BASP1	Suppression domain (residues 71–101)	[51]
FHL2	Activation domain (residues 182–250)	[201]
CBP	Zinc fingers	[124]
Par-4	Zinc fingers, + 17AA	[121,122]
Ciao1	Zinc fingers	[202]
(c) RNA metabolism and processing factors
Protein	WT1-interaction domain	Reference
Actin	Not determined	[104]
snRNP-U	Zinc fingers	[203]
PSF	Not determined	[95]
RBM4	Zinc fingers (+ KTS)	[99]
ACTN1 mRNA	Zinc fingers (+ KTS)	[95]
WTAP	Zinc fingers	[99]
U2AF65	Zinc fingers (3–4)	[97]
Splicing complexes (U2-B, U1-70K and coilin)	Not determined	[96]
(d) Epigenetic factors
Protein	WT1-interaction domain	Reference
DNMT1	Not determined	[125]
PcG complex	Not determined	[125]
Menin	Not determined	[125]
(e) Proteolytic factors
Protein	WT1-interaction domain	Reference
HtrA2	Suppression domain (residues 71–101)	[37]
Ubc9	N-Terminus	[53]
(f) Others
Protein	WT1-interaction domain	Reference
E1B55K	Zinc fingers	[204]
HCMV-1E2	Zinc fingers	[205]

Other DNA-binding transcription factors that interact with WT1 include p53 and p73, which have been shown to compete with WT1 in transcriptional regulation [106–108]. Specifically, WT1 interacts with p53 through its zinc fingers and through this interaction WT1 stabilizes p53 and enhances p53-mediated apoptosis by ultraviolet radiation (Figure 2A) [109]. Moreover, p53 interaction can switch the role of WT1 from an activator of transcription to a repressor (Figure 2B). The other member of the p53 family of products, p73, also binds to WT1 through the zinc fingers and reduces WT1 transcriptional activity and DNA binding by WT1 (Figure 2D) [109]. Work in mice suggests that p53 and WT1 can co-operate in tumorigenesis [110] and p53 andWT1 have been shown to associate at the Igf1r and Podxl promoters [107,111]. ThusWT1 binding to p53-like proteins links WT1 with the stress response, development and tumorigenesis through p53. A recent study using Wilms’ tumour cell lines revealed that mutant WT1 proteins antagonize transcriptional repression mediated by p53 [112]. Whether this effect required recruitment of WT1 to the promoter region by p53 is not clear, but suggests a gain of function by WT1 mutant proteins that lack a functional DNA-binding domain.

Figure 2. DNA-binding transcription factors that associate with WT1.

Mechanisms of action of complexes formed by DNA-binding transcription factors and WT1 are shown. The cofactors lead to either gene activation (left-hand panels) or repression (right-hand panels) by WT1. The small yellow circle in (I) indicates that the interaction is dependent on the splice isoform of WT1.

In relation to tumorigenesis, WT1 interacts with STAT3 (signal transducer and activator of transcription 3) and promotes cell proliferation [113]. Interestingly, WT1 and STAT3 interact in a variety of primaryWilms’ tumour cells, raising the possibility that activated STAT3 and WT1 may play a role in tumorigenesis. On the other hand, WT1 association with the inducible chaperone Hsp70 (heat-shock protein 70) causes inhibition of cellular proliferation [114]. WT1 and Hsp70 physically interact in rat embryonic kidney cells and co-localization of WT1 and Hsp70 is evident within podocytes of the developing kidney (Figure 2E). Thus, by inhibiting proliferation, the WT1–Hsp70 association might be important during kidney differentiation.

WT1 has been shown to associate with Pax2, which is also essential during kidney development [115]. WT1 modifies the Pax2 phenotype in mice and it has been proposed that this might require a direct interaction between WT1 and Pax2 [115]. In mammalian gonads, WT1 and SRY genes form a complex together to regulate transcription. SRY uses WT1 as a transcriptional co-activator and the WT1–SRY interaction is important in testis development (Figure 2C) [116]. WT1 interacts with SF1 (splicing factor 1), which is also essential for mammalian gonadogenesis, and SF1, like SRY, uses WT1 as a transcriptional co-activator (Figure 2G) [117]. The characterization of WT1-interacting partners provides new insights that lead to the identification of important pathways in growth and differentiation. The DNA-binding factors ofWT1 and their role in transcriptional activation or repression is shown in Figure 2.

Transcriptional co-regulators of WT1

Members of the second category of WT1-binding factors are proteins that regulate the transcriptional activity of WT1 (Figure 3). PINCH1 [particularly interesting new Cys-His protein 1; also known as LIMS1 (LIM and senescent cell antigen-like domains 1)], an adaptor protein, is a transcriptional regulator expressed in podocytes that interacts with WT1 and represses Podxl gene expression [118]. Another study has identified that the tumour suppressor WTX shuttles to the nucleus and interacts with WT1 and modulates its transcriptional activity. Specifically, WTX enhances the WT1-mediated transcriptional activation of amphiregulin (Figure 3G) [119]. WT1 can act as a transcriptional cofactor for other proteins such as WTIP (WT-interacting protein). WTIP is expressed at the cell surface in cell–cell junctions, but shuttles to the nucleus and interacts with WT1 when podocytes are subjected to damaging agents [120]. WTIP interaction with WT1 in the nucleus leads to transcriptional repression (Figure 3L). Thus WTIP acts as a transcriptional co-repressor ofWT1 after podocyte injury. It has been suggested that WTIP might monitor podocyte damage and modulate WT1 transcriptional activity to help in maintaining podocyte integrity [120]. Whether this involves an intermediary role for cytoplasmic WT1 is not known, but it has become clear that several WT1-interaction partners can be found in the cytoplasm and also at the cell periphery. Further biochemical studies of cytoplasmic WT1 will be needed to shed light on the functions of WT1 outside of direct gene regulation and potential roles in transducing signals within the cell.

Figure 3. Transcriptional co-regulators of WT1.

Transcriptional regulation by WT1 and its transcriptional cofactors are shown, leading to the formation of either an activator complex (left-hand panels) or a repressor complex (right-hand panels). The small yellow circle in (C) indicates that the interaction is dependent on the splice isoform of WT1. Histone modifications are indicated (Ac; H3K27me3).

WT1 contains separate regions that mediate transcriptional activation (A) and repression (R) (Figure 1). The transcriptional activation domain is suppressed by a ten-amino-acid region (SD) within the N-terminus of the repression domain of WT1. The SD interacts with a transcriptional co-repressor, BASP1, that converts WT1 from an activator into a repressor (Figure 3B) [51]. Like WT1, BASP1 is expressed during kidney development and in the adult podocyte. The functional role of WT1 and BASP1 in development is yet to be determined, but WT1 and BASP1 are co-expressed in several developing organs and tissues [51]. The mechanisms of action of the WT1–BASP1 complex are discussed below.

Par-4 (prostate apoptosis response-4) was initially identified as a WT1-binding protein in a yeast two-hybrid screen [121]. Par-4 interactswith the zinc fingers ofWT1 resulting in the repression of transcription at WT1 target genes (Figure 3D). Par-4 can engage in an interaction with another region of WT1 containing the 17AA-encoding alternatively spliced exon (Figure 3C) [122]. The 17AA region acts as an autonomous transcriptional activation domain that is dependent on a specific interaction with Par-4. Transcriptional activation by the WT1 + 17AA isoform through engagement with Par-4 is one of the mechanisms by which WT1 can inhibit apoptosis. Thus Par-4 can mediate different effects on WT1 function through interaction with distinct domains (and splice isoforms) of WT1.

Like Par-4, Cited2 [Cbp (CREB-binding protein)/p300-interacting transactivator, with Glu/Asp-rich C-terminal domain 2] can also function as either a co-activator or a co-repressor of WT1 transcriptional activity, depending on the promoter or cellular context (Figures 3E and 3F). Cited2 interacts with WT1 and stimulates SF-1 gene transcription, whereas it represses the WT1-mediated transcriptional activation of the amphiregulin gene [123]. Another cofactor of WT1 is the transcriptional co-activator and histone acetyl transferase CBP. WT1 binds to CBP via its first two zinc fingers and this association leads to synergistic activation of WT1 target genes (Figure 3A) [124]. Thus the function ofWT1 can be transcriptionally modulated by both co-repressors and co-activators.

Epigenetic factors and WT1

A previous study found that WT1 can associate with epigenetic factors such as menin and DNMT1 [125]. Menin is a tumour suppressor that has a critical role in the control of H3K4me3 (histone H3 trimethylated at Lys⁴) and gene transcription. Menin represses Pax2 expression, which is one of the early markers of kidney progenitor cells, via the WT1–PcG (polycomb group) complex. Menin does not bind directly to the Pax2 locus, instead it interacts with WT1 and up-regulates WT1 expression, possibly through post-transcriptional regulation, as WT1 mRNA expression was not affected by menin. WT1 then recruits the PcG complex to the Pax2 gene causing transcriptional repression of Pax2 [122]. Furthermore, WT1 interaction with DNMT1 leads to recruitment of DNMT1 to the Pax2 promoter resulting in hypermethylation of CpG in the Pax2 promoter (Figure 3H). Thus menin and DNMT1 act through WT1 to epigenetically repress genes by recruiting the PcG complex and promoting H3K27 (histone H3 Lys²⁷) methylation [125]. These findings suggest that WT1 controls epigenetic regulation by regulating H3K27 histone modification to repress gene transcription.

Proteolytic factors and WT1

We have reported previously that WT1 is a bona fide substrate of the HtrA2/Omi apoptotic protease [37]. HtrA2 binds to the SD of WT1 and cleaves WT1 at multiple sites (Figure 1). Several studies suggest that WT1 acts as an oncogene in tumours primarily through inhibition of apoptosis [83]. The treatment of tumour cell lines with cytotoxic drugs induces HtrA2 activity, which subsequently degrades WT1. This removes WT1 from the promoters of its target genes and leads to changes in gene expression that enhance apoptosis [37]. Thus the cleavage of WT1 by HtrA2 eliminates key anti-apoptotic functions that are mediated by WT1. These findings provide insights into the function of HtrA2 in the control of the oncogenic activities of WT1. Moreover, HtrA2-mediated cleavage of WT1 could be advantageous to adult cancers where WT1 acts as an oncogene. In contrast, the activities of WT1 as a tumour suppressor would be compromised by the induction of HtrA2. Diverse cytotoxic agents, including many drugs used in cancer treatment, can activate HtrA2. Further studies will be required to determine if chemotherapeutic regimes could be optimized to account for the action of HtrA2 on WT1. For example, drugs that are potent inducers of HtrA2 activity would probably be more effective in the treatment of tumours in which WT1 acts as an oncogene.

Ubc9 was found previously to interact with WT1 [53]. The authors suggested a possible link between the ubiquitin–proteasome proteolytic pathway and WT1. However, it was later identified that the primary function of Ubc9 is concerned with SUMOylation. Indeed, WT1 is directly SUMOylated on Lys⁷³ and Lys¹⁷⁷ (Figure 1), the former of which lies within the HtrA2- and BASP1-binding site [52]. This raises the possibility that SUMOylation of WT1 might regulate either or both the proteolysis by HtrA2 and transcriptional repression through BASP1. However, further studies are needed to determine the role of SUMOylation in the regulation of WT1 function.

TRANSCRIPTIONAL REGULATION BY THE WT1–BASP1 COMPLEX

The BASP1 protein and its role as a regulator of WT1 transcription function

The transcriptional activation domain of WT1 is subject to regulation by an SD in the N-terminus of WT1 (Figure 1). We identified BASP1 as a transcriptional co-repressor that can block the function of the transcriptional activation domain ofWT1 [51]. BASP1 is expressed in several cell types, but was first identified in neuronal cells where it binds to the cytoplasmic face of the cell membrane through an N-terminal myristoylation motif [126]. At first, BASP1 seemed an unusual candidate to play a role in transcription; however, BASP1 contains a functional bipartite nuclear localization sequence and BASP1 is found exclusively in the nucleus of some cells [20,51,127]. WT1 associates with BASP1 in cells that naturally express both proteins. Moreover, genome-wide analysis has shown that expression of BASP1 in K562 cells leads to transcriptional repression of >90% of WT1 target genes, suggesting that it is likely to be a significant regulator of WT1 transcription function [127].

Since BASP1 can be found in both the nucleus and cytoplasm, this suggests that BASP1 subcellular localization might be an important point of regulation [51]. Moreover, this property of BASP1 might help, at least in part, to explain the variability of WT1 function at target genes in different cells and tissues. WT1 and BASP1 are present in large complexes in cell lines suggesting that several other components may participate in transcriptional repression mediated by WT1 [54]. Indeed, we have recently reported that the transcriptional co-repressor prohibitin is a component of the BASP1 complex and is also required for WT1-mediated transcriptional repression [128].

Function of the WT1–BASP1 complex in growth and differentiation

WT1 and BASP1 are temporally and spatially co-expressed at several sites in the embryo and adult, including the kidney. Analysis of the WT1–BASP1 complex in a model cell line of podocyte differentiation (MPC5) showed that WT1 and BASP1 co-localized at the Bak, c-Myc and Podxl gene promoters in undifferentiated MPC5 cells [54]. Upon differentiation, the Podxl gene was up-regulated, whereas the Bak and c-Myc promoters were down-regulated. The same study also found that BASP1 could be SUMOylated and the SUMOylation of BASP1 was a dynamic process during the differentiation of podocytes. How SUMOylation of BASP1 affects its function as a WT1 transcriptional co-repressor is not yet clear, but may involve a change in the subnuclear distribution of BASP1. Upon differentiation-dependent SUMOylation of BASP1, BASP1 remained bound to the Bak and c-Myc promoters, but dissociated from the Podxl promoter. It is not clear, however, why BASP1 is selective with specific gene promoters of WT1 involved in the differentiation of the podocytes. Interestingly, WT1 and BASP1 can together divert the differentiation of K562 cells from the blood cell lineage to cells with a neuronal-like morphology that express specific neuronal genes and respond to the neurotransmitter ATP [127]. It is tempting to speculate that the WT1–BASP1 complex might co-operate in the formation of neuronal cell types where BASP1 is also highly expressed. This possibility could be explored further by analysis of WT1 and BASP1 in the development of sensory systems that have been shown to be WT1-dependent.

The mechanism of BASP1-mediated transcriptional repression by WT1

A previous study has associated BASP1 with localized histone the modification at the Wnt4 promoter [129]. Although WT1 activates the Wnt4 gene in the kidney, it represses Wnt4 in the epicardium. WT1 activates Wnt4 in the kidney by recruitment of the co-activator CBP, whereas it represses Wnt4 in the epicardium by recruitment of BASP1. The Wnt4 gene locus is flanked by CTCF (CCCTC-binding factor) sites. In the epicardium, the domain between CTCF sites is in a repressive chromatin state through BASP1 recruitment and WT1 depletion causes a flip to an activation state [130]. Thus WT1 represses Wnt4 in a BASP1-dependent chromatin flip–flop.

How does BASP1 mediate transcription repression by WT1? A surprising finding was that BASP1 is required to be myristoylated at its N-terminus to act as a transcriptional co-repressor [131]. The N-terminal myristoylation of BASP1 facilitates its interaction with PIP₂ (phosphatidylinositol 4,5-bisphosphate) [126,132–134]. Indeed, chromatin-bound BASP1 contains PIP₂ and this is required for the recruitment of HDAC1 (histone deacetylase 1) to the promoter regions of WT1 target genes to mediate transcriptional repression (Figure 3B) [131]. Moreover, we have recently reported that the transcriptional repressor prohibitin forms an integral component of the WT1–BASP1 complex and is required to recruit ATP-dependent chromatin remodelling activities to the promoters [128]. Prohibitin and BASP1 cooperate through PIP2 to repress transcription by recruitment BRG1 (brahma-related gene 1) and HDAC1 to WT1-dependent promoters [128]. How these chromatin remodelling complexes interface with the WT1–BASP1 complex, and the crucial role that phospholipid plays in mediating these interactions, is not yet clear.

CONCLUDING REMARKS

WT1 functions as a transcriptional regulator that represses or activates the expression of numerous target genes. In the past two decades, ChIP experiments and genome-wide analysis, such as microarray and ChIP-chip experiments, have provided a large wealth of information concerning the genes that WT1 regulates. However, many target genes still have to be validated as biological targets ofWT1 and which cellular pathways they regulate in vivo. Elucidating the biological role of the target genes of WT1 using organ culture morpholino systems, gene targeting and transgenic mice will help to further understand the many functions ascribed to WT1 in development and disease.

In addition, much work over the last 20 years has led to a better understanding of WT1 interaction partners and how they mediate the context-specific activities of WT1. Identification of WT1-interacting proteins is an ongoing task and proteomic screens are likely to yield new insights in the coming years. Further studies of epigenetic regulation by WT1 and how this is modulated by its interaction partners will lead to a better understanding of how WT1 establishes gene transcription profiles that regulate developmental and disease processes. Although not the focus of the present review, it is probable that in the next few years we will also learn more about the function of WT1 as an RNA-binding protein. Going forward, once the downstream mediators of WT1 and its cofactors have been characterized by biochemical and cellular approaches followed by the physiological relevance in vivo, the sometimes-confusing functions of WT1 will be resolved and its functions in development and disease will be better understood.

Abbreviations:

ACTN1

actinin α1

Amhr2

anti-Mullerian hormone receptor 2

Bak

BCL2-antagonist/killer

BASP1

brain acid-soluble protein 1

BCL2

B-cell CLL/lymphoma 2

BMP

bone morphogenetic protein

CBP

CREB-binding protein

Cited2

Cbp/p300-interacting transactivator, with Glu/Asp-rich carboxy-terminal domain 2

CTCF

CCCTC-binding factor

CX3CL1

chemokine (C-X3-C motif) ligand 1

CXXC5

CXXC finger protein 5

DDS

Denys–Drash syndrome

DNMT

DNA methyltransferase

EGF

epidermal growth factor

EGFR

EGF receptor

EGR

early growth response factor

GOF

gain of function

HDAC1

histone deacetylase 1

H3K27

histone H3 lysine 27

Hsp70

heat-shock protein 70

HtrA2

HtrA serine peptidase 2

IFI16

interferon-inducible 16

IGF

insulin-like growth factor

IGF1R

IGF1 receptor

JUNB

jun B proto-oncogene

LOF

loss of function

LSP1

lymphocyte-specific protein 1

MAPK

mitogen-activated protein kinase

MET

mesenchyme–epithelial transition

MIS

Mullerian-inhibiting substance

Par-4

prostate apoptosis response-4

PAX2

paired box 2

PcG

polycomb group

PIP₂

phosphatidylinositol 4,5-bisphosphate

RBM4

RNA-binding motif protein 4

RPS6KA3

ribosomal protein S6 kinase, 90 kDa, polypeptide 3

SCX

scleraxis homologue

SD

suppression domain

SF1

splicing factor 1

SOX

SRY box

SRY

sex-determining region Y

STAT3

signal transducer and activator of transcription 3

SUMO

small ubiquitin-related modifier

TGFβ

transforming growth factor β

Ubc9

ubiquitin-conjugating enzyme 9

WNT

wingless-type MMTV integration site family

WT1

Wilms’ tumour 1

WTIP

WT-interacting protein

WTX

Wilms’ tumour on the X