BIOLOGICAL OVERVIEW

Steroid hormones are key regulators of numerous physiological and developmental processes, including metastasis of breast and ovarian cancer. The Drosophila gene taiman encodes a steroid hormone receptor coactivator related to AIB1. Mutations in tai cause defects in the migration of specific follicle cells (the border cells) in the Drosophila ovary. Mutant cells exhibit abnormal accumulation of E-cadherin, ß-catenin, and focal adhesion kinase. Tai protein colocalizes with the ecdysone receptor in vivo and augments transcriptional activation by the ecdysone receptor in cultured cells. The finding of this type of coactivator required for cell motility suggests a novel role for steroid hormones, in stimulating invasive cell behavior, independent of effects on proliferation (Bai, 2000).

A small group of follicle cells in the Drosophila ovary, the border cells, has been studied as a model system for a forward genetic approach to the study of cell motility. The border cells originate within an epithelium of approximately 1100 follicle cells that surrounds a cluster of 16 germline cells to form an egg chamber. Early in oogenesis, a pair of specialized follicle cells, called polar cells, differentiates at each end of the egg chamber. The anterior polar cells recruit an additional four to eight cells, and this cluster then detaches from the follicle cell epithelium and invades the neighboring group of fifteen nurse cells (Bai, 2000).

A number of genes are known to be required to convert the border cells from stationary, epithelial cells to invasive, migratory cells. The first gene found to be required for border cell migration was slow border cells: slbo encodes a basic region/leucine zipper transcription factor related to the mammalian CCAAT enhancer binding protein (C/EBP) family. It has been proposed that expression of C/EBP is one factor controlling the timing of border cell migration during oogenesis (Bai, 2000).

A screen was carried out for mutations on the left arm of the second chromosome (2L) that cause border cell migration defects in mosaic clones. Border cell position was monitored using beta-galactosidase expression from an enhancer trap line, PZ6356. Of 2885 mutant lines screened for migration defects, one mutant named taiman^61G1 (tai^61G1, meaning 'too slow') was selected for further study. In egg chambers containing tai^61G1 mosaic clones, PZ6356 expression is unaffected. In some egg chambers containing tai^61G1 clones, migration is completely inhibited, and border cells remain at the anterior tip of the egg chamber. In other egg chambers, the border cells migrate partway. Border cell clusters that undergo partial migration are typically composed of a mixture of heterozygous and homozygous mutant cells (Bai, 2000).

Border cells mutant for tai express wild-type levels of the Slbo protein, indicating that the tai mutant phenotype is not due to reduced expression of Slbo. For example, in a border cell cluster composed of a mixture of wild-type cells and cells homozygous mutant for tai, Slbo protein is expressed similarly in all of the cells (Bai, 2000).

A second protein known to be required for border cell migration is Drosophila E-cadherin (Shotgun). To determine whether the tai migration defect might be due to reduction in Shotgun expression, egg chambers containing tai mutant clones were stained with antibodies against Shotgun. In all wild-type stages examined, Shotgun accumulates in the central, nonmigratory polar cells, as well as in the junctions between individual border cells. Shotgun colocalizes with cortical F-actin in these locations. Prior to migration, when the border cells are still part of the follicular epithelium, Shotgun also accumulates at the junctions between border cells and nurse cells. However, once the border cells leave the follicular epithelium and invade the neighboring germline cell cluster, much less Shotgun staining is evident at the junctions between the nurse cells and border cells, relative to the level between border cells or in the polar cells. When migration is complete, Shotgun accumulates again in the junctions between the border cells and the oocyte (Bai, 2000).

In tai mutant clusters, Shotgun staining is abnormally elevated at the border cell/nurse cell junctions. In contrast, in slbo mutants, Shotgun expression fails to rise at the time of migration and Shotgun immunoreactivity is only detected at high levels within the polar cells. Armadillo (Arm) colocalizes with Shotgun in wild-type and mutant border cells. The abnormal accumulation of Shotgun and Arm in tai mutants does not appear to result from increased transcription of Shotgun because overexpression of Shotgun in border cells causes neither a migration defect nor specific accumulation of cadherin staining at the border cell/nurse cell junctions. Nor does the abnormal accumulation of Shotgun and Arm appear to be simply a consequence of the migration failure. In addition to slbo, Shotgun and Arm expression were examined in border cells that fail to migrate due to mutations in the jing locus: no defect in either expression or localization of adhesion complexes was observed. Nor are defects in either Shotgun or Arm expression or localization found in border cells that fail to migrate due to expression of dominant-negative Rac (Bai, 2000).

The accumulation of Shotgun at the border cell/nurse cell boundary suggests that the role of tai in border cell migration might be to stimulate turnover of adhesion complexes during migration in order to allow forward movement. One protein believed to play a role in turnover of adhesion complexes is Focal adhesion kinase. Drosophila FAK (Fak56D) is highly enriched in the border cells during their migration, but not in the polar cells (Bai, 2000).

To determine whether Fak56D expression or localization is affected by mutations that disrupt border cell migration, wild-type and slbo mutant egg chambers were stained and the staining was compared to that of egg chambers containing tai mosaic clones. Fak56D expression is significantly reduced in slbo mutant border cells. Furthermore, the level of reduction correlates with the degree of inhibition of migration. That is, in some slbo egg chambers, border cell migration fails completely and the cells remain at the anterior tip. In such egg chambers, Fak56D expression is undetectable. In a minority of slbo mutant chambers, the cells migrate a little. In these egg chambers, Fak56D expression is reduced compared to wild type, but is detectable. In tai mutant border cells, Fak56D expression is present; however, its distribution is altered relative to wild type. Rather than being evenly distributed throughout the cytoplasm, Fak56D appears to accumulate at the would-be leading edge of the cluster. Some border cell clusters that are mutant for tai exhibit partial migration and in these clusters, the abnormal distribution of Fak56D is only slightly affected such that little Fak56D accumulation can be detected at the most posterior position within the cluster. Thus, the severity of the migration defect in tai mutants correlates with the severity of the defect in Fak56D localization (Bai, 2000).

The similarity of Taiman to steroid hormone receptor coactivators suggests that Tai might interact with one or more steroid hormone receptors. The only known steroid hormone in Drosophila is ecdysone, and the ovary is a major site of ecdysone synthesis, which peaks at stage 9. The functional ecdysone receptor is a heterodimer composed of Ultraspiracle (Usp), which is the fly retinoid X receptor (RXR) homolog, and the Ecdysone receptor. To determine whether the ecdysone receptor complex would be a good candidate for interaction with Tai, expression of ecdysone receptor subunits in egg chambers was examined using antibodies against Usp, EcR-A, and EcR-B. EcR-A and EcR-B are distinct isoforms of the EcR subunit, which are generated by alternative splicing. Usp, EcR-A, and EcR-B colocalize with Tai protein in migrating border cells; Usp and EcR-A are expressed generally, in both follicle cells and nurse cells (Bai, 2000).

These observations raise the possibility that the timing of border cell migration might be controlled by ecdysone. To test whether border cell migration is responsive to hormone, the effects of injecting hormone into female flies were examined. It was not expected that increasing the hormone concentration alone would be sufficient to cause precocious border cell migration because expression of the slbo gene and its targets are independently required for migration. Therefore, slbo was precociously expressed using transgenic flies carrying a heat-inducible slbo transgene, followed by injection of hormone. Border cell migration was assayed in stage 8 egg chambers dissected from flies treated with heat shock and hormone, and compared to control flies treated with heat shock and ethanol, or with hormone in the absence of heat shock. Precocious border cell migration was observed in 20% of egg chambers that were treated with both heat shock and hormone but not in controls. The observed effects are consistent with a role for ecdysone in regulating the timing of border cell migration (Bai, 2000).

If the rising ecdysone level at stage 9 is required to stimulate border cell migration, then reducing the ecdysone level should cause a delay in border cell migration. The ecdysoneless mutant ecd¹ is temperature sensitive for production of ecdysone. Females homozygous for ecd¹ are sterile when held at the nonpermissive temperature for 5 days, and egg chambers in these flies arrest development at stage 8 and subsequently degenerate. Border cells fail to develop in these arrested egg chambers. However, when ecd¹ mutants are held at the nonpermissive temperature for 2 days, some stage 10 egg chambers develop, in which border cells differentiate and express Slbo protein. Greater than 50% of these egg chambers exhibit delayed border cell migration (Bai, 2000).

Since the effects on border cell migration of increasing or decreasing ecdysone levels could have been indirect, whether there is a cell autonomous requirement for the ecdysone receptor in border cells was tested. The EcR locus is proximal to available FRT insertion sites, preventing mosaic analysis. Therefore, the analysis was carried out using mutations in usp. Border cells that were homozygous mutant for a null allele of usp exhibit inhibition of border cell migration, but no obvious defects in other follicle cells (Bai, 2000).

To assess whether Tai and the ecdysone receptor are likely to associate in a complex in vivo, Tai expression was examined in third instar larvae. Antibodies against Tai react specifically with the salivary gland nuclei, as well as other larval tissues. Polytene chromosome spreads were stained with antibodies against Tai and Usp proteins in a double labeling experiment. Anti-Tai antibody labels specific loci on the polytene chromosomes. Moreover, Usp and Tai proteins colocalize precisely. Since previous experiments have shown that Usp and EcR colocalize as a complex on polytene chromosomes, these results indicated that Tai colocalizes with the functional Ecdysone receptor complex at specific target sites (Bai, 2000).

Whether expression of Tai can enhance ecdysone receptor-dependent transcriptional activation in EcR-293 mammalian cells was tested. These cells respond to hormone, either ecdysone or an analog known as ponasterone, with a substantial increase in transcriptional activation of genes placed under the control of a cis-acting sequence known as an E/GRE. Transcriptional activation was tested in cells expressing varying amounts of Tai in transient transfection assays. Tai expression increases transcriptional activation up to 5-fold, in a dose-dependent manner, specifically in the presence of hormone (Bai, 2000).

Furthermore, a GST-fusion protein containing the region of Tai protein containing the LXXLL motifs predicted to interact with EcR (residues 1028 to 1235 of Tai) associates with in vitro translated EcR in a ligand-dependent manner. The same fusion protein does not associate detectably with Usp alone. However, in the presence of EcR and ligand, the Tai-GST fusion protein is able to coprecipitate Usp. Taken together, these results suggest that Tai is a bona fide ecdysone receptor coactivator (Bai, 2000).

Thus, Tai appears to be a coactivator of the p160 class based not only on amino acid sequence similarity and overall domain structure, but based also on its in vivo colocalization with EcR, its direct, ligand-dependent binding to EcR, and its ability to potentiate hormone-dependent transcription in cultured cells. The homology of Tai to SRC proteins suggests that Tai might interact with a steroid hormone receptor. Although there are more than 20 genes in Drosophila that code for proteins related to nuclear hormone receptors, ecdysone is the only known steroid hormone. Since SRC proteins require the presence of a ligand in order to interact with receptors, the ecdysone receptor seems like the best candidate partner for Tai. The colocalization of Tai protein with the ecdysone receptor complex at specific chromosomal loci in third instar larva, the direct and ligand-dependent binding of Tai to EcR in vitro, and the ability of Tai to potentiate the ecdysone response in cell culture lend substantial support to this proposal (Bai, 2000).

The ligand-dependent interaction of Tai with the ecdysone receptor suggests that ecdysone regulates border cell migration. The strongest evidence in support of this is that border cells lacking Usp are unable to migrate. Consistent with this observation, numerous unfertilized eggs were produced from females lacking usp function. Moreover usp is required specifically in somatic cells for production of a fertilizable egg. Defects in border cell migration are known to lead to the production of unfertilized eggs. Whether EcR loss of function mutations affect border cell migration could not be examined. This is because the EcR locus, at 42A, is proximal to available FRT insertions, making it impossible to make FLP-mediated mosaic clones. The frequency of X-ray induced mitotic clones is too low to be useful, and marking such clones is problematic. A temperature-sensitive allele of EcR exists and flies at the nonpermissive temperature exhibit a variety of defects in oogenesis, including arrest prior to border cell migration. Even though it was not possible to assess the effect of EcR mutations specifically in the border cells, the observations that hormone injections can lead to precocious border cell migration and that reduced ecdysone levels can lead to delayed migration provide additional support for the hormonal control of migration (Bai, 2000).

The rise in ecdysone after eclosion, specifically in females, occurs in response to adequate nutrition. In the absence of a rich diet, yolk protein synthesis is inhibited and oogenesis does not progress. Yolk protein synthesis can be restored in the absence of a rich diet by applying ecdysone or juvenile hormone (JH) to cultured ovaries. Recent studies indicate that functional ecdysone receptors are required in the germline for progression of oogenesis through vitellogenesis, the stages during which yolk is taken up by the oocyte. In summary, then, adequate nutrition appears to lead to elevated hormone levels, which in turn stimulate yolk protein synthesis and uptake, and progression of oogenesis beyond stage 8. Together with the results reported here, these findings suggest that a rising ecdysone titer coordinates a variety of events that occur in early vitellogenic egg chambers, including border cell migration (Bai, 2000).

These studies indicate a role for steroid hormones in cell motility that is independent of any role in cell proliferation or cell fate determination. If tai function were required for follicle cell proliferation, it would not be possible to generate homozygous mutant clones since the homozygous mutant cells would fail to proliferate. Many homozygous mutant tai clones are found in the follicular epithelium, some of which are quite large; therefore, there does not appear to be a requirement for tai function in follicle cell proliferation. In addition, tai mutant border cells clearly differentiate from the neighboring follicle cells, based on their morphology, and they continue to express all of the border cell markers tested. Therefore, there does not appear to be any detectable change of cell fate or differentiation state in these cells. Rather, they appear to have a relatively specific defect in their ability to migrate through the neighboring nurse cell cluster (Bai, 2000).

These findings may have significance for steroid hormone-dependent human cancers since hormones are known to promote progression of breast, ovarian, and prostate cancers, which includes acquisition of highly invasive characteristics. The prevailing view is that the hormones act to stimulate proliferation of the cancer cells, leading to an increased likelihood of mutation and appearance of an invasive phenotype. However, the data presented in this paper suggest that steroid hormones can also stimulate invasive behavior independently of any discernible effect on proliferation. Thus, steroid hormones, like many peptide growth factors, may possess both mitogenic and motogenic properties. This notion is supported by studies that show effects of an anti-estrogen on metastasis of prostate cancer cells in the rat. Raloxifene, an anti-estrogen, inhibits metastasis of PAIII adenocarcinoma cells to the lymph nodes and lungs, in vivo, without effects on growth of the primary tumor, or proliferation of the PAIII cells in vitro. The treatment also extends the survival of the animals (Bai, 2000 and references therein)

A number of genes that have been described, slbo, jing, breathless, shotgun, and PZ6356 define a slbo-dependent pathway required for border cell migration. Experiments reported in this study indicate that expression of Fak56D also depends upon the slbo pathway. In contrast, tai function appears to be independent of slbo, based on the lack of effect of slbo mutations on tai expression and the lack of effect of tai mutations on slbo expression. In addition, overexpression of tai fails to rescue even mild slbo migration defects and overexpression of slbo fails to rescue tai migration defects. Mutations in either slbo or tai affect cadherin and Fak56D. Whereas slbo function is required for expression of these two proteins, tai function is required for proper subcellular localization of both proteins (Bai, 2000).

The finding that Shotgun is required both in the border cells and in the nurse cells for normal border cell migration is surprising since the prevailing view has been that E-cadherin promotes formation of stable cell–cell adhesion belts and inhibits motility. However, there are numerous exceptions to the general correlation of decreased E-cadherin expression with increasing motility. One particularly interesting exception is the human ovarian epithelium. Normal cells within the human ovarian surface epithelium express relatively low levels of E-cadherin. However, carcinomas derived from this epithelium express high levels of E-cadherin, and overexpression of E-cadherin in T antigen transformed ovarian surface epithelium cells, causing them to become invasive and to form distant metastases in nude mice. Thus, in both human and Drosophila ovaries, E-cadherin seems to promote rather than inhibit motility (Bai, 2000).

Why then do some cells respond to increased cadherin expression by increasing invasiveness whereas other cells respond by decreasing invasiveness? It is proposed that the difference is that some cell types are capable of rapidly turning over E-cadherin-containing adhesion complexes whereas other cells are not. If the complexes can be turned over efficiently, the cells behave like wild-type border cells and become invasive. If the complexes cannot be turned over efficiently, the cells behave as tai mutant border cells, accumulate stable cell–cell adhesion complexes and do not migrate. It will be important to identify the critical downstream targets of Tai because one or more of these may be a protein important for turnover of adhesion complexes (Bai, 2000).

Steroid hormones may stimulate formation and turnover of cadherin-containing cell adhesion complexes in human breast cancer as well. In support of this, MCF7 human breast cancer cells have been found to respond to beta-estradiol treatment by extending motile lamellipodia, which make small, transient, E-cadherin-containing contacts with underlying cells. This behavior is prevented by treatment of the cells with anti-estrogens. Taken together with the proven effectiveness of anti-estrogens in preventing and reversing metastasis of hormone-dependent cancers, these findings suggest that steroid hormones may stimulate invasive cell behavior by facilitating rapid turnover of E-cadherin containing cell adhesion complexes. This could be one mechanism by which amplification of AIB1 contributes to the progression of breast and ovarian cancer (Bai, 2000).

GENE STRUCTURE

cDNA clone length - 7535

Bases in 5' UTR - 813

Bases in 3' UTR - 614

PROTEIN STRUCTURE

Amino Acids - 2035

Structural Domains

The tai locus encodes a protein with amino acid sequence similarity to steroid hormone receptor coactivator proteins of the p160 family (SRCs). SRCs bind to steroid hormone receptor (SHR) complexes in a ligand-dependent manner and potentiate hormone-induced transcriptional activation (Leo, 2000). The most related protein is AIB1, a steroid hormone receptor coactivator that is amplified in breast and ovarian cancer (Anzick, 1997). SRCs typically contain a basic helix-loop-helix (bHLH) domain near the N terminus, a PAS domain, LXXLL motifs, which are responsible for binding to the hormone bound receptor (McInerney, 1998), and one or more polyglutamine stretches that mediate transcriptional activation. The predicted Tai protein contained all of the domains featured in the p160 class of SRCs, and the top 19 BLAST scores are from members of this family. The highest level of amino acid sequence identity is found in the bHLH domain, which is 48% identical and 71% similar between AIB1 and Tai (Bai, 2000).

date revised: 15 April 2001

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.