UNIVERSITY OF CINCINNATI
______, 20 _____
I,______, hereby submit this as part of the requirements for the degree of:
______in: ______It is entitled: ______
Approved by: ______DOWNSTREAM EFFECTORS OF THE
HOMEOBOX TRANSCRIPTION FACTOR
HOXA 11
A dissertation submitted to the
Division of Research and Advanced Studies of the University of Cincinnati
in partial fulfillment of the requirements for the degree of
DOCTORATE OF PHILOSOPHY (Ph.D.)
in the Department of Developmental Biology of the College of Medicine
2000
by
M. Todd Valerius
B.S. University of Cincinnati, 1993
Committee Chair: S. Steven Potter Abstract
Homeobox transcription factors play critical roles in the genetic cascade of development. Critical to understanding their function is describing the expression and identifying the downstream targets of homeobox genes. The initial characterization of the dispersed homeobox gene Gsh-1 is reported in chapter 2. In situ hybridization showed developmental expression was limited to the CNS. In the hindbrain and neural tube stripes of Gsh-1 were seen in neuroepithelial tissue. The diencephalon and mesencephalon also showed Gsh-1 expression including the future thalamus and hypothalamus. Using a fusion protein approach, the in vitro consensus DNA binding site of Gsh-1 was determined.
Chapter 3 reports the characterization of kidney cell lines derived from Hoxa 11-
SV40 T antigen transgenic mice. Molecular marker analysis determined the mK3 and mK4 cell lines represent early metanephric mesenchyme and differentiated epithelial-like cells respectively. Co-culture experiments with isolated ureteric bud demonstrated mK3 cells retained early metanephric mesenchyme function by supporting ureteric bud growth.
Expression profile comparisons of the mK3 and mK4 cell lines identified 121 differentially expressed genes. Several of these were previously described in the differentiation of metanephric mesenchyme, validating the approach. The remaining genes, consisting of both known and unknown, are now implicated in this process.
Chapter 4 describes the identification of candidate downstream targets of Hoxa
11. The Hoxa 11-SV40 T antigen transgene described in chapter 3 was bred into the
Hoxa 11/Hoxd 11 double mutant mouse line. A kidney cell line, mK10, was isolated from mutant transgenic E18.5 embryos (Hoxa 11-/- Hoxd 11 -/- Hoxa 11/SV40 Tag +). The mK10 cells, as well as HEK293 cells, were transfected with expression constructs containing the Hoxa 11 cDNA to create cell populations with and without Hoxa 11 for each. Differential Display, GDA arrays, and GeneChip microarrays were used for expression profile comparisons. These screens identified nine genes that were reproducibly altered in expression with the addition of Hoxa 11 expression. Integrin alpha 8 (ItgA8) was altered in both cell lines. In situ hybridization of E13.5 Hoxa
11/Hoxd 11 mutant kidneys showed ItgA8 expression was altered, consistent with ItgA8 being a downstream target of Hoxa 11.
Acknowledgements
I would first like to thank my primary advisor Dr. S. Steven Potter. Without his mentoring and endless patience I would not have accomplished this lofty goal. I would also like to thank my committee, Dr. Robert Arceci, Dr. Jun Ma, Dr. Anil Menon, and Dr.
Jeff Robbins for their guidance and support.
Many thanks to the other faculty and staff of the Molecular and Developmental
Biology Program. My future accomplishments in science will in no small way be a result of the excellent training I have received here.
To past and present members of the Potter laboratory I am very appreciative of the friendship and help I have always received.
Finally, a most special thanks goes to my wife Wendy and the rest of my family.
I can not imagine having the will to finish this program without the love and support of these wonderful people. I am blessed to have such a family. Table of Contents
Chapter 1 Introduction 2 References 9 Chapter 2 Gsh-1 : A novel murine homeobox gene expressed in the central nervous system 14 Abstract 15 Introduction 16 Methods 20 Results 23 Discussion 39 Acknowledgements 44 References 45 Chapter 3 Defining the genetic basis of metanephric mesenchyme differentiation 55 Abstract 56 Introduction 57 Methods 61 Results 66 Discussion 80 Acknowledgements 83 References 84 Chapter 4 Hoxa 11 downstream target candidates identified by expression profile analysis 94 Abstract 95 Introduction 96 Methods 98 Results 101 Discussion 114 Acknowledgements 117 References 118 Chapter 5 Additional Discussion 127 References 132
- 1 - Chapter 1
Introduction
- 2 - Homeobox genes encode transcription factors that contain a 60 amino acid helix- turn-helix DNA-binding motif called the homeodomain. These genes function by regulating batteries of downstream targets responsible for organogenesis of specific structures (Lewis, 1978; McGinnis and Krumlauf, 1992). Some homeobox genes are found in clusters and in Drosophila this is known as the homeotic complex (HomC).
Mammalian homologs of the Drosophila homeotic genes, the Hox genes, are well conserved in homeodomain sequence and their clustered chromosomal organization
(Sharkey et al., 1997). While there is a single homeotic complex of 8 genes in
Drosophila, mammals have four Hox clusters and a total of 39 Hox genes located on different chromosomes with each cluster containing 9-11 genes. These four complexes are believed to have formed through evolution via duplication of a single ancestral homeobox cluster. The position of the Hox genes within a cluster is conserved allowing them to be aligned into 13 paralogous groups by homeodomain sequence homology. In addition, the Abdominal-B (Abd-B) type Hox genes are expanded in mammals with groups 9-13 sharing homology with the single Drosophila Abd-B gene. In Drosophila the homeotic genes act as master regulators specifying body segment identity along the
A/P axis of the embryo. This is illustrated by the classic antennapedia mutation in flies in which an extra pair of wings is formed in a body segment that normally would not have wings (Lewis, 1978). This phenomenon of homeotic tranformation is also seen in Hox gene mutations in mouse, notably in skeletal elements where A/P axis phenotypes are easily studied. Both loss of function (Post et al., 1993; Kanzler et al., 1998; Small and
Potter, 1993) and gain of function (Charite et al., 1994; Zhang et al., 1994) studies support this general role for A/P axis identity in vertebrates as well. The conserved
- 3 - chromosomal organization and role in segment identity in such distantly related species as Drosophila and mouse point to the important role Hox genes play in development.
In vitro DNA-binding studies have determined the consensus binding sequence for Hox genes is rather small, consisting of a TAAT core (Benson et al., 1995;
Maconochie et al., 1996; Pellerin et al., 1994; Valerius et al., 1995) (See Chapter 2). In addition, due to the high similarity between Hox homeodomains there is very little variation in the consensus element. In some cases specificity is conferred by Hox cofactors that may modify DNA-binding via protein-protein interactions or by cooperative DNA-binding, effectively expanding the consensus site or favoring one Hox protein over another (Castelli-Gair, 1998; Li et al., 1999; Mann and Chan, 1996; Shen et al., 1997; Shen et al., 1999). The small amount of binding site information from these in vitro DNA-binding studies has made the search for the downstream targets by sequence searches difficult. These studies on the whole demonstrate that the complete complex of cofactors must be understood to identify the proper DNA sequence target, and this may vary tissue to tissue with the complement of cofactors and Hox proteins. Furthermore, it would require knowledge of all Hox cofactors and their expression to use an in vitro approach to determine target sequence sites for a given cell or tissue, and this is not known today.
In Drosophila only about a dozen Hox targets are known and these can be broken into three categories including signaling molecules (e.g. dpp, wingless), transcription factors like empty spiracles (e.g. distaless), and structural or cell adhesion genes (Graba et al., 1997). Regulation of other Hox genes as well as autoregulation has also been described previously (Gould et al., 1997; Popperl et al., 1995). In mammals the number
- 4 - of known Hox targets has recently caught up to the fly. Early reports identified thyroid transcription factor 1 (TTF1) as a target of Hoxb 3 (Guazzi et al., 1994) and LCAM as a target of Hoxd 9 (Goomer et al., 1994). More recently studies have shown Hoxa 5 to directly regulate p53 and the progesterone receptor (Raman et al., 2000a; Raman et al.,
2000b). Promoter studies have suggested targets for several Hox genes. Hoxa 10 has been shown to bind the p21 promoter with MEIS1 and PBX1 and activate transcription
(Bromleigh and Freedman, 2000). It has also been shown that Hoxc 8 can bind the osteopontin promoter and repress transcription unless displaced by Smad1 signaling (Shi et al., 1999). Hoxc 6 was shown to regulate a NCAM promoter/reporter construct in osteoclast like cells upon activation by Bmp2 (Boersma et al., 1999). Finally, subtractive hybridization screens have identified candidate targets of Hoxa 1 in P9-10 teratocarcinoma cells including BMP4, cadherin6, HMG-1, Gbx-2, and Evx-2 (Shen et al., 2000). The dearth of downstream targets despite the efforts of many laboratories highlights the complexities of the DNA binding specificity of Hox genes and the difficulty of the problem.
The work presented in the following chapters is an effort to understand Hox gene function by identifying the downstream targets. In Chapter 2 the initial characterization of a dispersed homeobox gene, Gsh1, is described including determination of the consensus DNA-binding site using an in vitro approach. Realizing the limitations of this approach as outlined above, we decided to use genetic tools available in the lab to search for downstream targets as described in Chapter 4. For these studies, Hoxa 11/Hoxd 11 double knockout mice were used to screen for downstream targets of these paralogous
Hox genes. In addition, because of the undefined and varied involvement of cofactors,
- 5 - we focused our search in the context of kidney development. This also gave us the opportunity to study aspects of kidney development as described in Chapter 3. The kidney represents an excellent model system for the study of the general principles of organogenesis. Kidney formation involves stem cells, mesenchymal-epithelial inductive interactions, mesenchymal to epithelial conversion, proximal-distal segmentation, vascularization, and apoptosis. Development of the metanephric kidney begins in mouse at around E11.5 when the ureteric bud invades a region of intermediate mesoderm called the metanephric mesenchyme. The metanephric mesenchyme condenses around the bud tip, inducing it to branch and eventually form the collecting ducts. The mesenchymal stem cells in turn are induced to undergo further division and differentiation to form the nephron and stromal cell lineages. This interaction is reiterated with each branch leading to new bud tips, mesenchymal condensations and nephrons (Saxen, 1987). Gene knockout studies have implicated several genes as significant in this process, including wnt-4, pax-2, lim-1, bmp-7, WT-1, GDNF, and its receptor c-ret (Dudley et al., 1995;
Moore et al., 1996; Pichel et al., 1996; Sanchez et al., 1996; Shawlot and Behringer,
1995; Stark et al., 1994; Torres et al., 1995).
The Hoxa 11 gene is expressed in the metanephric mesenchyme just prior to invasion of the ureteric bud and continues in the condensing mesenchyme around the ureteric bud tips. As this mesenchyme differentiates, some cells undergo a mesenchymal to epithelial conversion and Hoxa 11 expression is lost, suggesting an early role for Hoxa
11 in the process of nephrogenesis. Gene knockout studies have demonstrated a redundant role for the paralogous Hox genes Hoxa 11 and Hoxd 11 in kidney development (Davis et al., 1995). Hoxd 11 is widely expressed in the developing kidney,
- 6 - overlapping Hoxa 11 expression domains. Neither gene is expressed in the ureteric bud or its derivatives. Mutation in either gene alone has no detectable renal phenotype while the Hoxa 11/Hoxd 11 double knockout results in a rudimentary kidney or no kidney at all.
The double null mutants have defects in nephrogenesis and in branching morphogenesis of the ureteric bud. This indicates these redundant Hox genes are critical for the interaction between the mesenchyme and the ureteric bud.
We have sought to understand how the Hoxa 11 gene carries out its function by identifying the downstream target genes. To accomplish this we wanted to design differential screens to examine gene expression changes in response to expression of
Hoxa 11. Suspecting that Hox cofactors are critical to Hox function as mentioned above, we established cell lines that most accurately reflect the normal in vivo environment in which Hoxa 11 functions by using a directed oncogenesis approach in transgenic mice.
Cell lines were made from several Hoxa 11 expressing tissues including the kidney, ductus deferens, uterus, and limb. The kidney derived cell lines were shown to represent slightly different cell types from the mesenchymal to epithelial conversion that occurs during nephron formation, consistent with the expression pattern of Hoxa 11. Molecular marker analysis and functional analysis in kidney co-culture indicated these cells retain embryonic kidney cell phenotype. This presented the opportunity to expand our study of this process by comparing the expression profiles of these two cell lines to identify genes involved in the conversion (Chapter 3). This transgene was also bred onto the Hoxa
11/Hoxd 11 mutant background and we have isolated a double mutant kidney cell line.
The mutant cell line and a human kidney derived cell line (HEK 293) were modified to add expression of Hoxa 11 and used in the difference cloning screens. We have done
- 7 - three screens total using differential display, GDA blotted cDNA arrays, and Affymetrix
GeneChip arrays. All three screens yielded confirmed differences.
- 8 - REFERENCES Benson, G. V., Nguyen, T. H., and Maas, R. L. (1995). The expression pattern of the murine Hoxa-10 gene and the sequence recognition of its homeodomain reveal specific properties of Abdominal B-like genes, Mol Cell Biol 15, 1591-601.
Boersma, C. J., Bloemen, M., Hendriks, J. M., van Berkel, E. A., Olijve, W., and van
Zoelen, E. J. (1999). Homeobox proteins as signal transduction intermediates in regulation of NCAM expression by recombinant human bone morphogenetic protein-2 in osteoblast-like cells, Mol Cell Biol Res Commun 1, 117-24.
Bromleigh, V. C., and Freedman, L. P. (2000). p21 is a transcriptional target of HOXA10 in differentiating myelomonocytic cells [In Process Citation], Genes Dev 14, 2581-6.
Castelli-Gair, J. (1998). The lines gene of Drosophila is required for specific functions of the Abdominal-B HOX protein, Development 125, 1269-74.
Charite, J., de Graaff, W., Shen, S., Deschamps, J. (1994). Ectopic expression of Hoxb-8 causes duplication of the ZPA in the forelimb and homeotic transformation of axial structures, Cell 78, 589-601.
Davis, A. P., Witte, D. P., Hsieh-Li, H. M., Potter, S. S., and Capecchi, M. R. (1995).
Absence of radius and ulna in mice lacking hoxa-11 and hoxd-11, Nature 375, 791-5.
Dudley, A. T., Lyons, K. M., and Robertson, E. J. (1995). A requirement for bone morphogenetic protein-7 during development of the mammalian kidney and eye, Genes
Dev 9, 2795-807.
- 9 - Goomer, R. S., Holst, B. D., Wood, I. C., Jones, F. S., and Edelman, G. M. (1994).
Regulation in vitro of an L-CAM enhancer by homeobox genes HoxD9 and HNF-1, Proc
Natl Acad Sci U S A 91, 7985-9.
Gould, A., Morrison, A., Sproat, G., White, R. A., and Krumlauf, R. (1997). Positive cross-regulation and enhancer sharing: two mechanisms for specifying overlapping Hox expression patterns, Genes Dev 11, 900-13.
Graba, Y., Aragnol, D., and Pradel, J. (1997). Drosophila Hox complex downstream targets and the function of homeotic genes, Bioessays 19, 379-88.
Guazzi, S., Lonigro, R., Pintonello, L., Boncinelli, E., Di Lauro, R., and Mavilio, F.
(1994). The thyroid transcription factor-1 gene is a candidate target for regulation by Hox proteins, Embo J 13, 3339-47.
Lewis, E. B. (1978). A gene complex controlling segmentation in Drosophila, Nature
276, 565-70.
Li, X., Veraksa, A., and McGinnis, W. (1999). A sequence motif distinct from Hox binding sites controls the specificity of a Hox response element, Development 126, 5581-
9.
Maconochie, M., Nonchev, S., Morrison, A., and Krumlauf, R. (1996). Paralogous Hox genes: function and regulation, Annu Rev Genet 30, 529-56.
Mann, R. S., and Chan, S. K. (1996). Extra specificity from extradenticle: the partnership between HOX and PBX/EXD homeodomain proteins [published erratum appears in
Trends Genet 1996 Aug;12(8):328], Trends Genet 12, 258-62.
- 10 - McGinnis, W., and Krumlauf, R. (1992). Homeobox genes and axial patterning, Cell 68,
283-302.
Moore, M. W., Klein, R. D., Farinas, I., Sauer, H., Armanini, M., Phillips, H., Reichardt,
L. F., Ryan, A. M., Carver-Moore, K., and Rosenthal, A. (1996). Renal and neuronal abnormalities in mice lacking GDNF, Nature 382, 76-9.
Pellerin, I., Schnabel, C., Catron, K. M., and Abate, C. (1994). Hox proteins have different affinities for a consensus DNA site that correlate with the positions of their genes on the hox cluster, Mol Cell Biol 14, 4532-45.
Pichel, J. G., Shen, L., Sheng, H. Z., Granholm, A. C., Drago, J., Grinberg, A., Lee, E. J.,
Huang, S. P., Saarma, M., Hoffer, B. J., et al. (1996). Defects in enteric innervation and kidney development in mice lacking GDNF, Nature 382, 73-6.
Popperl, H., Bienz, M., Studer, M., Chan, S. K., Aparicio, S., Brenner, S., Mann, R. S., and Krumlauf, R. (1995). Segmental expression of Hoxb-1 is controlled by a highly conserved autoregulatory loop dependent upon exd/pbx, Cell 81, 1031-42.
Post, L.C., Innis, J.W. (1999) Altered Hox expression and increased cell death distinguish
Hypodactyly from Hoxa13 null mice, Int J Dev Biol. 43, 287-9
Raman, V., Martensen, S. A., Reisman, D., Evron, E., Odenwald, W. F., Jaffee, E.,
Marks, J., and Sukumar, S. (2000a). Compromised HOXA5 function can limit p53 expression in human breast tumours, Nature 405, 974-8.
Raman, V., Tamori, A., Vali, M., Zeller, K., Korz, D., and Sukumar, S. (2000b). HOXA5 regulates expression of the progesterone receptor, J Biol Chem 275, 26551-5.
- 11 - Sanchez, M. P., Silos-Santiago, I., Frisen, J., He, B., Lira, S. A., and Barbacid, M. (1996).
Renal agenesis and the absence of enteric neurons in mice lacking GDNF, Nature 382,
70-3.
Saxen, L. (1987). Organogenesis of the kidney (Cambridge, UK, Cambridge University
Press).
Sharkey, M., Graba, Y., and Scott, M. P. (1997). Hox genes in evolution: protein surfaces and paralog groups, Trends Genet 13, 145-51.
Shawlot, W., and Behringer, R. R. (1995). Requirement for Lim1 in head-organizer function [see comments], Nature 374, 425-30.
Shen, J., Wu, H., and Gudas, L. J. (2000). Molecular cloning and analysis of a group of genes differentially expressed in cells which overexpress the Hoxa-1 homeobox gene,
Exp Cell Res 259, 274-83.
Shen, W. F., Montgomery, J. C., Rozenfeld, S., Moskow, J. J., Lawrence, H. J.,
Buchberg, A. M., and Largman, C. (1997). AbdB-like Hox proteins stabilize DNA binding by the Meis1 homeodomain proteins, Mol Cell Biol 17, 6448-58.
Shen, W. F., Rozenfeld, S., Kwong, A., Kom ves, L. G., Lawrence, H. J., and Largman,
C. (1999). HOXA9 forms triple complexes with PBX2 and MEIS1 in myeloid cells, Mol
Cell Biol 19, 3051-61.
Shi, X., Yang, X., Chen, D., Chang, Z., and Cao, X. (1999). Smad1 interacts with homeobox DNA-binding proteins in bone morphogenetic protein signaling, J Biol Chem
274, 13711-7.
- 12 - Small, K.M. and Potter, S.S. (1993) Homeotic transformation and limb defects in Hox
A11 mutant mice. Genes Dev. 7: 2318-2328.
Stark, K., Vainio, S., Vassileva, G., and McMahon, A. P. (1994). Epithelial transformation of metanephric mesenchyme in the developing kidney regulated by Wnt-
4, Nature 372, 679-83.
Torres, M., Gomez-Pardo, E., Dressler, G. R., and Gruss, P. (1995). Pax-2 controls multiple steps of urogenital development, Development 121, 4057-65.
Valerius, M. T., Li, H., Stock, J. L., Weinstein, M., Kaur, S., Singh, G., and Potter, S. S.
(1995). Gsh-1: a novel murine homeobox gene expressed in the central nervous system,
Dev Dyn 203, 337-51.
Zhang, M., Kim, H.J., Marshall, H., Gendron-Maguire, M., Lucas, D.A., Baron, A.,
Gudas, L.J., Gridley, T., Krumlauf, R., Grippo, J.F. (1994). Ectopic Hoxa-1 induces rhombomere transformation in mouse hindbrain, Development 120, 2431-42.
- 13 - Chapter 2
Gsh-1 : A novel murine homeobox gene expressed in the central
nervous system
M. Todd Valerius, Hung Li, Jeffrey L. Stock, Michael Weinstein, Satbir Kaur,
Gurparkash Singh, S. Steven Potter
Published in Developmental Dynamics, 1995
- 14 - ABSTRACT
We report the characterization of Gsh-1, a novel murine homeobox gene.
Northern blot analysis revealed a transcript of approximately 2 kb in size present at embryonic days 10.5, 11.5 and 12.5 of development. The cDNA sequence encoded a proline rich motif, a polyalanine tract, and a homeodomain with strong homology to those encoded by the clustered Hox genes. The Gsh-1 expression pattern was determined for days E8.5 to E13.5 by whole mount and serial section in situ hybridizations. Gsh-1 transcription was restricted to the central nervous system. Expression is present in the neural tube and hindbrain as two continuous, bilaterally symmetrical stripes within neural epithelial tissue. In the mesencephalon expression is seen as a band across the most anterior portion. There is also diencephalon expression in the anlagen of the thalamus and the hypothalamus as well as in the optic stalk, optic recess and the ganglionic eminence. Moreover, through the use of fusion proteins containing the Gsh-1 homeodomain, we have determined the consensus DNA binding site of the Gsh-1 homeoprotein to be GCT/CA/CATTAG/A.
- 15 - INTRODUCTION
The organization of the mammalian homeobox gene family is striking. Each gene contains a homeobox, which encodes the DNA-binding, helix-turn-helix motif homeodomain. Thirty-eight of these genes are arranged in four clusters which derived from the quadruplication of an original ancestral complex. These clustered homeobox genes show similarities in sequence and expression, and are designated Hox genes. Of particular interest is a feature referred to as colinearity. The genes located more 3’ on the clusters are expressed at earlier times in development, and with more anterior boundaries of expression, than genes placed more 5’ (Graham et al., 1989; Duboule and Dolle,
1989). The developmental importance of these Hox genes has been established by their homology to Drosophila genes of known developmental significance (McGinnis and
Krumlauf, 1992), by their suggestive embryonic expression patterns (Holland and Hogan,
1988), and by both dominant gain of function (Wolgemuth et al., 1989; Kessel et al.,
1990; Kaur et al., 1992; McLain et al., 1992; Jegalian and De Robertis, 1992; Pollock et al., 1992; Lufkin et al., 1992; Charite et al., 1994) and recessive loss of function (Chisaka and Capecchi, 1991; Lufkin et al., 1991; Chisaka et al., 1992; Mouellic et al., 1992;
Ramírez-Solis et al., 1993; Dolle et al., 1993; Small and Potter, 1993) mutational analyses in mice.
In addition to the clustered Hox genes there exists an even greater number of homeobox genes which are seemingly randomly dispersed in chromosomal position.
These genes can be grouped into subfamilies according to homeobox sequence comparisons. In general, it is found that there are several mammalian homologues of each dispersed Drosophila homeobox gene. Examples include the Dlx-1 and Dlx-2
- 16 - Distal-less genes (Price et al., 1991; Robinson et al., 1991; Porteus et al., 1991, 1992;
Dolle et al., 1992), the En-1 and En-2 engrailed genes (Joyner and Martin, 1987), the msx-1, msx-2 and msx-3 genes (Hill et al., 1989; Robert et al., 1989; Davidson et al.,
1991) and the cdx-1 and cdx-2 genes homologous to the Drosophila caudal gene (Duprey et al., 1988). These genes have expression patterns that distinguish them from the clustered Hox genes, with goosecoid, for example, expressed at the anterior end of the primitive streak during gastrulation (Blum et al., 1992). Targeted mutations for four dispersed homeobox genes have now been described (Joyner et al., 1991; Wurst et al.,
1994; Li et al., 1994; Satokata and Maas, 1994), with the resulting phenotypes demonstrating their developmental importance.
A formal effort to explain in a general theory the functions of homeobox genes in development was first described by Lewis (1978) in reference to Drosophila, and later extended to mammals (Kessel and Gruss, 1991). This Hox code hypothesis states that the identity of a segment or region is determined by the combination of homeobox genes expressed. That is, there exists a binary homeobox gene code that specifies the developmental destinies of groups of cells. Considerable experimental evidence supporting this hypothesis has accumulated, with many mutations in homeobox genes causing alterations in segment identities (Ramírez-Solis et al., 1993; Mouellic et al.,
1992; Jegalian and De Robertis, 1992; Small and Potter, 1993). Indeed, detailed Hox codes have now been defined for the determination of rhombomere identity in the developing hindbrain (Keynes and Krumlauf, 1994). Nevertheless, aspects of some mutant phenotypes have been difficult to interpret in terms of this model (Dolle et al.,
- 17 - 1993; Small and Potter, 1993; Davis and Capecchi, 1994), suggesting that it may not universally apply.
The roles of homeobox genes in the development of the mammalian brain are of particular interest for a number of reasons. This organ, as the source of thought, memory and emotion, is primarily responsible for distinguishing Homo sapiens from other species. The brain, with approximately 100 billion neurons and 1015 connections is an extremely complex, yet modular structure which we are only beginning to understand.
The clustered Hox genes are clearly implicated in the development of the hindbrain, but these genes are not expressed in the more rostral regions of the developing brain.
Increasing evidence suggests that the dispersed homeobox genes play critical roles in the formation of the forebrain and the midbrain, with several now reported to have restricted domains of expression in the developing brain (Porteus et al., 1991; Lu et al., 1992;
Simeone et al., 1993; Saha et al., 1993; Dush and Martin, 1992), and with a targeted mutation in the Engrailed-1 gene resulting in the deletion of certain midbrain structures
(Wurst et al., 1994).
In order to define the homeobox codes that may regulate brain development it will be necessary to isolate and analyze the complement of genes active during brain formation. In this report we characterize a dispersed murine homeobox gene that is expressed in discrete regions of the forebrain, midbrain and hindbrain during development. The cDNA for this gene, designated Gsh-1 for Genomic screen homeobox gene, is shown to have sequence characteristics that closely resemble the clustered Hox genes. Furthermore, two Gsh-1-glutathione-S-transferase fusion proteins were produced in E. coli and used in a random oligonucleotide selected binding and PCR amplification
- 18 - procedure to define the Gsh-1 DNA target binding sequence, as a first step in the identification of downstream genes regulated by Gsh-1.
- 19 - METHODS
Cloning and sequencing Gsh-1 cDNA
A 1.95 kb Bam H1 fragment from the original genomic clone including the homeobox was used as a probe to screen an E11.5 mouse cDNA library and one positive clone was isolated and sequenced. This cDNA was determined to be a chimeric cloning artifact, with only the 3’ end having genuine Gsh-1 cDNA sequence. Therefore, a 900 bp
Pst I fragment from this region was used to screen an E11.5 random primed cDNA library (Stratagene) as well as E10.5 (Novagen) and E12.5 (Stratagene) poly d(T) primed cDNA libraries. Positive clones were subcloned, restriction mapped, and the longest cDNAs were sequenced using Sequenase with 7-deaza-dGTP (U.S. Biochemical). These two cDNAs were nearly identical, differing by only 3 bases in length at the 5' end.
Northern blot
Superovulated B6C3F1 female mice (Hogan et al., 1986) were mated with
B6C3F1 males. Counting the day of plug as day 0.5 of gestation, embryos were removed by cesarean section at various time points and frozen in liquid nitrogen. Total RNA was isolated using RNAzol (TelTest Inc.). mRNA was then isolated using oligo d(T) cellulose (Celano et al., 1993). Following electrophoresis in 1% agarose/formaldehyde gels, the RNA was blotted onto Gene Screen Plus (DuPont) and hybridized according to manufacturer's protocol. A 450 bp Sfi I fragment from the 5' end of the cDNA (without homeobox) was labeled with [32P] dATP and used to probe the Northern blot. After exposure to Kodak XAR film with intensifying screens, the blot was subsequently stripped and reprobed with a Pst I/Xba I fragment of glyceraldehyde-3-phosphate dehydrogenase (Tso et al., 1985) to quantitate mRNA loading.
- 20 - In Situ hybridization
A 450 bp fragment was used initially as a template for [35S] UTP riboprobe production using a Gemini System II kit (Promega). Although this initial probe contained homeobox sequences, control Southern blots indicated no detectable cross hybridization to other genomic sequences at the stringencies employed. A second 470 bp fragment that did not contain any homeodomain was later used and the expression at all time points agreed with the initial results (both whole mounts and serial section in situs).
Furthermore, the observed expression pattern is distinct from that seen for the most closely related homeobox gene, Gsh-2 (data not shown). Riboprobe was purified using
Nu-Clean R50 spin columns (IBI) and ammonium acetate/ethanol precipitation, then resuspended in TE and incorporated radiolabel determined. Embryos were isolated by cesarean section at various time points, fixed in 4% paraformaldehyde/PBS overnight, and frozen in embedding oil. Sections of 7-10 microns were cut and placed onto silane coated slides (Histology Control Systems), dried, and post-fixed with 4% paraformaldehyde/PBS. The slides were then rinsed in 2X SSC and treated briefly with a pronase solution (Proteinase K 20 mg/ml). Rinses in Tris-Glycine and acetic anhydride solutions were used to stop the pronase reaction and were then followed with rinses in 2X
SSC and dehydration. The slides were then incubated in prehyb solution (50% formamide, 2X SSC, 750 mg/ml yeast tRNA, 1X Denhardt’s, 10% Dextran sulfate, 30 mg/ml Thio UMP, 100 mg/ml BSA, and 5mM DTT) at 45°C for 15 minutes. A hybridization solution (same as prehyb without Thio UMP and including probe) was then added such that the final probe concentration was 750,000 cpm/slide. The slides were covered with Sigmacote treated coverslips, sealed with rubber cement, and incubated
- 21 - overnight at 45°C. The coverslips were removed and the slides washed in 1X SSC+DTT for 30 minutes (numerous changes) at 50°C. A RNase A/T1 solution (50 mg/ml and 50 units/ml respectively) was used to reduce background (37°C for 30 minutes). The sections were then subjected to a series of washes in SSC+DTT(1X-.1X) at 50-55°C followed by dehydration in ethanol. Autoradiography was performed by dipping the slides into liquefied Kodak NTB photographic emulsion and allowing exposure for 10-20 days at 4°C. The slides were developed using Kodak developer and fixer according to manufacturer’s protocol. Following Hemotoxylin-Eosin staining, coverslips were added with Permount (Fisher Scientific).
Target binding sequence
Fusion proteins were constructed using the Glutathione S-Transferase (GST)
Gene Fusion system (Pharmacia). A total of 9 DNA fragments containing the homeodomain region were cloned in frame into pGEX vectors. Two of these, #2 (bp
407-1068 in Fig.1) and #13 (bp 620-848) resulted in the most efficient protein production in E. coli and therefore column purified and used for selection. Four rounds of selection were performed as previously described (Wilson et al., 1993). Individual selected oligonucleotides were cloned and sequenced using either Sequenase (USB) or dsDNA
Cycle Sequencing (BRL). A total of 32 clones from #2 examined for a possible consensus sequence, finding total of 45 sites present. Ten other sequences from #13 were also sequenced and found to be similar to those found from #2 (not represented in Fig. 8).
- 22 - RESULTS
Gsh-1 northern analysis
We previously described the results of a search for novel murine homeobox genes
(Singh et al., 1991). A degenerate oligonucleotide probe with a mixture of all nucleotide sequences capable of encoding the highly conserved KIWFQNRR component of the homeodomain was used to screen a mouse genomic DNA library. Sequence analysis of over seventy positive clones identified ten potential novel homeobox genes. Partial sequence analysis of the Gsh-1 genomic DNA, coupled with the recombination mapping results which localized these sequences to a position near the Kit gene on chromosome 5, suggested that Gsh-1 represented a novel dispersed homeobox gene. Nevertheless, as no
Gsh-1 transcripts had been identified, it remained possible that this was a pseudogene.
To examine temporal developmental expression of Gsh-1 and to determine transcript size, poly (A)+ RNA was isolated from mouse embryos at days 10.5, 11.5,
12.5, and 14.5 of gestation. Fig. 1 shows a northern blot of these mRNAs hybridized with a Gsh-1 specific probe. A single prominent band of hybridization was observed for days 10.5, 11.5 and 12.5 of development, corresponding to a Gsh-1 mRNA of approximately 2 kb in size. Expression was reduced at day 14.5, with no hybridization band detected.
Gsh-1 cDNA
To better define the coding potential of Gsh-1, and to initiate analysis of gene structure and regulation, mouse cDNA libraries representing days 10.5, 11.5 and 12.5 of gestation were screened with Gsh-1 probe. A total of ten Gsh-1 cDNA clones were identified and characterized. Six were isolated from an embryonic day 10.5 (E10.5)
- 23 - Figure 1. Northern blot analysis of Gsh-1 transcripts at various embryonic time points. Approximately 10 µg of poly A+ RNA from embryonic day 10.5 (lane 1), 11.5 (lane 2), 12.5 (lane 3), and 14.5 (lane 4) was probed with a Gsh-1 probe. Relative levels of loading and RNA intensity were checked by reprobing the blot with a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe.
library, three from two independent E11.5 libraries, and one from an E12.5 library. The two longest cDNAs were completely sequenced and one was found to be 3 bp longer at the 5’ end.
The Gsh-1 cDNA sequence and predicted amino acid sequence are shown in Fig.
2. The cDNA appears complete at the 3’ end, with polyadenylation signal and the beginning of the poly (A) tail. The size of this cDNA, 1.786 kb without the bulk of the poly (A) tail, approximates the 2 kb transcript detected by northern analysis, suggesting that this cDNA is full length or near full length. The ATG beginning at base 195 is immediately preceded by a CC dinucleotide, in agreement with the translation initiation consensus sequence determined by Kozak (1987). Nevertheless, no in frame termination codon is present upstream of the ATG, leaving open the possibility that the full open reading frame is not included.
The predicted amino acid sequence has several interesting characteristics. First, the amino portion is relatively rich in prolines, a common feature of clustered Hox encoded proteins. Proline rich sequences have been associated with both transcription
- 24 - - 25 - Figure 2. Nucleotide and predicted amino acid sequence of the Gsh-1 cDNA. The homeodomain is underlined, the prolines are highlighted in bold face, and the polyalanine stretch is both bold and underlined. This sequence represents a clone from the day 10.5 library and was the longest cDNA isolated from all screens. activation and repression function (Han and Manley, 1993a,b; Mitchell and Tjian, 1989).
In addition, there is an eight amino acid polyalanine tract encoded by bases 522 to 545.
Similar polyalanine runs have been previously observed in several developmentally important genes, including engrailed (Macdonald et al., 1986), even-skipped (Poole et al.,
1985), runt (Kania et al., 1990), cut (Blochinger et al., 1988) and ovo (Mevel-Ninio et al.,
1991) in Drosophila. This motif may form an alpha helix structure in transcription repression domains (Han and Manley, 1993a,b).
Homeodomain comparisons are shown in Fig. 3. The Gsh-1 encoded homeodomain is most similar to that of Gsh-2, another dispersed murine homeobox gene identified in the previously described degenerate oligonucleotide screen for novel homeobox genes (Singh et al., 1991; Hsieh-Li et al., 1995). With only two amino acid differences, or 97% identity, these two homeodomains show a level of homology equal to that observed for the homeodomains encoded by a single paralogous group of clustered
Hox genes. The Gsh-1 and Gsh-2 genes therefore appear to represent a distinct subfamily of dispersed homeobox genes. Interestingly, the Gsh-1 and Gsh-2 encoded proteins show very little amino acid sequence homology outside of the homeodomains
(data not shown).
The homeodomains of the Gsh-1 and Gsh-2 proteins are surprisingly similar those of the clustered Hox genes, as shown for HoxB-2, HoxB-3 and HoxA-11 in Fig. 3. The clustered Hox genes generally encode similar homeodomains that as a group are distinct
- 26 - Figure 3. Homeodomain comparison. The Gsh-1 deduced homeodomain amino acid sequence compared to murine genes Gsh-2, HoxB2, HoxB3, HoxA11. Also cnox2 from Hydra and the Drosophila genes Zen-1 and Ant. (All sequences are from Kappen et al., 1993). At the right is listed the percent identity and percent similarity, respectively, when comparing each gene to GSH-1. from those encoded by the dispersed homeobox genes. But in pair-wise comparisons the
Gsh-1 and Gsh-2 homeodomains fit into the clustered Hox grouping better than some of the clustered Hox genes themselves. In particular, the Abdominal B type clustered Hox genes, represented by HoxA-11 in Fig. 3, are more divergent from the standard
Antennapedia sequence than Gsh-1 and Gsh-2.
It is interesting to note that among the sequenced Drosophila homeobox genes
Gsh-1 shows the greatest homology to Zen-1 (Fig. 3). In Drosophila the Zen-1 and Zen-2 genes reside in the HOM-C cluster. In mammals, however, the Hox clusters have no homologues of the Zen genes. This suggests the possibility that the Gsh-1 and Gsh-2 genes are the cognates of the Drosophila Zen genes. This difference in homeobox gene organization would not be without precedent, as the mammalian Evx homeobox genes are associated with the A and D clusters (Dush and Martin, 1992), while in Drosophila the corresponding even-skipped gene is dispersed. Nevertheless, the degree of homology observed between the Gsh and Zen genes is too low, with only about 70% amino acid identity in the encoded homeodomains, for them to represent true orthologs.
- 27 - In situ hybridizations
To determine the developmental expression pattern of Gsh-1, whole mount in situ hybridizations were performed. As shown in Fig. 4A, initial expression can be seen at
E8.5 in the hindbrain (arrowheads). This early expression is expanded at E9.5 with two stripes stretching posteriorly from the hindbrain while new expression is also seen in the diencephalon (Fig. 4B). At E10.5 and 11.25 the two stripes of expression now stretch the length of the neural tube and continue into the hindbrain. At E10.5 the stripes of expression terminate in the hindbrain, but at E11.25 this expression extends into the mesencephalon (Fig. 4C, D). Expression in the mesencephalon can also be seen at both
E10.5 and E11.25 as a wide band across the anterior portion of the mesencephalon.
Though not visible in these photographs, there is a gap in this expression at the midline of the embryo (also see Fig. 5A). Expression in the diencephalon is observed in the form of a crescent at E10.5, later changing to form stripes that asymptotically converge in the
E11.25 embryo (large arrowheads in Fig. 4C, D).
Serial section in situ hybridizations were performed to examine the expression patterns of Gsh-1 in E11.5 and E13.5 embryos in more detail. Sense strand controls were used for both whole mount and serial section in situs, with no hybridization signal detected (data not shown). Two different antisense probes showed identical patterns at all time points. The expression of Gsh-1 transcripts is restricted to mitotically active neural epithelial tissue, seen as the darker staining ventricular regions in the brain. The various sections of an E11.5 embryo in Fig. 5 show the Gsh-1 expression pattern to be very discrete. Fig. 5A shows a transverse section with the broad band of expression that
“caps” the mesencephalic vesicle and the gap in this expression at the midline. In
- 28 - Figure 4. Spatial distribution of Gsh-1 transcripts viewed by whole mount in situ hybridizations. A: Lateral view of day 8.5 embryo. The arrowheads indicate the two small points of expression in the hindbrain region. B: A day 9.5 embryo. The hindbrain expression has extended caudally into the neural tube forming two bilaterally symmetrical stripes. There is also slight expression in the diencephalon. C: Two day 10.5 embryos. The stripes starting at the roof of the hindbrain now stretch the length of the neural tube. Note also the band of expression at the top of both embryos. The arrowhead indicates the more intense diencephalon expression. D: Lateral view of a day 11.25 embryo. The stripes in the neural tube and hindbrain now extend into the midbrain. The large arrowhead indicates a refinement in diencephalon expression in comparison to C. The small arrows approximate the angle of the section in Fig. 5E. addition, the two stripes of expression extending rostrally from the neural tube are apparent in the metencephalon (at the IV ventricle). Fig. 5B is a more posterior transverse section intersecting the well defined stripes in the myencephalon and showing
- 29 - - 30 - Figure 5. Day 11.5 - Distribution of Gsh-1 transcripts viewed by serial section in situ hybridizations. A: Darkfield of a transverse section in the head region showing expression in the metencephalon in the fourth ventricle and in the mesencephalon. B & C: Darkfield transverse section with corresponding brightfield showing myencephalon expression in the fourth ventricle as well as expression in the diencephalon. D: Darkfield frontal section that intersects the two bilateral stripes in the myencephalon. E & F: Darkfield frontal section with corresponding brightfield. Each stripe of expression in the hindbrain is intersected three times. The angle of this section is approximated by the arrows in Fig. 4D. G & H: Darkfield parasagittal section with corresponding brightfield. Note the expression in the optic stalk and the ganglionic eminence. d = diencephalon, dt = dorsal thalamus, et = epithalamus, ge = ganglionic eminence, h = hypothalamus, ms = mesencephalic vesicle, mt = metencephalon, my = myencephalon, te = telencephalic vesicle, r = Rathke's pouch, IV = fourth ventricle. the discrete diencephalon expression also seen in Fig. 4D. Fig. 5D is a frontal section through the hindbrain also intersecting the very well defined stripes in the myencephalon.
Fig. 5E is another frontal section that is more ventral than the previous. This section shows the stripes of expression in the myencephalon and metencephalon, intersecting each stripe three times. The angle of this section is approximated by the arrows in Fig.
4D. Fig. 5G is a parasagittal section showing expression in the anlage of the thalamus and the mesencephalon. Also visible is Gsh-1 expression that wraps around the optic recess, as well as weaker expression in the ganglionic eminence. The thoracic neural tube is shown in Fig. 6. Gsh-1 expression is restricted to the mid-ventricular area, with very fine boundaries throughout the neural tube. These stripes continue both rostral and caudal, extending from the tail through the hindbrain and into the midbrain.
Sagittal and transverse sections were done for E13.5 to examine Gsh-1 expression in more developed brain structures. Fig. 7A shows a parasagittal section with expression in similar regions as earlier time points. This includes the stripes of expression in the medulla and cerebellum, and the "cap" of expression in the midbrain. There is expression
- 31 - Figure 6. Gsh-1 expression in the thoracic neural tube. A: Serial section in situ hybridization shows the very restricted expression pattern of Gsh-1 in the midventricular region of the developing neural tube of a day 11.5 embryo. B: The brightfield of A. d = dorsal root ganglia, m = mantle layer, sc = spinal canal, v = ventricular zone.
in the hypothalamus immediately adjacent to the pituitary and intense signal in the optic stalk. Weaker expression is present around the medial ganglionic eminence in the telencephalic vesicle. Fig. 7C is more near the midline of the embryo and shows similar expression as 7A but now includes expression in a portion of the thalamus as well as expanded expression in the hypothalamus. Fig. 7E is a transverse section showing a portion of the band of expression that "caps" the midbrain. The two continuous stripes from the neural tube can also be seen in the neural epithelial tissue surrounding the
- 32 - - 33 - Figure 7. Day 13.5 - Distribution of Gsh-1 transcripts viewed by serial section in situ hybridizations. A&B: Darkfield & brightfield of a parasagittal section through the head region. Very restricted domains of expression are in all regions of the brain. Note the intense expression in the optic stalk and hypothalamus adjacent to the pituitary. C & D: A more central parasagittal with regions of expression similar to the above (A & B) but also showing some expression in the thalamus. E: Rostral transverse section with expression in the mesencephalon. F: Transverse section at the pontine flexure of the pons, intersecting each stripe of expression in the hindbrain twice. Faint expression can also be seen in the medulla and diencephalon. G & H: Darkfield and brightfield of a transverse section showing expression in the diencephalon. Note the faint signal in the medial ganglionic eminence (comparable to A & C). AP = anterior pons, AQ = aqueduct, C = cerebellum, D = diencephalon, E = ear, ET = epithalamus, H = hypothalamus, LG = lateral ganglionic eminence, M = medulla, MG = medial ganglionic eminence, O = optic stalk, P = pons, PP = posterior pons, R = Rathke's pouch, RM = roof of midbrain, T = thalamus, V = trigeminal nerve, IV = fourth ventricle. aqueduct. Fig. 7F is another transverse section showing slight Gsh-1 expression in the diencephalon and medulla and the well defined stripes of expression that run through the pons. This section intersects the stripes at the pontine flexure of the pons which separates the anterior and posterior pons. In Fig. 7G is a transverse section through the diencephalon showing the well defined expression of Gsh-1 in the hypothalamus and thalamus as well as expression in and around the ganglionic eminence.
Gsh-1 target DNA binding sequence
As a possible first step in the isolation of genes that are downstream targets of
Gsh-1 we have determined the DNA consensus binding sequence of the Gsh-1 encoded homeodomain. A Gsh-1 homeodomain-GST (glutathione-S-transferase) fusion protein was synthesized using the Pharmacia pGEX system. A 660 bp Sma I-Msc I restriction segment from the Gsh-1 cDNA, including the entire homeobox, was subcloned into the
Sma I restriction site of the pGEX-3X vector. Fusion protein was purified from IPTG induced E. coli using Glutathione-Sepharose 4B beads. PAGE analysis showed a single
- 34 - - 35 - Figure 8. Selected oligonucleotides and the resulting in vitro consensus sequence for the Gsh-1 homeodomain. The 32 selected clones sequenced are shown aligned by the core ATTA sequence. Some clones contained multiple sites and are therefore listed more than once so that each ATTA site may be aligned. A summary of the results is at the bottom with the resulting consensus sequence.
band of protein, with a mobility corresponding to the predicted molecular weight of 46 kd (data not shown).
The random oligonucleotide selected binding and PCR amplification procedure described by Blackwell and Weintraub (1990), and modified by Wilson et al. (1993), was used to define the binding sequence. A 74 base oligonucleotide with a 26 base random sequence core was synthesized and made double stranded, using a primer complementary to one end and Klenow fragment DNA polymerase. The fusion protein was added to the oligonucleotide mix, and bound sequences were purified with sepharose 4B beads, PCR amplified, and re-selected. After four complete cycles selected DNAs were subcloned and sequenced. Fig. 8 shows 45 target sequences vertically aligned. Interestingly, these
45 sequences were found from only 32 separate subclones, with 12 carrying multiple consensus binding sites, likely a result of their preferred selection. The bottom of Fig. 8 shows the 9 bp consensus binding site of GCT/CA/CATTAG/A. This sequence includes the ATTA core element which has been previously identified in a number of homeodomain target binding sequences (Hayashi and Scott, 1990; Odenwald et al.,
1989). Interestingly, however, five flanking bases also show significant sequence preferences.
In addition to the fusion construct described above, two additional constructs were made. One carried a random open reading frame subcloned into pGEX-3X, and served as
- 36 - Figure 9. Gel shift assay with the GSH-1-GST fusion protein. To determine the specificity a competition gel shift was done with oligonucleotide #3-25 as the specific target. Lane 1 is free probe alone, lane 2 with GST protein, lane 3 with GSH- 1-GST fusion protein and no competition. Lanes 4-7 are competed with 1X, 10X, 100X, and 1000X cold specific oligonucleotide. Lanes 8-11 are the same except with cold non- specific competitor.
a negative control, with subclones following selected binding and amplification showing no consensus sequence. The other clone included the homeobox, but with considerably less flanking sequence, having only cDNA bases 620 to 848. Interestingly, this fusion protein showed less sequence specificity in its binding. Twelve selected and amplified oligonucleotides were subcloned and sequenced. Two of these had a single base mis- match with the ATTA core, compared with only one in 45 for the fusion protein including more flanking amino acids. The sequences adjacent to the ATTA core showed even greater loss of specificity. Immediately 5’ of the ATTA core, for example, one now observed a random mix of bases, compared to 78% A or C for the longer fusion protein.
At other positions there was some conservation of the GCT/CA/CATTAA/G consensus, but in each case less sequence preference than observed with the longer fusion protein.
- 37 - A gel shift assay was performed to confirm that the sequences shown in Fig. 8 do bind to the GSH-1-GST fusion protein. As shown in Fig. 9, the #3-25 oligonucleotide sequence does produce a gel shift band, which is competed with specific competitor, but not with non-specific competitor.
- 38 - DISCUSSION
In this report we describe an initial characterization of the murine Gsh-1 homeobox gene. The Gsh-1 cDNA sequence encodes a proline rich amino terminal end, and a polyalanine tract that may represent a transcription repressor domain. The Gsh-1 homeodomain is 97% identical to that of Gsh-2, suggesting that these two genes represent a novel dispersed family of homeobox genes. The Gsh-1 and Gsh-2 homeodomains in turn are very similar to those encoded by the Hox cluster genes, with greater homology to
Antennapedia than the Hox Abd-B type paralogous groups.
In situ hybridizations showed Gsh-1 expression to be CNS restricted. This expression shows extreme spatial and temporal restriction. A pair of stripes of Gsh-1 transcribing cells course from the caudal end of the neural tube rostrally through the hindbrain. Restricted domains of expression are also observed in the forebrain and midbrain. The development of these regions of the brain is unique is certain respects, with the absence of underlying notochord which plays an important role in the induction of the hindbrain and neural tube (Placzek et al., 1990). The expression pattern of Gsh-1 suggests potential roles in the development of several CNS components. In addition to expression in the spinal cord and hindbrain, Gsh-1 is extensively expressed in the diencephalon and the ganglionic eminence. The diencephalon develops into the thalamus and hypothalamus, as well as regions of the brain that receive neural input from the eyes.
Gsh-1 expression is also present in these mature structures at E13.5. The thalamus is an important motor and sensory function integration center (Mojsilovic and Zecevic, 1991) and thalamic disturbances have been associated with Korsakoff syndrome (Squire and
Moore, 1979) and schizophrenia (Oke and Adams, 1987; Pakkenberg, 1990). The
- 39 - hypothalamus is also of critical importance, with the paraventricular nucleus, which secretes oxytocin, vasopressin and corticotropin releasing hormone, and the suprachiasmatic nucleus, which provides an important circadian pacemaker function
(Swaab et al., 1993). The nuclei of the hypothalamus are important in feeding, sexual behavior and aggression. The hypothalamus also plays a key role in the development of the pituitary, through interactions with Rathke’s pouch. At later times the releasing hormones produced by the hypothalamus regulate the further development and function of the pituitary.
The dispersed homeobox genes are postulated to play an important role in directing development of the forebrain and midbrain, with none of the clustered Hox genes expressed in positions more rostral than the hindbrain. In contrast, several of the dispersed homeobox genes show discrete patterns of transcription in more anterior regions of the developing brain. These dispersed homeobox genes may function in a combinatorial fashion to determine pattern formation in the neuromeric segmentation of the brain similar to the clustered Hox genes along the neural tube (Puelles and
Rubenstein, 1994; Stoykova and Gruss, 1994). The Dlx family members, for example, are expressed in intricate patterns in the forebrain, as well as in limbs, developing branchial arches, face and eyes (Robinson et al., 1991; Porteus et al., 1991, 1992; Price et al., 1991; Dolle et al., 1992; Bulfone et al., 1993). The Evx-1 and -2, Otx-1 and -2 and
NK family of homeobox genes are also expressed in the developing forebrain (Dush and
Martin, 1992; Saha et al., 1993; Simeone et al., 1993). Furthermore, the Dbx homeobox gene is expressed in the telencephalon, diencephalon, hindbrain and spinal cord (Lu et al.,
1992).
- 40 - The cloning and characterization of the complement of homeobox genes expressed in more rostral regions of the brain contributes to our eventual understanding of the development and function of this organ. As expression patterns are determined, potential homeobox codes that may determine regional identities are defined. And as these genes are subjected to targeted mutagenesis, thereby directly altering homeobox gene codes during development, the validity of the code model will be tested, and the in vivo developmental functions of these genes determined.
At the molecular level it is clear that one function of homeobox genes is to modulate expression patterns of downstream target genes. To determine the target sequence preference of Gsh-1 we used a fusion protein in conjunction with a selected binding and amplification procedure. When the fusion protein included 221 amino acids of Gsh-1 protein a surprising specificity was observed. The consensus binding sequence of GCT/CA/CATTAG/A is nine bases long and preference for these bases is strong.
Interestingly, when fewer amino acids flanking the Gsh-1 homeodomain were included in the fusion protein the target sequence specificity was reduced. This suggests that at least in some cases amino acid residues beyond the homeodomain can make important contributions to target site preferences.
These observations are of some interest since the mechanisms by which homeodomain proteins achieve their biological specificity in vivo are uncertain. Many
Drosophila homeoproteins have been shown to have similar in vitro DNA binding target sites (Hayashi and Scott, 1990), suggesting that they have similar or identical downstream target genes. The work of Walter et al. (1994) supports this conclusion, with their in vivo cross-linking studies showing that the even-skipped and fushi tarazu proteins
- 41 - have similar targets. Other evidence, however, suggests that homeodomain proteins bind to similar, yet distinct target sites. In Drosophila, for example, it is known that when two homeodomain proteins have the same target gene, they physically bind to different cis- regulatory elements (O’Hara et al., 1993).
The Gsh-1 consensus binding site presented in this report resembles those previously described for homeodomain proteins, with a well-conserved ATTA core.
Nevertheless, five positions flanking this core also show preferred bases, distinguishing the Gsh-1 target. The variability observed in individual bound sequences, however, suggests that the Gsh-1 target sequences likely overlap those of several Hox cluster encoded proteins, as might be predicted by their homologous homeodomains.
In vivo target specificity of homeoproteins is likely determined by a number of factors. First, there are differences in preferred binding sequences. Available data, however, suggests that these differences are often very subtle (Phelan et al., 1994).
Affinities for target sites can, however, vary dramatically among homeoproteins (Phelan et al., 1994; Pellerin et al., 1994). Moreover, the non-homeodomain components of these proteins are quite variable. These regions can determine transcription activator or repressor function as well as dimerization reactions with other proteins and cooperative binding to target sequences (Wilson et al., 1993). In Drosophila the extradenticle encoded divergent homeodomain protein undergoes cooperative interactions with other homeoproteins to alter their target site specificity (Chan et al., 1994; van Dijk and Murre,
1994). The sum of these different effects, coupled with the distinct developmental expression patterns for individual homeobox genes, likely generates the diverse biological activities of homeobox genes.
- 42 - These considerations argue that a defined consensus target binding sequence can serve only as a rough guide in identifying downstream genes. Potential targets identified by computer sequence searches can undergo preliminary confirmation by cell transfection studies. But these systems generally provide a poor representation of the complex milieu of potential interacting proteins present in the developing organism. Further verification of potential targets will therefore require analysis of expression patterns in embryos with a targeted mutation in the Gsh-1 gene.
- 43 - ACKNOWLEDGMENTS
We thank Claude Desplan for providing the procedure for identifying consensus target binding sequences prior to publication. We thank Jan Hagedorn for assistance in manuscript preparation. This work was supported by NIH grant HD29599.
- 44 - REFERENCES
Blackwell, T.K. and Weintraub, H. (1990) Differences and similarities in DNA-binding preferences of MyoD and E2A protein complexes revealed by binding site selection.
Science 250: 1104-1110.
Blochinger, K., Bodmer, R., Jack, J., Jan, L.Y. and Jan, Y.N. (1988) Primary structure and expression of a product from cut, a locus involved in specifying sensory organ identity in Drosophila. Nature 333: 629-635.
Blum, M., Gaunt, S.J., Cho, K.W.Y., Steinbeisser, H., Blumberg, B., Bittner, D. and
DeRobertis, E.M. (1992) Gastrulation in the mouse: the role of the homeobox gene goosecoid. Cell 69: 1097-1106.
Bulfone, A., Kim, H.-J., Puelles, L., Porteus, M.H., Grippo, J.F. and Rubenstein, J.L.R.
(1993) The mouse Dlx-2 (Tes-1) gene is expressed in spatially restricted domains of the forebrain, face and limbs in midgestation mouse embryos. Mech. Dev. 40: 129-140.
Celano, P., Vertino, P.M. and Casero, R.J. (1993) Isolation of polyadenylated RNA from cultured cells and intact tissues. Biotechniques 15: 26-28.
Chan, S., Jaffe, L., Capovilla, M., Botas, J. and Mann, R.S. (1994) The DNA binding specificity of ultrabithorax is modulated by cooperative interactions with extradenticle, another homeoprotein. Cell 78: 603-615.
Charite, J., de Graaf, W., Shen, S. and Deschamps, J. (1994) Ectopic expression of Hox
B-8 causes duplication of the ZPA in the forelimb and homeotix transformations of axial structures. Cell 78: 589-601.
- 45 - Chisaka, O. and Capecchi, M.R. (1991) Regionally restricted developmental defects resulting from targeted disruption of the mouse homeobox gene Hox-1.5. Nature 350:
473-479.
Chisaka, O., Musci, T.S. and Capecchi, M.R. (1992) Developmental defects of the ear, cranial nerves and hindbrain resulting from targeted disruption of the mouse homeobox gene Hox-1.6. Nature 355: 516-521.
Davidson, D.R., Crawley, A., Hill, R.E. and Tickle, C. (1991) Position-dependent expression of two related homeobox genes in developing vertebrate limbs. Nature 352:
429-431.
Davis, A.P. and Capecchi, M.R. (1994) Axial homeosis and appendicular skeleton defects in mice with a targeted disruption of hoxd-11. Development 120: 2187-2198.
Dolle, P., Dierich, A., LeMeur, M., Schimmang, T., Schuhbaur, B., Chambon, P. and
Duboule, D. (1993) Disruption of the Hox d-13 gene induces localized heterochrony leading to mice with neotenic limbs. Cell 75: 431-441.
Dolle, P., Price, M. and Duboule, D. (1992) Expression of the murine Dlx-1 homeobox gene during facial, ocular and limb development. Differentiation 49: 93-99.
Duboule, D. and Dolle, P. (1989) The structural and functional organization of the murine Hox gene family resembles that of Drosophila. EMBO J. 8: 1497-1505.
Duprey, P., Chowdhury, K., Dressler, G.R., Balling, R., Simon, D., Guenet, J. and Gruss,
P. (1988) A mouse gene homologous to the Drosophila gene caudal is expressed in epithelial cells from the embryonic intestine. Genes Dev. 2: 1647-1654.
- 46 - Dush, M.M. and Martin, G.R. (1992) Analysis of mouse Evx genes: Evx-1 displays graded expression in the primitive streak. Dev. Biol. 151: 273-287.
Graham, A., Papalopulu, N. and Krumlauf, R. (1989) The murine and Drosophila homeobox gene complexes have common features of organization and expression. Cell 5:
367-378.
Han, K. and Manley, J. (1993) Transcriptional repression by the Drosophila Even- skipped protein: definition of a minimal repression domain. Genes Dev. 7: 491-503.
Han, K. and Manley, J.L. (1993) Functional domains of the Drosophila Engrailed protein.
EMBO J. 12: 2723-2733.
Hayashi, S. and Scott, M.P. (1990) What determines the specificity of action of
Drosophila segmentation gene fushi tarazu. Cell 63: 883-844.
Hill, R.E., Jones, P.F., Rees, A.R., Christina, M.S., Justice, M.J., Copeland, N.G.,
Jenkins, N.A., Graham, E. and Davidson, D.R. (1989) A new family of mouse homeobox-containing genes: molecular structure, chromosomal location, and developmental expression of Hox-7.1. Genes Dev. 3: 26-37.
Hogan, B., Constantini, F. and Lacy, E. (1986) "Manipulating the mouse embryo. A laboratory manual." Cold Spring Harbor, NY: Cold Spring Harbor Press.
Holland, P.W.H. and Hogan, B.L.M. (1988) Expression of homeobox genes during mouse development: a review. Genes Dev. 2: 773-782.
Hsieh-Li, H.M., Witte, D., Szucsik, J., Weinstein, M., Li, H. and Potter, S. (1995) Gsh-2, a murine homeobox gene expressed in the developing brain. Mech. Dev. ,In press.
- 47 - Jegalian, B.G. and De Robertis, E.M. (1992) Homeotic transformation in the mouse induced by overexpression of a human Hox 3.3 transgene. Cell 71: 901-910.
Joyner, A.L., Herrup, K., Auerbach, B.A., Davis, C.A. and Rossant, J. (1991) Subtle cerebellar phenotype in mice homozygous for a targeted deletion of the En-2 homeobox.
Science 251: 1239-1242.
Joyner, A.L. and Martin, G.R. (1987) En-1 and En-2, two mouse genes with sequence homology to the Drosophila engrailed gene: expression during embryogenesis. Genes
Dev. 1: 29-38.
Kania, M.A., Bonner, A.S., Duffy, J.B. and Gergen, J.P. (1990) The Drosophila segmentation gene runt encodes a novel nuclear regulatory protein that is also expressed in the developing nervous system. Genes Dev. 4: 1701-1713.
Kappen, C., Schughart, K. and Ruddle, F.H. (1993) Early evolutionary origin of major homeodomain sequence classes. Genomics 18: 54-70.
Kaur, S., Singh, G., Stock, J., Schreiner, C., Kier, A.B., Yager, K., Mucenski, M.L.,
Scott, W.J. and Potter, S.S. (1992) Dominant mutation of the murine Hox-2.2 gene results in developmental abnormalities. J. Exp. Zool. 264: 323-336.
Kessel, M., Balling, R. and Gruss, P. (1990) Variations of cervical vertebrae after expression of a Hox-1.1 transgene in mice. Cell 61: 301-308.
Kessel, M. and Gruss, P. (1991) Homeotic transformations of murine vertebrae and concomitant alteration of Hox codes induced by retinoic acid. Cell 67: 89-104.
Keynes, R. and Krumlauf, R. (1994) Hox genes and regionalization of the nervous system. Annu. Rev. Neurosci. 17: 109-132.
- 48 - Kozak, M. (1987) An analysis of 5'-noncoding sequences from 699 vertebrate messenger
RNAs. Nucl. Acids Res. 15: 8125-8148.
Lewis, E.B. (1978) A gene complex controlling segmentation in Drosophila. Nature 276:
565-570.
Li, H., Witte, D.P., Branford, W.W., Aronow, B.J., Weinstein, M., Kaur, S., Wert, S.,
Singh, G., Schreiner, C.M., Whitsett, J.A., Scott, W.J. and Potter, S.S. (1994) Gsh-4 encodes a LIM-type homeodomain, is expressed in the developing central nervous system and is required for early postnatal survival. EMBO J 13: 2876-2885.
Lu, S., Bogorad, L.D., Murtha, M.T. and Ruddle, F.H. (1992) Expression pattern of a murine homeobox gene, Dbx, displays extreme spatial restriction in embryonic forebrain and spinal cord. Proc. Natl. Acad. Sci. USA 89: 8053-8057.
Lufkin, T., Dierich, A., LeMeur, M., Mark, M. and Chambon, P. (1991) Disruption of the
Hox-1.6 homeobox gene results in defects in a region corresponding to its rostral domain of expression. Cell 66: 1105-1119.
Lufkin, T., Manuel, M., Hart, C.P., Dolle, P., LeMeur, M. and Chambon, P. (1992)
Homeotic transformation of the occipital bones of the skull by ectopic expression of a homeobox gene. Nature 359: 835-841.
Macdonald, P.M., Ingham, P. and Struhl, G. (1986) Isolation, structure, and expression of even-skipped: a second pair-rule gene of Drosophila containing a homeo box. Cell 47:
721-734.
McGinnis, W. and Krumlauf, R. (1992) Homeobox genes and axial patterning. Cell 68:
283-302.
- 49 - McLain, K., Schreiner, C., Yager, K.L., Stock, J.L. and Potter, S.S. (1992) Ectopic expression of Hox-2.3 induces craniofacial and skeletal malformations in transgenic mice. Mech. Dev. 39: 3-16.
Mevel-Ninio, M., Terracol, R. and Kafatos, F.C. (1991) The ovo gene of Drosophila encodes a zinc finger protein required for female germ line development. EMBO J. 10:
2259-2266.
Mitchell, P.J. and Tjian, R. (1989) Transcriptional regulation in mammalian cells by sequence-specific DNA binding proteins. Science 245: 371-378.
Mojsilovic, J. and Zecevic, N. (1991) Early development of the human thalamus: Golgi and Nissl study. Early Hum. Dev. 27: 119-144.
Mouellic, H.L., Lallemand, Y. and Brulet, P. (1992) Homeosis in the mouse induced by a null mutation in the Hox-3.1 gene. Cell 69: 251-264.
O'Hara, E.O., Cohen, B., Cohen, S. and McGinnis, W. (1993) Distal-less is a downstream gene of Deformed required for ventral maxillary identity. Development 117: 847-856.
Odenwald, W.F., Garbern, J., Arnheiter, H., Tournier-Lasserve, E. and Lazzarini, R.A.
(1989) The Hox 1.3 homeobox protein is a sequence specific DNA binding phosphoprotein. Genes Dev. 3: 158-172.
Oke, A.F. and Adams, R.N. (1987) Elevated thalamic dopamine: possible link to sensory dysfunctions in schizophrenia. Schizophr. Bull. 13: 589-604.
Pakkenberg, B. (1990) Pronounced reduction of total neuron number in mediodorsal thalamic nucleus and nucleus accumbence in schizophrenics. Arch. Gen. Psychiatry 47:
1023-1029.
- 50 - Pellerin, I., Schnabel, C., Catron, K.M. and Abate, C. (1994) Hox proteins have different affinities for a consensus DNA site that correlate with the position of their genes on the hox clusters. Mol. Cell. Biol. 14: 4532-4545.
Phelan, M., Sadoul, R. and Featherstone, M. (1994) Functional differences between HOX proteins conferred by two residues in the homeodomain N-terminal arm. Mol. Cell. Biol.
14: 5066-5075.
Placzek, M., Tessier-Lavigue, M., Yamada, T., Jessel, T. and Dodd, J. (1990)
Mesodermal control of neural cell identity: floor plate induction by the notochord.
Science 250: 985-988.
Pollock, R.A., Jay, G. and Bieberich, C.J. (1992) Altering the boundaries of Hox 3.1 expression: evidence for antipodal gene regulation. Cell 71: 911-923.
Poole, S.J., Kauvar, L.M., Drees, B. and Kornberg, T. (1985) The engrailed locus of
Drosophila: structural analysis of an embryonic transcript. Cell 40: 37-43.
Porteus, M.H., Brice, A.E.J., Bulfone, A., Usdin, T.B., Ciaranello, R.D. and Rubenstein,
J.L.R. (1992) Isolation and characterization of a library of cDNA clones that are preferentially expressed in the embryonic telencephalon. Mol. Brain Res. 12: 7-22.
Porteus, M.H., Bulfone, A., Ciaranello, R.D. and Rubenstein, J.L. (1991) Isolation and characterization of a novel cDNA clone encoding a homeodomain that is developmentally regulated in the ventral forebrain [published erratum appears in Neuron
1992 Jul;9(1):187]. Neuron 7: 221-229.
- 51 - Price, M., Lemaistre, M., Pischetola, M., DiLauro, R. and Duboule, D. (1991) A mouse gene related to Distal-less shows a restricted expression in the developing forebrain.
Nature 351: 748-751.
Puelles, L. and Rubenstein, J.L. (1993) Expression patterns of homeobox and other putative regulatory genes in the embryonic mouse forebrain suggest a neuromeric organization. [Review]. Trends in Neurosciences 16: 472-479.
Ramírez-Solis, R., Zheng, H., Whiting, J., Krumlauf, R. and Bradley, A. (1993) Hox B4
(Hox-2.6) mutant mice show homeotic transformation of a cervical vertebrae and defects in the closure of sternal rudiments. Cell 73: 279-294.
Robert, B., Sassoon, D., Jacq, B., Gehring, W. and Buckingham, M. (1989) Hox-7, a mouse homeobox gene with a novel pattern of expression during embryogenesis. EMBO
J. 8: 91-100.
Robinson, G.W., Wray, S. and Mahon, K.A. (1991) Spatially restricted expression of a member of a new family of murine Distal-less homeobox genes in the developing forebrain. New Biologist 3: 1183-1194.
Saha, M.S., Michel, R.B., Gulding, K.M. and Grainger, R.M. (1993) A Xenopus homeobox gene defines dorsal-ventral domains in the developing brain. Development
118: 193-202.
Satokata, I. and Maas, R. (1994) Msx1 deficient mice exhibit cleft palate and abnormalities of craniofacial and tooth development. Nature Genetics 6: 348-356.
Simeone, A., Acampora, D., Mallamaci, A., Stornaiuolo, A., D'Apice, M.R., Nigro, V. and Boncinelli, E. (1993) A vertebrate gene related to orthodenticle contains a
- 52 - homeodomain of the bicoid class and demarcates anterior neuroectoderm in the gastrulating mouse embryo. EMBO J. 12: 2735-2747.
Singh, G., Kaur, S., Stock, J.L., Jenkins, N.A., Gilbert, D.J., Copeland, N.G. and Potter,
S.S. (1991) Identification of 10 murine homeobox genes. Proc. Natl. Acad. Sci. USA 88:
10706-10710.
Small, K.M. and Potter, S.S. (1993) Homeotic transformation and limb defects in Hox
A11 mutant mice. Genes Dev. 7: 2318-2328.
Squire, R. and Moore, R. (1979) Dorsal thalamic lesion in note case of human memory disfunction. Ann. Neurol. 6: 503-506.
Stoykova, A. and Gruss, P. (1994) Roles of Pax-genes in developing and adult brain as suggested by expression patterns. J. Neurosci. 14: 1395-1412.
Swaab, D.F., Hofman, M.A., Lucassen, P.J., Purba, J.S., Raadsheer, F.C. and Van de
Nes, J.A.P. (1993) Functional neuroanatomy and neuropathology of the human hypothalamus. Anat. Embryol. 187: 317-330.
Tso, J.Y., Sun, X.H., Kao, T.H., Reece, K.S. and Wu, R. (1985) Isolation and characterization of rat and human glyceraldehyde-3-phosphate dehydrogenase cDNAs: genomic complexity and molecular evolution of the gene. Nucl. Acids Res. 13: 2485-
2502. van Dijk, M.A. and Murre, C. (1994) extradenticle raises the DNA binding specificity of homeotic selector gene products. Cell 78: 617-624.
- 53 - Walter, J., Dever, C.A. and Biggin, M. (1994) Two homeodomain proteins bind with similar specificity to a wide range of RNA sites in Drosophila embryos. Genes Dev. 8:
1678-1692.
Wilson, D., Sheng, G., Lecuit, T., Dostatni, N. and Desplan, C. (1993) Cooperative dimerization of paired class homeo domains on DNA. Genes Dev. 7: 2120-2134.
Wolgemuth, D.J., Behringer, R.R., Mostoller, M.P., Brinster, R.L. and Palmiter, R.D.
(1989) Transgenic mice overexpressing the mouse homeobox-containing gene Hox-1.4 exhibit abnormal gut development. Nature 337: 464-467.
Wurst, W., Auerbach, A.B. and Joyner, A.L. (1994) Multiple developmental defects in
Engrailed-1 mutant mice: an early mid-hindbrain deletion and patterning defects in forelimbs and sternum. Development 120: 2065-2075.
- 54 - Chapter 3
Defining the genetic basis of metanephric mesenchyme differentiation.
M. Todd Valerius, Larry T. Patterson, and S. Steven Potter
Submission planned to Mechanisms of Development
- 55 - ABSTRACT
Clonal cell lines representing different developmental stages of the metanephric mesenchyme were made from transgenic mice with the SV40 T antigen gene driven by the Hoxa 11 promoter. The mK3 cells represented early metanephric mesenchyme, prior to induction by the ureteric bud. These cells showed a spindle-shaped, fibroblast morphology. They expressed genes characteristic of early mesenchyme, including Hoxa
11, Hoxd 11, Emx-2, collagen I and vimentin. Moreover, the mK3 cells displayed early metanephric mesenchyme biological function. In organ co-culture experiments they were able to induce growth and branching of the ureteric bud. Another cell line, mK4, represented later, induced metanephric mesenchyme undergoing epithelial conversion.
These cells were more polygonal, or epithelial in shape, and expressed genes diagnostic of late mesenchyme, including Pax-2, Pax-8, Wnt-4, Cadherin-6, collagen IV and LFB3.
To better define the gene expression patterns of kidney metanephric mesenchyme cells at these two stages of development labeled RNAs from the mK3 and mK4 cells were hybridized to Affymetrix GeneChip probe arrays. Over 4,000 expressed genes were identified as present and thereby implicated in kidney formation. Comparison of the mK3 and mK4 gene expression profiles revealed 121 genes showing greater than a ten- fold difference in expression level. Several had been involved in metanephric mesenchyme differentiation, but most had not been previously associated with this process.
- 56 - INTRODUCTION
The kidney is a powerful model system for studying mammalian organogenesis
(Davies and Bard, 1998; Kuure et al., 2000). The developing kidney employs many mechanisms of organogenesis, including budding, reciprocal inductive interactions, stem cell growth and differentiation, conversion of mesenchyme into epithelia, branching morphogenesis, apoptosis, fusion (nephrons to collecting ducts), and proximodistal segmentation (along the length of the nephron).
Multiple reciprocal inductive interactions drive the process of kidney formation.
The presence of the metanephric mesenchyme is necessary for ureteric bud formation, growth, and branching to form the collecting duct system. In turn the ureteric bud induces the mesenchyme cells to condense, divide and undergo an epithelial conversion to eventually form the tubules of the nephrons which join to the collecting ducts. Without the presence of the ureteric bud, or an ‘artificial’ inducer such as embryonic spinal cord, the metanephric mesenchyme fails to differentiate and undergoes apoptotic death. In turn, the bud fails to grow and bifurcate in the absence of the metanephric mesenchyme.
We are still in the early phases of understanding the genetic regulation of kidney development, but significant advances have been made. The uninduced metanephric mesenchyme produces glial cell line derived neurotrophic factor (GDNF), which interacts with the ret receptor on the ureteric bud to promote outgrowth. Mice with mutation of either the GDNF or ret gene have severe failure of kidney development (Moore et al.,
1996; Schuchardt et al., 1994). The initial signaling from the bud to the metanephric mesenchyme is less well understood. FGF2, BMP7, and Wnts have been considered as candidate early inducers, but there are problems for each. However, it is clear from
- 57 - mouse gene targeting experiments that BMP7 is essential for continued kidney growth
(Dudley et al., 1995; Luo et al., 1995), and Wnt4 is required for the formation of the nephron from the initial mesenchymal condensation (Stark et al., 1994). In total, over
300 genes have now been implicated (most by expression only) in kidney development.
Gene targeting studies have shown that Lim1 (Shawlot and Behringer, 1995), Pax-2
(Torres et al., 1995), Hoxa 11/Hoxd 11 (Davis et al., 1995), Eya1 (Xu et al., 1999), WT-1
(Kreidberg et al., 1993), Emx-2 (Miyamoto et al., 1997), Integrin a8 (Muller et al., 1997), are all essential for kidney development. Mutations in BF-2 (Hatini et al., 1996), RARa and RARb (Mendelsohn et al., 1999) perturb signaling from the stroma. Leukemia inhibitory factor (LIF) appears to be an important signal from the late ureteric bud to the differentiating nephrons (Barasch et al., 1999). Some progress has been made in defining the genetic regulatory network of kidney development (Cook et al., 1996; Dehbi et al.,
1996; Dehbi and Pelletier, 1996; Hewitt et al., 1995; Lee et al., 1999; Maheswaran et al.,
1993; Phelps and Dressler, 1996; Torban and Goodyer, 1998), but much work remains to be done.
To better understand the genetic basis of kidney development it will be necessary to identify more of the genes expressed during this process. The approximately 300 genes now known to be active in kidney development represent an extremely small fraction of the 50,000-100,000 genes of the mammalian genome. It is likely that hundreds, or even thousands of genes expressed in the developing kidney remain to be recognized, many of which may be key regulators.
In the relatively near future it may become possible to define complete gene expression profiles of the different cell types of the forming kidney as a function of
- 58 - developmental time. This will provide a remarkable molecular portrait of kidney organogenesis. The changing expression levels of every gene in every cell will be known.
Cell lines can provide useful experimental systems for the study of developmental processes. Cell lines offer the advantages of unlimited material, reproducibility, sample homogeneity, and are amenable to manipulations (e.g. transfection) not always available in vivo. The adult Madin-Darby canine kidney cell line, for example, has been used to study the factors influencing branching morphogenesis (Sakurai and Nigam, 1997; Santos et al., 1994; Santos et al., 1993), cell polarity (Hobert et al., 1999), and ion pumps
(Gagnon et al., 1999). The human embryonic kidney cell line HEK 293 has been used to study ion channels (Kupper, 1998; Pedersen et al., 1999; Tong et al., 1999), upstream gene regulation (Jahroudi et al., 1990; Meroni et al., 1997; Wick et al., 1999), and the characterization of downstream targets of transcription factors (Torban and Goodyer,
1998).
The precise spatio-temporal expression of Simian Virus 40 T-antigen (SV40 Tag) in transgenic mice, directed by specific promoter elements, can result in the immortalization and developmental arrest of desired cell types (Lew et al., 1993; Paul et al., 1988; Windle et al., 1990). For example, a series of spatially and temporally restricted promoters were used to generate cell lines that represent different stages of pituitary development (Alarid et al., 1998; Alarid et al., 1996). The cell lines appeared developmentally frozen at approximately the point of initial SV40 Tag expression (Lew et al., 1993; Mellon et al., 1991; Tsutsumi et al., 1992; Windle et al., 1990).
- 59 - In this report we describe making four kidney cell lines and using them to better understand the genetic program of metanephric mesenchyme differentiation. The mK3 cells represent early uninduced metanephric mesenchyme as determined by morphology, expression of multiple early mesenchyme marker genes, and by their ability to induce ureteric bud branching morphogenesis in organ co-culture experiments. The mK4 cells represent later induced mesenchyme as determined by their more epithelial morphology, expression of late mesenchyme marker genes, and their inability to induce ureteric bud branching. Affymetrix GeneChip probe arrays were used to determine gene expression profiles for the mK3 and mK4 cells. Over four thousand expressed genes were identified.
Comparisons between the mK3 and mK4 profiles showed 121 genes with a greater than ten-fold difference in gene expression level. This work provides new experimental systems for the study of kidney development, and better defines the genetic program of metanephric mesenchyme differentiation.
- 60 - METHODS
Creation of Hoxa 11/SV40 Tag transgenic animals
An 11.5 kb SpeI fragment containing the Hoxa 11 gene was cloned into Bluescript plasmid (Stratagene, La Jolla, CA). A 30bp duplex oligonucleotide with a SnaBI site was cloned into the BamHI site near the transcriptional start of Hoxa 11
(5’-GATCCGCTTCAAAGAGGCAGCTGCATACGTA-3’ and
5’-GATCTACGTATGCAGCTGCCTCTTTGAAGCG-3’). This was then digested with
SnaBI and AscI to allow insertion of the SV40 Tag gene (StuI at 5190 to BamHI at 2533) by blunt end ligation. The resulting transgene construct contained 5.1 kb of sequence upstream of the Hoxa 11 transcription start site, part of the first exon, the single intron, the second exon and 3.8 kb o downstream sequence (Fig. 1). Correct cloning was confirmed by DNA sequencing. The final construct was digested from vector with SalI, and purified by low gel temperature agarose electrophoresis and CsCl isopycnic centrifugation.
Transgenic mice were made according to standard pronuclear microinjection procedures using B6C3F1 eggs. Transgenic founders were identified by Southern blot analysis using a 32P labeled (~500bp) SV40 Tag specific probe and by PCR with SV40
Tag specific primers (SV1 5’-CATCAACCTGACTTTGGAGGCTTCT-3’,
SV2!5’-CACTC-TATGCCTGTGTGGAGTAAGA-3’). Skeletons were prepared and stained as previously described (Kuczuk and Scott, 1984).
Isolation and establishment of kidney cell lines
Kidney tissue was carefully dissected from 4-12 week old transgenic mice, minced with scissors, dissociated with trypsin, and the cells plated on 100 mm tissue
- 61 - culture dishes. All cells were cultured in DMEM (Gibco BRL, Rockville, MD) with 10%
o Fetal Bovine Serum (FBS) (Gibco, Rockville, MD) at 37 C in 5% CO2 on standard tissue culture plates (Falcon/BDL, Franklin Lakes, NJ). Media was changed every 2-3 days and cells were split when confluent. After 20+ passages, clones were established by dilution cloning and have subsequently been grown for over 50 additional passages, suggesting these clones are immortalized.
RT-PCR assays
Total RNA was isolated from trypsinized cells using RNAzol (TelTest,
Friendswood, TX). Reverse transcription was performed with Superscript II (Gibco
BRL) and a random hexamer mix according to manufacturer’s directions. Following reverse transcription, PCR was done using Vent polymerase (New England Biolabs,
Beverly, MA). The oligonucleotide primers SV-1 and SV-2 (sequence above) detect a
408bp mRNA specific product, as they amplify around the SV40 intron. Any DNA contamination or unprocessed mRNA in the reaction would result in a larger 745bp product. Primer sets used for RT-PCR were as follows:
Emx-2,!5’TCGCCGTCCCAGCTTTTAAG3’
5’TCCGTAAGACTGAGACTGTGAGCC3’; Bf-2, !5’TGTCCAGTGTGGAGAACT-
TTACTG3’ 5’CTCTACACCTCAAAAAGGGCTTAG3’; Pax!8,!5’ATTCCAGCAT-
TGCCGCTCAC3’ 5’TGGTCCCATCTGTTGTGCTTCC3’; Unigene mm.80595
5'TCACAGCGATGAGATACATAGGTCC3' 5'TGGTAACTCGTTCCCTTAT-; Shroom
5'CTCTGTGAGTAGTAATGGCACCGTC3' 5'TGATTGGAACACGCTCTGCTG3';
Peg-3 5’AACTCAGACTCCAGCGAGCATC3' 5'CTTTTCCTTCATAGGTGTTG-
CCTC3'; Cadherin-6 5'TTTGTGGTCCAAGTCACGGC3' 5'CATCGGCATCAC-
- 62 - TGGCTTTG3'; Aquaporin-1 5'TGCCTGGGATTCTGCCAATG3' 5'GGAGACTG-
GAGGACCGAAATAAAC3'; IGF-2 5' TGAGCAGGGACAGTTCCATCAC3'
5'AAAGAGTTTGGGCGGCTATTG3'; sFRP-1 5' TCTGCCTTAGATAGA-
CCATCGCC3' 5'ACCCTGACTTGTGACCTGCTTGAC3'; GDNF 5'CGGGACTCTAA-
GATGAAG3' 5'CAGGAATGCAATACACAG3'; Fibulin-2 5'TCTATCTACACT-
GCCACCTGACCG3' 5'TCTCCATCTCTGAAACTCTGCGAG3'; Cadherin-16
5'CCTATCAGGAGACTCAAACACGGC3' 5'AATGTGGAAGCGAAGGTCGGAG3';
HGF 5’ CACCGTCATATCTTCTGGGAGC 3’ 5’ GAACATCGTGGATGCCAAGC 3’
Immunocytochemistry
Cells were cultured on chamber slides (Nalge, Naperville, IL), washed with PBS, then fixed with 4% paraformaldehyde in PBS. The cells were then treated with 50 mM
NH4Cl in PBS 15 min, washed, then treated with 0.1% Triton in PBS for 15 min. After pre-blocking in 20% FBS in PBS for 15 min, diluted primary antibody was added (PBS with 20% FBS) and the cells incubated one h at room temperature. Secondary antibodies
(Vector Labs, Burlingame CA) were used according to manufacturer’s protocols. Slides were then washed and mounted in Vectashield (Vector Labs) for viewing. All washes were done 3 times with PBS between each step. The primary antibodies used were as follows: Pax-2, 1:400, (Babco, Richmond, CA), SV40, 1:500, (Sigma, St. Louis, MO),
E-cadherin, 1:1600, (Sigma).
Northern blot analysis
Cells of 10 confluent or near confluent 100mm plates were trypsinized, pelleted, and then washed with PBS. Total RNA was isolated using RNAzol, (TelTest) followed by Poly A+ selection using the Oligotex mRNA isolation kit (Qiagen, Carlsbad, CA).
- 63 - Approximately equal amounts of mRNA (1-5 µg) from each cell line were loaded onto
1.5% agarose formaldehyde denaturing gel and electrophoresed at 5 V/cm for 2-3 hours.
Following photo-documentation of the ethidium bromide stained RNA, the gel was rinsed in distilled water for 15 min then transferred to GeneScreen Plus (NEN, Boston, MA) with 10X SSC overnight. The resulting Northern blots were rinsed in 2X SSC and air- dried prior to hybridization. All hybridizations were performed at 65oC overnight according to GeneScreen manufacturer protocols and then washed in 2X SSC/1% SDS at
65oC for 5-6 h. Hybridized blots were exposed 1-2 days using X-ray film or phosphoimager screen before being stripped and re-hybridized with a GAPDH probe to control for equal loading.
Affymetrix GeneChip probe arrays
The Affymetrix Mu6500, Mu11K, and Mu74 chip sets were used to define gene expression profiles of the mK3 and mK4 cell lines according to manufacturer’s protocols.
Briefly, biotinylated RNAs were synthesized and hybridized to the GeneChip probe arrays, which were then washed, stained with streptavidin-phycoerythrin and scanned.
Analysis was performed using Affymetrix GeneChip software.
Co-culture experiments
Aggregates of each cell line was made by trypsinizing monolayers, recombining free cells in a hanging drop culture, and incubating 4 hrs to overnight. Ureteric buds were isolated at the “T” bud stage from E11.5 mouse embryos by dissection of the urogenital ridge, after brief trypsinization. Isolated ureteric buds were recombined with cell line aggregates in 3-dimensional culture using either Matrigel or Collagen Matrix
(Becton Dickinson, Bedford, MA). When included, the GDNF concentration was 50
- 64 - o ng/ml. These were cultured (DMEM with 10% fetal calf serum, 37 C, 5% CO2) for up to
10 days with media changes every 2-3 days.
DBA staining facilitated visualization of ureteric buds. Ureteric buds were fixed in 2% paraformaldehyde in PBS following co-culture with mK3 cells. After washing twice with PBS buds were blocked with PBS containing 3% BSA and 0.05% Triton for one h at 37° C. FITC labeled DBA lectin (1:40 dilution in blocking solution) was then added and incubated for one h at 37º C. Following two brief five-minute washes in PBS with 0.05% Triton, a third wash was continued overnight. Buds were post-fixed with 2% paraformaldehyde, rinsed in PBS, and imaged by confocal microscopy.
Whole mount in situ hybridizations
Whole mount in situ hybridizations were carried out precisely as previously described (Hogan et al., 1994). Digoxygenin probes were generated from the PCR products above by cloning into pCR BLUNT II (Invitrogen) and performing in vitro transcription with T7 or SP6 RNA polymerase according to manufacturers protocol
(Boerhinger Mannheim).
- 65 - RESULTS
Hoxa 11-SV40 Tag construct and transgenic mouse phenotype.
To make cell lines representative of the early kidney metanephric mesenchyme we used the Hoxa 11 promoter to drive expression of SV40 Tag in transgenic mice. The
Hoxa 11 gene is expressed in the metanephric mesenchyme prior to bud invasion and continues in the induced condensing mesenchyme around the ureteric bud tips (Hsieh-Li et al., 1995) (Patterson et al., in preparation). Hoxa 11 is also expressed in the developing limbs, vertebral column, uterus and ductus deferens (Gendron et al., 1997;
Haack and Gruss, 1993; Hsieh-Li et al., 1995; Small and Potter, 1993). The transgene
DNA construct included 5.1 Kb upstream of the Hoxa 11 transcription start site, 3.8 Kb downstream, and the intact Hoxa 11 intron (Fig. 1). Ten founder Hoxa 11-SV40 Tag transgenic animals were made. Four of these animals had shortened hindlimbs while the remaining six appeared normal. The four with shortened hindlimbs suffered declining health over a 4-12 week period and were sacrificed. Gross examination revealed a unilateral fluid filled cystic kidney in two of the males while the third had an apparently normal kidney, but a massively distended bladder. The female exhibited grossly normal kidneys, but an abnormally thin, pale uterus similar to that seen in the Hoxa 11 knockout mouse (Gendron et al., 1997; Hsieh-Li et al., 1995). Skeletal stains showed the transgenic tibia and fibula failed to fuse and were approximately 1/3 shorter than in the age matched wild type control (data not shown), resembling limbs of Hoxa 11/Hoxd 11 double mutants (Davis et al., 1995). Using RT PCR, SV40 Tag transcripts were detected in the developing kidney, uterus, and the ductus deferens indicating the transgene at least partially recapitulated normal Hoxa 11 expression (data not shown).
- 66 - Figure 1. Hoxa 11-SV40 Tag construct. The SV40 Tag gene cassette, including the intron, was cloned into the first exon of the Hoxa!11 genomic clone, replacing the coding region in this exon. Arrowheads mark the positions of the primers used for genotyping and RT-PCR.
Establishment of Hoxa 11/SV40 Tag immortalized cell lines.
The four Hoxa 11-SV40 Tag transgenic mice with shortened hindlimbs were used to make cell lines. Cells were cultured from kidneys, uterus, and ductus deferens, all tissues that express Hoxa 11. All the primary cell cultures were heterogeneous in cell morphology. After 10-15 passages, however, a single predominant cell morphology emerged in each case. This presumably reflected senescence and death of cells not expressing SV40 Tag, and continued growth of those that did. After several months (20+ passages), dilution cloning was used to make clonal cell lines from each primary isolation. The resulting four kidney cell lines, one from each affected mouse, were designated mK1, mK2, mK3, and mK4. In addition two of the males yielded ductus deferens cell lines, mDD1 and mDD2, and the female a uterine cell line, mU1. These reproductive tract cell lines will be characterized elsewhere.
The mK1, mK2, and mK3 kidney derived cells all exhibited a similar spindle- shaped morphology with several irregular cytoplasmic projections (Fig. 2). The mK4 cells were smaller and polygonal in shape with a cobblestone-like appearance at
- 67 - Figure 2. Kidney cell line morphologies. The mK1, mK2 and mK3 cells were spindle-shaped. The mK4 cells were smaller, showed fewer projections, and were more cobblestone-like in appearance.
confluence. The morphologies suggested the mK4 cells represent a later stage of nephrogenesis, when the mesenchymal cells are converting to epithelia.
Gene expression patterns of kidney cell lines determined by immunocytochemistry, northern hybridizations and RT-PCR.
The mK1-4 kidney cell lines were further characterized by examining their expression of 17 genes, using a combination of immunocytochemistry, northern hybridizations and RT-PCR. Representative northern blots are shown in Fig. 3 and the data summarized in Table 1. The results indicate the mK1, mK2 and mK3 cells represent early metanephric mesenchyme, while the mK4 cells represent metanephric mesenchyme at a later developmental stage, as it undergoes epithelial conversion to form the structures of the nephron.
All of the cell lines expressed genes diagnostic of metanephric mesenchyme, including Hoxa 11 (Hsieh-Li et al., 1995)(Patterson et al., in preparation), Hoxd 11
(Dolle et al., 1991), Hoxa!10, and Emx-2 (Miyamoto et al., 1997; Pellegrini et al., 1997).
Hoxa 11, for example, is expressed in the metanephric mesenchyme of the developing
- 68 - Figure 3. Northern blot analysis of the mK1, mK2, mK3, and mK4 cell lines. (A) Expression of Hoxa 11 sense and antisense transcripts was observed in all of the cell lines. (B) Expression of WT- 1, Pax-2 and Hoxc 9. Each blot was stripped and re- probed with GAPDH to control for loading. The locations of the 18S and 28S ribosomal RNA bands are indicated.
kidney, but not in the ureteric bud or in mature epithelial structures of the nephron. All of the cell lines also expressed SV40 Tag, consistent with their being selectively immortalized by transcription of the Hoxa 11-SV40 Tag transgene. None of the cell lines expressed the ureteric bud marker c-ret (Pachnis et al., 1993).
The mK4 cells expressed several genes characteristic of metanephric mesenchyme cells undergoing epithelial conversion, including E-cadherin (Davies and
Brandli, 1997), Wnt4 (Kispert et al., 1998; Stark et al., 1994), Pax-2 (Dressler, 1996;
Rothenpieler and Dressler, 1993; Torres et al., 1995), lim-1 (Shawlot and Behringer,
1995), Pax-8 (Davies and Brandli, 1997; Igarashi, 1994), and Bmp-7 (Dudley et al.,
1995). The mK1-3 cell lines, which appear to represent an earlier stage of metanephric mesenchyme development, did not express any of these genes. The molecular data
- 69 - correlate well with the morphology of the mK4 cells, which appeared more epithelial in character (Fig. 2). The mK4 cells did not, however, appear to be terminally differentiated because Pax-2 and Wnt4 are transiently expressed during the conversion process and not in the mature epithelial cells of the nephron (Kispert et al., 1998).
Table 1. Summary of genes expressed in the kidney cell lines. GENE NAME mK1 mK2 mK3 mK4 Method HoxA11 + + + + Northern SV40 LTA + + + + RT-PCR/ICC c-ret - - - - Northern WT-1 - + - + Northern Pax-2 - - - + Northern Bmp-7 - - - + RT-PCR HoxC9 - + + - Northern Wnt-4 - - - + Northern E-Cadherin - - - + ICC Emx-2 + + + + RT-PCR Lim-1 - - - + Northern Evi-1 + + + + Northern HoxD11 + + + + Northern HoxA10 + + + + Northern BF-2 + + + + RT-PCR Pax-8 - - - + RT-PCR HoxA11 Anti + + + + Northern ICC, immunocytochemistry. HoxA11 Anti indicates the presence of the endogenous HoxA11 antisense transcripts (Hsieh-Li, et al., 1995).
Several additional features of the gene expression patterns are worthy of note.
WT-1 expression was seen in mK2 and at low levels in mK4 cells (Fig. 3). WT-1 expression is found in uninduced mesenchyme, stem cells, and condensing mesenchyme, but not in ureteric bud derivatives or in stromal cells (Armstrong et al., 1993). Hoxc!9 expression, a marker of stem cells and condensing mesenchyme, was only detected in
- 70 - mK2 and mK3 cells. Hoxc 9 is not expressed in stromal cells (Erselius et al., 1990;
Suemori et al., 1995). RT-PCR detected very low levels of BF-2 transcript in all of the kidney cell lines. This is somewhat surprising since BF-2 has been thought to be a kidney stromal cell marker (Hatini et al., 1996). And finally, all of the cell lines expressed Hoxa 11 antisense transcripts, which we have previously reported are expressed in the same kidney cells expressing Hoxa 11 sense RNA (Potter and Branford,
1998). mK3 cells retain metanephric mesenchyme biological function.
The gene expression analysis suggested that the mK1-4 cells represent mesenchymal cells developmentally arrested at an early stage of kidney development. To test for metanephric mesenchyme biological function we performed co-culture experiments. The mK3 cells were able to induce the growth and branching of the ureteric bud in culture (Fig. 4). Aggregates of cells for each cell line were formed in hanging drop culture and then positioned adjacent to isolated E11.5 ureteric bud in Matrigel matrix. Each cell line was tested at least twice and only the mK3 cells were able to support significant growth and branching of the ureteric bud. The mK1 cells promoted minimal growth and branching, while the presence of mK2 or mK4 cells resulted in ureteric buds no different from controls. For confirmation the mK3 co-culture experiments were repeated twenty times in seven independent experiments, with the same results. The mK3 cells were observed to promote ureteric bud growth and branching with or without GDNF added to the media. The experiments were repeated in Collagen I matrix to determine if the Matrigel matrix influenced the results. The results were again the same, suggesting that the mK3 cells alone were able to provide signals necessary for
- 71 - Figure 4. The mK3 cell line supports growth and branching of isolated ureteric buds in 3-D culture. (A) Two mK3 co-culture experiments and one control experiment are shown. The ureteric bud was placed next to mK3 aggregates in Matrigel with added GDNF. After one day the mK3 cells had started to grow and disperse, rapidly filling in the field of view. Several branches of the ureteric bud can been seen after seven days of culture. The control bud was placed in the same conditions without the mK3 cells and failed to maintain its integrity. (B) DBA-lectin staining of one week old co-cultures shows more clearly the growth and branching of the ureteric bud when cultured with the mK3 cell line. ureteric bud growth and branching. Fig. 4 shows the typical progression of the ureteric bud in co-culture with mK3 cells. To define the resulting ureteric bud more clearly, cultures were stained with DBA lectin, which binds to ureteric bud epithelia (Fig. 4). The extensive growth and branching from the initial “T” bud stage was clear. The control ureteric bud cultured alone always lost structural integrity and the cells dispersed or died.
Another control experiment was performed to insure that the growth of the ureteric bud was not the result of metanephric mesenchyme contamination during the dissection of the bud. Recombination experiments were done with ureteric buds dissected from ROSA26 mice, which ubiquitously express b-galactosidase, and the resulting cultures stained for b-galactosidase activity. No contaminating mesenchymal cells were detected, indicating that the growth and branching were the result of signals from the mK3 cells. These
- 72 - results demonstrate that the mK3 cells retain early metanephric mesenchyme properties and can substitute for metanephric mesenchyme in co-culture. The mK3 cells appear developmentally frozen, however, and do not form epithelia in response to the ureteric bud. We did not observe any condensations around the bud tips or any mature nephron structures.
Affymetrix GeneChip probe array analysis of the mK3 cells
In many respects the mK3 cells appear to represent the early metanephric mesenchyme of the developing kidney. Their morphology (Fig. 2), their expression of a number of metanephric mesenchyme genes (Table 1), and their ability to induce ureteric bud branching morphogenesis in organ co-culture (Fig. 4), are all consistent with this conclusion.
Affymetrix GeneChip probe arrays were used to create a molecular portrait of the transcription pattern of the mK3 cells. The experiments were performed in duplicate, using the Mu11K A and B chips, and using the u74A chip, providing expression data for over 11,000 known genes and ESTs. The expression data was culled based on reproducibility of the expression calls (called present in both experiments). Over four thousand expressed genes were identified. The complete list, presented in order of transcript abundance, can be found as an addendum at http://genome.chmcc.org/potter.
This work provides an extensive gene expression profile of the early kidney metanephric mesenchyme cell.
The Affymetrix GeneChip probe array data identified over 70 transcription factors, for example, expressed in the mK3 cells. Most are known, but some are only tentatively assigned transcription factor function based on structural motifs. The six Hox
- 73 - genes expressed, in order of decreasing transcript abundance, were Hoxc 9, Hoxa 11,
Hoxc 8, Hoxa 9, Hoxa 10 and Hoxa 5. The precise expression patterns of these Hox genes in the developing kidney has not been previously reported. Their expression in mK3 cells suggests they are expressed in early, uninduced, metanephric mesenchyme.
The cells also expressed ENX-1 and eed, members of the Polycomb group of genes that stabilize repression of inactive Hox genes, and HRX, a Trithorax family member gene involved in the positive regulation of Hox genes. In addition they expressed non- clustered homeobox genes such as Cux-1, which has previously been shown to be expressed in both the uninduced and condensed metanephric mesenchyme (Vanden
Heuvel et al., 1996), and Msx-1, which has been previously implicated in epithelial- mesenchymal interactions (Dassule and McMahon, 1998). The mK3 cells also expressed many other types of transcription factors, including GATA-3 and GATA-6, which have been associated with development of the glomerulus (Morrisey, 2000; Van Esch et al.,
2000), and Sox11, which is transcribed in both uninduced metanephric mesenchyme and in comma and S-shaped bodies (Hargrave et al., 1997). These experiments also identified the expression of dozens of other transcription factor genes that had not been previously implicated in kidney development.
The gene expression profile further confirmed the early metanephric mesenchyme character of the mK3 cells. Selected data is summarized in Table 2. The mK3 cells, for example, expressed collagen I and vimentin, markers of early mesenchyme, but did not express Wnt7b or Wnt-11, markers of ureteric bud, or Pax-8, Wnt4 or collagen IV, markers of late induced mesenchyme.
- 74 - Table 2. Brief summary of GeneChip analysis. Genes tested by RTPCR (RT) or Northern Blot (NB) are indicated. Gene Name mK3 mK4 Test Transcription Hoxa 7 - + Factors Pax-8 - + RT LFB3 - + LFB1 - - BF-2 - - RT* Hoxa 10 + + NB Hoxa 11 + + NB Hoxa 11Anti + + NB Hoxb 7 + + Hoxd 13 - - WT-1 - - NB mm.80595 - + RT Eya-2 - + RT Pax-2 - + NB Meis1b + - Peg-3 - + RT Signaling Wnt-4 - + NB Pathway Notch-1 - + Molecules Follistatin - + Igf-2 - + RT sFRP-1 - + RT PDGF A + + GDNF - + RT* HGF - - RT Neurturin - - FGF-7 + - Wnt-7 - - Wnt-11 - - FGF-2 - - Epimorphin + + Cell Cadherin 6 - + RT Adhesion Cadherin 16 - + RT Extracellular Fibronectin + + Matrix Col 4a-1 - + Molecules Col 4a-2 - + Col 1a-2 + - Syndecan-2 + + Fibulin-2 + - RT Cytoskeletal Shroom - + RT Elements Vimentin + - Transporters Aquaporin-1 - + RT KCC4 - +
- 75 - The Affymetrix chips were generally more sensitive than northern blots, but less sensitive than RT-PCR. BF-2, a stromal cell marker (Hatini et al., 1996), and GDNF, ligand for the ret receptor expressed in the ureteric bud, were listed as not expressed by the Affymetrix chips, although RT-PCR detected low abundance transcripts for both.
Gene Expression profile of mK4 cells.
The gene expression profile of the mK4 cells was also determined, in duplicate, with the same Mu11KA and B and U74A chips used for the mK3 cells. Over 4000 expressed genes were identified, with the complete list included in the addendum
(http://genome.chmcc.org/potter/).
The results further confirmed the late metanephric mesenchyme character of the mK4 cells. The gene chips identified expression of Pax-8, Pax-2, Wnt-4, Cadherin-6, collagen IV, and LFB3. All of these genes are expressed in the late induced condensing mesenchyme and forming nephron epithelia, but none are expressed in the early, uninduced mesenchyme. In addition the mK4 cells expressed some genes associated with more differentiated tubule cells, such as Aquaporin-1 and the cation-chloride transporter KCC4 (Mount et al., 1999). Nevertheless, it appears that expression of these genes can precede the fully differentiated state. Wnt-4, for example, is expressed in mK4 cells and is a marker of condensing cells, but not of differentiated cells of the nephron following fusion. The mK4 cells also express LFB3, which is first expressed in induced mesenchyme, but not LFB1, which appears later in S-shaped bodies (Lazzaro et al.,
1992). It is also interesting to note that Aquaporin-1 and Cadherin 6 are both markers of the proximal tubule, suggesting that mK4 cells represent precursors to proximal tubule cells (Cho et al., 1998; Mah et al., 2000). Cadherin 16 is associated with the renal tubule
- 76 - epithelia but excluded from mesenchymal and stromal cell populations (Thomson et al.,
1999). This further supports the epithelial character of the mK4 cells
Comparison of mK3 and mK4 gene expression profiles.
The mK3 cells represent early uninduced metanephric mesenchyme, while the mK4 cells represent later mesenchyme condensing to form epithelia. Comparison of their gene expression profiles could provide better understanding of the genetic basis of this differentiation process.
Reproducible gene expression differences between mK4 and mK3 were identified from the duplicate gene chip hybridization experiments and sorted according to fold change. The complete list is included in the online addendum
(http://genome.chmcc.org/potter) and the raw DAT image files are available on request.
Table 2 and Figure 5 include a short list of differences. In total there were 121 genes observed to change in expression level over ten fold, with 66 upregulated and 55 downregulated in the mK4 cells compared to the mK3 cells. We performed RT-PCR for selected genes to confirm the gene expression differences. 100% of tested genes identified in duplicate by the Affymetrix gene chip probe arrays as differently expressed were verified (Fig. 5).
Figure 5. RTPCR of identified expression differences confirms GeneChip comparison of the mK3 and mK4 cell lines.
- 77 - A number of genes known to change in expression during the condensation and epithelialization of mesenchyme were recognized. These included the Pax-8, Pax-2,
Wnt-4, Cadherin 6, Cadherin 16, collagen IV, and LFB3 genes mentioned above, which are expressed in late mesenchyme, but not early mesenchyme. This serves to validate the utility of the comparison.
Figure 6. Whole mount in situ hybridization shows the restricted expression patterns of expression differences identified by the GeneChip comparison.
In addition the comparison identified many genes not previously implicated in metanephric mesenchyme differentiation. This list included Hoxa 7, Notch-1, Follistatin,
Igf-2, sFRP-1, shroom, and many other genes, as well as ESTs. Some of these have been reported expressed in the kidney, but with their detailed developmental expression patterns not described (Davies and Brandli, 1997). Others have not been previously associated with kidney development. For two genes, Shroom and the Unigene cluster mm.80595, we performed whole mount in situ hybridizations to E13.5 kidneys.
- 78 - Hybridization signal was seen in each case in the condensing mesenchyme of the forming nephrons (Fig. 6).
- 79 - DISCUSSION
The mK3 cells represent early metanephric mesenchyme that has not yet been induced by the ureteric bud to condense and epithelialize. These cells were fibroblastic in appearance, not epithelial. They expressed genes diagnostic of uninduced mesenchyme, such as collagen I and vimentin, and did not express genes associated with induced mesenchyme. Most striking, however, the mK3 cells were able to induce ureteric bud growth and branching in co-culture, showing that they maintained early metanephric mesenchyme function.
The mK4 cells represent later, induced metanephric mesenchyme that has initiated epithelial conversion. This was shown by their more polygonal, epithelial morphology, and by their expression of a number of genes, including Pax-8, Pax-2, Wnt-
4, Cadherin-6, collagen IV, and LFB3, which are expressed in late, induced mesenchyme, but not early, uninduced mesenchyme.
Co-culture techniques have long been used to study metanephric mesenchyme- ureteric bud inductive interactions (Saxen, 1987). Recent applications have focused on the identification of specific signaling molecules, with results suggesting roles for GDNF,
Neurturin, HGF, EGF and their receptors (Ehrenfels et al., 1999; O'Rourke et al., 1999;
Qiao et al., 1999; Sakurai et al., 1997). We did not detect Neurterin or HGF expression in the mK3 cells. A very low level of GDNF expression was seen by RT-PCR, but was undetectable by Northern or GeneChip. Of interest, the mK4 cells, which did not induce ureteric bud growth and branching in co-culture, expressed a higher level of GDNF, readily detected by Affymetrix GeneChip. This suggests that the low level GDNF expression by mK3 cells is not solely responsible for their ability to induce bud
- 80 - branching. The molecular mechanism of bud induction by mK3 cells remains to be determined.
It is remarkable that the use of SV40 Tag allowed the isolation of cell lines with developing metanephric mesenchyme characteristics from the kidneys of 4-12 week old transgenic mice. This was likely the result of mosaic expression of SV40 Tag. Sufficient numbers of metanephric mesenchyme cells escaped SV40 Tag expression to allow the development of a functional kidney. Some cells, expressing SV40 Tag, were developmentally arrested within the transgenic kidney, and gave rise to the cell lines.
This could have resulted from mosaic Fo transgenic mice, with not every cell carrying the transgene. It has also been observed that gene repeats, such as found in transgene concatemers, can give highly variegated expression patterns resulting from poorly understood repeat induced gene silencing (Garrick et al., 1998). It should also be noted that different chromosomal integration sites could influence the timing of transgene expression in the different transgenic mice, resulting in developmental arrest of the cell lines at different stages of development.
In any event, the mK3 and mK4 cells appear to provide freeze frame pictures of the uninduced and induced metanephric mesenchyme during kidney development. These pure, clonal cells allow an intimate view of these developmental time points. The
Affymetrix gene chip analysis of the mK3 and mK4 cells generated a list of over 4,000 expressed genes, indicating kidney development function for a large number of ESTs and genes not previously implicated in this process. Furthermore, the mK3-mK4 expression profile comparison identified over one hundred genes showing more than a ten-fold difference in expression level. These are excellent candidates to play a role in the
- 81 - differentiation of the metanephric mesenchyme. This work moves us towards the ultimate goal of a full understanding of the gene expression patterns of early developing kidney cells and how these patterns change during nephrogenesis.
- 82 - ACKNOWLEDGEMENTS
We thank Affymetrix, and particularly the members of the academic user center, for their help. We also thank the CHMCC Transgenic Core for production of transgenic mice, and Dr. Mark Sussman for confocal imaging. This work was supported by NIH grant HD32061 to SSP.
- 83 - REFERENCES
Alarid, E. T., Holley, S., Hayakawa, M., and Mellon, P. L. (1998). Discrete stages of anterior pituitary differentiation recapitulated in immortalized cell lines, Mol Cell
Endocrinol 140, 25-30.
Alarid, E. T., Windle, J. J., Whyte, D. B., and Mellon, P. L. (1996). Immortalization of pituitary cells at discrete stages of development by directed oncogenesis in transgenic mice, Development 122, 3319-29.
Armstrong, J. F., Pritchard-Jones, K., Bickmore, W. A., Hastie, N. D., and Bard, J. B.
(1993). The expression of the Wilms' tumour gene, WT1, in the developing mammalian embryo, Mech Dev 40, 85-97.
Barasch, J., Yang, J., Ware, C. B., Taga, T., Yoshida, K., Erdjument-Bromage, H.,
Tempst, P., Parravicini, E., Malach, S., Aranoff, T., and Oliver, J. A. (1999).
Mesenchymal to epithelial conversion in rat metanephros is induced by LIF, Cell 99,
377-86.
Cho, E.A., Patterson, L.T., Brookhiser, W.T., Mah, S., Kintner, C., Dressler, G.R. (1998)
Differential expression and function of cadherin-6 during renal epithelium development,
Development 125, 803-12.
Cook, D. M., Hinkes, M. T., Bernfield, M., and Rauscher, F. J., 3rd (1996).
Transcriptional activation of the syndecan-1 promoter by the Wilms' tumor protein WT1,
Oncogene 13, 1789-99.
Dassule, H. R., and McMahon, A. P. (1998). Analysis of epithelial-mesenchymal interactions in the initial morphogenesis of the mammalian tooth, Dev Biol 202, 215-27.
- 84 - Davies, J. A., and Bard, J. B. (1998). The development of the kidney, Curr Top Dev Biol
39, 245-301.
Davies, J. A., and Brandli, A. (1997). The Kidney Development Database, World Wide
Web, URL: http://www.ana.ed.ac.uk/anatomy/database/kidbase/kidhome.html.
Davis, A. P., Witte, D. P., Hsieh-Li, H. M., Potter, S. S., and Capecchi, M. R. (1995).
Absence of radius and ulna in mice lacking hoxa-11 and hoxd-11, Nature 375, 791-5.
Dehbi, M., Ghahremani, M., Lechner, M., Dressler, G., and Pelletier, J. (1996). The paired-box transcription factor, PAX2, positively modulates expression of the Wilms' tumor suppressor gene (WT1), Oncogene 13, 447-53.
Dehbi, M., and Pelletier, J. (1996). PAX8-mediated activation of the wt1 tumor suppressor gene, Embo J 15, 4297-306.
Dolle, P., Izpisua-Belmonte, J. C., Brown, J. M., Tickle, C., and Duboule, D. (1991).
HOX-4 genes and the morphogenesis of mammalian genitalia, Genes Dev 5, 1767-7.
Dressler, G. R. (1996). Pax-2, kidney development, and oncogenesis, Med Pediatr Oncol
27, 440-4.
Dudley, A. T., Lyons, K. M., and Robertson, E. J. (1995). A requirement for bone morphogenetic protein-7 during development of the mammalian kidney and eye, Genes
Dev 9, 2795-807.
Ehrenfels, C. W., Carmillo, P. J., Orozco, O., Cate, R. L., and Sanicola, M. (1999).
Perturbation of RET signaling in the embryonic kidney, Dev Genet 24, 263-72.
- 85 - Erselius, J. R., Goulding, M. D., and Gruss, P. (1990). Structure and expression pattern of the murine Hox-3.2 gene, Development 110, 629-42.
Gagnon, F., Dulin, N. O., Tremblay, J., Hamet, P., and Orlov, S. N. (1999). ATP-induced inhibition of Na+, K+, Cl- cotransport in Madin-Darby canine kidney cells: lack of involvement of known purinoceptor-coupled signaling pathways, J Membr Biol 167, 193-
204.
Garrick, D., Fiering, S., Martin, D. I., and Whitelaw, E. (1998). Repeat-induced gene silencing in mammals [see comments], Nat Genet 18, 56-9.
Gendron, R. L., Paradis, H., Hsieh-Li, H. M., Lee, D. W., Potter, S. S., and Markoff, E.
(1997). Abnormal uterine stromal and glandular function associated with maternal reproductive defects in Hoxa-11 null mice, Biol Reprod 56, 1097-105.
Haack, H., and Gruss, P. (1993). The establishment of murine Hox-1 expression domains during patterning of the limb, Dev Biol 157, 410-22.
Hargrave, M., Wright, E., Kun, J., Emery, J., Cooper, L., and Koopman, P. (1997).
Expression of the Sox11 gene in mouse embryos suggests roles in neuronal maturation and epithelio-mesenchymal induction, Dev Dyn 210, 79-86.
Hatini, V., Huh, S. O., Herzlinger, D., Soares, V. C., and Lai, E. (1996). Essential role of stromal mesenchyme in kidney morphogenesis revealed by targeted disruption of Winged
Helix transcription factor BF-2, Genes Dev 10, 1467-78.
Hewitt, S. M., Hamada, S., McDonnell, T. J., Rauscher, F. J., 3rd, and Saunders, G. F.
(1995). Regulation of the proto-oncogenes bcl-2 and c-myc by the Wilms' tumor suppressor gene WT1, Cancer Res 55, 5386-9.
- 86 - Hobert, M. E., Friend, L. A., and Carlin, C. R. (1999). Regulation of EGF signaling by cell polarity in MDCK kidney epithelial cells, J Cell Physiol 181, 330-41.
Hogan, B., Beddington, R., Costantini, F., and Lacy, E. (1994). Manipulating the Mouse
Embryo, 2nd edn (Plainview, NY, Cold Spring Harbor Laboratory Press).
Hsieh-Li, H. M., Witte, D. P., Weinstein, M., Branford, W., Li, H., Small, K., and Potter,
S. S. (1995). Hoxa 11 structure, extensive antisense transcription, and function in male and female fertility, Development 121, 1373-85.
Igarashi, P. (1994). Transcription factors and apoptosis in kidney development, Curr
Opin Nephrol Hypertens 3, 308-17.
Jahroudi, N., Foster, R., Price-Haughey, J., Beitel, G., and Gedamu, L. (1990). Cell-type specific and differential regulation of the human metallothionein genes. Correlation with
DNA methylation and chromatin structure, J Biol Chem 265, 6506-11.
Kispert, A., Vainio, S., and McMahon, A. P. (1998). Wnt-4 is a mesenchymal signal for epithelial transformation of metanephric mesenchyme in the developing kidney,
Development 125, 4225-34.
Kreidberg, J. A., Sariola, H., Loring, J. M., Maeda, M., Pelletier, J., Housman, D., and
Jaenisch, R. (1993). WT-1 is required for early kidney development, Cell 74, 679-91.
Kuczuk, M. H., and Scott, W. J., Jr. (1984). Potentiation of acetazolamide induced ectrodactyly in SWV and C57BL/6J mice by cadmium sulfate, Teratology 29, 427-35.
Kupper, J. (1998). Functional expression of GFP-tagged Kv1.3 and Kv1.4 channels in
HEK 293 cells, Eur J Neurosci 10, 3908-12.
- 87 - Kuure, S., Vuolteenaho, R., and Vainio, S. (2000). Kidney morphogenesis: cellular and molecular regulation, Mech Dev 92, 31-45.
Lazzaro, D., De Simone, V., De Magistris, L., Lehtonen, E., and Cortese, R. (1992).
LFB1 and LFB3 homeoproteins are sequentially expressed during kidney development,
Development 114, 469-79.
Lee, S. B., Huang, K., Palmer, R., Truong, V. B., Herzlinger, D., Kolquist, K. A., Wong,
J., Paulding, C., Yoon, S. K., Gerald, W., et al. (1999). The Wilms tumor suppressor
WT1 encodes a transcriptional activator of amphiregulin, Cell 98, 663-73.
Lew, D., Brady, H., Klausing, K., Yaginuma, K., Theill, L. E., Stauber, C., Karin, M., and Mellon, P. L. (1993). GHF-1-promoter-targeted immortalization of a somatotropic progenitor cell results in dwarfism in transgenic mice, Genes Dev 7, 683-93.
Luo, G., Hofmann, C., Bronckers, A. L., Sohocki, M., Bradley, A., and Karsenty, G.
(1995). BMP-7 is an inducer of nephrogenesis, and is also required for eye development and skeletal patterning, Genes Dev 9, 2808-20.
Mah, S.P., Saueressig, H., Goulding, M., Kintner, C., Dressler, G.R. (2000) Kidney development in cadherin-6 mutants: delayed mesenchyme-to-epithelial conversion and loss of nephrons, Dev Biol. 223, 38-53.
Maheswaran, S., Park, S., Bernard, A., Morris, J. F., Rauscher, F. J. d., Hill, D. E., and
Haber, D. A. (1993). Physical and functional interaction between WT1 and p53 proteins,
Proc Natl Acad Sci U S A 90, 5100-4.
Mellon, P. L., Windle, J. J., and Weiner, R. I. (1991). Immortalization of neuroendocrine cells by targeted oncogenesis, Recent Prog Horm Res 47, 69-93.
- 88 - Mendelsohn, C., Batourina, E., Fung, S., Gilbert, T., and Dodd, J. (1999). Stromal cells mediate retinoid-dependent functions essential for renal development, Development 126,
1139-48.
Meroni, G., Reymond, A., Alcalay, M., Borsani, G., Tanigami, A., Tonlorenzi, R., Nigro,
C. L., Messali, S., Zollo, M., Ledbetter, D. H., et al. (1997). Rox, a novel bHLHZip protein expressed in quiescent cells that heterodimerizes with Max, binds a non-canonical
E box and acts as a transcriptional repressor [published erratum appears in EMBO J 1997
Oct 1;16(19):6055], Embo J 16, 2892-906.
Miyamoto, N., Yoshida, M., Kuratani, S., Matsuo, I., and Aizawa, S. (1997). Defects of urogenital development in mice lacking Emx2, Development 124, 1653-64.
Moore, M. W., Klein, R. D., Farinas, I., Sauer, H., Armanini, M., Phillips, H., Reichardt,
L. F., Ryan, A. M., Carver-Moore, K., and Rosenthal, A. (1996). Renal and neuronal abnormalities in mice lacking GDNF, Nature 382, 76-9.
Morrisey, E. E. (2000). GATA-6: The Proliferation Stops Here : Cell Proliferation in
Glomerular Mesangial and Vascular Smooth Muscle Cells, Circ Res 87, 638-640.
Mount, D. B., Mercado, A., Song, L., Xu, J., George, A. L., Jr., Delpire, E., and Gamba,
G. (1999). Cloning and characterization of KCC3 and KCC4, new members of the cation- chloride cotransporter gene family, J Biol Chem 274, 16355-62.
Muller, U., Wang, D., Denda, S., Meneses, J. J., Pedersen, R. A., and Reichardt, L. F.
(1997). Integrin alpha8beta1 is critically important for epithelial-mesenchymal interactions during kidney morphogenesis, Cell 88, 603-13.
- 89 - O'Rourke, D. A., Sakurai, H., Spokes, K., Kjelsberg, C., Takahashi, M., Nigam, S., and
Cantley, L. (1999). Expression of c-ret promotes morphogenesis and cell survival in mIMCD-3 cells, Am J Physiol 276, F581-8.
Pachnis, V., Mankoo, B., and Costantini, F. (1993). Expression of the c-ret proto- oncogene during mouse embryogenesis, Development 119, 1005-17.
Paul, D., Hohne, M., Pinkert, C., Piasecki, A., Ummelmann, E., and Brinster, R. L.
(1988). Immortalized differentiated hepatocyte lines derived from transgenic mice harboring SV40 T-antigen genes, Exp Cell Res 175, 354-62.
Pedersen, K. A., Schroder, R. L., Skaaning-Jensen, B., Strobaek, D., Olesen, S. P., and
Christophersen, P. (1999). Activation of the human intermediate-conductance Ca(2+)- activated K(+) channel by 1-ethyl-2-benzimidazolinone is strongly Ca(2+)-dependent,
Biochim Biophys Acta 1420, 231-40.
Pellegrini, M., Pantano, S., Lucchini, F., Fumi, M., and Forabosco, A. (1997). Emx2 developmental expression in the primordia of the reproductive and excretory systems,
Anat Embryol (Berl) 196, 427-33.
Phelps, D. E., and Dressler, G. R. (1996). Identification of novel Pax-2 binding sites by chromatin precipitation, J Biol Chem 271, 7978-85.
Potter, S. S., and Branford, W. W. (1998). Evolutionary conservation and tissue-specific processing of Hoxa 11 antisense transcripts, Mamm Genome 9, 799-806.
Qiao, J., Sakurai, H., and Nigam, S. K. (1999). Branching morphogenesis independent of mesenchymal-epithelial contact in the developing kidney, Proc Natl Acad Sci U S A 96,
7330-5.
- 90 - Rothenpieler, U. W., and Dressler, G. R. (1993). Pax-2 is required for mesenchyme-to- epithelium conversion during kidney development, Development 119, 711-20.
Sakurai, H., and Nigam, S. K. (1997). Transforming growth factor-beta selectively inhibits branching morphogenesis but not tubulogenesis, Am J Physiol 272, F139-46.
Sakurai, H., Tsukamoto, T., Kjelsberg, C. A., Cantley, L. G., and Nigam, S. K. (1997).
EGF receptor ligands are a large fraction of in vitro branching morphogens secreted by embryonic kidney, Am J Physiol 273, F463-72.
Santos, O. F., Barros, E. J., Yang, X. M., Matsumoto, K., Nakamura, T., Park, M., and
Nigam, S. K. (1994). Involvement of hepatocyte growth factor in kidney development,
Dev Biol 163, 525-9.
Santos, O. F., Moura, L. A., Rosen, E. M., and Nigam, S. K. (1993). Modulation of HGF- induced tubulogenesis and branching by multiple phosphorylation mechanisms, Dev Biol
159, 535-48.
Saxen, L. (1987). Organogenesis of the kidney (Cambridge, UK, Cambridge University
Press).
Schuchardt, A., D'Agati, V., Larsson-Blomberg, L., Costantini, F., and Pachnis, V.
(1994). Defects in the kidney and enteric nervous system of mice lacking the tyrosine kinase receptor Ret [see comments], Nature 367, 380-3.
Shawlot, W., and Behringer, R. R. (1995). Requirement for Lim1 in head-organizer function [see comments], Nature 374, 425-30.
Small, K. M., and Potter, S. S. (1993). Homeotic transformations and limb defects in Hox
A11 mutant mice, Genes Dev 7, 2318-28.
- 91 - Stark, K., Vainio, S., Vassileva, G., and McMahon, A. P. (1994). Epithelial transformation of metanephric mesenchyme in the developing kidney regulated by Wnt-
4, Nature 372, 679-83.
Suemori, H., Takahashi, N., and Noguchi, S. (1995). Hoxc-9 mutant mice show anterior transformation of the vertebrae and malformation of the sternum and ribs, Mech Dev 51,
265-73.
Thomson, R.B., Aronson, P.S. (1999) Immunolocalization of Ksp-cadherin in the adult and developing rabbit kidney, Am J Physiol. 277, F146-56.
Tong, J., Du, G. G., Chen, S. R., and MacLennan, D. H. (1999). HEK-293 cells possess a carbachol- and thapsigargin-sensitive intracellular Ca2+ store that is responsive to stop- flow medium changes and insensitive to caffeine and ryanodine, Biochem J 343 Pt 1, 39-
44.
Torban, E., and Goodyer, P. R. (1998). Effects of PAX2 expression in a human fetal kidney (HEK293) cell line, Biochim Biophys Acta 1401, 53-62.
Torres, M., Gomez-Pardo, E., Dressler, G. R., and Gruss, P. (1995). Pax-2 controls multiple steps of urogenital development, Development 121, 4057-65.
Tsutsumi, K., Belaguli, N., Qi, S., Michalak, T. I., Gulliver, W. P., Pater, A., and Pater,
M. M. (1992). Human papillomavirus 16 DNA immortalizes two types of normal human epithelial cells of the uterine cervix, Am J Pathol 140, 255-61.
Van Esch, H., Groenen, P., Nesbit, M. A., Schuffenhauer, S., Lichtner, P., Vanderlinden,
G., Harding, B., Beetz, R., Bilous, R. W., Holdaway, I., et al. (2000). GATA3 haplo- insufficiency causes human HDR syndrome, Nature 406, 419-22.
- 92 - Vanden Heuvel, G. B., Bodmer, R., McConnell, K. R., Nagami, G. T., and Igarashi, P.
(1996). Expression of a cut-related homeobox gene in developing and polycystic mouse kidney, Kidney Int 50, 453-61.
Wick, M., Zubov, D., and Hagen, G. (1999). Genomic organization and promoter characterization of the gene encoding the human telomerase reverse transcriptase
(hTERT), Gene 232, 97-106.
Windle, J. J., Weiner, R. I., and Mellon, P. L. (1990). Cell lines of the pituitary gonadotrope lineage derived by targeted oncogenesis in transgenic mice, Mol Endocrinol
4, 597-603.
Xu, P. X., Adams, J., Peters, H., Brown, M. C., Heaney, S., and Maas, R. (1999). Eya1- deficient mice lack ears and kidneys and show abnormal apoptosis of organ primordia,
Nat Genet 23, 113-7.
- 93 - Chapter 4
Hoxa 11 downstream target candidates identified by
expression profile analysis.
M. Todd Valerius, Yuxin Feng, and S. Steven Potter
Planned submission to PNAS.
- 94 - ABSTRACT
Homeobox genes are critical developmental transcriptional regulators that are thought to act on batteries of downstream targets. Because of the high degree of conservation among the cluster members and the shared DNA binding consensus site,
Hox genes interact with cofactors that confer further DNA binding specificity. These properties have slowed the identification of the downstream targets of Hox genes. We present a strategy using cell lines and expression profile comparisons to identify candidate Hoxa 11 downstream targets. A Hoxa 11/Hoxd 11 double mutant kidney cell line was established using a Hoxa 11-SV40 Large T antigen transgene. These cells were then transfected with a Hoxa 11 expression construct to create a “plus” or “rescued” mutant cell line to be compared with non-transfected mutant cells. Differential Display and Gene Discovery Arrays were used to identify gene expression differences due to
Hoxa 11 expression in the mutant kidney cell line. Another kidney cell line, 293 Tet-Off, was used to express an epitope tagged Hoxa 11 cDNA under the control of the Tet
Transactivator (tTA). Treated and untreated 293 – Tet-Hoxa 11 cells were profiled using
Affymetrix GeneChips. Collectively, nine candidate downstream targets were identified from the three screens. Integrin Alpha 8 (ItgA8) showed that most dramatic expression difference in both the 293 cells and the Hoxa 11/Hoxd 11 mutant kidney cells. ItgA8 expression was then examined in E13.5 Hoxa 11/Hoxd 11 mutant kidneys and found to have an aberrant expression pattern consistent with being under transcriptional control by
Hoxa 11.
- 95 - INTRODUCTION
Homeobox genes encode important developmental transcription factors that all share the highly conserved homeodomain protein motif. In Drosophila, homeobox genes are organized in a split cluster and act as master regulators specifying body segment identity along the A/P axis of the embryo. In mammals there are 39 homeobox genes assembled into 4 similar clusters thought to have arisen from a common ancestral homeobox complex. It is believed that the duplication and expansion of the Hox genes allowed a more sophisticated utilization of Hox gene function for the development of more elaborate structures. Hox genes function by controlling the expression of groups of downstream targets. In vitro DNA-binding studies have determined the consensus binding sequence for Hox genes is rather small, consisting of a TAAT core with some weak additional nucleotides (Benson et al., 1995; Pellerin et al., 1994). In some cases, functional specificity is determined by cofactors or by cooperative DNA-binding, effectively expanding the consensus site or favoring one Hox protein over another
(Castelli-Gair, 1998; Li et al., 1999b; Shen et al., 1997; Shen et al., 1999). Our limited understanding of the DNA-binding specificity of Hox proteins and how this influences gene expression has hindered the search for downstream genes, a critical step to understand how Hox genes carry out their role and to map the genetic cascades of organogenesis.
In the developing kidney Hoxa 11 is expressed in the metanephric mesenchyme prior to bud invasion and continues in the condensing mesenchyme around the ureteric bud tips (Hsieh-Li et al., 1995; Patterson and Potter, 2000). As the condensing mesenchyme further differentiates and some cells undergo a mesenchymal to epithelial
- 96 - conversion, Hoxa 11 expression is lost, suggesting an early role for Hoxa 11 in the process of nephrogenesis. Gene knockout studies have demonstrated a redundant role for the paralogous Hox genes Hoxa 11 and Hoxd 11 in kidney development (Davis et al.,
1995). Hoxd 11 is expressed in overlapping domains in the mesenchyme and neither gene is expressed in the ureteric bud or its derivatives (Patterson and Potter, 2000).
Though the single knockout of each gene has no detectable renal phenotype, the Hoxa
11/Hoxd 11 double knockout results in renal agenesis or the formation of a small kidney rudiment that contains a very reduced number of mature nephrons (Davis et al., 1995).
The double null mutants have defects in nephrogenesis and in branching morphogenesis of the ureteric bud indicating the defect caused by the double knockout alters the interaction between the mesenchyme and the ureteric bud (Patterson and Potter, 2000).
The complexity of the kidney and the highly restricted expression pattern of transcription factors represent major hurdles in studying the role of genes like Hoxa 11.
To overcome this problem and simplify the system we chose to use a cell culture approach that provided a homogeneous cell population and the ability to modify the cells by transfection with expression constructs. The use of cell lines and organ culture in the study of kidney development has been successful in identifying several important growth factors able to support branching morphogenesis (Barasch et al., 1999; Santos et al.,
1994). We previously described a directed oncogenesis approach using the Hoxa 11 promoter to drive expression of SV40 Large Tag in transgenic mice (Valerius et al.,
2000). This transgene was bred onto the Hoxa 11/Hoxd 11 double mutant background to establish immortalized cell lines from mutant kidneys. These cells serve as a model of
Hoxa 11/Hoxd 11 mutant kidney tissue.
- 97 - We have sought to better understand Hoxa 11 function by identifying potential downstream gene targets using universal screens. We report the modification of a Hoxa
11/Hoxd 11 double mutant kidney cell line by transfection to regulate Hoxa 11 expression and create Hoxa 11 plus and minus cell populations. This modified mutant cell line was used for Differential Display and Gene Discovery Array (GDA) analysis. A human embryonic kidney cell line (HEK293) was used with a Tet-inducible Hoxa 11 cDNA to screen for gene expression differences using Affymetrix GeneChips. These screens have identified several candidate downstream targets of Hoxa 11 including transcription factors, growth factors, and cell membrane genes. Integrin alpha 8 (ItgA8) was up regulated in both cell lines in response to Hoxa 11 expression. In situ hybridization analysis of ItgA8 in Hoxa 11/Hoxd 11 mutant kidney reveals ItgA8 expression is altered in the double mutants.
METHODS
Isolation and establishment of kidney cell lines
Hoxa 11-SV40 Large Tag transgenic mice were bred with Hoxa 11+/- Hoxd
11+/- double heterozygous mice. Triple heterozygous mice (Hoxa 11+/- Hoxd 11+/-
Hoxa 11-SV40 Large Tag+) were bred to produce embryos of the desired genotype.
Kidney tissue was carefully dissected from E18.5 embryos, dissociated with trypsin, and the cells plated on 100 mm tissue culture dishes. Genotype was determined by PCR on head tissue from the embryo. The cells were cultured in DMEM (Gibco BRL, Rockville,
o MD) with 10% Fetal Bovine Serum (FBS) (Gibco, Rockville, MD) at 37 C in 5% CO2 on standard tissue culture plates (Falcon/BDL, Franklin Lakes, NJ). Media was changed every 2-3 days and cells were split when confluent. After 10+ passages, clones were
- 98 - established by dilution cloning and have subsequently been grown for over 20 additional passages, suggesting these clones are immortalized. HEK293 Tet-Off cells were purchased from Clontech (Palo Alto CA) and maintained according to manufacturer’s protocol.
Modification of cell lines
The plasmid pcDNA 3.1/Hygro (Invitrogen, Carlsbad CA) was used to drive the expression of full length Hoxa 11 cDNA in mK10 cells. For the 293 Tet-Off cells, an epitope tagged Hoxa 11 cDNA was cloned into the pBI Tet expression vector (Clontech) and introduced by transfection. One day after calcium chloride transfection, cells were subjected to Hygromycin selection and colonies picked later. FLAG epitope tagging was introduced with PCR using the following primers: 5'-
CCCCAAGCTTCCACCATGGCTCCAAAGAAGAAGCGTAAGGTAGACTACAAGG
ACGACGATGACAAGGATTTTGATGAG-3' and 5’-
AGGAAGCTTAACCACGGAGATCTG-3’. The FLAG epitope was detected with the
M5 monoclonal antibody (Sigma) according to manufacturers protocol. Doxycycline
(2ug/mL) was added to the media for two days with media changes every 12 hours to regulate Hoxa 11 expression in the 293 Tet-Off cells.
Gene expression difference profiles
Differential display was performed by manufacturers protocol (GeneHunter
Corporation, Nashville, TN). The first 5 pairs of randomly designed hexamers were used on each sample and done in triplicate. PCR products were cloned into the pTRAP vector
(GeneHunter Corporation, Nashville, TN) after isolation from a PAGE gel. For Gene
Discovery Arrays (Genome Systems), probe labeled from cell line mRNA using a poly
- 99 - dT oligonucleotide in a reverse transcription reaction with 33P-UTP was used to probe the arrays according to manufacturer’s protocol. Resulting phosphoimage files were sent to
Genome Systems for analysis. Affymetrix Human 6800 gene chip probe arrays were hybridized with biotinylated cRNA prepared according to the Affymetrix protocol.
Northern blots and RTPCR
Total RNA was isolated from trypsinized cells using RNAzol (TelTest,
Friendswood, TX). Poly A+ selection was done using the Oligotex mRNA isolation kit
(Qiagen, Carlsbad, CA). Reverse transcription was performed with Superscript II (Gibco
BRL) and a random hexamer mix according to manufacturer’s directions. Following reverse transcription, PCR was done using Taq polymerase (Qiagen, Carlsbad, CA).
Semiquantitative RTPCR was done by adding 1uCi of 32P-dATP to each 50ul PCR reaction. The linear amplification phase was determined for each primer pair by checking incorporation at 18, 20, 22, 24, and 26 cycles. Gels were exposed to phosphoimager screens and the plots of intensity vs cycle number were generated. For each primer pair, the fewest cycles necessary for detection was selected for the comparison analysis to ensure measurements in the linear range. All gene expression was normalized to GAPDH. Primers used: Hoxa 11 RT 5'
AAAACCTCGCTTCCTCCGACTACC 3' 5' CGCAATGTGGCTTGACCTTGTC 3' ;
GCMa 5'- GAACCTGACGACTCTGATTCTGAAG -3' 5'-
TCACAGTTGGGACAGCGTTTCC -3' ; ItgA8 5'
AGCTACTTCGGCTACTCACTGGAC -3' 5'- TCCTCCCACTATAAGGTCTCCATTC
3' ; PDGF-A 5'- GTGCTTTATTGCCAGTGTGCG -3' 5'-
- 100 - AAGACATTCCTGCTTCCTGCG -3' ; GAPDH 5’- ACCACAGTCCATGCCATCAC-
3’ 5’- TCCACCACCCTGTTGCTGTA -3’
In situ hybridization
The genitourinary (GU) block was isolated from the E13.5 embryos from a Hoxa
11/Hoxd 11 double heterozygous mating. Head tissue from each embryo was used to determine the genotype by PCR. Double mutant and littermate control GU blocks were probed according to established protocols (Hogan et al., 1994). The images in Fig.6 were constructed using Adobe Photoshop (Adobe Systems, San Jose, CA).
RESULTS
Strategy
Hoxa 11 is primarily expressed transiently in the condensing mesenchyme of the developing kidney. This cell population represents a portion of all the cells present in the kidney. Therefore the representation of RNA from Hoxa 11 expressing cells is under- represented in RNA samples isolated from whole kidneys, yet this specific RNA population is what we are trying to study. Furthermore, cells undergoing condensation also begin the process of converting from mesenchyme to epithelia, presumably due to gradual changes in gene expression. The fractional representation combined with potential temporal differences of gene expression thus requires a fine dissection to enrich for the appropriate tissues. To overcome this problem we decided to use an alternative approach utilizing cell lines that represent specific Hoxa 11 expressing tissues (Figure 1).
This also permitted modification of the cells by transfection. We previously reported the establishment of several cell lines from Hoxa 11/SV40 transgenic mice (Valerius et al.,
- 101 - 2000). These transgenic mice express the SV40 large T antigen gene from a Hoxa 11 promoter, directing expression to Hoxa 11 expressing cells and permitting the
Express SV40 Large Tag in transgenic mice using the Hoxa 11 promoter
Make Hoxa 11/SV40 transgenic mouse
“wt” Transgenic Hoxa 11-SV40
Breed onto double mutant background
Double mutant Hoxa 11-/- Hoxd 11-/- Transgenic Hoxa 11-SV40
Isolate cells from appropriate Hoxa 11 expressing tissues
Hoxa 11-/-;Hoxd 11-/- double mutant cell line
Add cmv-Hoxa 11 to create + and - Add Hoxa 11 RNA populations Pcmv - Hoxa 11 in mutant cell line
Figure 1. Strategy for creating appropriate cell lines as tissue models for the Hoxa 11 gene. establishment of these cell lines in culture. This transgene was bred onto the Hoxa
11/Hoxd 11 double mutant background to generate embryos mutant for both endogenous genes (Hoxa 11 -/- and Hoxd 11 -/-) and positive for the transgene (Hoxa 11/SV40 +).
From the mutant mice a kidney cell line was established and used in the subsequent
- 102 - studies as the “minus“ cell population, that is, minus Hoxa 11 expression. This provides a model of Hoxa 11/Hoxd 11 double mutant kidney tissue. Using an expression construct, Hoxa 11 expression was then added back to these cells by transfection to create the “plus” cell population. The use of these cell lines permitted the isolation of enough
RNA for expression profiling and Northern blots. The plus and minus expression profiles were compared to identify those genes responding to the introduction of Hoxa 11 expression.
Establishment of Hoxa 11/Hox d11 double mutant cell lines
Previously we used a Hoxa 11/SV40 construct in transgenic animals to establish cell lines from various tissues including the uterus, the ductus deferens, and the kidney, all sites of Hoxa 11 expression (Chapter 3). These transgenics were first bred with Hoxa
11 +/- Hoxd 11 +/- double heterozygotes to introduce the transgene into the mutant line.
Double heterozygotes were used because of the further reduced fertility when three alleles (e.g. Hoxa 11 +/- Hoxd 11 -/-) are missing. The Hoxa 11/SV40 mice also suffer from reduced fertility due to the transgene and this problem appeared additive when the transgene was bred onto the mutant background resulting in reduced litter sizes and failed pregnancies. Triple heterozygotes (Hoxa 11+/- Hoxd 11+/- Hoxa 11/SV40+) were bred to produce the desired genotype of mutant null for both Hox genes and positive for the
Hoxa 11/SV40 allele (Hoxa 11-/- Hoxd 11-/- Hoxa 11/SV40+). Embryos were isolated at
E18.5 and mutants initially identified by the forelimb phenotype previously described
(Davis et al., 1995). Kidneys were dissected from each embryo, trypsinized, and cultured in standard tissue culture conditions. The phenotype of the mutant kidneys was as previously described with a small, rudimentary kidney present with a normal adrenal
- 103 - Figure 2. Genotype and morphology of mK10 double mutant cells. (A) Mutant embryos used for kidney cell isolation were genotyped by PCR. (B) Gross morphology of the mutant mK10 cells and HEK293 Tet-Off cells. Note the flattened appearance of the fibroblast like mutant cells compared to the rounded epithelial HEK293 cells
gland also present (Davis et al., 1995). DNA was isolated from head tissue and used for genotyping the embryos. Figure 2A shows the results of PCR genotyping of two embryos. The positive control used was a double heterozygote with the SV40 transgene
(Hoxa 11+/- Hoxd 11+/- Hoxa 11-SV40+). The Hoxa 11 targeting is a deletion and results in a smaller band when the mutant allele is present. The Hoxd 11 targeting is an insertion and results in a larger product from the mutant allele. The negative control is without template added. Both of these animals were double mutant however only one also harbored the Hoxa 11-SV40 transgene and kidney cells from this embryo were then maintained in culture. Initially the culture appeared heterogeneous but after 10 passages a dominant cell type was evident. At this point dilution cloning was done to establish a single line of cells for study. One clone, mK10, was selected for use in subsequent studies. The double mutant cells have a very irregular shape in culture, appearing fibroblastic with long extensions and flatten as the cells stretch out on the plate (Fig. 2B).
- 104 - These cells have been cultured for over 30 passages, suggesting they are immortalized.
Surprisingly, despite the confirmed presence of the transgene in mK10 cell lines, RTPCR analysis was unable to detect expression of SV40 Large Tag (data not shown). Similar instances, loss of SV40 expression but maintained imortalization, have been reported in the literature (Mora et al., 1986; Salewski et al., 1999). However, in these cases the cell lines were transformed using SV40 virus. Furthermore, the inserted DNA was completely lost from the genome, indicating rearrangements in the DNA. In our case the
SV40 transgene is still present though we have lost expression.
Modification of the cell lines to create + and - Hoxa 11 populations
The mK10 cells were then transfected with a Pcmv-Hoxa 11 expression construct that drives expression of Hoxa 11 from the CMV promoter. Transfected cells were selected for presence of the Hygromycin resistance gene present in the construct and single colonies were picked and cultured. These transfected clones were tested for Hoxa
11 expression. The clone with the most robust expression was selected for future studies as the “plus Hoxa 11” cell population (mK10/Pcmv-Hoxa 11) (Fig.5).
To complement our screen using the mutant cells the human kidney cell line
HEK293 was also used for screening. This cell line does not express detectable levels of
Hoxa 11 (data not shown). HEK293 cells are epithelial cells used extensively for the functional study of channel and transporter proteins (Kupper, 1998; Lee et al., 1999;
Pedersen et al., 1999). They have also been used to study the effect of Pax-2 expression on the expression of key kidney development markers (Torban and Goodyer, 1998), a study similar to our own in principle. We decided to use Tet-inducible HEK293 cells in our study of Hoxa 11 function. A FLAG-tagged Hoxa 11 cDNA was cloned into the Tet-
- 105 - responsive vector and transfected into the modified HEK 293 cells that express the Tet-
Transactivator (tTA), referred to as the “Tet-Off” system. Clones were picked after
Hygromycin selection and FLAG-Hoxa 11 expression levels were tested. In the Tet-Off system the addition of doxycycline to the culture media blocks the tTA protein from binding to the promoter sites in the expression vector. Removal of doxycycline then permits binding and activation of FLAG-Hoxa 11. The inclusion of the FLAG tag on the
N-terminal end of the Hoxa 11 protein permitted Western analysis of Hoxa 11 expression with commercially available monoclonal antibodies against the FLAG epitope (Sigma).
Figure 3 shows a representative induction of FLAG-Hoxa 11 in the HEK293 cells with an
Actin loading control. Small amounts of Hoxa 11 expression were always detectable in this system as has been reported previously (REF). RNA was isolated from Hoxa 11 +
(no doxycycline) and - (with added doxycycline) for expression profiling and comparisons.
Figure 3. Western blot analysis of 293 Tet-Off-FLAG-Hoxa 11 cells. Epitope tagged Hoxa 11 expression was detected with the M5 monoclonal antibody. Actin was probed as a loading control.
+ and – Hoxa 11 cell line populations were used in screens for Hoxa 11 responsive genes
The double mutant cell line was used for two differential screens, Differential
Display and Gene Discovery Array (GDA) analysis. Differential display is a PCR based
- 106 - approach used to compare the RNA fingerprint of two or more RNA sources. This approach has been used to identify gene expression differences in modified cell lines (Jo et al., 1998; Li et al., 1999a; Zhang et al., 2000) and induced and uninduced kidney mesenchyme in culture (Leimeister et al., 1999; Plisov et al., 2000). RNA from mK10 cells and mK10/Pcmv-Hoxa 11 cells was isolated and used for PCR in Differential
Display with the first five sets of randomly generated anchor oligonucleotides provided by the manufacturer. Each PCR was done with three different RNA isolations to reduce sample variability. Over 100 differences were identified and initially tested by reverse
Northern (RNA labeled, PCR products blotted) which identified 33 clones that appeared differentially expressed. Northern blots of cell line mRNA were then probed with each of these to confirm expression level differences. Resulting blots were later stripped and probed with GAPDH as a loading control. Quantization was done with a phosphoimager, normalizing the expression to GAPDH expression to then produce the ratios of expression. Most expression differences had very small changes in expression level and only two clones proved differentially expressed in repeated experiments at a 2 fold difference, both down regulated with the addition of Hoxa 11 expression to the mK10 cells (Table 1 and Figure 4). Cathepsin L is a widely expressed cysteine protease found in lysosomes that is sometimes secreted by expressing cells (Ishidoh and Kominami,
1998). Cathepsin L has been implicated in tumor invasion (Dohchin et al., 2000), trophoblast invasion during implantation (Afonso et al., 1997; Hemberger et al., 2000), and endochondral bone formation (Nakase et al., 2000). Annexin I (Lipocortin I) is a member of the calcium- phospholipid-binding family of proteins thought to be involved in signal transduction (Alldridge et al., 1999). Annexin I is expressed in restricted
- 107 - epithelial kidney structures in the adult including Bowman’s capsule, macula densa, and collecting ducts (McKanna et al., 1992).
The double mutant cell line was also used in a cDNA array screen using Gene
Discoveries Arrays (GDA). These arrays consist of over 18,000 cDNA clones blotted
Figure 4. Northern blot and semi-quantitative RTPCR analysis of candidate Hoxa 11 downstream targets. (A) Northern blots confirm expression differences in mutant mK10 cells (top row) with and without a Hoxa 11 expression construct and 293 Tet-inducible Hoxa 11 treated and untreated with doxycycline (bottom row). “+” indicates with expression of Hoxa 11. Blots were stripped and then re- probed with GAPDH as a loading control. (B) Semi-quantitative RTPCR was also used to confirm differences in the 293 Tet-Off Hoxa 11 cells. onto a nylon filter. Duplicate arrays are probed with 33P labeled cDNA made by reverse transcription of each mRNA sample. Analysis of the phosphoimage files by Genome
Systems identified 705 cDNAs altered at least two fold between the + and – Hoxa 11 sample (420 up, 285 down). Each blotted cDNA was examined manually to eliminate blotting errors that may lead to incorrect expression level reports by the software. Ten of these cDNAs with the greatest fold change were selected for confirmation by Northern blot analysis. Expression levels were normalized to GAPDH expression as above using a
- 108 - phosphoimager. Two clones showed a reliable expression difference though very modest
(Table 1). One clone represents an EST cluster designated Mm.24736 that contains weak homology to EHD1, an EH domain-containing protein expressed in the developing limb buds, testis, and kidney (Mintz et al., 1999). EH domain proteins are implicated in signaling and ligand induced endocytosis (Tong et al., 2000). The Est2 repressor factor
(ERF) gene is an ets-domain transcription factor expressed widely in the adult (Liu et al.,
1997) that acts as an antagonist to other ets-domain transcriptional activators (Sgouras et al., 1995).
Using the Tet-inducible-Hoxa 11 in HEK293 cells, RNA was prepared from induced and uninduced cell populations and used for Affymetrix GeneChip analysis on the Human 6800 gene set. A total of 309 genes (155 up regulated and 154 down) were differentially expressed at 1.5 fold or greater. The greatest fold differences were no greater than 2.8 fold in either direction. This is perhaps expected do to the small amount of Hoxa 11 expression in the uninduced cells due to leakiness of the system. Genes were selected for confirmation by quantitative Northern blot or quantitative RTPCR (Figure 4 and Table 1). The fold change for each of the confirmed expression differences was generally similar to the calculated fold change on the GeneChip. Integrin alpha 8 (ItgA8) is a cell membrane protein that interacts with the extracellular matrix (ECM). Integrins regulate cell growth and differentiation by interacting with several important signaling pathways, creating a link between the ECM environment and cell behavior (Boudreau and Jones, 1999). Platelet Derived Growth Factor A chain (PDGF-A) is a well- characterized growth factor expressed in epithelial cells in many tissues including the kidney. There, PDGF-A expression is initiated in newly formed epithelial cells as they
- 109 - are formed from condensed mesenchymal cells around the ureteric bud tips while the receptors for PDGF are expressed in mesenchymal cells (Seifert et al., 1998). Hoxa 11 is expressed in the condensing mesenchymal cells that are beginning the epithelialization process and shuts down as the cells differentiate further. If Hoxa 11 were responsible for repressing PDGF-A in mesenchymal cells, directly or otherwise, the same repression would be expected when the mesenchymal Hoxa 11 is expressed in an epithelial cell type
Figure 5. ItgA8 expression is up regulated in double mutant mK10 cells upon expression of Hoxa 11. (A) Standard RTPCR indicated ItgA8 expression was increased with expression of Hoxa 11. (B) Semi-quantitative RTPCR confirmed the strong increase in ItgA8 expression with the expression of Hoxa 11. Intensities were normalized to GAPDH.
such as HEK293 cells. Glial Cells Missing alpha (GCM1 in mouse) is a transcription factor cloned by homology to the Drosophila gene. Though responsible for neuronal cell fate decisions in the fly, both known mammalian homologs are expressed in E16.5 kidneys and in the placenta suggesting other roles in mammalian development (Kim et al., 1998). Gene knockout of GCM1 disrupted branching morphogenesis of trophoblast populations in the placenta, resulting in lethality at E10 prior to kidney formation
(Anson-Cartwright et al., 2000). Neuronal cell adhesion molecule (NrCAM) plays an important role in the growth cone during neural development (Davis and Bennett, 1994).
- 110 - Though no role for NrCAM in kidney development has been ascribed, other neural cell adhesion molecules have been described in the ureteric epithelium and ductal epithelial portions of the nephron (Debiec et al., 1998). Integrin associated protein (IAP/CD47) is a membrane protein important in the inflammatory response to infection (Lindberg et al.,
1996). Zinc finger protein 192 is a Kruppel family member expressed in the adult kidney
(Lee et al., 1997).
Table 1. Confirmed gene expression differences from each screen. The cell lines used in each screen are indicated. “+” indicates increased expression with Hoxa 11 expression. Screen/Cells Used Unigene Cluster Description Fold Chg. Differential Display Mm.930 Cathepsin L -2.0 mK10 & Pcmv-Hoxa 11 Mm.14860 Annexin A1 -2.0 Gene Discovery Array Mm.24736 ESTs -1.8 mK10 & Pcmv-Hoxa 11 Mm.8068 Est2 repressor factor -1.4 Hs.91296 Integrin a8 +23.0 Hs.37040 PDGF-A -2.0 Affymetrix GeneChip Hs.28346 GCMa +1.6 Tet inducible Hoxa 11/ HEK 293 Hs.7912 hBRAVO/NrCAM + GB Z25521 IAP/CD47 + Hs.57679 Zinc finger protein 192 +
Integrin Alpha 8 is upregulated upon expression of Hoxa 11 in double mutant mK10 cells
Integrin alpha 8 (ItgA8) was identified as a putative Hoxa 11 gene target in the expression profile comparison of FLAG-Hoxa 11/HEK293 cells. To further characterize this activity, ItgA8 expression was examined in the double mutant mK10 cells with and without Hoxa 11 (Pcmv-Hoxa 11). Standard PCR revealed an increase in ItgA8 expression in the mK10/PcmvHoxa 11 cells compared to the non-transfected mK10 cells without Hoxa 11 expression (Figure 5A). Semi-quantitative RTPCR was then performed and revealed ItgA8 expression was 23 fold increased in the Hoxa 11 positive cells (Figure
5B). ItgA8 is expressed in overlapping patterns with Hoxa 11 in the developing kidney,
- 111 - notably the condensing mesenchyme during early nephrogenesis (Muller et al., 1997).
Both Hoxa 11 and ItgA8 down regulate as the condensing mesenchyme begins to convert to the epithelial cells of the nephron. The overlapping expression pattern and timing in the developing kidney supported the cell line data suggesting ItgA8 is a downstream target of Hoxa 11 in the kidney.
Integrin Alpha 8 expression is altered in Hoxa 11/Hoxd 11 double mutant kidneys
The gene knockout of ItgA8 exhibited a renal phenotype remarkably similar to the renal phenotype observed in the Hoxa 11/Hoxd 11 double mutants (Davis et al., 1995;
Muller et al., 1997). The kidneys, when present, were small rudiments that contained few functional nephrons. Furthermore, branching morphogenesis was aberrant in the ItgA8 mice, similar to that reported in the Hoxa 11/hoxd 11 double mutants (Patterson and
Potter, 2000). If Hoxa 11 plays a critical role in the expression of ItgA8, we would predict the expression pattern might be altered in the Hoxa 11/Hoxd 11 double mutants.
Figure 6. ItgA8 shows an altered expression pattern in Hoxa 11/Hoxd 11 double mutant kidney. Whole mount in situ hybridization was done on E13.5 mutant and littermate wild type control to examine Itg expression. Note the missing/reduced expression on the ventral side of the mutant kidneys.
- 112 - To test this, E13.5 kidneys were isolated from Hoxa 11/Hoxd 11 double mutant embryos and wild type littermates and examined for ItgA8 expression. ItgA8 was altered in the double mutants most severely on the ventral side of each kidney (Figure 6). In the wild type, ItgA8 expression can be seen in the condensing mesenchyme around the ureteric bud tips. While some condensations are present and express ItgA8 in the mutant, there are large regions that lack expression. This was most evident on the ventral side of the kidney (Figure 6). The formation of reduced numbers of nephrons in the Hoxa 11/Hoxd
11 mutants is presumably due to the redundancy of the Hox family (Davis et al., 1995).
The altered expression of ItgA8 in the mutant kidney is consistent with ItgA8 being a target of Hoxa 11 function.
- 113 - DISCUSSION
To map the transcriptional cascade controlled by Hox genes several efforts have been made to identify the gene targets of homeobox proteins. Some progress has been made in Drosophila and in mammals using approaches such as immunoprecipitation
(Chauvet et al., 2000; Gould and White, 1992; Graba et al., 1992). Other successful approaches include gel shift experiments (Care et al., 1996) and co-transfection studies
(Jones et al., 1993). Recently both p53 and the progesterone receptor have been identified as direct targets of Hoxa 5 (Raman et al., 2000a; Raman et al., 2000b). Though fruitful, these approaches have resulted in the identification of only a few target genes for
Hox gene family members. The use of Differential Display, GDA cDNA arrays, and
GeneChip microarrays allows the examination of thousands of genes simultaneously.
This massively parallel profiling can efficiently identify gene expression differences and elucidate new relationships between groups of genes (Reinke et al., 2000). Limitations of the GeneChip approach include incomplete gene representation and the need for substantial starting RNA template, though both of these technical problems have steadily improved.
The very restricted cell populations of the developing kidney that express Hoxa
11 meant dissection of these expressing tissues for RNA isolation would be a daunting if not impossible task. Using whole kidneys from mutant and wild type embryos presents other difficulties. The cells expressing Hoxa 11 in the kidney make up a small percentage of the whole kidney. This makes expression differences more difficult to detect and results in fewer genes being identified as tissue complexity increases
(Geschwind, 2000). Furthermore, efficient isolation of tissues for confirmation of gene
- 114 - expression differences would also be very problematic due to poor fertility of the compound heterozygotes. For these reasons we chose to use a cell line model of Hoxa 11 expressing kidney tissues in which we can modulate Hoxa 11 expression in a homogeneous system.
Several of the candidate target genes are of interest in kidney development. The growth factor PDGF A is expressed in newly formed epithelial cells derived from the condensing mesenchyme (Seifert et al., 1998). Hoxa 11 is expressed in the condensing mesenchyme until the cells begin to convert to epithelial cells. The repression of PDGF
A by Hoxa 11 in the HEK293 cells fits with this endogenous expression pattern, suggesting the down regulation of Hoxa 11 in the mesenchymal cells is necessary to permit PDGF A expression. Gene targeting studies of PDGF A focused on the lung
(Bostrom et al., 1996) and the CNS (Fruttiger et al., 1999) and did not report any obvious kidney defects. However, it is possible that loss of PDGF A function is compensated by the overlapping expression of a new family member, PDGF C (Li et al., 2000). The transcription factor GCM1 is expressed in E16.5 kidneys and the placenta, however spatial data on the kidney expression is not available (Kim et al., 1998). The knockout of
GCM1 is lethal at E10 due to placental defects (Anson-Cartwright et al., 2000). This is prior to kidney formation, therefore a possible role for GCM1 in kidney development is unknown. A conditional knockout of GCM1 in the kidney would be of great interest because the placental defect results from disrupted branching morphogenesis of the chorioallantoic villi. This type of branching morphogenesis initiates due to interactions between the chorionic plate, made up of trophoblast cells, and the mesodermal allantois.
A similar process of branching morphogenesis, from the interaction of the ureteric bud
- 115 - and the metanephric mesenchyme, is disrupted in the kidney of Hoxa 11/Hoxd 11 double mutants.
The Integrin Alpha 8 (ItgA8) gene is expressed in the condensing mesenchyme during kidney development and is down regulated upon epithelialization of those mesenchymal cells (Muller et al., 1997). This expression pattern overlaps well with the expression of Hoxa 11 in the kidney both temporally and spatially. Additionally, gene targeting studies report a similar renal phenotype as the Hoxa 11/Hoxd 11 double mutant renal phenotype (Muller et al., 1997). Our results showed ItgA8 expression increased upon expression of Hoxa 11 in both the HEK293 and mK10 cell lines. In situ hybridization analysis of Hoxa 11/Hoxd 11 double mutant kidneys revealed altered expression of ItgA8. These data are consistent with a model of Hoxa 11 regulating the expression of ItgA8 in condensing mesenchyme of the developing kidney.
During kidney development the metanephric mesenchyme condenses and differentiates in response to interactions with the ureteric bud epithelium. In response the bud undergoes branching morphogenesis. The morphogenetic movements of both the mesenchymal cells and the ureteric bud require complex interactions that result in modifications of cell behavior, especially in how the cell interacts with the local ECM.
This is influenced by the extracellular composition, the expression of cell adhesion molecules such as integrins, as well as ECM degrading enzymes. This permits a cell to appropriately invade or move through a tissue to form complex structures. The candidate targets identified here suggest Hoxa 11 plays a role in controlling cell behavior by modulating the expression of several types of genes known to be important in this process.
- 116 - ACKNOWLEDGEMENTS
We thank Affymetrix, and particularly the members of the academic user center, for their help. This work was supported by NIH grant HD32061 to SSP.
- 117 - REFERENCES
Afonso, S., Romagnano, L., and Babiarz, B. (1997). The expression and function of cystatin C and cathepsin B and cathepsin L during mouse embryo implantation and placentation, Development 124, 3415-25.
Alldridge, L. C., Harris, H. J., Plevin, R., Hannon, R., and Bryant, C. E. (1999). The annexin protein lipocortin 1 regulates the MAPK/ERK pathway, J Biol Chem 274,
37620-8.
Anson-Cartwright, L., Dawson, K., Holmyard, D., Fisher, S. J., Lazzarini, R. A., and
Cross, J. C. (2000). The glial cells missing-1 protein is essential for branching morphogenesis in the chorioallantoic placenta [see comments], Nat Genet 25, 311-4.
Barasch, J., Yang, J., Ware, C. B., Taga, T., Yoshida, K., Erdjument-Bromage, H.,
Tempst, P., Parravicini, E., Malach, S., Aranoff, T., and Oliver, J. A. (1999).
Mesenchymal to epithelial conversion in rat metanephros is induced by LIF, Cell 99,
377-86.
Benson, G. V., Nguyen, T. H., and Maas, R. L. (1995). The expression pattern of the murine Hoxa-10 gene and the sequence recognition of its homeodomain reveal specific properties of Abdominal B-like genes, Mol Cell Biol 15, 1591-601.
Bostrom, H., Willetts, K., Pekny, M., Leveen, P., Lindahl, P., Hedstrand, H., Pekna, M.,
Hellstrom, M., Gebre-Medhin, S., Schalling, M., et al. (1996). PDGF-A signaling is a critical event in lung alveolar myofibroblast development and alveogenesis, Cell 85, 863-
73.
- 118 - Boudreau, N. J., and Jones, P. L. (1999). Extracellular matrix and integrin signalling: the shape of things to come, Biochem J 339, 481-8.
Care, A., Silvani, A., Meccia, E., Mattia, G., Stoppacciaro, A., Parmiani, G., Peschle, C., and Colombo, M. P. (1996). HOXB7 constitutively activates basic fibroblast growth factor in melanomas, Mol Cell Biol 16, 4842-51.
Castelli-Gair, J. (1998). The lines gene of Drosophila is required for specific functions of the Abdominal-B HOX protein, Development 125, 1269-74.
Chauvet, S., Maurel-Zaffran, C., Miassod, R., Jullien, N., Pradel, J., and Aragnol, D.
(2000). dlarp, a new candidate Hox target in Drosophila whose orthologue in mouse is expressed at sites of epithelium/mesenchymal interactions, Dev Dyn 218, 401-13.
Davis, A. P., Witte, D. P., Hsieh-Li, H. M., Potter, S. S., and Capecchi, M. R. (1995).
Absence of radius and ulna in mice lacking hoxa-11 and hoxd-11, Nature 375, 791-5.
Davis, J. Q., and Bennett, V. (1994). Ankyrin binding activity shared by the neurofascin/L1/NrCAM family of nervous system cell adhesion molecules, J Biol Chem
269, 27163-6.
Debiec, H., Christensen, E. I., and Ronco, P. M. (1998). The cell adhesion molecule L1 is developmentally regulated in the renal epithelium and is involved in kidney branching morphogenesis, J Cell Biol 143, 2067-79.
Dohchin, A., Suzuki, J. I., Seki, H., Masutani, M., Shiroto, H., and Kawakami, Y. (2000).
Immunostained cathepsins B and L correlate with depth of invasion and different metastatic pathways in early stage gastric carcinoma, Cancer 89, 482-7.
- 119 - Fruttiger, M., Karlsson, L., Hall, A. C., Abramsson, A., Calver, A. R., Bostrom, H.,
Willetts, K., Bertold, C. H., Heath, J. K., Betsholtz, C., and Richardson, W. D. (1999).
Defective oligodendrocyte development and severe hypomyelination in PDGF-A knockout mice, Development 126, 457-67.
Geschwind, D. H. (2000). Mice, microarrays, and the genetic diversity of the brain [In
Process Citation], Proc Natl Acad Sci U S A 97, 10676-8.
Gould, A. P., and White, R. A. (1992). Connectin, a target of homeotic gene control in
Drosophila, Development 116, 1163-74.
Graba, Y., Aragnol, D., Laurenti, P., Garzino, V., Charmot, D., Berenger, H., and Pradel,
J. (1992). Homeotic control in Drosophila; the scabrous gene is an in vivo target of
Ultrabithorax proteins, Embo J 11, 3375-84.
Hemberger, M., Himmelbauer, H., Ruschmann, J., Zeitz, C., and Fundele, R. (2000). cDNA subtraction cloning reveals novel genes whose temporal and spatial expression indicates association with trophoblast invasion, Dev Biol 222, 158-69.
Hogan, B., Beddington, R., Costantini, F., and Lacy, E. (1994). Manipulating the Mouse
Embryo, 2nd edn (Plainview, NY, Cold Spring Harbor Laboratory Press).
Hsieh-Li, H. M., Witte, D. P., Weinstein, M., Branford, W., Li, H., Small, K., and Potter,
S. S. (1995). Hoxa 11 structure, extensive antisense transcription, and function in male and female fertility, Development 121, 1373-85.
Ishidoh, K., and Kominami, E. (1998). Gene regulation and extracellular functions of procathepsin L, Biol Chem 379, 131-5.
- 120 - Jo, H., Zhang, H., Zhang, R., and Liang, P. (1998). Cloning oncogenic ras-regulated genes by differential display, Methods 16, 365-72.
Jones, F. S., Holst, B. D., Minowa, O., De Robertis, E. M., and Edelman, G. M. (1993).
Binding and transcriptional activation of the promoter for the neural cell adhesion molecule by HoxC6 (Hox-3.3), Proc Natl Acad Sci U S A 90, 6557-61.
Kim, J., Jones, B. W., Zock, C., Chen, Z., Wang, H., Goodman, C. S., and Anderson, D.
J. (1998). Isolation and characterization of mammalian homologs of the Drosophila gene glial cells missing, Proc Natl Acad Sci U S A 95, 12364-9.
Kupper, J. (1998). Functional expression of GFP-tagged Kv1.3 and Kv1.4 channels in
HEK 293 cells, Eur J Neurosci 10, 3908-12.
Lee, M. G., Wigley, W. C., Zeng, W., Noel, L. E., Marino, C. R., Thomas, P. J., and
Muallem, S. (1999). Regulation of Cl-/ HCO3- exchange by cystic fibrosis transmembrane conductance regulator expressed in NIH 3T3 and HEK 293 cells, J Biol
Chem 274, 3414-21.
Lee, P. L., Gelbart, T., West, C., Adams, M., Blackstone, R., and Beutler, E. (1997).
Three genes encoding zinc finger proteins on human chromosome 6p21.3: members of a new subclass of the Kruppel gene family containing the conserved SCAN box domain,
Genomics 43, 191-201.
Leimeister, C., Bach, A., Woolf, A. S., and Gessler, M. (1999). Screen for genes regulated during early kidney morphogenesis, Dev Genet 24, 273-83.
Li, H., Schrick, J. J., Fewell, G. D., MacFarland, K. L., Witte, D. P., Bodenmiller, D. M.,
Hsieh-Li, H. M., Su, C. Y., and Potter, S. S. (1999a). Novel strategy yields candidate
- 121 - Gsh-1 homeobox gene targets using hypothalamus progenitor cell lines, Dev Biol 211,
64-76.
Li, X., Ponten, A., Aase, K., Karlsson, L., Abramsson, A., Uutela, M., Backstrom, G.,
Hellstrom, M., Bostrom, H., Li, H., et al. (2000). PDGF-C is a new protease-activated ligand for the PDGF alpha-receptor [see comments], Nat Cell Biol 2, 302-9.
Li, X., Veraksa, A., and McGinnis, W. (1999b). A sequence motif distinct from Hox binding sites controls the specificity of a Hox response element, Development 126, 5581-
9.
Lindberg, F. P., Bullard, D. C., Caver, T. E., Gresham, H. D., Beaudet, A. L., and Brown,
E. J. (1996). Decreased resistance to bacterial infection and granulocyte defects in IAP- deficient mice, Science 274, 795-8.
Liu, D., Pavlopoulos, E., Modi, W., Moschonas, N., and Mavrothalassitis, G. (1997).
ERF: genomic organization, chromosomal localization and promoter analysis of the human and mouse genes, Oncogene 14, 1445-51.
McKanna, J. A., Chuncharunee, A., Munger, K. A., Breyer, J. A., Cohen, S., and Harris,
R. C. (1992). Localization of p35 (annexin I, lipocortin I) in normal adult rat kidney and during recovery from ischemia, J Cell Physiol 153, 467-76.
Mintz, L., Galperin, E., Pasmanik-Chor, M., Tulzinsky, S., Bromberg, Y., Kozak, C. A.,
Joyner, A., Fein, A., and Horowitz, M. (1999). EHD1--an EH-domain-containing protein with a specific expression pattern, Genomics 59, 66-76.
- 122 - Mora, P. T., Parrott, C. L., Baksi, K., and McFarland, V. (1986). Immunologic selection of simian virus 40 (SV40) T-antigen-negative tumor cells which arise by excision of early SV40 DNA, J Virol 59, 628-34.
Muller, U., Wang, D., Denda, S., Meneses, J. J., Pedersen, R. A., and Reichardt, L. F.
(1997). Integrin alpha8beta1 is critically important for epithelial-mesenchymal interactions during kidney morphogenesis, Cell 88, 603-13.
Nakase, T., Kaneko, M., Tomita, T., Myoui, A., Ariga, K., Sugamoto, K., Uchiyama, Y.,
Ochi, T., and Yoshikawa, H. (2000). Immunohistochemical detection of cathepsin D, K, and L in the process of endochondral ossification in the human [In Process Citation],
Histochem Cell Biol 114, 21-7.
Patterson, L. T., and Potter, S. S. (2000). Hoxa 11 and hoxd 11 regulate branching morphogenesis through a proliferation independent pathway. In preparation.
Pedersen, K. A., Schroder, R. L., Skaaning-Jensen, B., Strobaek, D., Olesen, S. P., and
Christophersen, P. (1999). Activation of the human intermediate-conductance Ca(2+)- activated K(+) channel by 1-ethyl-2-benzimidazolinone is strongly Ca(2+)-dependent,
Biochim Biophys Acta 1420, 231-40.
Pellerin, I., Schnabel, C., Catron, K. M., and Abate, C. (1994). Hox proteins have different affinities for a consensus DNA site that correlate with the positions of their genes on the hox cluster, Mol Cell Biol 14, 4532-45.
Plisov, S. Y., Ivanov, S. V., Yoshino, K., Dove, L. F., Plisova, T. M., Higinbotham, K.
G., Karavanova, I., Lerman, M., and Perantoni, A. O. (2000). Mesenchymal-epithelial
- 123 - transition in the developing metanephric kidney: gene expression study by differential display, Genesis 27, 22-31.
Raman, V., Martensen, S. A., Reisman, D., Evron, E., Odenwald, W. F., Jaffee, E.,
Marks, J., and Sukumar, S. (2000a). Compromised HOXA5 function can limit p53 expression in human breast tumours, Nature 405, 974-8.
Raman, V., Tamori, A., Vali, M., Zeller, K., Korz, D., and Sukumar, S. (2000b). HOXA5 regulates expression of the progesterone receptor, J Biol Chem 275, 26551-5.
Reinke, V., Smith, H. E., Nance, J., Wang, J., Van Doren, C., Begley, R., Jones, S. J.,
Davis, E. B., Scherer, S., Ward, S., and Kim, S. K. (2000). A global profile of germline gene expression in C. elegans [In Process Citation], Mol Cell 6, 605-16.
Salewski, H., Bayer, T. A., Eidhoff, U., Preuss, U., Weggen, S., and Scheidtmann, K. H.
(1999). Increased oncogenicity of subclones of SV40 large T-induced neuroectodermal tumor cell lines after loss of large T expression and concomitant mutation in p53, Cancer
Res 59, 1980-6.
Santos, O. F., Barros, E. J., Yang, X. M., Matsumoto, K., Nakamura, T., Park, M., and
Nigam, S. K. (1994). Involvement of hepatocyte growth factor in kidney development,
Dev Biol 163, 525-9.
Seifert, R. A., Alpers, C. E., and Bowen-Pope, D. F. (1998). Expression of platelet- derived growth factor and its receptors in the developing and adult mouse kidney, Kidney
Int 54, 731-46.
Sgouras, D. N., Athanasiou, M. A., Beal, G. J., Jr., Fisher, R. J., Blair, D. G., and
Mavrothalassitis, G. J. (1995). ERF: an ETS domain protein with strong transcriptional
- 124 - repressor activity, can suppress ets-associated tumorigenesis and is regulated by phosphorylation during cell cycle and mitogenic stimulation, Embo J 14, 4781-93.
Shen, W. F., Montgomery, J. C., Rozenfeld, S., Moskow, J. J., Lawrence, H. J.,
Buchberg, A. M., and Largman, C. (1997). AbdB-like Hox proteins stabilize DNA binding by the Meis1 homeodomain proteins, Mol Cell Biol 17, 6448-58.
Shen, W. F., Rozenfeld, S., Kwong, A., Kom ves, L. G., Lawrence, H. J., and Largman,
C. (1999). HOXA9 forms triple complexes with PBX2 and MEIS1 in myeloid cells, Mol
Cell Biol 19, 3051-61.
Tong, X. K., Hussain, N. K., Adams, A. G., O'Bryan, J. P., and McPherson, P. S. (2000).
Intersectin can regulate the Ras/MAP kinase pathway independent of its role in endocytosis [In Process Citation], J Biol Chem 275, 29894-9.
Torban, E., and Goodyer, P. R. (1998). Effects of PAX2 expression in a human fetal kidney (HEK293) cell line, Biochim Biophys Acta 1401, 53-62.
Valerius, M. T., Li, H., Stock, J. L., Weinstein, M., Kaur, S., Singh, G., and Potter, S. S.
(1995). Gsh-1: a novel murine homeobox gene expressed in the central nervous system,
Dev Dyn 203, 337-51.
Valerius, M. T., Patterson, L. T., and Potter, S. S. (2000). Molecular and functional properties and expression profiles of kidney cell lines from Hoxa 11-SV40 Tag transgenic mice. In preparation.
Zhang, R., Tan, Z., and Liang, P. (2000). Identification of a novel ligand-receptor pair constitutively activated by ras oncogenes, J Biol Chem 275, 24436-43.
- 125 - Chapter 5
Additional Discussion
- 126 - Hoxa 11 during nephron formation
In Drosophila the homeobox selector genes are named such because they control complete developmental programs responsible for the formation of whole structures
(Lewis, 1978). These programs are cascades of gene activities that affect cell state (the complement of gene activities), cell proliferation, cell death, cell shape, and the morphogenetic movements necessary to form structures. Our studies of Hoxa 11 in the developing kidney suggest this gene plays an important role in controlling mesenchymal cell state during condensation. The candidate downstream targets presented paint a picture of Hoxa 11 as a controller of this induced response. By modifying the expression of proteinases (Cathepsin L), ECM receptors (Integrin alpha 8), and growth factors
(PDGF-A), and transcription factors (GCM1), Hoxa 11 controls metanephric mesenchymal cell behavior. This is the essence of patterning during development – the spatial and temporal organization of cell behavior to form ordered structures (Wolpert,
1998).
To place Hoxa 11 into a larger cascade of development we must draw on both studies of the kidney and in vitro work focused on individual genes. As a cell surface receptor, Integrin alpha 8 (ItgA8) may lead the mesenchymal cells toward a localized ligand source, producing the mesenchymal condensation. Alternatively, ItgA8 may serve to solidify the cells in the current position if ligand were present throughout the metanephric mesenchyme. Known ligands of ItgA8, include fibronectin, vitronectin, tenascin-C, and osteopontin (OPN) (Muller et al., 1995; Schnapp et al., 1995). Of these only osteopontin is expressed appropriately to mediate ItgA8 function in the condensing mesenchyme, specifically in the ureteric bud epithelium (Denda et al., 1998; Muller et al.,
- 127 - 1997). A localized source of OPN ligand in the ureteric bud fits well with a model of
ItgA8 expression leading to cell condensation about the ureteric bud epithelium.
Functional analysis also supports this model. In ItgA8 mutant animals, epithelialization of the metanephric mesenchyme is disrupted, leading to defects in branching morphogenesis (Muller et al., 1997). The authors comment that mesenchymal condensates are present though not well organized, and that few comma- or S-shaped bodies were evident (Muller et al., 1997). Disrupting OPN function in rat kidney cultures using anti-OPN antibodies blocked condensation of the mesenchyme and subsequent tubulogenesis (Rogers et al., 1997). Thus, an interaction between ItgA8 in the metanephric mesenchyme and OPN in the ureteric bud epithelium is important for proper condensation. This is a critical step for subsequent tubulogenesis in the mesenchyme, and branching morphogenesis of the ureteric bud.
Cathepsins are cysteine proteinase involved in tumor invasiveness (Dohchin et al.,
2000; Ishidoh and Kominami, 1998) and trophoblast invasion during implantation
(Afonso et al., 1997). A role for cathepsin L may be to balance or regulate intracellular space in the mesenchyme by degrading ECM components, such as Collagen I.
Repression by Hoxa 11 would aid condensation of the mesenchyme by permitting a build up of the ECM. Interestingly, cathepsin K has been shown to cleave Collagen I and also
OPN (Yamashita and Dodds, 2000). It has not been determined whether cathepsin L has a similar activity on OPN or not. However, regulation of the OPN ligand by degradation in the mesenchymal space is an attractive addition to the model.
- 128 - Reducing the problem
Wolpert (1998) lists 5 elements that determine cell behavior; cell state, shape and movement, cell-to-cell communication, proliferation, and programmed cell death. To gain understanding as to how a given protein ultimately affects a process and produces a phenotype, we require knowledge of how that protein affects first cell behavior, then the developmental processes that result from that cell behavior. In studies of development we often start at the very end, or the beginning, and work our way to the middle. That is, with a phenotype (the end) or a gene (the beginning). After examining the two extremes, we narrow the scale first to discover the role a particular gene plays in a given tissue or cell, then the molecular details of how that gene executes its role (e.g. binding DNA and influencing transcription). We did this in two ways in these studies. First, we focused on a specific tissue by studying the role of Hoxa 11 in the kidney and further focused on a cell type by establishing immortalized cell lines from Hoxa 11 expression domains in the kidney. Secondly, by using immortalized cell lines we “froze” cell state with the expression of SV40 T Antigen, resulting in a snapshot of the current cell state, and stabilization of proliferation and cell death such that these two variables were removed.
This permited us to observe and describe the three other elements of cell behavior using microarrays to determine cell state, observation to determine cell shape and movement, and co-culture with ureteric bud epithelium to examine cell-to-cell communication. Of future interest is that while we have shown that the wildtype mK3 cell line can promote ureteric bud growth and branching, we do not yet know if the mutant mK10 cell line
(Hox a11-/-; Hoxd 11 -/-) has similar ability. The next step will be to determine if the
- 129 - mK10 cells have this ability and furthermore, if this is altered (i.e. enhanced) with the addition of Hoxa 11 (a “rescue” experiment).
Expanding the analysis
A feature of the work presented here is the use of microarrays. Microarrays give us the tremendous power to analyze the expression of thousands of genes simultaneously.
This has two immediate benefits. First, as a tool to describe the “state” of transcription in a cell or tissue (the “transcriptome”), microarrays are dramatically more efficient than traditional marker analysis by offering data on hundreds of markers from a single experiment. Secondly, for differential screening microarrays again offer a tremendous efficiency improvement over a candidate gene, or best guess approach, by permitting the analysis of many more genes. The drawbacks include both temporary technical problems and lasting limitations of the approach. The temporary include design problems with the
Affymetrix arrays. The design of the probesets on the chip is based on a certain version of publicly available sequence data using automated methods that are not yet proven.
This will be eventually overcome with the sequencing of the genome and a full description of the gene transcripts, and with experience using these arrays to refine the design algorithm. The problems with probeset design are evident in the datasets because certain known markers are missing though Northern and/or RTPCR analysis confirms their presence. This could be due to poor sequence choice or alternative splicing that was not previously recognized. Furthermore, until the complete genome is sequenced and described, we cannot be sure we are examining the expression of all genes or all transcripts. Perhaps the most important things to remember about microarrays are the limitations. An array provides data on the transcriptional state of a cell or tissue. This is
- 130 - only part of a complete description of cell state that should include protein status such as translation (is it made?), phosphorylation state, subcellular localization, protein-protein interactions, as well as information on the signals a cell is exchanging with its neighbors.
While microarrays are the current favored approach, clearly additional complementary approaches will be necessary to complete the picture of cell state as described here. This includes information on individual proteins as well as other types of data (such as mutant data) that contribute to our understanding of a process. For example, processes well understood in one system may aid in the understanding of the analogous process in another system. This information, while easily examined ad hoc, must be somehow incorporated into microarray approaches in a more automated way.
One approach would be to create modules of genes whose expression status collectively describes the status of a process, such that data on say 20 individual genes is reduced to a single “on/off” output for that process. This reduces the information into a more manageable form. As one can see, the goal here is to include more information in an automated fashion to speed the analysis of these huge datasets and the eventual distillation of the data into a useful bit of knowledge.
- 131 - REFERENCES
Afonso, S., Romagnano, L., and Babiarz, B. (1997). The expression and function of cystatin C and cathepsin B and cathepsin L during mouse embryo implantation and placentation, Development 124, 3415-25.
Denda, S., Reichardt, L. F., and Muller, U. (1998). Identification of osteopontin as a novel ligand for the integrin alpha8 beta1 and potential roles for this integrin-ligand interaction in kidney morphogenesis, Mol Biol Cell 9, 1425-35.
Dohchin, A., Suzuki, J. I., Seki, H., Masutani, M., Shiroto, H., and Kawakami, Y. (2000).
Immunostained cathepsins B and L correlate with depth of invasion and different metastatic pathways in early stage gastric carcinoma, Cancer 89, 482-7.
Ishidoh, K., and Kominami, E. (1998). Gene regulation and extracellular functions of procathepsin L, Biol Chem 379, 131-5.
Lewis, E. B. (1978). A gene complex controlling segmentation in Drosophila, Nature
276, 565-70.
Muller, U., Bossy, B., Venstrom, K., and Reichardt, L. F. (1995). Integrin alpha 8 beta 1 promotes attachment, cell spreading, and neurite outgrowth on fibronectin, Mol Biol Cell
6, 433-48.
Muller, U., Wang, D., Denda, S., Meneses, J. J., Pedersen, R. A., and Reichardt, L. F.
(1997). Integrin alpha8beta1 is critically important for epithelial-mesenchymal interactions during kidney morphogenesis, Cell 88, 603-13.
- 132 - Rogers, S. A., Padanilam, B. J., Hruska, K. A., Giachelli, C. M., and Hammerman, M. R.
(1997). Metanephric osteopontin regulates nephrogenesis in vitro, Am J Physiol 272,
F469-76.
Schnapp, L. M., Hatch, N., Ramos, D. M., Klimanskaya, I. V., Sheppard, D., and Pytela,
R. (1995). The human integrin alpha 8 beta 1 functions as a receptor for tenascin, fibronectin, and vitronectin, J Biol Chem 270, 23196-202.
Wolpert, L. (1998). Principles of Development (New York, Oxford University Press).
Yamashita, D. S., and Dodds, R. A. (2000). Cathepsin K and the design of inhibitors of cathepsin K, Curr Pharm Des 6, 1-24.
- 133 -