Cell fate specification along the anterior–posterior axis of the intermediate mesoderm

Authors


Abstract

The vertebrate intermediate mesoderm (IM) is highly patterned along the anterior–posterior (A-P) axis. In the chick embryo, the kidney tissue, which is a derivative of the IM, is generated only from IM located posterior to the sixth somite axial level, which also marks the border between cranial and trunk segments. The cellular and molecular mechanisms that govern the formation of the anterior border of the kidney morphogenetic field are currently unknown. In this study, we asked whether specific A-P patterning information is conveyed by the movement of cells through the primitive streak (PS) at different time points that consequently affects the expression of kidney genes, or by the environment that these cells encounter during their migration to the IM. In this study, we show that kidney-inductive signals are present along the whole axis, including anterior non–kidney-generating regions. These inductive signals are generated by tissues that are located medial to the anterior IM. We also demonstrate that cells that migrate through the PS of early embryonic stages (Hamburger and Hamilton stage 3–4 and earlier), which will give rise to anterior nonkidney IM, are competent to respond to these inductive factors. This prospective anterior IM tissue loses its competence to respond to kidney inducing signals during its migration from the PS to its final location in the anterior IM. We present here a model in which changes in cell competence determine the formation of the anterior border of kidney gene expression and discuss the possible evolutionary implications of this developmental mechanism. Developmental Dynamics 232:901–914, 2005. © 2005 Wiley-Liss, Inc.

INTRODUCTION

Organs and structures arise at particular locations along the anterior–posterior (A-P) embryonic axis. The kidney is an example of an organ that develops in a specific position along this axis. Vertebrate kidney tissues are derived from a narrow strip of tissue called the intermediate mesoderm (IM), which is located between the paraxial and lateral plate mesoderms of the developing embryo (Patten,1929). In chick embryos, this strip of tissue can be observed from at least the first formed somite and posteriorly (Patten,1929; Eichler,1978; Bellairs and Osmund,1998). Hiruma and Nakamura's (2003) study in the chick embryo, defined the IM differently, as a clear condensation of cells that appears from the sixth somite level and posteriorly, consistent with the molecular gene expression domains of the kidney morphogenetic field (see below). As development proceeds, the IM differentiates sequentially from anterior to posterior into several types of kidney tissue (Saxén,1987) that in birds and mammals are called pronephros, mesonephros, and metanephros (anamniote embryos form only a pronephros and mesonephros). In birds and mammals, the definitive adult kidney is the metanephros, whereas the pro-and mesonephros largely degenerate or are incorporated into the developing reproductive system. Whereas the pronephros appears to arise de novo from cells in the IM, the mesonephros arises along the course of the nephric duct and the metanephros is generated by mutual inductions between duct and mesenchyme (Goodrich,1930; Saxén,1987). Recently, it has been shown that expression patterns and the function of many genes that characterize metanephric development appear to be conserved in pronephros development, including Lim-1, Pax-2, Pax-8, WT-1, Sim-1, c-ret, and Wnt-4 (Vize et al.,1997; Chan and Asashima,2000; Kuure et al.,2000; Jones,2003).

The transcription factors Pax-2/8 and Lim-1 have been shown to have a major role in kidney development (Dressler et al.,1990; Plachov et al.,1990; Taira et al.,1994; Carroll and Vize,1999; Carroll et al.,1999). Knockout of these genes produces severe defects in pronephros formation and in later stages of kidney development (Torres et al.,1995; Shawlot and Behringer,1995; Bouchard et al.,2002). In Xenopus, misexpression of Lim-1 and Pax-8 can produce ectopic kidney tubule formation, suggesting that the interaction between Lim-1 and Pax-8 is a key step in the establishment of the pronephric primordium (Carroll and Vize,1999). In the chick, IM cells that express Pax-2 and Lim-1 contribute primarily to the nephric duct and once the duct begins to migrate over the nephrogenic cord, it induces the expression of Pax-2 in the underlying cord (James and Schultheiss,2003). In chick embryos, the expression pattern of Pax-2 and Lim-1 along the A-P axis of the IM shows a sharp border at the level of the sixth somite as was described previously by Mauch et al. (2000) and James and Schultheiss (2003). There is no information regarding the cellular and molecular mechanisms that control the formation of this border.

Little is known regarding the specification of the IM. Several studies in recent years have demonstrated the role of neighboring tissues in IM specification, mainly in relation to the mediolateral axis. Mauch et al. (2000) have shown that axial and rostral tissues, lateral plate mesoderm, endoderm, and ectoderm neither induce nor inhibit pronephros formation in the IM, whereas Obara-Ishihara et al. (1999) gave the ectoderm and bone morphogenetic protein-4 (BMP-4) an inductive role. James and Schultheiss (2003), in turn, suggested that the ectoderm is essential for cell survival but not for kidney gene induction and that the lateral plate mesoderm in different experimental conditions can induce kidney properties in the paraxial mesoderm. Finally, although trunk paraxial mesoderm has been shown to have a critical role in kidney induction, the molecular mechanisms that control this induction remain unknown (Seufert et al.,1999; Mauch et al.,2000; James and Schultheiss,2003).

Even less is known regarding the patterning of the IM along the A-P axis. The above data suggest an important role for the Lim-1 and Pax-2 genes in the establishment of the kidney morphogenetic field and early kidney formation posterior to the sixth somite level. However, nothing is known regarding the specific expression and function of the genes involved in the differentiation of IM, anterior to the sixth somite level. This observation suggests that at least two major types of cells arise along this axis: one anterior to the sixth somite level, which gives rise to nonkidney tissue, and the other posterior to this axial level, which gives rise to kidney tissue. The main focus of the current study was to investigate the embryological mechanisms by which expression of kidney genes is restricted to IM posterior to the sixth somite level. Specifically, we asked when and where the two major cell groups (anterior and posterior to the anterior border of kidney gene expression) are specified and what the positional information required for this specification is. The results of these studies have implications for understanding the patterning of other derivatives of the IM along the A-P axis and for a more general understanding of A-P patterning in the early embryo.

RESULTS

Transplantation experiments (Fig. 1) were performed to determine two major parameters of the A-P specification of the IM: (1) the time and (2) the location at which IM cells become specified to give rise to kidney vs. nonkidney tissue. Tissue manipulation experiments will reveal the state of specification of the transplanted tissues and will also characterize the environment into which such tissues were transplanted. Such experiments will reveal the timing and positional information that is essential for specifying the A-P axis of IM cells.

Figure 1.

Scheme showing the 12 transplantation experiments performed in these studies, which use three embryonic stages (3–4, 6, and 8) and two locations: the region in the primitive streak where the prospective intermediate mesoderm (IM) is located (red box) and the studied region in the IM itself (blue boxes). The yellow box next to the black line represents transplanted tissue behind a barrier. Blue arrows indicate control transplantations.

A comprehensive set of experiments was performed, using cells taken from two locations along the migratory route of IM cells and from three embryonic stages. The two locations were (1) the region in the primitive streak (PS) that contains the prospective IM region, and (2) the regions in the prospective IM to which these PS cells migrate. The three embryonic stages examined were Hamburger and Hamilton stages 3–4, 6, and 8. According to previous studies (Garcia-Martinez and Schoenwolf,1992; Psychoyos and Stern,1996; James and Schultheiss,2003), cells in the PS at stage 3–4 will migrate to a position lateral to the first prospective somite by stage 6, anterior to the border of the sixth somite, a presumptive nonkidney tissue. Cells in the PS at stage 6 will migrate to a position lateral to the prospective sixth somite or posterior at stage 8 and will terminate in the presumptive pronephros region (Fig. 1).

Posterior IM From Stage 8 (posterior IM8) Is Committed to Its Fate and Is Autonomous

Posterior IM8 is a tissue that develops posterior to the sixth somite border and expresses kidney genes. A previous study has already revealed that this IM region is committed to its kidney fate with regard to the mediolateral axis (James and Schultheiss,2003). However, it is still unclear whether this tissue is also committed to its fate in the context of the A-P axis in relation to the sixth somite border. When posterior IM8 from quail embryos was transplanted into the anterior IM of stage 6 (anterior IM6) chick embryos, ectopic expression of Lim-1 and Pax-2 in anterior IM regions was detected by in situ hybridization (ISH) and immunohistochemistry, respectively (Fig. 1, transplantation 1; Fig. 2; Table 1). Sectioning followed by immunostaining of QCPN revealed an overlapping of quail cells, Lim-1 and Pax-2 expression (Fig. 2C–E).

Figure 2.

Posterior intermediate mesoderm (IM) 8 is already determined to its kidney fate and is autonomous. A: Posterior IM8 from quail embryos was transplanted into anterior IM6 of chick embryos (Fig. 1, transplantation 1). B: In situ hybridization of Lim-1 showed ectopic expression (indicated by an arrow) on the operated side. C–E: Sectioning through the graft region (C) followed by double immunostaining using Pax-2 (D) and QCPN antibodies (E). The white lines mark the graft border and clearly shows an overlapping of quail cells, Lim-1 expression, and Pax-2 staining. S, somite.

Table 1. Summary of Tissue Transplantation Experimentsa
Transplantation numberDonor quailHost chick% with expression of kidney gene markers% with no expression of kidney gene markersTotal (n)
  • a

    Rows indicate the type of tissue manipulation. St., stage of embryo; K, tissue that will give rise to intermediate mesoderm (IM) region posterior to the sixth somite border, a kidney-generating IM; N.K, tissue that will give rise to IM region anterior to the sixth somite border, a non–kidney-generating IM. n, total number of successful experiments. Successful experiment was defined as an experiment in which the graft reached the correct position in relation to the sixth somite axial border of the IM.

1St. 8 IM (K)St. 6 IM (N.K)57 (4/7)43 (3/7)7
2St. 6 PS (K)St. 6 IM (N.K)81 (42/52)19 (10/52)52
3St. 6 PS (K)St. 8–9 IM (N.K)78 (7/9)22 (2/9)9
4Control side: St. 6 PS (K)St. 7 IM (N.K.)Lim-1 Control side: 89 (17/19)Lim-1 Control side: 11 (2/19)19
 Experimental side: St. 6 PS (K) + Barrier Experimental side: 0 (0/19)Experimental side: 100 (19/19) 
   Pax-2 Control side: 100 (17/17)Pax-2 Control side: 0 (0/17)17
   Experimental side: 65 (11/17)Experimental side: 35 (6/17) 
5St. 6 IM (N.K)St. 8 IM (K)7 (1/14)93 (13/14)14
6St. 6 IM (N.K)St. 6 PS (K)100 (11/11)0 (0/11)11
7St. 3–4 PS (N.K)St. 6 PS (K)88 (7/8)12 (1/8)8
8St. 3–4 PS (N.K)St. 8 IM (K)93 (14/15)7 (1/15)15
9St. 3–4 PS (N.K)St. 6 IM (N.K)92 (11/12)8 (1/12)12
10St. 6 PS (K)St. 6 PS (K)100 (7/7)0 (0/7)7
11St. 6 IM (N.K)St. 6 IM (N.K)0 (0/9)100 (9/9)9
12St. 8 IM (K)St. 8 IM (K)100 (6/6)0 (0/6)6

To clarify whether the expression of Lim-1 and Pax-2 in the grafted cells is the result of the autonomy of the grafted tissue or the result of inductive signals emanating from surrounding tissues in anterior IM regions, we cultured chick posterior IM of that stage on collagen cushions. Consistent with previous observation made by James and Schultheiss (2003), Pax-2 and Lim-1 were detected in the cultured tissue (data not shown), suggesting that the posterior IM8 is already specified. Previous studies have found that paraxial tissue plays a role in patterning the IM (Seufert et al.,1999; Mauch et al.,2000; James and Schultheiss,2003). Therefore, to confirm the absence of any paraxial tissue in these cultures, Pax-7 was served as a paraxial marker, because this transcription factor is up-regulated in the paraxial mesoderm by the surface ectoderm (Reshef et al.,1998; Wilm et al.,2004). In that sense, this gene is an excellent marker for the presence of lateral paraxial mesoderm cells in these cultures, which contain surface ectoderm as a source of factors necessary for the survival of the mesoderm. Immunostaining of Pax-7 in these cultures showed no discernible expression of this gene compared with cultures of somite plus ectoderm of the same axial level (data not shown). The results of both the in vivo and the in vitro experiments suggest that the posterior IM8 is committed to its kidney fate and is already autonomous in relation to the A-P axis.

Prospective Posterior IM From PS of Stage 6 (PS6) Is Committed to Its Fate but Requires Extrinsic Signals to Turn on Kidney Genes in an Anterior Environment

By having found that the posterior IM8 is committed to its kidney fate, we next sought to understand whether the source of this tissue too is committed to the kidney fate. The prospective posterior IM from PS6 of quail embryos was transplanted into anterior chick IM at the same stage (Fig. 1, transplantation 2). After incubation, donor cells stained by CM-DiI were detected in IM regions anterior to the sixth somite border, a non–kidney-generating IM. ISH, followed by sectioning and immunostaining, revealed ectopic expression of Lim-1 and Pax-2 in the anterior IM overlapping with the QCPN antibody (Fig. 3A–E; Table 1). These results are consistent with three possibilities. The prospective posterior IM of PS6 may already be autonomous. Alternatively, inductive signals in the anterior IM may induce kidney markers in PS6 tissue. The third possibility derives from an experimental technique: when dissecting out the prospective IM region of the PS, there is a chance that cells of prospective paraxial mesoderm are included in the graft and, therefore, might induce kidney genes.

Figure 3.

Prospective posterior intermediate mesoderm (IM) of primitive streak (PS) 6 requires extrinsic inductive signals to express kidney genes in anterior nonkidney region. A: Prospective posterior IM of PS6 from quail embryos was transplanted into chick anterior IM at the same stage (Fig. 1, transplantation 2). B: In situ hybridization of Lim-1 revealed ectopic expression of this gene anterior to the sixth somite border (indicated by an arrow). C–E: Section at the level of the implanted tissue shows ectopic expression of Lim-1 within the graft (C), Pax-2 (D), and QCPN (E) antibodies. White lines mark the graft borders and clearly shows that quail cells ectopically express Lim-1 and Pax-2. H,G: Cultures of the prospective posterior IM of PS6 alone show Pax-2 (H) but not Lim-1 (G) expression. I: Pax-7 served as control for the presence of paraxial mesoderm. F: The cultures were stained with DAPI to mark nuclei. Nt, neural tube; S, somite; Lp, lateral plate mesoderm.

To distinguish between these possibilities we cultured prospective posterior IM from PS6 for 24 hr isolated from any surrounding tissues. Expression of Pax-2 but not Lim-1 or Pax-7 was detected (Fig. 3F–I; 100%, n = 8). Positive controls of Lim-1 in parallel stage matched whole embryo and Pax-7 in cultures of somite plus ectoderm show expression of these genes (data not shown). The observation that Lim-1 is detected when prospective posterior IM from PS6 is transplanted anteriorly but not when placed in culture suggests that this tissue is committed to its kidney fate but requires inductive signals that are present in the region of anterior non–kidney-generating IM and probably emanate from surrounding tissues. The lack of Pax-7 expression in the cultures excludes the third possibility, that of the presence of prospective paraxial mesoderm in the culture of PS6 tissue. Moreover, from the in vitro experiments it is clear that Pax-2 and Lim-1 are regulated differentially.

Kidney Inductive Signals Are Present in Anterior Non–Kidney-Generating IM of Stages 8–9

In the previous experiment, we demonstrated that the environment anterior to the sixth somite border can promote kidney gene expression in the prospective posterior IM of PS6. The region of anterior IM6 (in which the graft was transplanted in that experiment, transplantation 2) is still in a migratory state before the formation of the first somites. Previous studies have demonstrated a role for the adjacent paraxial mesoderm in positive regulation of kidney genes in posterior IM (Seufert et al.,1999; Mauch et al.,2000; James and Schultheiss,2003). To understand whether the anterior inductive signals are time dependent and are associated with events along the paraxial mesoderm that lead to somite formation, we transplanted the same tissue (prospective posterior IM of PS6) from quail origin into anterior IM at later stages (8–9) of chick embryo in parallel to the first already existing somites (Fig. 1, transplantation 3). After incubation, the graft was detected next to the most anterior somites. ISH for Lim-1, followed by sectioning and double-immunostaining using Pax-2 and QCPN, showed overlapping of the two kidney genes and QCPN in ectopic regions (Fig. 4; Table 1). These results indicate that at relatively late stages when somites are already formed, inductive signals for kidney genes in anterior non–kidney-generating IM are present and sufficient for the induction of kidney properties in prospective posterior IM of PS6.

Figure 4.

Kidney-inductive signals are present in anterior non–kidney-generating intermediate mesoderm (IM) of stages 8–9. A: Prospective posterior IM of primitive streak (PS) 6 from quail embryos was transplanted into anterior IM of stages 8–9 of chick embryos (Fig. 1, transplantation 3). B: In situ hybridization of Lim-1 revealed ectopic expression on the operated side (indicated by an arrow). C–E: Sectioning through the graft region (C) followed by double immunostaining by using Pax-2 (D) and QCPN (E) antibodies. White lines mark the graft borders and clearly show ectopic expression of Lim-1 and Pax-2 in the graft cells. Nt, neural tube; S, somite.

Kidney Inductive Signals Are Secreted From Medial Tissues in an Anterior Region

In the two previous experiments, we demonstrated that the environment anterior to the sixth somite border can induce kidney gene expression in grafts of prospective posterior IM of PS6. To determine which of the adjacent neighboring tissues of the anterior IM is the source of these inductive factors, we transplanted two parallel regions of prospective posterior IM of PS6 from quail embryos into both sides of anterior IM7 of chick embryo. On the left side we inserted a cellophane barrier between the graft and the paraxial mesoderm (Fig. 1, transplantation 4). After incubation and ISH for Lim-1 (Fig. 5; Table 1), the graft in the control side showed clear expression of Lim-1, while in the other side that contained the barrier, the graft showed no expression of this gene. Cross-section through this region followed by double immunostaining using Pax-2 and QCPN, clearly indicated that the quail cells were expressing Pax-2 on both the experimental and control sides. This differential expression of Lim-1 and Pax-2 is consistent with the culture experiments. The 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI) staining of this section showed that these cells survived, despite the presence of the barrier (Fig. 5C–J). Because Lim-1 and Pax-2 are transiently expressed in the forming somites as well as in the IM (James and Schultheiss,2003), one could argue that some of the Lim-1 and Pax-2 expression in the transplants is derived from paraxial mesoderm. However, because this same tissue, when placed in culture, expresses Pax-2 but not the paraxial marker Pax-7 (Fig. 3), the Lim-1/Pax-2–expressing tissue in Figure 5 appears to be intermediate mesoderm. This result indicates that kidney-inductive signals are secreted from tissues (paraxial mesoderm, neural tube, or notochord) that are medial to the IM also in anterior regions where non–kidney-generating IM is formed.

Figure 5.

Kidney-inductive signals are secreted from tissues that are located medial to the anterior intermediate mesoderm (IM). A–J: In barrier experiments, two parallel regions of prospective posterior IM of primitive streak (PS) 6 from quail embryos were transplanted into anterior IM7 on both sides in chick embryos. A: A cellophane barrier was inserted between the graft and the paraxial mesoderm on the left side (Fig. 1, transplantation 4). B: In situ hybridization of Lim-1 revealed no expression of the gene on the experimental side compared with the control (indicated by an arrow). The yellow line marks the barrier. The dashed line marks the location of the transplanted tissue. C–J: Sectioning through the graft regions (C–F, experimental side; G–J, control side) followed by double immunostaining using Pax-2 (D,H) and QCPN (E,I) antibodies show expression of Pax-2 (D) but not Lim-1 (C) in the experimental graft compared with expression of both genes in the control graft (H,G). The 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI) staining (F,J) indicates the viability of grafted cells. White lines mark the graft borders. S, somite.

Anterior IM6 Can Be Reprogrammed to a Kidney Fate Depending on the Local Environment

The experiments to date indicated the presence of kidney-inductive signals in the environment of the prospective anterior non–kidney-generating IM, raising the next question: what normally directs the anterior IM not to express kidney genes? To achieve this goal, we sought to understand the state of specification of stage 6 IM that will give rise to the IM anterior to the sixth somite axial level. We thus transplanted anterior IM6 into posterior IM8 (Fig. 1, transplantation 5), whose adjacent paraxial mesoderm has been shown to secrete kidney-inductive signals (Mauch et al.,2000). After incubation, cells of donor origin reached a location in the IM posterior to the sixth somite border. ISH revealed a gap in kidney marker expression, indicating that the quail tissue did not express Pax-2 (Fig. 6A,B; Table 1) or Lim-1 (data not shown). Sectioning followed by immunostaining of QCPN clearly confirmed that the gap exclusively contained quail cells (Fig. 6C,D). These results suggest that anterior IM6 under these experimental conditions cannot be altered into kidney-generating IM. Whether these results indicate the commitment of anterior IM6 to an anterior IM fate by not being competent to respond to kidney-inductive signals in the region of posterior IM8 was examined in the next experiment.

Figure 6.

Anterior intermediate mesoderm (IM) 6 can be reprogrammed into kidney fate depending on the local environment. A: Anterior IM6 of quail embryos was transplanted into posterior IM8 of chick embryos (Fig. 1, transplantation 5). B: In situ hybridization of Pax-2 revealed a gap in this gene expression consistent with the final location of the donor cells (indicated by an arrow). C,D: Sectioning through the graft region shows lack of Pax-2 expression on the right side (C) overlapped in a merged image with QCPN antibody (D), indicating the graft origin. E,F: Anterior IM6 from quail embryos was transplanted into prospective posterior IM of primitive streak (PS) 6 from chick embryos (A, Fig. 1, transplantation 6); sectioning through the graft region and double immunostaining using Pax-2 (E) and QCPN (F) antibodies show in merged images that quail cells that migrated through prospective posterior IM of PS6 to become posterior IM (marked by white lines) express the kidney gene Pax-2. G–I: Culture of anterior IM6 stained with 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI) to mark nuclei (G), shows no discernible expression of Lim-1 (H) or Pax-2 (I). Nt, neural tube; S, somite.

In this experiment, we transplanted anterior IM6 from quail donor into prospective posterior IM of PS6, the precursor tissue to posterior IM8, of chick embryo (Fig. 1, transplantation 6; Fig. 6A) to ascertain whether the PS environment and the migratory pathway are able to affect the grafted tissue. After incubation, donor cells were integrated into posterior IM tissue and, in contrast to the previous experiment, they changed their fate and expressed kidney genes. Double immunostaining, using Pax-2 and QCPN, clearly indicated that the quail cells were expressing Pax-2 (Fig. 6E,F; Table 1). These two latter experiments suggest that anterior IM6 is not competent to respond to kidney-inductive signals in the environment of posterior IM8 or in its original location, which was shown to contain kidney-inductive signals (see Fig. 1, transplantations 2, 3, 4). However, when transplanted into prospective posterior IM of PS6, this special environment reprogrammed grafted cells and changed their ability to respond to inductive signals emanating from surrounding tissues along the migratory pathway and, therefore, changed their original fate. Control experiments in which isolated IM tissue from the anterior region of stage 6 embryos was cultured for 24 hr showed no expression of Pax-2, Lim-1 (Fig. 6G–I; 100%, n = 7) and Pax-7 (data not shown) in cultured cells. DAPI staining confirmed cell viability.

Prospective Anterior IM of PS3–4 Is not Committed to Its Nonkidney Fate

The results obtained with anterior IM6 (Fig. 1, transplantations 5, 6) showed a dual response, dependent on the environment into which this tissue was transplanted. In posterior IM8 environment it preserved its nonkidney fate and when transplanted into prospective posterior IM of PS6 this tissue was reprogrammed and expressed kidney markers. This phenomenon may indicate that developmental events along the migratory route from its source patterned this tissue to become non–kidney-generating IM. We investigated prospective anterior IM of PS3–4, the source tissue of anterior IM6 to understand these events and the commitment of PS cells to their nonkidney destiny at that time point. In this set of experiments, we transplanted prospective anterior IM from PS3–4 of quail donors into the prospective posterior IM of PS6 or into posterior IM8 of chick embryos, prospective kidney regions (Fig. 1, transplantations 7, 8). After incubation, in both experiments, the cells reached locations along the IM posterior to the sixth somite border and expressed the kidney genes Pax-2 (Fig. 7; Table 1) and Lim-1 (data not shown). From these results, we can conclude that prospective anterior IM of PS3–4 is not yet committed to its final fate, which can be changed through such manipulations in a new instructive environment. Note that, in contrast to anterior IM6, which in the inductive environment of posterior IM8 did not change its fate, prospective anterior IM of PS3–4 did express kidney genes. These results indicate that, along the migratory pathway from PS3–4 to anterior IM6 (the normal migratory pathway of these cells), the cells undergo a process that changes their competence to respond to inductive signals.

Figure 7.

Prospective anterior intermediate mesoderm (IM) of primitive streak (PS) 3–4 is not yet specified to its nonkidney fate and is competent to respond to signals that can induce the tissue to express kidney genes. A: Prospective anterior IM of PS3–4 of quail embryos was transplanted into posterior IM8 of chick embryos (Fig. 1, transplantation 8). B,C: Sectioning through the graft region following double immunostaining using Pax-2 (B) and QCPN (C) antibodies revealed in merged images that quail cells reached a location posterior to the sixth somite level express Pax-2 (marked by white lines). Nt, neural tube; S, somite.

Competence of Anterior IM Cells to Respond to Kidney Inductive Signals Changes Along the Migratory Pathway From the PS

The findings from the previous experiments suggested the presence of kidney-inductive signals in regions that are anterior to the sixth somite axial level (Fig. 1, transplantations 2, 3, 4; Figs. 3, 4, 5). Because prospective anterior IM of PS3–4 is competent to respond to inductive signals in posterior IM regions (Fig. 1, transplantations 7, 8; Fig. 7) and normally this tissue terminates in a non–kidney-generating IM that loses its ability to express kidney genes (Fig. 1, transplantation 5; Fig. 6), we hypothesized a role for the migratory pathway from prospective anterior IM of PS3–4 to anterior IM6 in altering this competence to respond to inductive signals. We thus asked whether bridging over this normal migratory pathway would affect the competence to respond to inductive signals in anterior regions. Prospective anterior IM from PS3–4 of quail donor was transplanted into the anterior IM6 of chick embryo (Fig. 1, transplantation 9). After incubation, the transplanted tissue reached a location in the anterior IM region. ISH for Lim-1, followed by sectioning and double-immunostaining using QCPN and the Pax-2 antibodies, revealed an overlapping ectopic expression of these three markers in the anterior region (Fig. 8; Table 1). This finding suggests that prospective anterior IM of PS3–4, which is the source tissue of IM6, by omitting its normal migratory pathway (this experiment), maintains its competence to respond to inductive signals.

Figure 8.

Prospective anterior intermediate mesoderm (IM) of primitive streak (PS) 3–4 expresses kidney genes when transplanted into anterior IM6. A: Prospective anterior IM of PS3–4 of quail embryos was transplanted into anterior IM6 of chick embryos (Fig. 1, transplantation 9). B: In situ hybridization for Lim-1 revealed ectopic expression of this gene in anterior non–kidney-generating IM (arrow). C–E: Sectioning through the graft region (C) and double immunostaining using Pax-2 (D) and QCPN (E) antibodies, confirmed the quail origin of the graft (marked by white lines).

Control transplantation experiments (Fig. 1, transplantations 10, 11, 12; Table 1) in which the transplanted tissue was inserted into the same region of origin in the host, were performed to exclude the possibility that the manipulation process itself had affected the results. No change in original fate was observed in these experiments (data not shown). Another internal control, in which grafts of prospective anterior IM from PS3–4, prospective posterior IM from PS6 and anterior IM6 transplanted into PS and migrated into non-IM tissues (such as paraxial and lateral mesoderms), resulted in no expression of kidney genes. However, in contrast to this observation, IM8, which terminated in non-IM regions, expressed kidney markers, confirming its autonomy (data not shown).

DISCUSSION

Gene expression patterns that exhibit clear borders in the developing embryo raise the fundamental question of how such borders are determined and what are the regulatory mechanisms that govern cell fate specification within the same type of tissue. The anterior expression border of Hox genes is probably the most discussed phenomenon along the A-P axis, yet little is known regarding the complex cellular and molecular regulatory mechanisms of these anterior borders (Gellon and McGinnis,1998; Burke and Nowicki,2001).

Burke (2000) distinguishes between two types of “positional information” (as termed by Wolpert,1969). One type is that of local information, which influences the final differentiation of particular cells in the context of an overall pattern of the embryo. The second type is that of global information, which determines global patterning. Global patterning belongs to the higher levels of the developmental hierarchy and may serve as a type of information determining the location and timing of the activity of the local information. In that sense, creating borders is part of global information, whereas differentiation of tissues on each side of the border is within the scope of local information. Accordingly, we suggest that the activity of a particular tissue in response to local signals that result in differentiation is part of a process that is the consequence of local information. However, creating the ability to respond to those local signals is the result of global information and global patterning. In this context, the competence of cells to respond to inductive or inhibitory signals (the local information) reflects one of the ways in which global information can be translated into a specific pattern. Consequently, we can say that this study is about the competence of cells to respond to local signals (inductive and inhibitory); a competence that creates the anterior border of the kidney field. The conclusions that arise from our results are summarized in Figure 9.

Figure 9.

Patterning of the anterior non–kidney-generating intermediate mesoderm (IM). Diagram depicting two alternative models of the change in cell competence during cell migration from primitive streak (PS) 3–4 to the anterior IM. In the left pathway (model A), cells that migrate through prospective anterior IM of PS3–4 are competent to respond to two opposing signals: kidney-inductive signals and kidney inhibitory signals. By the time the cells reach the anterior IM6, they are no longer competent to respond to kidney-inductive signals present in anterior nonkidney IM. According to this model, the inhibitory signals change the competence of migrating cells to respond to kidney-inductive signals; therefore, despite the presence of such inductive factors in the anterior environment, this region does not express the kidney program. As can be seen in this model, the kidney inhibitory signals are expressed in a narrow time window, which restricts the inhibitory effect only to early stages. In the middle pathway (model B), kidney inducing factors may not be expressed in the embryo until some time after the prospective anterior IM has gastrulated (approximately stage 6). By the time PS cells reach the anterior IM, they are no longer competent to respond to kidney-inducing factors. Prospective posterior IM of PS6 (right pathway) is competent to respond to kidney-inductive signals that are present along their migratory pathway to become posterior IM8. By that time, this tissue is determined to its kidney fate and is autonomous.

Posterior IM8, Which Gives Rise to Kidney-Generating IM, Is Already Committed to Its Kidney Fate

Consistent with previous studies (Mauch et al.,2000; James and Schultheiss,2003), the above observation is based on three complementary results: (1) ectopic expression of kidney genes in anterior nonkidney IM (transplantation 1), (2) ectopic expression of kidney genes in different mesoderms (paraxial and lateral plate mesoderms) and in premature IM parallel to the presegmented mesoderm where kidney genes are normally not expressed (data not shown), and (3) kidney marker expression in the same tissue that was isolated from surrounding tissues and cultured for 24 hr (data not shown). From the in vivo experiments, we concluded that different mesodermal environments cannot alter kidney gene expression in posterior IM8. From the culture experiment, it was clear that no extrinsic signals are required to turn on kidney gene expression, although both surface ectoderm and endoderm were included in the explants. Obara-Ishihara et al. (1999) showed that the overlaying ectoderm is a source of kidney-inductive signals, whereas two later studies by Mauch et al. (2000) and James and Schultheiss, (2003) demonstrated that the ectoderm is required for survival and proliferation of the mesoderm but not for the induction of kidney genes. Mauch et al. (2000) also provided evidence that the endoderm has no role at these stages in kidney induction. It is important to note that the study by Obara-Ishihara et al. (1999) was performed on later stages than those used in the present study, in which BMP-4 is expressed in the surface ectoderm. Based on these other reports and on our own experiments, we can conclude that posterior IM8 is already autonomous in relation to the A-P axis too.

Prospective Posterior IM of PS6 Is Committed to its Kidney Fate but Requires Extrinsic Signals to Turn on Kidney Genes

In the in vivo experiments, prospective posterior IM of PS6 expressed kidney genes upon reaching anterior IM, a non–kidney-generating IM (Fig. 1, transplantations 2, 3, 4). The requirement for signals from surrounding tissues to induce kidney genes was confirmed by cultures of isolated tissues (Fig. 3) and the barrier experiments (Fig. 5), in both of which Lim-1 expression was not observed. However, in both experiments, Pax-2 expression was up-regulated, indicating that these two genes are controlled differently. The inductive signals anterior to the sixth somite level are not limited only to early stages but also persist in later stages when anterior somites are already formed (Fig. 1, transplantation 3). From these observations, we suggest the presence of kidney-inductive signals in anterior nonkidney regions at least for the induction of Lim-1.

Holtfreter (1933, cited in Fales,1935) showed that presumptive gastrula stage epidermis that was transplanted into the presumptive neurula stage kidney field in three different urodele species was induced to form pronephric tubules. Those experiments found that a diffuse nephric induction field extends from the gills to the anus, with highest induction properties in the pronephric region (Holtfreter, 1933, cited in Fales,1935). However, it is not clear from these experiments whether this anterior inductive field is anterior to the normal kidney morphogenetic field that extends from the most anterior somites (are parallel to the gills) posteriorly. The current study investigates a region that is significantly anterior to the known kidney morphogenetic field. Previous studies have suggested that kidney-inductive signals emanate from the paraxial mesoderm (Seufert et al.,1999; Mauch et al.,2000), although none of these studies have shown kidney induction anterior to the normal kidney morphogenetic field. Moreover, the study by Mauch et al. (2000) provides evidence that signaling from the paraxial mesoderm is both required and sufficient for the induction of the pronephros and that the neural tube and notochord are not a source of such inductive signals consistent with another study in Xenopus (Seufert et al.,1999). In our barrier experiments, we also show that kidney-inductive signals in anterior regions emanate from tissues that are located medial to the IM. Taken together, our results and the above-mentioned studies suggest that kidney-inductive signals secreted from the paraxial mesoderm also exist in anterior nonkidney regions.

The commitment of prospective posterior IM of PS6 to its kidney fate contrasts with that made by James and Schultheiss (2003), in whose study the prospective IM region from PS5–6 was transplanted into other regions of the PS. In their experiments, the transplanted tissue changed its fate and did not express the kidney marker Pax-2. When interpreting the results of both studies, it is apparent that they are the consequence of two different experimental conditions, which indicate the different abilities of different environments to interact with the same tissue. These differences may reflect the different molecular states of the interacting tissues. Therefore, we believe that the term “committed” is relative to experimental conditions and type of tissue interaction and, therefore, cannot present an absolute state (for definitions, see Slack,1983, p.18).

If, indeed, kidney-inductive signals are secreted from the paraxial mesoderm along the migratory pathway of cells that normally terminate in anterior non–kidney-generating IM, then the question must be, what kind of positional information exists in that time window that patterns the cells to become nonkidney tissue? First, however, it is necessary to understand the state of specification of anterior IM6, the result of the discussed migratory pathway.

Prospective Posterior IM of PS6 Can Alter the Commitment of Anterior IM6 to Its Nonkidney Fate and Transform It Into Kidney-Generating IM

We have shown that anterior IM6 when transplanted into the posterior IM8 (transplantation 5) did not express kidney genes. However, when this tissue was transplanted into prospective posterior IM of PS6 (transplantation 6), after reaching a location posterior to the sixth somite axial border, it changed its original fate and expressed kidney genes (Fig. 6). Cultures of the same tissue clearly showed that alone, without surrounding tissues, anterior IM6 did not express kidney markers. The findings suggest that the environment of the prospective posterior IM of PS6 is capable of changing the competence of cells in the anterior nonkidney IM and reprogram this tissue to respond to kidney-inductive signals. The ability of certain locations in the PS and in Hensen's node to reprogram other locations of the streak and the node, respectively, has already been demonstrated in several studies (Selleck and Stern,1992; Garcia-Martinez and Schoenwolf,1992; Garcia-Martinez et al.,1997; James and Schultheiss,2003). However, it is still too early to speculate on the molecular mechanisms in which such activity occurs and the meaning of the ability to dominate other specified tissues.

Cells That Migrate Through Prospective Anterior IM of PS3–4 Lose Their Competence to Respond to Kidney-Inducing Signals

It has already been demonstrated that prospective anterior IM of PS3–4 normally does not express Lim-1 or Pax-2 (Chapman et al.,2002; H. Barak and R. Reshef, unpublished data). The competence to respond to kidney-inductive signal was evident from two experiments in which cells of prospective anterior IM of PS3-4 expressed kidney genes whether exposed to the PS environment (Fig. 1, transplantation 7) or directly transplanted into posterior IM8 (Fig. 1, transplantation 8; Fig. 7), a kidney-generating IM. From these experiments, we can conclude that anterior IM6 (nonkidney tissue) that originated in prospective anterior IM of PS3–4 was competent to respond to kidney-inductive signals and that competence was changed during the migration of cells from the PS to become anterior IM tissue. Taking together the results obtained in transplantations 7, 8 and transplantations 5, 6, we can speculate that the ability of prospective posterior IM of PS6 to reprogram anterior IM6 and turn it into kidney tissue is based on the previous “ontogenic competence” of prospective anterior IM of PS3–4 to respond to kidney-inductive signals.

The main evidence for the time window and location in which cells of prospective anterior IM of PS3–4 lose their competence to respond to kidney-inductive signals is found in transplantation 9. Prospective anterior IM from PS3–4 was “bridged-over” its own migratory pathway and transplanted directly into anterior IM6. This manipulation terminated in ectopic expression of kidney genes in nonkidney IM (Fig. 8). The loss of competence to respond to kidney-inducing signals could be due to several mechanisms (Fig. 9). One possibility is a change in cell competence occurring right after cells gastrulate in early stages, which causes these cells to lose their ability to respond to kidney-inductive signals. This change in cell competence may involve the localized activity of inhibitors along the migratory pathway from the primitive streak of early stages to the prospective anterior nonkidney IM. As a result, the anterior border of the kidney field is situated in a more posterior position, despite the presence of kidney-inductive signals in the anterior segments. An alternative interpretation is that kidney inducing factors may not be expressed in the embryo until some time after the prospective anterior IM has gastrulated. If cells lose competence to respond to kidney-inducing signals within a certain time window after gastrulation, then the prospective anterior IM may have lost its competence to respond before kidney inducing factors are expressed. Note that this latter model does not require that the prospective anterior IM be any different from other prospective IM tissue. The more posterior IM could also undergo progressive loss of competence but, because it gastrulates later, it would simply encounter kidney inducing signals while it is still competent to respond to them.

Two Main Themes Emerge From the Results of This Study (Fig. 9)

The first postulates that kidney-inductive signals are present in the area of the anterior non–kidney-generating IM. These signals are probably secreted from the adjacent paraxial mesoderm, as previously described for more posterior regions (Mauch et al.,2000). The induction of kidney properties by kidney-inductive signals in anterior nonkidney IM is dependent on the competence of cells to respond to such signals. The second theme suggests that IM that lies anterior to the sixth somite axial level does not express kidney genes, as a consequence of a change in the competence of these cells to respond to kidney-inductive signals as proposed above in two alternative mechanisms. This result is a reminiscent of previous observations in lithium-dorsalized Xenopus embryos, showing the loss of capacity to form cell competence to respond to kidney-inductive signals emanating from dorsal tissues (Seufert et al.,1999). These conclusions may provide a cellular mechanism to explain the formation of the sharp anterior border of kidney genes at the sixth somite axial level.

Several Evolutionary Implications Arise From This Study

As previously described in marine vertebrates such as fishes and the cyclostomes group, the anterior border of the embryonic kidney field is situated in the most anterior segments of the body (Romer,1970; Hildebrand,1974). Moreover, in the prechordates such as the cephalochordate amphioxus (Branchiostoma lanceolatum), the anterior-most, unpaired, Hatschek's nephridium, develops in association with the larval mouth (Ruppert,1994). Although some studies, including Goodrich (1930) and Romer (1970), suggest that this nephridium is not homologous to the vertebrate kidney, recent studies in Amphioxus have found expression of AmphiPax2/5/8 and AmphiNotch (homologs of the subfamily of Pax-2, Pax-5, Pax-8, and Notch, respectively) in the rudiment of the developing Hatchek's nephridium (Kozmik et al.,1999; Holland et al.,2001). These findings strengthen the notion that the kidney morphogenetic field in prechordates lies in the most anterior segment of the body. In hagfishes, the pronephros (also called head kidney), develops from the head posteriorly and persists throughout life (Romer,1970; Hildebrand,1974). During the transition from sea to land, a new trend of posteriorization of the kidney morphogenetic field appeared, as evident in terrestrial vertebrates that exhibit only the mesonephric- or metanephric-type of kidney.

Although the developmental mechanisms that generated the posteriorization of the kidney field during vertebrate evolution are not known, the current study suggests possible ways in which the kidney developmental program could be altered to create a posteriorization of the kidney field. The change in cell competence of early gastrulating IM cells to respond to kidney-inductive signals in anterior regions can serve as a plausible way in which an evolutionary mechanism operates. The two developmental mechanisms that were suggested above (Fig. 9) can explain such a change in cell competence and, hence, explain this posteriorization shift. Both of these potential mechanisms can be considered instances of heterochrony (Gould,1977). Such an event takes place at the higher level of global patterning, and processes that execute such patterning occur, as observed in our study, during the course of gastrulation.

EXPERIMENTAL PROCEDURES

Embryos

Fertile white Leghorn chicken and Japanese quail (Coturnix coturnix japonica) eggs were incubated at 38°C in a humidified incubator until embryos had reached stages 3–4, 6, and 8 (Hamburger and Hamilton,1951).

Microsurgery and Culture

The chick–quail chimera technique was used in the in vivo experiments (Le Douarin,1969). The existence of a monoclonal antibody (QCPN) that recognizes quail but not chick tissue allows us to distinguish between donor and host cells. In these transplantation experiments, the chick host embryo was cultured using a modified form of New culture (New,1955; Chapman et al.,2001; James and Schultheiss,2003). The donor quail graft (approximately 100μ × 80μ) was cut by using tungsten needles and labeled by using cell tracker CM-DiI (Molecular Probes) in 0.3 M glucose. A small piece of tissue from the chick embryo was removed and the graft piece was placed into the vacated site, approximately 600 μ posterior to the posterior end of Hensen's node or into the IM (James and Schultheiss,2003). In experiments for which a cellophane barrier was used (Heanue et al.,1999), the grafts were placed as described above into two parallel sides of the embryo and the barrier was inserted next to one of the grafts. The embryo was cultured for 18–24 hr, judged by using a fluorescent stereomicroscope (Leica, MZFLIII) to detect and photograph (DC-290 Kodak digital camera) DiI-labeled migrating cells. This strategy was to detect their migratory pathway and final location after incubation. A successful experiment was defined as one in which grafted cells migrated into the IM and terminated in locations along the A-P axis comparable to those expected in normal development. The experiments were performed on one side of the embryo, while the contralateral side remained as control. Embryos were analyzed for different IM markers. Pax-2 and Lim-1 expression were used to ascertain the fate of the transplanted tissue. Embryos were analyzed by whole-mount RNA ISH and immunohistochemistry to detect expression at the single cell level. The embryo was fixed in 4% paraformaldehyde (PFA) overnight at 4°C, dehydrated in graded methanol, and frozen at −20°C until analysis.

Culture experiments were performed to understand the ability of the investigated tissues in isolated conditions to express kidney genes. Chick embryonic tissue was cultured in type I collagen gels (Münsterberg et al.,1995) suffused with defined tissue culture medium: DMEM-F12 (Gibco), 5 μr/ml human transferrin (Gibco), 100 μr/ml conalbumin (Sigma), 1× insulin–transferrin–selenium (Gibco), 1% Pen-Strep, and 1% L-glutamine. The cultures were incubated up to 24 hr (equivalent to the in vivo experiments) and fixed in 4% PFA. In cultures of IM regions, we used all three germ layers. According to James and Schultheiss (2003), the ectoderm is required for IM survival in cultures, but it does not play a specific role in induction of kidney genes at these embryonic stages. Furthermore, Mauch et al. (2000) demonstrated that neither ectoderm nor endoderm are required for pronephros formation at early stages of IM development.

Whole-Mount RNA ISH and Sectioning

ISH was performed as described previously (Schultheiss et al.,1995) by using probes for chick Lim-1 (Tsuchida et al.,1994) and chick Pax-2 (Burrill et al.,1997; Herbrand et al.,1998). All probes recognized both chick and quail mRNAs in control embryos (data not shown). Sectioning was carried out by using 10-μm cryostat sections (Leica) on gelatin-embedded embryos, as previously described (Schultheiss et al.,1995). Images were collected from an Axioskop (Zeiss) microscope with a DC-290 Kodak digital camera.

Immunohistochemistry

For chick–quail chimeric experiments after ISH, we used the following immunohistochemistry method. The sections were washed in phosphate buffered saline (PBS) for 10 min, washed in 0.25% Triton X-100 in PBS for 20 min, soaked in blocking solution (2% sheep serum, 0.1% Tween 20 in PBS) for 30 min, then incubated with primary antibodies for 2 hr at room temperature. The primary antibodies were the monoclonal antibody QCPN (undiluted) that recognizes specifically quail cells (Developmental Studies Hybridoma Bank) and the polyclonal anti-rabbit Pax-2 (1:50) (BAbCO) antibodies. The sections were washed three times for 5 min, and then incubated with the secondary antibodies for 1.5 hr at room temperature. The secondary antibody was donkey anti-mouse, rhodamine (1:100; Jackson Laboratories) and goat anti-rabbit, fluorescein isothiocyanate (1:50; Sigma). To visualize nuclei, the sections were washed once in 1 μg/ml DAPI (Sigma) in PBS.

Culture experiments were fixed in 4% PFA for 30 min at room temperature and then washed three times for 5 min each in PBS at room temperature. The cultures were gelatin-embedded and cryosectioned in 10-μm sections. Because two hybridoma antibodies were used, we separated sequential sections into two slides. For immunohistochemistry, we used the same protocol as previously described with some modifications: for all the washes, we used PBS with 0.02% Tween 20 (PBTween), and the blocking solution contained 1% bovine serum albumin (BSA), 1% sheep serum, 1% horse serum in PBTween. The primary antibodies were the polyclonal rabbit anti–pax-2 (1:50) and the monoclonal antibodies 4F2 against lim-1 (1:45; Developmental Studies Hybridoma Bank) and the monoclonal Pax-7 antibodies (undiluted; Developmental Studies Hybridoma Bank). After incubation with the primary antibodies, the sections were washed three times in PBTween for 5 min at room temperature, then incubated with 1% BSA in PBTween for 30 min at room temperature, followed again by three washes with PBTween for 5 min each. The secondary antibodies were as previously described.

The Developmental Studies Hybridoma Bank antibodies are maintained by the Department of Pharmacology and Molecular Sciences, The Johns Hopkins University School of Medicine (Baltimore, MD) and the Department of Biological Sciences, University of Iowa (Iowa City, IA), under contract N01-HD- 2-3144 from NICHD.

Acknowledgements

The authors thank Claudio Stern, Richard James, Einat Halperin, Yuval Rinkevich, Ella Preger Ben-Noon, and Haggar Harduf for discussion on various aspects of this study. We also thank Orna Halevy for the Pax-7 antibodies. This study was supported by the Forchheimer Foundation student fellowship to H.B., Fogarty International Research Collaboration Award (FIRCA) to T.M.S. and R.R., and grants to R.R. from the Mallat Family Fund, Israel Cancer Research Fund, Israel Science Foundation, and Matilda Barnett Revocable Trust.

Ancillary