Engraftment and Differentiation of Human Metanephroi into Functional Mature Nephrons after Transplantation into Mice Is Accompanied by a Profile of Gene Expression Similar to Normal Human Kidney Development

Abstract

ABSTRACT. Metanephroi, the embryonic precursors of the adult kidney, can be induced in vivo to grow and develop. Despite their potential clinical utility for transplantation, the ability of human metanephroi to differentiate after transplantation into functional mature nephrons is mostly unknown. To address this, 70-d human metanephroi were transplanted into NOD/SCID mice; global gene expression patterns that underlie development of human metanephric transplants were analyzed and compared with normal human kidney development. In addition, functionality of the grafts was assessed by dimercaptosuccinic acid radioisotope scans at different times after transplantation. The results of hybridization to cDNA arrays when RNA was derived from normal human kidneys at 8, 12, 16, and 20 wk gestation demonstrated that a subset of 240 genes changed substantially with time. The induced genes can be classified as cell cycle regulators, transcription and growth factors, and signaling, transport, adhesion, and extracellular matrix molecules. Strikingly, clustering analysis of global gene expression at 2, 6, and 10 wk after metanephric transplantation revealed an expression profile that characterizes normal human kidney development. Moreover, maturation of the transplants was accompanied by an increased uptake of dimercaptosuccinic acid. Nevertheless, expression levels of specific genes were mostly found to be suppressed in the transplants compared with the normal kidneys. These data provide insights into human kidney development and support the possibility of the transplantability of human metanephroi. Understanding of the molecular regulation of the transplanted developing metanephroi might lead to the development of strategies aimed at increasing the levels of specific genes, nephron endowment, and graft function.

Kidney transplantation has been one of the major medical advances of the past 30 yr; however, it is becoming increasingly apparent that the supply of organs is limited and will not improve with current medical practice. The morbidity, poor quality of life, and extensive burden on the health of patients receiving chronic dialysis are major factors that prompt the development of alternatives for patients with chronic renal failure. The possible use of human renal precursor cells or tissue (i.e., human metanephroi) seems worthy of consideration but has hardly been explored.

Several reports have documented development of murine metanephroi after transplantation (^1–3). Rogers et al.⁽⁴⁾ studied the feasibility of transplanting murine metanephroi in the omentum of rat hosts sufficiently near one of the ureters so as to render connection between the transplant and the host’s urine system possible. When the murine transplants were connected to the host’s urine system, they were shown to be functional, and clear inulin was infused into the host’s circulation. Such anastomosis was made possible after transplanted murine metanephroi had developed.

We have previously established human adult and fetal kidney grafts in immunodeficient rat hosts (^{5, 6}). The fetal transplants achieved rapid growth and long-term survival in vivo. Moreover, when studying the transplant immunology of the human kidney precursors, our small-animal model demonstrated a reduced alloimmune response of allogeneic human peripheral blood mononuclear cells to the metanephric transplants when compared with that of transplants of adult kidney tissue (^5–7). Thus, the immunologic advantage of human metanephroi over the adult kidney in avoiding graft rejection was shown ⁽⁸⁾. Clearly, in vivo differentiation of human metanephric transplants into functional mature nephrons is critical if such were to be applicable as donor tissue in clinical practice. Moreover, a better understanding of the molecular signals that stimulate this process might lead to improvement in the differentiative potential after transplantation.

Therefore, in this report, with the use of cDNA arrays, we have examined the global mRNA expression of cell cycle and DNA replication regulators, transcription and growth factors, and signaling, transport, and adhesion and extracellular matrix (ECM) molecules in human fetal kidneys at various times during gestation. Moreover, we were able to monitor gene expression patterns in vivo in our model of transplanted human metanephroi and to determine to what extent they resemble those involved in the induction of the normal kidney or the transformation into an embryonic kidney malignancy (i.e., a Wilms tumor). In addition, we assessed the functional capacity of the newly formed nephrons in the developing human metanephric transplants.

Materials and Methods

Mice

NOD/SCID mice were bred and maintained under defined flora conditions at the Weizmann Institute of Science (Rehovot, Israel) in sterile microisolator cages. All of the experiments were approved by the animal-care committee of the Weizmann Institute. Host mice used for transplantation of metanephroi were 8 wk old.

Human Fetal Kidneys

Normal human kidneys at 8, 12, 16, and 20 wk gestation were obtained after curettage of elective abortions. Studies with embryonic kidney tissue were approved by the Helsinki ethical committee. The kidneys were fixed in 10% paraffin, sectioned, and mounted on slides coated with poly-L-lysine, and sections were stained with hematoxylin and eosin for histologic evaluation.

Establishment of Human Kidney Grafts into NOD/SCID Mice

Human embryonic kidneys, 70 d gestation, were kept in sterile conditions at 4°C (for approximately 2 h) in either RPMI 1640 or Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum (Biological Industries, Bet HaEmek, Israel). Transplantation of the human metanephroi was performed with the mice under general anesthesia (2.5% Avertin in phosphate-buffered saline, 10 ml/kg intraperitoneally). Both kidneys were exposed through a bilateral incision. A 1.5-mm incision was made at the caudal end of the kidney capsule, and a 1-mm³ fragment of human kidney was implanted under the left capsule. Mice that received transplants were treated postoperatively with ciprofloxacin in their drinking water for 7 d. Animals were killed at 2, 6, and 10 wk after transplant. Human renal tissue was initially assessed for engraftment by macroscopic examination (size, color) of the transplants at the subcapsular site. Kidneys and their capsules were then removed and fixed in 10% paraffin. Human grafts were sectioned and mounted on slides coated with poly-L-lysine, and sections were stained with hematoxylin and eosin for histologic evaluation.

RNA Preparation

Normal human embryonic kidney tissue and human embryonic kidney grafts dissected carefully from the subcapsular site of NOD/SCID mice were homogenized with a glass-Teflon tissue tearer in 1 ml of Tri-reagent (Molecular Research Center, Cincinnati, OH). RNA was isolated as described previously ⁽⁹⁾.

cDNA Array Hybridization

We used cDNA arrays of human cDNA spotted on a nylon membrane (Atlas filter arrays; Clontech, Palo Alto, CA) as previously described ⁽⁹⁾. We used the human 1.2 expression arrays (broad-coverage arrays), which contain 1176 cDNA. The filters also include housekeeping control cDNA. A complete list of the cDNA and controls and their accession numbers is available on the Internet (http://www.clontech.com/atlas/genelists/index.html).

Total RNA were treated with DNAse I according to the manufacturer’s instructions and used for cDNA synthesis. A 3-μl mix containing 5 μg of total RNA and 1 μl of 10× cDNA synthesis primer mix (specific for each filter array, provided by Clontech) was incubated at 70°C for 2 min, followed by incubation at 48°C for 2 min. To this mix, 8 μl of master mix containing 2 μl 5× reaction buffer, 1 μl 10× dNTP mix, 3.5 μl [³²P]-dATP (3000 Ci/mmol, 10 mCi/ml; Amersham Pharmacia Biotech, Buckingham, UK), 0.5 μl 100 mM dithiothreitol, and 1 μl Moloney murine leukemia virus reverse transcriptase (50 U/μl) was added, mixed, and incubated for 25 min at 48°C. The reaction was terminated by adding 1 μl of 10× termination mix at room temperature.

The radioactively labeled cDNA mix was fractionated on a Chroma Spin-200 column (Clontech), and fractions that comprised the first peak of radioactivity were pooled for each cDNA synthesis reaction. In each set of hybridization, equal counts were taken for control and experimental labeled cDNA probes. The labeled cDNA probe was then mixed with 1/10 volume of 10× denaturing solution (1 M NaOH, 10 mM ethylenediaminetetraacetate) and incubated at 68°C for 20 min, followed by the addition of 5 μl (1 μg/μl) of cot-1 DNA and an equal volume of 2× neutralizing solution (1 M NaH₂PO₄, pH 7.0), and incubated at 68°C for 10 min. Denatured, labeled cDNA was then added to 5 ml of ExpressHyb solution (Clontech) with 1 mg of sheared salmon sperm DNA (Sigma Chemical Co., St. Louis, MO) and mixed. This hybridization solution was added to the Atlas cDNA Expression Array membrane, which was prehybridized in 10 ml of ExpressHyb hybridization solution at 68°C for 1 h. Hybridization proceeded overnight at 68°C in a roller bottle. Membranes were washed once with prewarmed 2× standard saline citrate/1% sodium dodecyl sulfate for 30 min and once or twice with 0.5× standard saline citrate/0.5% sodium dodecyl sulfate for 30 min at 68°C with constant agitation. The membranes were exposed to Fuji (Tokyo, Japan) x-ray films at 70°C with intensifying screens.

Analysis of Hybridization Signals

The cDNA microarray autoradiograms were scanned, and the images were analyzed by AtlasImage software, version 1.05 (Clontech). The background was calculated by default external background calculation, which takes into consideration the background signals at the blank space between the different panels of the arrays. Signal threshold was set as background-based signal threshold. The signal intensities were normalized globally by the sum method (AtlasImage; Clontech). A report of differentially expressed genes was generated on the basis of ratio and intensity differences.

Reverse Transcriptase-PCR Amplification

Total RNA was determined, and 1 μg of total RNA was reverse-transcribed into cDNA and amplified with the Access reverse transcriptase-PCR kit (Promega, Madison, WI) and with specific primers for human neu-differentiation factor (NDF), homeobox protein hLim1, mitogen-activated protein kinase p38, kidney glomeruli chloride channel, and our housekeeping gene, β-actin. The primers used were as follows: 5′-ACC-ATC-AAG-CTC-TGC-GTG-ACT-G-3′ (sense) and 5′-GCA-GGT-CAG-TTC AGT-TCC-AGG-TC-3′ (antisense) for β-actin (310 bp); 5′-CAA-AGA-AGG-CAG-AGG-CAA-AG-3′ (sense) and 5′-ACC-ACT-TGA-ATC-TGA-GAG-AGG-3′ (antisense) for human NDF (340 bp); 5′-CGT-CGT-CTT-CTT-CTC-3′ (sense) and 5′-CAG-GTT-GCA-TTT-ACA-TTC-3′ (antisense) for human Lim1 (506 bp); 5′-CCC-AGG-AGT-CCG-TAA-GTA-G-3′ (sense) and 5′-ACT-GGA-GAC-AGG-TTC-TTG-3′ (antisense) for human mitogen-activated protein (p38) kinase (200 bp); and 5′-TGA-TGT-GAT-ATG-GCT-GCA-AG-3′ (sense) and 5′-TGT-AGG-CAA-AGG-CTC-CCT-C-3′ (antisense) for human kidney chloride channel (700 bp). After amplification, the sample was separated on an agarose gel containing ethidium bromide, and bands were visualized and photographed with an ultraviolet transilluminator. The image densities were measured with Image software, version 1.62 (National Institutes of Health, Bethesda, MD), and semiquantitative results were expressed as a ratio of each reverse transcriptase-PCR product to β-actin density ⁽¹⁰⁾. Three independent experiments were performed for each PCR determination.

Immunohistology

Immunohistochemistry for the stem cell leukemia (SCL) protein was performed via a monoclonal anti-SCL/TAL1 antibody (2TL242; provided by Dr. Karen Pulford, Oxford, England, UK). Briefly, 5-μm sections of were mounted on Super Frost/Plus glass (Menzel, Glazer, Braunschweig, Germany) and processed by the labeled streptavidin-biotin method with a Histostain Plus kit (Zymed, San Francisco, CA). Heat-induced antigen retrieval was performed by controlled microwave treatment with a model H2800 processor (Energy Bean Sciences, Agawan, MA) in 10 mM citrate buffer, pH 6.0, for 10 min at 97°C. The sections were treated with 3% H₂O₂ for 5 min. Consecutive sections were incubated for 1 h with the 2TL242 antibody. Negative control incubations were performed by substituting nonimmune serum for the primary antibody (the control sections were entirely negative).

Biotinylated second antibody was applied for 10 min, followed by incubation with horseradish peroxidase-conjugated streptavidin for 10 min. After each incubation, the slides were thoroughly washed with OptiMax wash buffer (Biogenex, San Ramon, CA). The immunoreaction was visualized by a horseradish peroxidase-based chromogen/substrate system, including diaminobenzidine (brown) chromogen (liquid diaminobenzidine substrate kit; Zymed). The sections were then counterstained with Mayer’s hematoxylin, dehydrated, and mounted for microscopic examination.

Statistical Analyses

Comparisons between groups were evaluated by the unpaired t test. Data were expressed as mean ± SEM and were considered statistically significant at P ≤ 0.05.

Dimercaptosuccinic Acid Radioisotope Scans after Intra-abdominal Transplantation of Human Metanephroi

In some experiments, metanephroi were implanted and sutured (5-0 suture) onto the testicular fat of NOD/SCID mice while they were under general anesthesia (2.5% Avertin in phosphate-buffered saline, 10 ml/kg intraperitoneally). During the same surgery, host mice had one kidney removed. Dimercaptosuccinic acid (DMSA) scans were performed at different times after transplantation. Briefly, each mouse was injected intravenous in a tail vein with 1 × 10⁻³ mCi technetium-99m DMSA. Animals were sedated 2 h after injection with 2.5% Avertin in phosphate-buffered saline, 10 ml/kg intraperitoneally. Whole-body images were acquired in the posterior view on a gamma camera equipped with a pinhole collimator (Elscint Sp4, Haifa, Israel). Images were acquired to a total of 100 K. After imaging, animals were killed, and the transplants were excised and imaged. The mice were further imaged in the same manner after the transplants had been removed.

Results

Histology of Human Metanephric Transplants Established in NOD/SCID Mice

Figure 1 illustrates the histology of normal human kidneys at 10 and 20 wk gestation (Figure 1, A through C), as well as of human metanephroi derived from 70-d-old aborted fetuses 10 wk after transplant into an immunodeficient host (Figure 1, D and E). The human metanephroi, the precursor of the adult kidney, appears at 5 wk gestation, when it consists of nephrogenic mesenchyme, which condenses around the ureteric bud epithelium (Figure 1, A and B). Mutually inductive events cause the ureteric bud to branch serially to form the collecting ducts and urothelium of the renal pelvis, ureter, and bladder trigone. The renal mesenchyme undergoes epithelial conversion to form early nephron precursors (comma- and S-shaped bodies) and eventually mature nephrons (glomeruli, proximal tubules, loops of Henle) (Figure 1C).

Figure 1. : Photomicrographs of hematoxylin and eosin-stained sections illustrating the histology of normal renal development and of transplanted human metanephroi. Developing human kidneys are shown at 10 wk gestation (A and B) and 20 wk gestation (C), and human metanephroi are shown 10 wk after transplant into host mice (D and E). (A) Embryonic kidney showing ureteric buds and peripheral branches (ub) surrounded by condensing mesenchyme (original magnification, ×10). (B) Higher-power view (original magnification, ×20) of ureteric bud branches (ub) surrounded by condensing mesenchyme (mc) and early nephron precursors, including S-shaped bodies (s). (C) Fetal kidney showing well developed glomeruli (g) and tubuli (original magnification, ×10). (D) Overview of the developed transplant showing layers of human glomeruli and tubuli, and nephrogenesis in the outer rim of the transplant (original magnification, ×40). (E and F) Higher-power views (original magnification, ×20 and ×40, respectively) of human glomeruli (g), proximal tubules (p), and distal tubule (d) in the transplant.

The first human metanephric glomeruli form at 9 wk. Branching and nephrogenesis continue to occur in the outer rim of the kidney, the nephrogenic cortex, until 34 wk ⁽¹¹⁾. As shown in Figure 1, D through E, the human metanephroi transplanted under the kidney capsule of NOD/SCID mice undergo growth and differentiation with time, including the appearance of new layers of mature glomeruli and tubuli. We also observed a higher proportion of mesenchyme in the developing transplants compared with normal kidney (Figure 1). The appearance of increasing numbers of human nephrons indicated that nephron development or endowment is regulated in the developing metanephric transplants.

Functionality of Human Metanephric Transplants Established in NOD/SCID Mice

After establishment of human metanephric transplants in mice, we assessed their tubular function with ⁹⁹technetium-DMSA renography. We performed intra-abdominal rather than renal subcapsular transplantation to discriminate between uptake of native kidney and transplanted metanephroi (see Materials and Methods). DMSA was administered at different time points after transplant, and animals were then killed. Table 1 summarizes the DMSA experiments. We could not detect DMSA in the younger transplants (3 to 4 wk after transplant). However, positive uptake was clearly demonstrated in the older metanephric grafts (8 to 10 wk after transplant) and declined upon their removal (Figure 2, A through C). Positive uptake of DMSA was found only when unilateral nephrectomy was performed and not when both native kidneys were left intact (Table 1). This might be related to decreased perfusion of the transplant (vascularized by peripheral vessels and not by end-to-end anastomosis) compared with the native kidneys (steal effect). Thus, maturation of the human metanephric transplants with time (3 to 10 wk) was associated with increased uptake of DMSA.

Figure 2. : Renograms illustrating functional assessment of developing human metanephric transplants after intravenous injection of dimercaptosuccinic acid (DMSA) radioisotope. (A) Eight weeks after transplantation into abdomens of host mice. The whole mouse is displayed. k, host kidney; b, host bladder; t, human transplant. (B) After removal of the transplant, reduced intensity levels are now similar to background levels (circled). (C) Radioisotope uptake in the isolated transplant.

Table 1: Functional assessment of intra-abdominal human metanephric transplants after administration of ⁹⁹technecium-dimercaptosuccinic acid (DMSA)

Profile of Gene Expression during Development of Normal and Transplanted Human Metanephroi

We determined temporal programs of gene expression during development of the normal human kidney and compared them to those of developing human embryonic kidney grafts. For human kidney development, cDNA arrays were used to analyze gene expression patterns in normal kidneys obtained at 8, 12, 16, and 20 wk gestation. For the transplantation experiments, human metanephroi originating from a 70-d-old fetus were grafted under the kidney capsule of NOD/SCID mice. Hybridization to cDNA arrays was then performed when RNA used for probes was derived from 2-, 6-, and 10 wk-old metanephric transplants (designated t12, t16, and t20 wk and corresponding to human kidneys at 12, 16, and 20 wk gestation, respectively).

Figure 3 illustrates the results of hybridization to cDNA arrays when RNA used for probes was derived from the 12-, 16-, and 20-wk-old kidneys as well as from the t12-, t16-, and t20-wk transplants. A representative section (all of the E coordinates) from the atlas human 1.2 array (Clontech) representing 196 cDNA that profile transcription factors and DNA-binding proteins, ECM and adhesion proteins, cell surface antigens, and growth factors is shown. Expression levels of 1176 genes in 12-, 16-, 20-, t12-, t16-, and t20-wk kidneys were determined against baseline 8-wk-old kidneys (Figure 3G), as assessed by screening the differences in hybridization results with AtlasImage software. Only when expression levels deviated from that in the 8-wk-old kidneys by a factor >2.0 or <0.5 in at least 2 of the samples from the 12-, 16-, and 20-wk old kidneys or from the t12-, t16-, and t20-wk transplants were they considered substantial.

Figure 3. : Radiographs illustrating differential hybridization to cDNA profiling transcription factors, DNA-binding proteins, cell surface antigens, extracellular matrix and adhesion proteins, and growth factor receptors. RNA used for probes was derived from normal human kidneys at 12, 16, and 20 wk gestation (A through C) and from human metanephroi 2, 6, and 10 wk after transplantation (t12, t16, and t20 wk) (D through F). Note the upregulation of many genes in both normal and transplanted metanephroi profiles (black spots represent expressed genes) when compared with (G), where RNA was derived from 8-wk-old kidneys. (H) RNA used for probes was derived from a Wilms tumor. Note similar expression profile to that obtained for the embryonic 8-wk-old kidney (G).

This analysis identified a subset of 240 genes whose expression changed substantially in both normal and transplanted metanephroi. Only 20 genes changed in either the normal or transplanted kidney groups according to the selection criteria. Thus, the majority of genes induced during the maturation of the normal kidney are found in the developing transplanted metanephroi. The complete list of significantly changed genes is available on the Internet (http://www.weizmann.ac.il/immunology/dekel/). These genes can be grouped into categories on the basis of their functional role (Table 2).

Table 2: Examples of genes induced in developing normal and transplanted metanephroi that have been placed into functional categories

Table 2A: Continued

One measure of the reliability of the changes we observed is inherent in the expression profiles of the genes. For most genes whose expression levels changed, we could see a gradual change over several time points, which thus effectively provided independent measurements for almost all of the observations. Furthermore, an additional check was provided by utilizing three separate samples of RNA for three independent hybridizations to different macroarrays at each time (12, 16, 20, t12, t16, t20 wk). For example, an analysis of expression levels of multiple genes performed with separate RNA samples from 12-wk kidneys showed good correlation (Figure 4A). As an independent test, we measured the expression levels of several genes, NDF, homeobox protein hLim1 (hLIM), mitogen-activated protein kinase p38 (MAPK), and kidney glomeruli chloride channel (CIC-5), one of which has never been described to be involved in kidney development (NDF) by means of a semiquantitative PCR assay ⁽¹⁰⁾. The expression profiles of the genes, as measured by these two independent methods, were similar (Figure 4B). Moreover, the presence of the SCL gene, which has not yet been described in developing metanephroi, was confirmed at the protein level with immunostaining (Figure 4C).

Figure 4. : Reproducibility and verification of the macroarray data. (A) Bar graph illustrating the levels of multiple genes for 3 independent hybridizations comparing RNA isolated from human kidneys at 12 wk gestation. The ratio was calculated by comparison to the hybridization signals in the human kidneys at 8 wk gestation. (B) Line graph illustrating independent verification of the macroarray quantitation. Relative mRNA levels of the indicated genes (neu differentiation factor [NDF], homeobox protein hLim1 [hLIM], mitogen-activated protein kinase p38 [MAPK], and kidney chloride channel [CIC-5]) were measured with a semiquantitative reverse transcriptase-PCR (RT-PCR) assay (right) in the same samples that were used to prepare cDNA probes for macroarray hybridizations (left). The steady-state levels of mRNA measured by semiquantitative RT-PCR were normalized against the β-actin mRNA control and plotted relative to the levels in normal 8-wk-old kidneys so that results could be compared with those from the macroarray hybridizations. (C) Micrograph illustrating verification of the presence of the stem cell leukemia (SCL) protein in the developing metanephroi. Immunostaining for human SCL in human kidneys at 12 wk gestation (left) and 18 wk gestation (right); note staining in mesenchymal cells (left, inset) and glomerular capillary loop (right, inset).

The 240 genes that passed our filtering parameters were further selected for cluster analysis according to previously published parameters (^12–14). Strikingly, unbiased clustering of the experimental groups (human kidneys at 12, 16, and 20 wk gestation, and t12-, t16-, and t20-wk transplants) on the basis of the similarity of their expression profiles corresponded the normal kidney samples with their respective metanephric transplants (i.e., 12-t12, 16-t16, and 20-t20 wk) (Figure 5A). To determine the temporal changes of gene expression in normal and transplanted metanephroi, we applied a self-organizing map algorithm, previously described by Tamayo et al.⁽¹⁵⁾, where each group of experiments was normalized independently. Impressively, the majority of the genes behaved similarly in both groups (Figure 5B).

Figure 5. : Cluster analyses of gene expression profiles of normal and transplanted metanephroi. (A) Similarity index illustrating how experimental groups were clustered hierarchically on the basis of the similarity of their expression profiles by the procedure of Iyer et al. ⁽¹³⁾. Strikingly, the unbiased method clustered the normal kidneys with their corresponding metanephric transplants (i.e., 12-t12, 16-t16, and 20-t20 wk). (B) Self-organizing map algorithm applied to developing normal and transplanted metanephroi for the determination of temporal changes of global gene expression. For independent analysis of normal (left plot in each box) and transplanted kidneys (right plot in each box), the gene ratio was normalized against the 12- and t12-wk time points, respectively. Most genes (numbers are given in each box) were grouped in cluster 0 (c0) and cluster 3 (c3) (highlighted in yellow), demonstrating similar temporal expression profiles. (C and D) Pie diagrams illustrating temporal expression profiles and gene function. Genes in c1 and c3 were categorized into functional groups, and diagrams showing the percentage of genes belonging to each functional category were plotted for c1 (C) and c3 (D). ECM, extracellular matrix.

Thus, we identified in both normal and transplanted metanephroi a cluster of genes (c0) that were downregulated with time (n = 103). This cluster included mostly genes associated with transcription, replication and DNA repair, cell cycle, neuronal development, growth, and signaling (Figure 5C). A second cluster (c3) consisted of genes that were upregulated throughout normal kidney and transplant development (n = 92). Interestingly, this cluster contained mainly genes that function in cell adhesion and shaping of ECM, transport, metabolism, and protein turnover, as well as in growth and signaling (Figure 5D). The temporal profiles of genes in clusters c1 (n = 26) and c2 (n = 19) were somewhat different among normal and transplanted metanephroi. Genes in c1, which were initially diminished in the transplants, consisted of almost all categories, including signaling, transcription, cell cycle, growth and adhesion, and ECM (data not shown).

In contrast, c2 contained genes that were relatively upregulated after transplantation of metanephroi, most of which were related to cell and oxidative stress (9 of 19; glutathione-S-transferases 1, 3, and 12, heat shock proteins 90, 27, 40, and 60 kD, and glutathione and thioredoxin reductase) or angiogenesis (5 of 19; vascular endothelial growth factor [VEGF], VEGF receptors 1 and 2, angiopoietin 1 receptor, and endothelial monocyte-activating polypeptide 2). Thus, our approach could not only determine similar global patterns of gene expression in both normal and transplanted developing kidneys, but also could pinpoint sets of genes that are increased in the metanephric transplants and fall under specific functional categories.

Although cluster analysis provides for profiles of gene expression and the behavior of groups of genes, it does not identify differences at the level of a single gene. We therefore set out to determine differences between average mRNA levels of specific genes in the normal and transplanted kidneys. An example of such an analysis, here applied to specific genes that function in transport and growth and whose expression changed substantially in the induction of the normal and transplanted kidney, is illustrated in Figure 6. Analysis of genes that encode transport proteins (potassium [kv12] and chloride channel [ICLN] proteins, glucose [GLUT1], zinc, and Golgi transporters) demonstrated that although they increased steadily throughout development in both normal and transplanted metanephroi, thus showing similar temporal expression profiles (see cluster c3), mRNA levels were significantly decreased in the developed transplant (t20) compared with levels in the corresponding kidney at 20 wk gestation (Figure 6).

Figure 6. : Temporal expression profiles of specific genes encoding transporters (left column) and growth factors (right column) in normal and transplanted developing human metanephroi. Average intensity levels were determined in human kidneys at 12, 16, and 20 wk gestation (open squares) and in human metanephroi 2, 6, and 10 wk after transplantation (solid triangles). For each gene, intensity levels were normalized to its levels in the normal 8-wk-old kidneys. Three independent experiments were performed for each determination. *, P < 0.05 compared with mRNA level in corresponding transplant. PTN, pleiotrophin; GLUT1, glucose transporter 1; PDGFR, platelet-derived growth factor receptor; FGF7, fibroblast growth factor 7; ICLN, chloride channel protein; IGF-II, insulin growth factor 2; KV12, voltage-gated K⁽⁺⁾ channel 12; RARB, retinoic acid receptor B.

Analysis of growth factor genes, which have recognized roles in kidney development (pleiotrophin [PTN], platelet-derived growth factor [PDGFRβ] and retinoic acid [RARβ] receptors, fibroblast [FGF7] and insulin [IGF-II] growth factors), allows for the determination of factors that possibly could induce transplanted cells to differentiate more efficiently and improve nephron endowment. We found mostly reduced expression levels of these growth factors in the transplanted metanephroi throughout development, especially in the earlier time points after transplantation (t12 and t16) (Figure 6). Furthermore, average expression levels of several genes, which are categorized in different groups but are linked functionally, could also be compared. For instance, molecules that the actions of hepatocyte growth factor (HGF) are dependent on the following (^16–19)): HGF activator (growth), focal adhesion kinase (cell adhesion), follistatin, and activin-β (ECM modifying); GRB-2 and PLC (downstream signaling) were mostly suppressed in the transplants (Figure 7) and therefore might benefit from a supplement of exogenous HGF. Thus, our approach could identify specific factors involved in the differentiation of the embryonic transplants, including those that are relatively lacking at a given time point after transplantation.

Figure 7. : Temporal expression profiles of specific genes encoding growth factors, adhesion and extracellular matrix, and signaling molecules related to the action of hepatocyte growth factor (HGF) in normal and transplanted developing human metanephroi. Average intensity levels were determined in human kidneys at 12, 16, and 20 wk gestation (open squares) and human metanephroi 2, 6, and 10 wk after transplantation (solid triangles). For each gene, intensity levels were normalized to its levels in the normal 8-wk-old kidneys. Three independent experiments were performed for each determination. *, P < 0.05 compared with mRNA level in corresponding transplant. FAK, focal adhesion kinase; GRB-2, growth factor to bound protein 2; PLC, phospholipase C.

Profile of Gene Expression Associated with the Development of Human Metanephric Transplants Does Not Mimic a Wilms Tumor Genotype

The results of hybridization to cDNA arrays when RNA used for probes was derived from a Wilms tumor showed that unlike the embryonic transplants, the expression profile was not similar to that observed during normal human kidney development (Figure 3H). Thus, when Wilms tumor expression levels were normalized to those of 8-wk-old kidneys, we could hardly detect the subsets of genes shown in Table 2 and Figs. 3 and 5 to be induced in normal kidney development and subsequently in the developing metanephric transplants. Moreover, genes that were not induced or upregulated during normal kidney development were significantly upregulated in the tumor specimen (Table 3). For instance, although the β-thymosins were expressed in the normal and transplanted metanephroi, they were significantly upregulated only in the Wilms tumor. These actin G-sequestering peptides, which regulate actin dynamics, have been shown to be overexpressed in neoplastic transformation ⁽²⁰⁾. The clear differences in expression patterns imply that the embryonic kidney transplants do not acquire the Wilms tumor genotype and undergo malignant transformation.

Table 3: Genes induced or upregulated in Wilm’s tumor versus normal human kidneys at 8 wk

Discussion

The clinical use of human embryonic kidney precursors will require the formation of adequate numbers of mature functional nephrons at specific sites for therapeutic effect. This requires insight into regulation of metanephric growth and differentiation after transplantation and comparison to the normal developmental process. Thus, molecular aspects of normal human kidney development had to be defined first. Despite extensive characterization of human kidney development, global expression profiles have not yet been elucidated.

We therefore took a fresh look at the temporal program associated with the induction of the human kidney by use of cDNA arrays representing approximately 1200 known human genes. A coordinated induction of groups of genes was demonstrated during normal nephron formation, including DNA-binding and DNA transcription factors, proto-oncongenes, signaling molecules, growth factors or hormones and their receptors, cell adhesion and cytoskeletal molecules, and ECM glycoproteins and receptors, and genes associated with transport and metabolism. Genes induced in this program encode products that can (1) participate in the molecular regulation of growth and branching of the ureteric duct (e.g., PTN, c-ret, EGF-R, FGF-R, BMP-2, BMP-4, TGF-β, activin, follistatin, MMP-2) (^{17, 21–23}); (2) participate in the molecular regulation of mesenchymal-to-epithelial conversion, formation of basal lamina, cell polarization, and postinductive tubulogenesis (e.g., hLim1, c-ret, c-ros-1, collagen type IV, integrins α6, α3, β1, ZO-1, junction plakoglobin, BDNF/NT3) (^24–27); and (3) modulate angiogenesis, neovascularization, and glomerulogenesis (e.g., VEGF, VEGF-R, EMAP-2, angiopoietin-1 receptor, ephrin receptors, PDGFRα, PDGFRβ, integrin α3) (^{25, 28–30)}).

In addition, we observed genes that have no recognized role in any aspect of kidney development. The NDF gene, for instance, encodes a protein that acts either as a differentiation factor for neuronal development or as a mitogen for mammary tumor cells ⁽³¹⁾. A recent in vitro study revealed that mesenchymally derived HGF, kerotinacyte growth factor could activate autocrine NDF signaling in their epithelial targets ⁽³²⁾. Thus, the NDF protein might act in concert with these growth factors to direct epithelial proliferation, morphogenesis, and differentiation. Another example is the SCL gene. SCL, an established regulator of hematopoiesis, has been shown to play a pivotal role in hemangioblast formation and endothelial development ⁽³³⁾. Thus, its presence in the developing metanephroi might suggest a role in vasculogenesis.

Although many genes were induced during the developmental process, the biologic relevance of our data became apparent only after assigning the genes into functional categories and clustering them into groups on the basis of the similarity of their temporal expression profiles. We could demonstrate the induction of nonspecific regulatory replicative machinery (genes that function in cell cycle, replication and DNA repair, transcription, and growth) that characterizes early development, and thereafter the predominance of groups of genes committed to a more specialized task and terminal differentiation (adhesion and ECM molecules, transporters, and genes related to metabolism and protein turnover). The general outline of these results is similar to that found by Stuart et al.⁽³⁴⁾ when describing the transcriptional program associated with rat kidney maturation. Given the complex nature of the molecular program we obtained for several stages of normal kidney growth, it is likely that redundancies among growth factors and other molecules may exist in term of the role they play in kidney development in vivo.

Strikingly, many of the details of the molecular program required to generate and build a nephron after transplantation of human kidney precursors into living animals are similar, in a global sense, to normal kidney induction. Comparison of temporal expression profiles demonstrated that the time course for development of normal human kidneys is applicable to that for development of transplanted human metanephroi. Comparison of the expression profiles of developing metanephric transplants to a Wilms tumor specimen revealed no similarity and reassured us that the threat of malignant transformation after transplantation of human embryonic kidney precursors does not exist. A functional test devised by intravenous administration of DMSA, a radioisotope known to be extracted and secreted by mature, functional renal tubules, revealed that this function is maintained in the transplanted metanephroi that had developed. Thus, after transplantation, human metanephroi integrate into host tissue microenvironment (including host blood vessels), survive, and function with available growth and regulatory signals.

Nevertheless, our analysis has revealed several limitations after transplantation of human metanephroi. First, early after grafting, we found elevated levels of genes that function in cell stress and angiogenesis, a response that might be related to both ischemia and neovascularization of the embryonic kidney graft ⁽³⁵⁾. Relative ischemia and consequent injury might account for the mostly low expression levels of human genes in the early time points after transplantation compared with normal kidneys. Minimizing the time to grafting or developing a preservation system for the human metanephroi ⁽³⁶⁾ could reduce the magnitude of ischemia. Second, the failure of genes encoding for various transporters to sufficiently increase their levels after transplantation, similar to normal kidney development, could possibly limit transplant function with time. Kim et al.⁽³⁷⁾ have demonstrated diminished renal expression of aquaporin water channels in rats with experimental ureteral obstruction. We have not yet been able to establish adequate anastomosis of the developing ureter of the human metanephric transplants, and thus gene expression of at least some of the channels in the transplants might have been similarly altered.

Finally, a comparison of a subset of genes encoding growth factor proteins, which have a recognized and important role in kidney development (^{21, 24, 25}), detected mostly decreased transcript levels in the transplants compared with normal kidneys. There is a possibility that the xenogeneic microenvironment can compensate and support the differentiation of the transplanted human metanephroi by secreting diffusible proteins, or alternatively via growth factors present in the peripheral circulation of host mice. Nevertheless, these growth factors can be easily administered alone or in combination at a specific time after transplantation to further stimulate donor human metanephroi (Figure 6). On the basis of our results, an attractive option would be for instance to administer insulin growth factor early after transplantation and administer HGF and fibroblast growth factor later on.

The concept of gene dosage and nephron development/endowment has emerged from recent studies (thoroughly reviewed by Clark and Bertram [²⁴]). These data demonstrate an enormous range in nephron number in human kidneys and suggest that low nephron numbers might be associated with acquired kidney disease (^{38, 39}). Several processes, which are regulated by specific genes, ultimately govern the number of nephrons during the development of human metanephroi—hence molecular regulation of nephron endowment. Furthermore, the dosage of specific genes also appears critical to the establishment of adequate nephron endowment. Thus, if transplanted human metanephroi maintain a temporal profile of gene expression similar to normal kidney development and the dosage of specific genes is reduced, nephrogenesis evidently pursues, but nephron endowment is limited. Future studies aimed at increasing gene dosage in the transplanted human metanephroi could determine whether nephron endowment improves.

We suggest that our approach of screening global gene expression in transplanted differentiating embryonic kidney cells versus normally developing cells might be relevant to other systems that follow the fate of embryonic cells after transplantation and elucidating factors that might drive them toward a specific lineage and organ. Furthermore, as far as directing embryonic cells solely toward human kidneys in vivo, our results indicate that it might be worthwhile to first use human embryonic precursors that have already differentiated into the nephric lineage rather than more pluripotent cells ⁽⁴⁰⁾.

This work was supported in part by an award from the American Physicians Fellowship (BD) and by the Arison Dorsman family’s donation to the Center of DNA Chips in Pediatric Oncology.