An intact genome is essential for kidney growth and differentiation, but less is known about whether, and how, an altered fetal milieu modifies these processes. Maternal low-protein diets perturb growth of the metanephros, the precursor of the mature kidney. Fetal corticosteroid overexposure may, in part, mediate this, because such diets downregulate placental 11β-hydroxysteroid dehydrogenase-2, which degrades maternal corticosteroids. We report that glucocorticoid and mineralocorticoid receptors are expressed in mouse metanephric epithelia. Metanephroi maintained in organ culture with hydrocortisone (1.4 or 14 μM) underwent a dose-dependant deceleration of overall growth accompanied by cyst formation. Dexamethasone, a glucocorticoid, reproduced these outcomes, but aldosterone, a mineralocorticoid, did not. Hydrocortisone upregulated transcripts levels of cadherin-11 and downregulated prospero-related homeobox-1, hence mimicking reported effects of maternal low-protein diet. Hydrocortisone also upregulated transcripts encoding Na+-K+-ATPase subunits and ligands for the epidermal growth factor receptor, all previously implicated in renal cyst growth. The most upregulated transcript, however, was indian hedgehog, and the encoded protein was immunodetected in metanephric cysts. Furthermore, in the presence of hydrocortisone, cystogenesis, but not whole organ growth, was significantly reduced by cyclopamine, a drug downregulating hedgehog signaling. Finally, both glucocorticoid receptor and indian hedgehog proteins were detected by immunohistochemistry in cystic tubules within human dysplastic kidneys, consistent with the hypothesis that these molecules modify the severity of this congenital malformation. Collectively, our observations raise the possibility that enhanced hedgehog signaling is an important stimulus for renal cyst formation. Furthermore, pharmacological inhibition of this pathway should be explored as a potential therapy for renal cystic diseases, starting with relevant animal models.
- gene expression
- renal cystogenesis
the metanephros is the precursor of the mature mammalian kidney (48, 61). In mice, it arises at embryonic day 11 (E11), equivalent to 5 wk of human gestation, when the ureteric bud grows into a region of intermediate mesoderm called the renal mesenchyme. By E13, equivalent to 6–7 wk of human gestation, the bud has branched several times to form the first kidney collecting ducts, and mesenchyme in the organ's periphery forms the first layer of nephron precursors. In mice, branching and nephron formation continue to the first postnatal week, whereas in humans, these processes cease around 34 wk of gestation. After these main morphogenetic steps have occurred, epithelia undergo further structural and molecular differentiation. For example, nephron precursors undergo segmentation into glomerular, proximal tubule, and loop of Henle epithelia (48, 61), and tubule epithelia undergo further polarization into apical and basolateral domains (6). Numerous genes coding for growth factors, transcription factors, and cell adhesion molecules (15, 61) are expressed in the metanephros, and mutation of several of these are known to perturb kidney development (24, 63). Furthermore, mutations of genes controlling proteins located in primary cilia are associated with renal cyst formation (39, 43). Primary cilia protrude from the apical surfaces of tubule epithelial cells and are thought to bend when glomerular filtrate passes by. This, in turn, triggers intracellular signaling pathways, which maintain cell differentiation (39, 43).
Whereas an intact genome is essential for normal kidney growth and differentiation, alterations in the fetal milieu can modify kidney development. For instance, fetal kidneys become distorted and damaged when fetal urinary flow is impaired (7, 62). In fact, nephrogenesis can probably be perturbed by more subtle, biochemical alterations of the fetal milieu. The “fetal programming hypothesis” proposes that prenatal undernourishment affects organogenesis, predisposing to postnatal disease (4, 17). Maternal low-protein diet (LPD) in rats causes reduced numbers of nephrons to form in the offspring kidneys and alters gene expression in the just-formed metanephros (57, 58). In humans, low birth weight babies have fewer nephrons than normal (35), and adults with fewer glomeruli than normal, presumed to reflect a congenital deficit, are prone to essential hypertension (23).
Maternal LPD is thought to result in fetal corticosteroid overexposure, because it downregulates placental 11β-hydroxysteroid dehydrogenase-2 (HSD11B2), an enzyme degrading maternal corticosteroids (27). Interestingly, mouse fetuses with null mutation of the gene encoding HSD11B2 are growth restricted (18), and human HSD11B2 mutations are associated with low birth weight (11). Administration of dexamethasone, a HSD11B2-resistant glucocorticoid, through rat pregnancy leads to nephron deficits in offspring (8). Another study (2) found that administration of high doses of hydrocortisone (HC), a molecule with glucocorticoid and mineralocorticoid activities (41), caused cysts to form in cultured embryonic mouse kidneys. This report (2), however, predated the definition of genes implicated in normal nephrogenesis and pathological cyst formation; furthermore, it had yet to be reported that maternal LPD could affect metanephric gene expression.
In the present study, we demonstrated that glucocorticoid and mineralocorticoid receptors (GR and MR) are expressed in mouse metanephric epithelia from the organ's inception. Addition of HC (14 and 1.4 μM) to metanephric kidneys in organ culture caused formation of metanephric cysts, an effect mediated predominantly by its glucocorticoid activity, alongside a dose-dependent deceleration of overall explant growth. HC exposure altered the expression of numerous metanephric transcripts. The most upregulated transcript was indian hedgehog (Ihh), and IHH protein was immunodetected in cystic metanephric epithelia. Strikingly, HC-induced cystogenesis was significantly reduced by pharmacological downregulation of hedgehog signaling. Furthermore, GR and IHH proteins were immunodetected in cystic tubules within human dysplastic kidneys, consistent with the hypothesis that these molecules modify the severity of this congenital malformation.
MATERIALS AND METHODS
Reagents were obtained from Sigma Chemical (Poole, UK), unless stated. Animal studies using wild-type CD1 mice from our local colony were approved by the UK Home Office following ethical review. The morning after mating was designated E0, and freshly isolated metanephroi (E11 and E13) were explanted on Millicell inserts (Millipore, Bedford, MA) and grown in defined, serum-free medium (DMEM/F12; GIBCO BRL, Gaithersburg, MD) supplemented with insulin (10 mg/l), sodium selenite (5 μg/l), and transferrin (5.5 mg/l) at 37°C in a humidified atmosphere of air-5% CO2 (26). In some experiments, this control medium was supplemented with HC (0.14, 1.4, or 14 μM); the glucocorticoid dexamethasone (Dex; 0.47 μM); the mineralocorticoid aldosterone (Aldo; 23 nM); cyclopamine (Cyclo; 10 μM), an inhibitor of hedgehog signaling (19, 49); or recombinant IHH (0.1, 1.0, and 10 nM; R&D Systems, Abingdon, UK). HC (14 μM) previously induced cystogenesis in organ culture (2, 3), and the doses of Dex and Aldo mimic the potency of 14 μM HC for the GR and MR, respectively (41). Media were replenished after 3 days. For dissolving Aldo, ethanol vehicle was used; preliminary experiments showed that in the concentration used in culture, ethanol did not affect metanephric growth (data not shown). Metanephroi were photographed every day during the 6-day culture period, and the number of pixels that the kidney spanned was converted into an area measurement using ImageJ software (http://rsbweb.nig.gov.ij/). Three blinded observers (D. A. Long, A. S. Woolf, and S.-K. Chan) noted the number of abnormal, large, round structures in each explant: these represented the largest cysts induced by HC and Dex (see results).
Explants were fixed in paraformaldehyde and paraffin-embedded, and 5-μm sections were cut, dewaxed, and rehydrated. Histochemistry was performed as described (9) for the following probes: peroxidase-conjugated Dolichos biflorus lectin (DBL), goat antibody to IHH (Santa Cruz Biotechnology, Santa Cruz, CA), rabbit antibody to GR (sc-1004; Santa Cruz Biotechnology), goat antibody to MR (sc-6861; Santa Cruz Biotechnology); and rabbit antibody to Tamm-Horsfall protein (THP; Santa Cruz Biotechnology). Some sections were stained with hematoxylin and/or periodic acid-Schiff (PAS). PAS staining was utilized to quantify glomerular number and total cyst area and number in experiments where explants grown in HC alone were compared with those exposed to both HC and Cyclo. We noted that control explants contained a small subset of tubules with prominent lumens, as is normal in metanephrogenesis, the largest of which was 667 μm2. Thus, for this analysis, a cyst was defined as a dilated tubule >700 μm2. Human normal fetal kidneys (n = 4; gestational ages between 20 and 23 wk) and cystic dysplastic kidney malformations (n = 3; gestational ages between 20 and 24 wk), collected with appropriate Local Ethical Committee approval as described previously (56), were immunoprobed for GR and IHH. Negative controls consisted of omission of primary antibodies or substitution with preimmune serum.
RT2 Profiler PCR array.
On day 6 of culture, controls and explants exposed to 14 μM HC (n = 4 pools for each condition, with 12–20 organs/pool) were collected and RNA was extracted using the RNeasy kit (Qiagen, Crawley, UK). RNA quality was assessed on a Bioanalyzer 2100 (Agilent Technologies, Palo Alto, CA), and transcript levels were measured using the RT2 Profiler PCR array system (SA Biosciences, Frederick, MD) (5, 37). Each custom-designed plate permitted measurement of 88 transcripts (Tables 1 and 2) coding for: 1) growth factors, transcription factors, and adhesion molecules expressed in normal metanephroi (61); 2) primary cilia molecules, which prevent cyst formation (39, 43); 3) segment-specific markers of kidney tubules, Na+-K+-ATPase subunits, and transcripts encoding GR and MR; and 4) transcripts deregulated in metanephroi exposed to maternal LPD (57). Gene expression levels were normalized against levels of a group of five housekeeping genes. RNA (500 ng) from each pool was used to prepare cDNA using the RT2 first-strand kit (SA Biosciences) and then added to the qPCR master mix, and thermal cycling was performed using a 7900HT FAST 96-well block (Applied Biosystems, Foster City, CA). Data were analyzed using software provided by the manufacturer (http://www.superarray.com/pcrarraydataanalysis.php). For transcripts found to be significantly up- or downregulated by HC, we used the ClustalW tool (http://www.ebi.ac.uk/Tools/clustalw/index.html) to determine whether the 1,000 bp upstream of the transcription start site, in addition to either the complete 5′-untranslated region (UTR), or the first exon of the 5′-UTR when this is split over several exons, contained sequences with at least 75% homology to the glucocorticoid response element (GRE) (G/T)GTACANNNTGTTCT (40).
Other PCR assays.
To measure expression of Ptch1 and Gli1, transcripts upregulated by hedgehog signaling (49), and Atp1a and Atp1b, encoding Na+-K+-ATPase subunits, in explants exposed to either HC alone or HC together with Cyclo, we undertook quantitative PCR assays, as described previously (47), using three pools of organs for each condition. For each primer (sequences available on request), a standard curve was generated and Hprt used as a housekeeping gene to allow quantification. In addition, nonquantitative RT-PCR was performed to seek expression of transcripts for GR and MR on freshly dissected E11 and E13 metanephroi (primer sequences available on request).
Growth curves were compared using linear mixed-effects modeling based on restricted maximum likelihood methods (45). Poisson regression analysis with gamma random effects was used to analyze the occurrence of cysts counted by blinded observers (accounting for overdispersed data and repeated measures of the 3 observers). Parametric cyst area data and gene expression array data were analyzed using two-tailed Student's t-tests, and nonparametric qPCR data were analyzed using one-tailed Mann-Whitney U-tests. Significance was accepted at P < 0.05, unless otherwise stated.
Early studies reported that 14 μM HC induced cysts in mouse E13 metanephric cultures (2, 3). Accordingly, the current study started by attempting to confirm, and further explore, this effect. Figure 1 depicts whole mount images (A–F) and surface areas (G) of control metanephroi and those in media supplemented with HC (0.14, 1.4, or 14 μM). In controls, average explant area increased to a maximum of 217% by day 5. The increases of area with 14 and 1.4 μM HC were significantly less compared with controls, with 14 μM HC producing a greater reduction in overall growth than 1.4 μM. With the use of pools of 50–60 organs cultured for 6 days, average protein/explant was 12.5 μg in controls and 8.5 μg in those exposed to 14 μM HC, strengthening the conclusion, derived from explant area data, that high concentrations of HC reduce overall growth of embryonic kidneys.
As assessed by nonquantitative RT-PCR, transcripts for GR and MR were expressed in E11 and E13 metanephroi (Fig. 2, A and B). With the use of immunohistochemistry (Fig. 2, C–F), GR was detected in the ureteric tree and widely expressed in metanephric tubules of explants following culture. MR expression in E13 metanephroi was less prominent, with occasional staining observed in the branching ureteric tree. Because HC binds to GR and MR with equal affinity (41), further cultures were undertaken with a specific glucocorticoid (Dex) or mineralocorticoid (Aldo) at doses that mimic the respective potencies found in 14 μM HC (41). Dex (0.47 μM) produced an inhibition of area growth, whereas Aldo (23 nM) resulted in a growth trajectory not significantly different from that of controls (Fig. 1H).
Initial observations of whole mounts showed organs exposed to 14 μM HC contained large, round structures (arrows in Fig. 1F) lacking in controls (Fig. 1C). Three blinded observers found a significant (P < 0.001) and numerically similar occurrence of such structures compared with controls in explants exposed to HC (14 μM) or Dex (0.47 μM) but not in those exposed to Aldo (23 nM). Histologically, E13 metanephroi when explanted contained a central ureteric bud that had branched a few times (Fig. 2, C–E). After 6 days in culture, controls (Fig. 3, A–E) had undergone considerable morphogenesis and differentiation, with formation of glomeruli, proximal tubules, and loops of Henle interspersed by collecting ducts (Fig. 3, B–D). Organs exposed to 14 μM HC for 6 days also contained differentiated nephron segments (Fig. 3, F–J), but histology was dominated by cysts (compare Fig. 3F with 3A), the largest ones presumably corresponding to the abnormal structures visualized in whole mounts. A subset of cysts had PAS-positive brush borders (arrows in Fig. 3G), consistent with proximal tubule derivation, whereas others bound DBL (Fig. 3H), consistent with collecting duct identities. Cysts did not express THP (Fig. 3I), a loop of Henle marker.
We next explored the possible effects of lower doses of HC, as well as other steroids, on cyst formation. Cysts were not apparent as assessed by simple inspection of whole mounts of explants exposed to either 1.4 or 0.14 μM HC. However, when assessed using histology, explants exposed to 1.4 μM HC for 6 days contained plentiful cysts (Fig. 4, A and B), although they were usually smaller than in organs exposed to 14 μM HC (compare Fig. 4A with 3F). Metanephroi exposed to 0.14 μM HC were devoid of cysts, but their tubules contained prominent vacuoles not apparent in controls (Fig. 4, C and D). A spectrum of cysts similar to that occurring in metanephroi exposed to 14 μM HC for 6 days was produced by Dex (0.47 μM; Fig. 4, E and F). Histology of organs cultured in 23 nM Aldo for 6 days were indistinguishable from controls (Fig. 4, G and H). E13 explants cultured for 6 days were also exposed to 14 μM HC for a window of 3 days (either 0–3 days or 3–6 days, with remainder of experimental period in basal medium). Both of these conditions led to cyst formation but to a lower degree than that of explants cultured for 6 days in 14 μM HC (data not shown).
E11 explants (which contain a ureteric bud that has branched just once or twice) were grown with basal medium alone or supplemented with 14 μM HC. After 6 days in culture, explants in both conditions differentiated into proximal tubules and glomeruli and underwent ureteric bud branching, but the steroid did not generate cysts (data not shown).
We undertook RT2 Profiler PCR array analyses in E13 explants cultured for 6 days for selected genes implicated in normal nephrogenesis and cystogenesis or that had been reported to be deregulated by maternal LPD. Tables 1 and 2, respectively, show genes up- and downregulated by 14 μM HC. The most significantly changed (P < 0.01) transcripts included those coding for transcription factors (Wt1, Prox1, Six1, and Tbx18), growth factors (Ihh, Hgf and its Met receptor, Tgfa, and Wnt11), an adhesion molecule (Cdh11), and a primary cilia molecule (Mks1). We found more modest changes in transcripts coding for Na+-K+-ATPase subunits (Atp1a and Atp1b); Areg, a growth factor that, like Tgfa activates the epidermal growth factor (EGF) receptor; and fibroblast growth factor 2 (Fgf2). Ihh was the most markedly deregulated transcript (upregulated 4.7-fold, P = 0.0006) in HC-exposed vs. control explants. IHH was immunodetected in control tubules (Fig. 3E) and cystic epithelia in explants exposed to 14 μM HC (Fig. 3J). Whereas Shh transcripts, encoding sonic hedgehog, were not significantly changed, Ptch1, coding for the hedgehog cell surface receptor and a surrogate marker for hedgehog signaling (49), was modestly upregulated (1.6-fold; P = 0.016). Several of the significantly upregulated transcripts also contained a putative GRE, specifically Agtr2, Areg, Atp1a1, Atp1b, Cdh11, Fgf2, Met, Tshz1, Tshz2, and Wt1.
To test whether enhanced hedgehog signaling is involved in HC-mediated perturbation of metanephrogenesis, we undertook further E13 cultures with Cyclo (10 μM), an inhibitor of hedgehog signaling (19, 49) (Fig. 5A). Addition of Cyclo to basal medium produced a nonsignificant growth deceleration compared with controls but did not induce any cysts. When medium was supplemented with both HC and Cyclo, explant growth showed a nonsignificant increase compared with HC alone. Furthermore, the prevalence of large, round structures, presumed to be the largest cysts, was significantly (P < 0.001) lower in HC with Cyclo vs. HC alone cultures. Histology (Fig. 5, B–H) showed that addition of Cyclo to HC was associated with markedly reduced cystogenesis because 1) the proportion of explant section area occupied by cysts fell from 4.5 ± 0.9 (mean ± SE; n = 9 in each condition) to 1.3 ± 0.2% (P < 0.01); 2) the average number of cysts per unit area fell from 11.2 ± 1.6 to 6.0 ± 1.0 cysts/mm2 (P < 0.05); and 3) the average area of individual cysts fell from 3,860 ± 334 to 2,130 ± 190 μm2 (P < 0.001). Addition of Cyclo to HC did not significantly alter the number of glomeruli per unit area in noncystic areas (HC without CYCLO, 21.8 ± 1.8 glomeruli/mm2; with Cyclo, 24 ± 1.3 glomeruli/mm2). When HC cultures were supplemented with Cyclo, there were reductions (n = 3, P = 0.05) of transcript levels for both Ptch1 (∼3-fold) and Gli1 (∼4-fold) vs. explants cultured in HC alone, providing biochemical evidence of downregulated hedgehog signaling. There were no differences in levels of Atp1a and Atp1b transcripts between these two conditions.
To ascertain whether normal human fetal kidneys and fetal cystic dysplastic kidneys might be targets for glucocorticoids, we performed immunohistochemistry for GR. In normal fetal kidneys, a strong immunohistochemical signal for GR was detected in a minor subset of cortical tubules with profiles consistent with them being either distal convoluted tubules or collecting ducts (Fig. 6A). In malformed kidneys, GR protein was detected in dysplastic tubules and cysts (Fig. 6B and data not shown); the dysplastic tubules bound DBL but did not express THP (Fig. 6C and data not shown), consistent with a collecting duct phenotype (36). In normal human fetal kidneys, a weak signal for IHH protein was detected in glomeruli and renal arteries (Fig. 6D and data not shown), whereas dysplastic kidney epithelia were immunoreactive for IHH with an apical immunostaining pattern (Fig. 6E).
Steroid-induced growth restriction of the developing kidney.
Mouse E13 metanephroi differentiate over 6 days in culture, generating organs with layers of glomeruli anatomically equivalent to a late first trimester human kidney. Using this model, we found that either 14 or 1.4 μM HC led to significant restriction of kidney growth, with the higher concentration producing a greater effect, whereas 0.14 μM HC did not alter overall growth. Because HC binds to the GR and MR with equal affinity (41), we performed further experiments with specific agonists for these receptors and concluded that HC's general metanephric growth-decelerating property resided predominantly in its glucocorticoid activity. We found that E13 metanephroi expressed both MR and GR mRNAs and that the encoded proteins were detectable in ureteric bud collecting duct branches. The pattern of immunodetected GR protein is similar to that reported in mouse E14 metanephroi (51).
Exposure of metanephroi to 14 μM HC upregulated transcripts for Cdh11, an adhesion molecule expressed by renal mesenchyme (10), and downregulated Prox1, encoding a transcription factor. In this respect, the steroid mimicked previously reported changes in metanephric gene expression associated with maternal LPD (57) and provided further evidence that some of the fetal programming effects elicited by maternal LPD are mediated by steroid overexposure. Indeed, both the Cdh11 and the Prox1 genes contain a putative GRE. In a previous study, rat E14.5 metanephroi, anatomically equivalent to mouse E13 kidneys, were exposed to Dex (10 μM) for 2 days, and this caused inhibition of growth (50) accompanied by a decrease of transcripts encoding glial cell line-derived neurotrophic factor (GDNF), bone morphogenic protein-4 (BMP4), and transforming growth factor-β1 (TGF-β1). These changes in growth factors would inhibit branching morphogenesis (61). Our analyses of mouse E13 metanephroi cultured for 6 days, with or without 14 μM HC (equivalent to 0.47 μM Dex), showed no significant change in transcripts for Gdnf or Bmp4, whereas there was a small downregulation (1.4-fold; P = 0.01) for Tgfb1. We did, however, detect modest downregulation for Hgf (1.5-fold; P = 0.008) and Wnt11 (1.7-fold; P = 0.006), each of which enhances branching (34, 60). Also of possible relevance in the growth retardation produced by HC was the downregulation of Six1 transcripts, which enhance nephrogenesis (25).
Steroid-induced cystogenesis within the developing kidney.
HC induced cyst formation in E13 explants, confirming historical observations (2, 3). Although an early study (2) reported cysts arose mainly from proximal tubules, we noted that although a subset of cysts did contain brush borders consistent with a proximal tubule origin, others bound DBL, suggesting that they had properties of collecting ducts. Our data suggest that HC's cystogenic property resided predominately in its glucocorticoid activity, because addition of Dex led to cyst formation, whereas Aldo did not generate cysts. However, it is possible that there may be modifying effects of MR activation through Aldo or HC; therefore, further experiments could be performed to demonstrate whether the effects of HC can be reversed using specific antagonists of GR and MR. Although we found that GR transcripts expressed at the inception of the metanephros, exposure of E11 organs to 14 μM HC for 6 days failed to generate metanephric cysts, with no excessive tubular dilatation seen on histology. This suggests that a critical degree of metanephric maturation is necessary for glucocorticoids to generate the cystic phenotype. Indeed, in normal mouse metanephroi, Ellis et al. (14) noted that GR activity, as assessed by Dex-binding assays, increased between E13 and late gestation (E18). Accordingly, in E13 organ cultured for 6 days, we detected widespread GR immunoreactivity in tubules. In fact, in mature murine kidneys, GR expression is rather widespread, being found along the nephron tubule and in collecting ducts (30).
We did not detect significant effects of HC on the majority of primary cilia proteins transcripts. The only one significantly altered was Mks1 (P = 0.005); inactivating mutations of the human homolog cause a cystic kidney malformation in Meckel syndrome type 1 (12). Mks1 was, however, upregulated (1.4-fold), a change predicted to minimize cyst formation. Hence, our gene expression profiling data are consistent with the conclusion that HC does not perturb ciliary integrity. Indeed, intact cilia are considered essential for hedgehog signaling (20), which appears operative in our model of HC-induced cystogenesis, as explained below. In vivo, primary cilia on the apical surface renal tubule are envisaged to sense tubular flow and then transduce signals into these epithelia, preventing deregulated proliferation, which would lead to cyst formation (43). However, tubular flow is primarily generated by glomerular filtration, a process absent in organ culture because no blood flows through explants and glomeruli formed in culture are avascular (32). Thus, a priori, other mechanisms must be implicated in cystogenesis within culture metanephroi.
Other molecules also have been implicated in renal cyst formation. In polycystic kidney epithelia, Na+-K+-ATPase mislocalizes to apical epithelial membranes, rather than the basolateral position found in normal, adult cells (59). The former scenario may cause secretion of solutes into tubule lumens, which, if followed by water, would cause cyst expansion. Interestingly, Na+-K+-ATPase is present apically in normal, metanephric epithelia (6). Avner et al. (3) reported that HC-induced cyst formation in E13 mouse metanephroi was reduced by ouabain, a blocker of Na+-K+-ATPase. In fact, glucocorticoids induce Na+-K+-ATPase activity in rat isolated proximal tubule cells and whole kidneys in vivo (21, 55), and in fetal sheep, they upregulate kidney levels of transcripts and proteins for the α1- and β1-subunits of Na+-K+-ATPase (44). These effects are mediated by enhancement of transcription of genes coding for these molecules (55), and indeed, Atp1a and Atp1b each have a putative GRE. Consistent with these observations, we found that HC upregulated transcripts encoding epithelial Na+-K+-ATPase subunits.
Growth factors and cystogenesis.
Several growth factors are considered cystogenic, including those activating the EGF receptor (13, 33), and we noted that HC led to upregulation of transcripts encoding two such ligands, Tgfa and Areg. Similarly, FGF2 causes overgrowth of immature kidney tubules (31), and HC caused a modest upregulation of Fgf2 transcripts. Of interest, both the Areg and Fgf2 genes contain a putative GRE. In our study, however, the most upregulated transcript coding for a growth factor was Ihh, not previously implicated in renal cyst formation. The gene coding for Ihh lacks a GRE, so upregulation is likely to be indirect. In normal mouse embryos, kidney collecting ducts and urothelia express Shh (7, 38), and SHH enhances maturation of ureteric smooth muscle cells (7) and branching growth in the metanephros (19). Shh is not the only hedgehog family member to be expressed in normal metanephroi, because Ihh expression increases during prenatal maturation (53) with transcripts prominent in cortical tubules (38). In control E13 metanephroi cultured for 6 days, we immunodetected IHH in tubules, and the protein also immmunolocalized in cystic epithelia in HC-exposed metanephroi.
In the presence of HC, cystogenesis, but not whole organ growth or glomerular density, was significantly reduced by Cyclo. Precisely how blocking hedgehog signaling ameliorates HC-induced dysmorphogenesis remains to be established. In the presence of HC, Cyclo did not alter levels of transcripts for Na+-K+-ATPase subunits. In other experiments, recombinant IHH was added to E13 cultures, but cysts did not form (data not shown). Therefore, although IHH is likely to be permissive, other changes induced by HC (e.g., upregulation of Na+-K+-ATPase and/or overexpression of other cystogenic growth factors) must be present for robust cyst growth to occur. Moreover, we do not know whether hedgehog signaling enhances cyst growth by direct actions on epithelia or on adjacent cells. Urothelium-derived SHH acts as a paracrine factor on surrounding mesenchyme (7), and, in the current experiments, HC altered levels of transcripts (i.e., upregulated Cdh11 and Tshz2 and downregulated Acta2 and nestin) characteristically expressed by renal tract mesenchymal cells. Moreover, when Ojeda et al. (42) used methylprednisolone to induce collecting duct cysts in neonatal rabbits, prominent matrix vesicles were detected around cystic epithelia. Thus cystogenesis induced by glucocorticoids may in part be secondary to changes in adjacent cells and matrix.
Heine et al. (16) recently reported that in glucocorticoid-induced injury to the developing mouse cerebellum, SHH antagonized the steroid's effects, partly by upregulating HSD11B2. The latter molecule is expressed in normal metanephroi (51), and, in future, it may be informative to measure levels in our model system. Finally, we note that although ours is the first report linking corticosteroids and Ihh expression, Lee et al. (29) found that Ihh expressed in uterine epithelia is controlled by another steroid, progesterone; in this context, IHH mediates epithelium-stroma interactions required for embryo implantation.
Implications for human congenital disease.
Human cystic kidney diseases fall into two categories (48). The first includes the polycystic kidney diseases caused by mutations in genes encoding primary cilia proteins (39, 43), where tubules form normally and only later become cystic (48). In the other category, typified by congenital dysplastic kidneys, cysts arise as one aspect of abnormal metanephric development. We reasoned that the growth-restricted, cystic embryonic mouse kidneys generated by HC and Dex resembled the latter category of human disease. Accordingly, we undertook immunostaining in human fetal cystic dysplastic kidneys and found that GR and IHH were present in dysplastic tubules. Of interest, microdissection studies have shown that such tubules may represent malformed collecting ducts (46), and, consistent with this, they were found to bind DBL (36), as did most of the HC-induced cysts in metanephric organ culture (see Fig. 3H). These observations raise the possibility that glucocorticoids enhance cystogenesis in human renal malformations via IHH. Some human cystic dysplastic kidneys are associated with HNF1B mutations, and the severity of malformation can vary markedly between family members carrying the same mutation (1). We speculate that variations in the fetal milieu (e.g., glucocorticoid levels) that existed between such individuals may account for some of the variation in renal phenotype. It is unclear as to what levels of glucocorticoids a first trimester metanephric kidney would be exposed, but umbilical vein concentrations of cortisol from 25- to 35-wk gestations can reach 0.3 μM (28). This lies between the two lowest concentrations (i.e., 0.14 and 1.4 μM) of HC used in our study. Exposure to 1.4 μM HC led to a reduction in overall explant growth and the formation of plentiful cysts, whereas 0.14 μM HC did not alter growth but explants contained prominent vacuoles in tubules, which were not apparent in controls. A further consideration is that we only exposed metanephroi to HC levels for 6 days, whereas human kidney development spans many months starting at 5 wk of gestation and ending at 34 wk of gestation (61). Therefore, one could speculate that longer term exposure to modest increases in HC may be harmful to the fetal kidney. Furthermore, low birth weight and intrauterine growth restriction increase the risk of end-stage renal disease later in life, associations that hold in individuals with congenital renal malformations and cystic kidneys (54).
Collectively, our observations raise the possibility that enhanced hedgehog signaling is an important stimulus for renal cyst formation. Furthermore, given the recent interest in novel therapies for cystic kidney diseases (52), we suggest that pharmacological inhibition of this pathway within the kidney should be explored as a potential treatment for such conditions, starting with relevant animal models.
This work was supported by Kidney Research UK (Studentship to S.-K. Chan and Senior Fellowship to D. A. Long).
No conflicts of interest are declared by the authors.
We thank Angie Wade and Mario Cortina Borja (Centre for Paediatric Epidemiology and Biostatistics, UCL Institute of Child Health) for support with statistical analyses.
- Copyright © 2010 the American Physiological Society