Megalin Functions as an Endocytic Sonic Hedgehog Receptor*

Embryos deficient in the morphogen Sonic hedgehog (Shh) or the endocytic receptor megalin exhibit common neurodevelopmental abnormalities. Therefore, we have investigated the possibility that a functional relationship exists between the two proteins. During embryonic development, megalin was found to be expressed along the apical surfaces of neuroepithelial cells and was coexpressed with Shh in the ventral floor plate of the neural tube. Using enzyme-linked immunosorbent assay, homologous ligand displacement, and surface plasmon resonance techniques, it was found that the amino-terminal fragment of Shh (N-Shh) bound to megalin with high affinity. Megalin-expressing cells internalized N-Shh through a mechanism that was inhibited by antagonists of megalin, viz. anti-receptor-associated protein and anti-megalin antibodies. Heparin also inhibited N-Shh endocytosis, implicating proteoglycans in the internalization process, as has been described for other megalin ligands. Use of chloroquine to inhibit lysosomal proteinase activity showed that N-Shh endocytosed via megalin was not efficiently targeted to the lysosomes for degradation. The ability of megalin-internalized N-Shh to bypass lysosomes may relate to the finding that the interaction between N-Shh and megalin was resistant to dissociation with low pH. Together, these findings show that megalin is an efficient endocytic receptor for N-Shh. Furthermore, they implicate megalin as a new reg-ulatory component of the Shh signaling pathway. Sonic hedgehog (Shh) 1 is a

Sonic hedgehog (Shh) 1 is a secreted signaling molecule that is expressed in spatially restricted patterns during embryonic development. Shh signaling has been shown to regulate a wide range of developmental patterning events in Drosophila and vertebrates involved in lung (1), nervous system (2), eye (3), midbrain (4), and forebrain and facial (5,6) morphogenesis. During early vertebrate development, Shh signaling at the midline leads to patterning of the ventral neural tube and adjacent somites. Mice lacking Shh activity have anomalies of midline structures such as the notochord and floor plate of the early brain (7). Later, these mice display an absence of ventral neuronal cells and cranial motor neurons (8). The result of errant Shh signaling in humans has been directly linked to basal cell carcinoma (9,10) and holoprosencephaly (11).
Post-translational modification of the 45-kDa Shh polypeptide produces an ϳ19-kDa amino-terminal fragment (designated N-Shh) that has palmitic acid and cholesterol moieties covalently coupled to its amino and carboxyl termini, respectively (12)(13)(14). N-Shh is secreted and represents the biologically active form of the protein, capable of initiating signaling. The current model for Shh signaling involves a pair of multiple-pass plasma membrane proteins, Patched (Ptc or Ptc-1) and The expression of Ptc-1, Gli-2, HNF3␤, Nkx2.2, and netrin-1 has been shown to be activated by Shh, and genes including pax-3, gli-3, and ephrin A5 have been shown to be suppressed by Shh (16,17).
Megalin (also known as gp330 and low density lipoprotein receptor-related protein (LRP)-2) is an endocytic receptor belonging to the low density lipoprotein receptor (LDLR) family (18). The receptor is expressed on apical surfaces of numerous epithelia, where it functions to mediate endocytosis of ligands, targeting them for lysosomal degradation or transcytosis (18). Mice deficient in the expression of megalin demonstrate the critical neurodevelopmental role for this protein (19). These mice display numerous craniofacial abnormalities, including absence of olfactory bulbs, absence of the corpus callosum, and fusion of forebrain hemispheres, collectively an holoprosencephaly phenotype (19). During development, megalin-deficient embryos (9.5 days postcoitus) have pronounced cell death in several structures, including cranial nerves, the neural crest, and the optic vesicle (19). The spectrum of defects that constitute the megalin-deficient phenotype suggests that megalin expression is required for normal viability of the neural epithelium at an early embryonic stage.
The phenotype of megalin-deficient mice suggests a role for megalin in regulating cell fate specification in the patterning of the neural tube and is consistent with phenotypes observed in mice deficient in Shh and the Shh signal transducer, Smo (8,20). For example, Shh-deficient embryos lack cranial motor neurons (8). Inhibition of Shh signaling in the neural tube has been shown to result in extensive apoptosis of neural epithelial cells (21). Shh has also been shown to regulate proliferation and to inhibit differentiation of central nervous system precursor cells (22). Smo mutants also display neural tube-related defects, including increased apoptosis of cells within the neural tube, absence of secondary motor neurons, synopthalmia, and ventral forebrain defects (20,23). The shared aspects of the megalin-, Shh-, and Smo-deficient phenotypes suggest that Shh and megalin impact common mechanisms that underlie central nervous system development. Here, we report findings from experiments directed at determining whether a functional relationship exists between megalin and Shh.

EXPERIMENTAL PROCEDURES
Cells-Murine sarcoma virus-transformed Brown Norway rat yolk sac cells (BN cells) were provided by Dr. Pierre Verroust (Hospital Tenon, Paris, France). Mouse embryonic teratocarcinoma F9 cells (ATCC CRL1720) were differentiated by treatment with retinoic acid and dibutyryl cAMP for 6 days as previously described (24). C3H10T1/2 cells (ATCC CCL226) were obtain from the American Type Culture Collection (Manassas, VA).
Antibodies-Rabbit polyclonal and mouse monoclonal antibodies to megalin (rb6286 and 1H2) have been described previously (25). Rabbit anti-megalin IgGs were purified by protein G-Sepharose and megalin-Sepharose chromatography (26). Mouse monoclonal anti-receptor-associated protein (RAP) antibody 7F1 has been described previously (27). Mouse monoclonal anti-N-Shh antibody 5E1 IgG was isolated from the conditioned culture medium of a hybridoma cell line obtained from the Developmental Studies Hybridoma Bank (Johns Hopkins University School of Medicine and University of Iowa). Goat anti-glutathione Stransferase (GST) antibody was obtained from Amersham Biosciences. Fluorescein isothiocyanate-and indocarbocyanine (Cy3)-labeled secondary IgGs were purchased from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA).
Proteins-Megalin was purified from porcine kidney as described previously (28). Human RAP was expressed in bacteria and purified as described by Kounnas et al. (29). Recombinant murine N-Shh (residues  was obtained from R&D Systems (Minneapolis, MN). A plasmid construct was created to express GST-N-Shh fusion protein in bacteria. Briefly, this involved using reverse transcription-PCR to generate a cDNA encoding amino acids 20 -198 of Shh from a cDNA template prepared from day 9.5 postcoitus mouse embryo RNA. The Shh cDNA fragment was inserted into the bacterial expression vector pGEX-2TK (Amersham Biosciences) such that the resulting plasmid encoded a fusion protein composed of GST followed by a thrombin cleavage site (LVPRGS), a five-amino acid phosphorylation target site (RRASV), and the N-Shh polypeptide. The construct was transformed into BL21 bacteria, and the fusion protein was isolated using glutathione-Sepharose affinity chromatography. Recombinant GST was produced from cells transformed with the empty pGEX-2TK vector. Both recombinant protein preparations were adsorbed onto a Detoxigel endotoxin removing gel (Pierce). The biological activity of GST-N-Shh was assayed in C3H10T1/2 cells using the method of Williams et al. (30).
Whole-mount Embryo Immunolabeling-Zebrafish were maintained, and embryos were collected by standard methods (32). Embryos were fixed for 15 min in 4% paraformaldehyde in phosphate-buffered saline Megalin expression was intense on the interior epithelial surface of the otic vesicle of the developing ear (E). Megalin was also expressed in the paired pronephric ducts (F, arrowheads) of the forming kidney, where it was also associated with the luminal side of the epithelium. A frontal view of the developing oral region (G) shows that megalin was expressed in the ridge of the frontonasal process and maxillary processes (arrowheads). (PBS). Embryos were washed two times with PBS, permeabilized by washing three times with PBS containing 0.1% saponin at 37°C, and blocked for 30 min at 37°C in PBS containing 5% goat serum and 0.1% saponin (goat serum/saponin/PBS). Embryos were incubated with primary antibody (5 g/ml) in goat serum/saponin/PBS first for 1 h at 37°C, then overnight at 4°C, and finally for an additional 1 h at 37°C with rocking. Embryos were washed three times with goat serum/ saponin/PBS at 37°C and incubated with Cy3-coupled goat anti-mouse or anti-rabbit secondary antibody (1.5 g/ml) in goat serum/saponin/ PBS for 1 h at 37°C. Samples were washed three times with goat serum/saponin/PBS and dehydrated in methanol, followed by clearing in Murray's Clear (1:2 benzyl alcohol/benzyl benzoate). Laser scanning confocal microscopy was performed using a Bio-Rad MRC-1400 confocal microscope and Bio-Rad LaserSharp2000 software.
Immunoblotting and Ligand Overlay Assay-Detergent extraction of cells and immunoblot detection of megalin were performed as described previously (24). RAP ligand blot overlay assay was performed as described by Battey et al. (27).
For homologous ligand competition assays, 32 P-labeled GST-N-Shh (1 nM) was incubated in microtiter wells coated with megalin (3 g/ml) in the presence of increasing concentrations of unlabeled competitor (GST-N-Shh or RAP). All other conditions were similar to those described by Williams et al. (34). The algorithm Ligand (35) within Sig-maPlot 7.101 was used to analyze the competition data and to determine dissociation (K d ) and inhibition (K i ) constants for receptor-ligand interactions.
Kinetic Analysis of N-Shh-Megalin Binding-Kinetic analysis of the interaction of GST-N-Shh with purified megalin was performed using surface plasmon resonance (SPR) measurements made on a BIAcore 3000 instrument. BIAcore sensor chips (type CM5) were activated with a 1:1 mixture of 0.2 M N-ethyl-NЈ-(3-dimethylaminopropyl)carbodiimide and 0.05 M N-hydroxysuccinimide in water. Megalin (50 g/ml, 83 nM in 10 mM sodium acetate at pH 4.8) was immobilized on a CM5 sensor chip using the amine coupling kit (BIAcore) as described by the supplier. Unreacted sites were blocked with 1 M ethanolamine (pH 8.5). The SPR signal from immobilized megalin generated BIAcore response units ranging from 20,000 to 28,000. Control flow cells were activated and blocked in the absence of protein. Binding was evaluated over a range of GST-N-Shh concentrations (25-500 nM) in 150 mM NaCl, 0.005% polysorbate 20, and 100 mM HEPES (pH 7.4) with and without 1 mM CaCl 2 at 25°C. Binding of GST-N-Shh to megalin-immobilized flow cells was corrected for binding to control flow cells. Binding data were fitted to a 1:1 Langmuir binding model using BIAevaluation Version 3.1 software (BIAcore).
To evaluate the effect of pH on the dissociation of megalin-ligand complexes, GST-N-Shh or RAP (each at 3 M in buffer A (100 mM HEPES (pH 7.4) and 150 mM NaCl)) were passed at 10 l/min for 2 min over sensor chips containing immobilized megalin. Subsequently, protein-free buffer A or sodium acetate buffer (pH 4.5; sodium ion concen-FIG. 4. Enzyme-linked immunosorbent assay and competitive radioligand binding assay demonstrate that N-Shh binds to megalin and that RAP inhibits the binding. In A, enzyme-linked immunosorbent assay showed that both GST-N-Shh (E) and commercially available N-Shh (q) bound to megalin. In B, homologous ligand displacement assay (E) was used to demonstrate the interaction between 32 P-labeled GST-N-Shh and megalin, and heterologous ligand displacement assay (q) was used to show that RAP inhibited binding of 32 P-labeled GST-N-Shh to megalin. The curves shown in B were based on fits of the data calculated using the computer program Ligand. tration adjusted to 150 mM) was applied for 5 min. The kinetic dissociation profiles obtained under neutral and acid pH conditions were used to calculate off-rates (k off ) using the BIAevaluation Version 3.1 program. Between replicate experiments, the chip surface was regenerated with a 10-s pulse of 10 mM glycine (pH 2.2) at 5 l/min. Confocal Microscopic Analysis of N-Shh Uptake-BN cells were plated at 0.25 ϫ 10 5 cells/cm 2 in eight-well chamber slides (Nunc Nalge, Naperville, IL) in Eagle's minimal essential medium containing 10% fetal bovine serum, nonessential amino acids, 100 units/ml penicillin, and 100 g/ml streptomycin (complete medium). Cells were grown for 16 h at 37°C and 5% CO 2 , and the medium was replaced with serumfree medium (Eagle's minimal essential medium containing nonessential amino acids, 100 units/ml penicillin, 100 g/ml streptomycin, 5 g/ml insulin, 5 g/ml transferrin, and 5 g/ml selenic acid). After a 1.5-h incubation, the medium was replaced with serum-free medium containing 1.5% BSA and either GST-N-Shh (20 nM) or GST (20 nM) with or without competitors and cultured for 2 h. Competitors included RAP (1 M) and GST (1 M).
For immunological detection, GST-N-Shh-and GST-treated cells were rinsed in Dulbecco's phosphate-buffered saline (DPBS) (pH 7.4), fixed for 20 min in 3.7% paraformaldehyde with 0.2% Triton X-100 in DPBS, and then rinsed in DPBS. Cells were incubated with 2% BSA in DPBS for 1 h, treated with goat anti-GST IgG at 1 g/ml in 2% BSA in DPBS for 1 h and then with fluorescein isothiocyanate-labeled donkey anti-goat IgG at 3 g/ml in DPBS for 1 h, and rinsed in DPBS. For nuclear staining, cells were treated with RNase A (100 g/ml) for 20 min at 37°C, rinsed in DPBS, and then treated with TOTO-3 (Molecular Probes, Inc. Eugene, OR) at 1 g/ml in DPBS for 10 min at 37°C. Cells were rinsed in DPBS, mounted in Vector Shield mounting solution (Vector Laboratories, Burlingame, CA), and then examined by laser scanning confocal microscopy.
Cellular Internalization and Degradation Assays-BN cells were seeded into wells of 24-well plates at 0.5 ϫ 10 5 cells/cm 2 and grown for 16 h at 37°C and 5% CO 2 in complete medium and then for 1.5 h in serum-free medium. The medium was replaced with serum-free medium plus 1.5% BSA and 32 P-labeled GST-N-Shh (3 nM) with or without the indicated agents (i.e. RAP, IgG, or heparin), and cells were grown for 2-6 h. For experiments measuring the effect of chloroquine treatment, chloroquine was added at 0.1 mM concomitantly with radiolabeled ligands, and uptake was allowed to proceed for 6 h. Quantification of the amount of bound, internalized, and degraded ligands was performed as described previously (36). Radioactivity in the cell medium that was soluble in 10% trichloroacetic acid was taken to represent degraded ligand. Total ligand degradation was corrected for the amount of degradation that occurred in radioligand-containing medium in the absence of cells. To determine the amount of 32 P-labeled ligand that was bound and internalized, cells were washed three times with DPBS and then treated with serum-free medium containing 0.5 mg/ml trypsin, 0.5 mg/ml proteinase K (Sigma), and 0.5 mM EDTA for 2-4 min at 4°C. The cell suspension was centrifuged at 6000 ϫ g for 4 min, and the amount of radioactivity in the supernatant was taken to represent the bound fraction, whereas the amount in the cell pellet was taken as the internalized fraction. Uptake experiments with differentiated and control F9 cells were performed as described above with the exception that cells were seeded at 1.0 ϫ 10 5 cells/cm 2 , and the growth medium was Dulbecco's modified Eagle's medium and 10% fetal bovine serum (or insulin, transferrin, and selenic acid) containing 100 units/ml penicillin and 100 g/ml streptomycin.  6. N-Shh is endocytosed by BN cells, and uptake is inhibited by the megalin antagonist RAP. In A, BN cells were incubated with GST-N-Shh or GST (20 nM) in the presence of absence of RAP (1 M) for 2 h and immunostained with anti-GST antibody and fluorescein isothiocyanate-labeled anti-goat IgG (green). Nuclei were stained using TOTO-3 (blue). RAP treatment did not affect binding of GST-N-Shh to the cell, but inhibited its internalization. B shows that megalin was the principal RAP-binding protein present in detergent extracts of BN cells. Aliquots of BN cell extract were immunoblotted with anti-megalin IgG (lane 1) or were incubated with RAP (1 M), and the bound RAP was then detected with mouse monoclonal anti-RAP IgG (lane 2). No other RAP-binding proteins were evident even after prolonged exposure of the RAP overlay blot.   (19), little was known about the expression of the receptor during early development. Laser scanning confocal microscopic analysis of 16-h zebrafish embryos revealed that megalin was prominent in the floor plate of the neural tube (Fig. 1A, arrow) and on the apical surface of the optic cup (arrowhead). By 24 h, megalin expression was detected in cells of the ventral floor plate (Fig.  1B, arrow) and on the apical surface of cells lining the lumen of the neural tube (arrowhead). At 33 h, ventral floor plate expression persisted, and megalin was also extensively expressed on cells comprising the luminal surfaces of the forebrain and midbrain (Fig. 1D, arrowhead), with strong expression at the midbrain-hindbrain border (Fig. 1C, arrowhead). At the base of the midbrain, intense staining for megalin was seen at the most anterior extent of the floor plate (Fig. 1D, arrow). Outside the central nervous system, megalin was detected on the apical surfaces of cells lining the lumen of the otic vesicle of the developing ear (Fig. 1E, arrowheads). In the area of the developing mouth of 48-h embryos, megalin was distributed medially and laterally in the frontonasal and maxillary processes, respectively (Fig. 1G). These findings demonstrate that early embryonic expression of megalin occurs at specific organizing centers for morphogenesis, including the ventral neural tube, optic and otic vesicles, and orofacial regions.

Neurodevelopmental Expression of Megalin-Despite indications that megalin is critical to neurodevelopment
Many of the observed embryonic sites of megalin expression were the same as those known to express Shh, including the ventral floor plate, eye, otic vesicle, and frontonasal process (2)(3)(4)(5)37). A notable exception was the absence of megalin expression in the notochord (Fig. 2, A and B, arrowheads). Also, megalin expression in the neural tube extended more dorsally than Shh (Figs. 2 (insets) and 1B), detected in areas of the neural tube known to express the receptors for Shh, Ptc-1, and Ptc-2 (38). The results indicate that megalin is expressed in tissues that express Shh or in adjacent tissues regulated by Shh signaling. These observations support the possibility that a functional relationship exists between megalin and Shh during early neurodevelopment.
Megalin Is an N-Shh-binding Receptor-The similarity of megalin-and Shh-null phenotypes and the early embryonic distribution of megalin in relation to sites of Shh production led us to investigate whether megalin and N-Shh are capable of directly binding to one another. Enzyme-linked immunosorbent assay showed that recombinant GST-N-Shh (Fig. 3) and a commercial preparation of N-Shh bound to purified megalin with similar apparent affinities (Fig. 4A). Binding between GST-N-Shh and megalin was also tested using a homologous ligand competition assay. 32 P-Labeled GST-N-Shh bound to megalin, and the binding was inhibited in a dose-dependent manner by the addition of unlabeled GST-N-Shh (Fig. 4B). A K d of 81.3 nM was obtained from fitting the data to a one-site model using the Ligand algorithm. Binding of 32 P-labeled GST-N-Shh to megalin was also inhibited by RAP, a well established antagonist of megalin-ligand interaction (39). Interestingly, the RAP competition data could best be fit to a two-site model with K i values of 3.0 and 2341.9 nM. One interpretation of these findings is that RAP binds to megalin at multiple sites and that one of these binding interactions is a stronger inhibitor of N-Shh binding to megalin. Such an interpretation is consistent with the fact that the megalin family member LRP has multiple RAP-binding sites (34).
Binding of Shh to megalin was also evaluated using SPR. As shown in Fig. 5, GST-N-Shh bound to megalin immobilized on a sensor chip. GST alone displayed no measurable binding to megalin (data not shown). Optimal fitting of SPR data obtained from measuring the binding of various concentrations of GST-N-Shh to immobilized megalin was best achieved using a single class binding site model. As a result, an affinity constant (K D ) of 21 nM (n ϭ 7; 2 of fit Ͻ 10) was determined for GST-N-Shh binding to megalin in the presence of calcium. Recombinant N-Shh cleaved with thrombin to remove the amino-terminal GST moiety and commercial N-Shh were both found to bind megalin immobilized on a sensor chip with affinities similar to those observed for the fusion protein (data not shown).
Megalin Mediates Endocytosis of N-Shh-The role of megalin in mediating endocytosis of N-Shh was next evaluated. As shown in Fig. 6, confocal analysis of BN cells cultured in the presence of GST-N-Shh showed intracellular GST-N-Shh staining in a punctate pattern consistent with vesicular localization. Cells incubated with GST showed little to no intracellular staining (Fig. 6A). When BN cells were cultured in the presence  1 mM). Measurements of bound, internalized, and degraded radiolabeled ligands were made after a 3-h incubation. Note that RAP inhibited binding and internalization of 32 P-labeled GST-N-Shh and 32 P-labeled RAP. By contrast, RAP and chloroquine both blocked degradation of labeled RAP, but not of labeled N-Shh. of both GST-N-Shh and RAP, little if any intracellular staining was observed (Fig. 6A). Instead, RAP-treated cells displayed punctate foci of staining located on the cell periphery. This staining pattern is consistent with a plasma membrane or pericellular localization. Therefore, when megalin activity is abrogated, N-Shh appears to bind to the pericellular matrix or cell surface, and uptake is blocked.
We subsequently evaluated the ability of BN cells to mediate endocytosis of radiolabeled N-Shh. As shown in Fig. 7A, BN cells internalized 32 P-labeled GST-N-Shh. The uptake of 32 Plabeled GST-N-Shh could be blocked by either RAP or antibodies to megalin. The observed inhibitory effects support the interpretation that megalin mediates N-Shh endocytosis. Furthermore, inhibition by anti-megalin antibodies alleviates a concern that the inhibitory effects of RAP might not have been megalin-specific. In this regard, it is also important to note that we established that megalin is the only detectable RAP-binding member of the LDLR family in BN cells (Fig. 6B). Therefore, RAP can be considered a specific inhibitor of megalin in BN cells.
Uptake of GST-N-Shh was also evaluated in murine F9 cells. F9 cells express little or no megalin, but can be differentiated with retinoic acid and dibutyryl cAMP, causing induced megalin expression and decreased expression of other LDLR family members (24). As shown in Fig. 7B, differentiated cells exhibited an increased capacity to internalize 32 P-labeled GST-N-Shh. RAP effectively inhibited internalization of 32 P-labeled GST-N-Shh in differentiated F9 cells, but had little effect on the relatively low level of internalization in undifferentiated cells. These findings further support the interpretation that megalin mediates endocytosis of N-Shh.
N-Shh Endocytosis Involves Proteoglycans-In light of the fact that cell-surface proteoglycans have been implicated as partners with megalin and other LDLR family members in the uptake of numerous ligands (18), we investigated their possible involvement in N-Shh endocytosis. Heparin was an effective inhibitor of the uptake of 32 P-labeled GST-N-Shh by both BN cells and differentiated F9 cells (Fig. 8, A and B). This suggests the involvement of cell-surface proteoglycans in the process of N-Shh endocytosis.
N-Shh Is Not Efficiently Targeted to Lysosomes by Megalin-One well characterized consequence of megalin-mediated endocytosis is targeting of ligands to the lysosome for degradation. Inhibition of lysosomal proteinase activity using the drug chloroquine did not inhibit 32 P-labeled GST-N-Shh degradation in BN cells (Fig. 9). By contrast, in control experiments, chloroquine efficiently inhibited the degradation of 32 P-labeled GST-RAP (Fig. 9), a megalin ligand that is targeted to the lysosomes following megalin-mediated endocytosis. Interestingly, there was a significant level of chloroquine-insensitive N-Shh degradation, suggesting that degradation of N-Shh may occur extracellularly.
Evaluation of Lowered pH upon Dissociation of the N-Shh-Megalin Complex-The effect of low pH on the dissociation of the N-Shh-megalin complex was evaluated by SPR on a BIAcore instrument. Little difference was evident in the dissociation rate constants (k off ) for the N-Shh-megalin interaction under acidic versus neutral pH conditions: 1.3 ϫ 10 Ϫ3 and 1.28 ϫ 10 Ϫ3 s Ϫ1 , respectively (Fig. 10A). By contrast, dissociation of the RAP-megalin complex increased ϳ3-fold from 3.1 ϫ 10 Ϫ3 s Ϫ1 under neutral pH conditions to 8.36 ϫ 10 Ϫ3 s Ϫ1 under acidic pH conditions (Fig. 10B). These findings indicate that the N-Shh-megalin interaction is resistant to dissociation by acidic pH as low as 4.5 and suggest that N-Shh may not readily dissociate from megalin under acidic pH within endosomes. DISCUSSION Here, we have established that a functional relationship exists between the endocytic receptor megalin and the morphogen N-Shh. Specifically, we found that N-Shh binds to megalin with high affinity and that the interaction is resistant to dissociation by low pH. We have also shown that one consequence of the interaction is endocytosis of N-Shh. Megalin-mediated uptake of N-Shh can be blocked by heparin, suggesting the involvement of heparan sulfate proteoglycans in the internalization process.
Heparan sulfate proteoglycans have been implicated in N-Shh signaling (40,41) and in the process of megalin-mediated endocytosis of a number of its ligands (18). In the latter case, evidence suggests that heparan sulfate proteoglycans serve to sequester ligands at or near the cell surface and thereby either facilitate presentation of ligands to megalin or augment the affinity of ligands for megalin (18). Our observation that N-Shh appeared to accumulate pericellularly on BN cells after blocking the ligand-binding activity of megalin suggests the existence of an additional cell-surface or pericellular N-Shh-binding molecule. Considering recent evidence that Ptc is not detected at significant levels on the cell surface (42), this other N-Shhbinding component may very well be heparan sulfate proteoglycans.
The likely significance of the interaction of N-Shh with megalin is that it impacts Shh signaling. Three possibilities are that the interaction leads to 1) direct signal transduction by megalin, 2) modulation of the availability of N-Shh for its receptors, or 3) transcytosis of N-Shh important for long-range N-Shh signaling. Direct signal transduction by megalin is supported by recent evidence that other members of the LDLR family mediate signaling (43). For example, LRP has been shown to interact with the heparin-binding growth factor midkine and to regulate midkine-dependent survival of embryonic neurons (44). LRP has also been shown to interact with platelet-derived growth factor-BB and to function as a co-receptor in the process of platelet-derived growth factor signaling (45,46). Additionally, the very low density lipoprotein receptor and apoE receptor-2 interact with the neuronal protein reelin and mediate signaling through the cytoplasmic adaptor protein Dab1 (47). With respect to the second possibility, megalinmediated endocytosis of N-Shh may modulate the extracellular levels of N-Shh and thereby regulate availability to Ptc. For example, megalin might compete with Ptc for limiting levels of N-Shh and thereby reduce Ptc dissociation from Smo, leading to decreased Smo signaling. Alternatively, megalin may deliver N-Shh to vesicular pools of Ptc and thus regulate the potential of this Ptc to complex with Smo. The third possibility is consistent with the emerging role of megalin as a mediator of transepithelial transport of various macromolecules. For example, thyroglobulin, the transcobalamin-vitamin B 12 complex, and retinol-binding protein in complex with retinol/vitamin A are internalized by megalin, but avoid lysosomal degradation and are delivered to the basolateral membrane, from which they get released (48 -50). The mechanism by which megalin ligands bypass lysosomal degradation is not known. One possibility is that the interaction between megalin and these ligands might not be readily dissociated by acidic pH, such as occurs in endocytic vesicles. As a consequence, the ligands traffic together with the receptor and are transported to either apical or basolateral aspects of the cell. Our finding that the N-Shh-megalin interaction is insensitive to low pH suggests that N-Shh may also traffic in complex with megalin and thus be recycled and/or transcytosed. This possibility is further supported by our findings from chloroquine experiments indicating that endocytosed N-Shh bypasses lysosomes.
Megalin-mediated transcytosis of N-Shh may facilitate longrange signaling by N-Shh during early development. For example, N-Shh expressed in the floor plate may bind to megalin expressed on the apical surface of the neural tube epithelium and mediate transepithelial transport of N-Shh (Fig. 11). This process could account for delivery of N-Shh to cells in the ventral region of the neural tube that undergo differentiation to form ventral nerves, a process dependent on both N-Shh signal transduction and megalin expression (17,19). A similar process has been described in Drosophila involving transport of the morphogen Wingless protein over large distances through imaginal disc epithelia (51). In this case, membrane vesicles called argosomes, derived from the basolateral membranes, are transported throughout imaginal disc epithelia. The argosomes are thought to originate from either multivesicular endosomes and/or endosome transcytosis. Importantly, Wingless signaling has been shown to involve a megalin family member, LRP6/ arrow, although its exact role in the process remains to be determined (52,53).
In addition to megalin mediating long-range signaling via transcytosis of N-Shh, as discussed above, its ability to endocytose N-Shh may also impact N-Shh signaling in the early neural epithelium directly. Whether the mechanism for this involves effects on the bioavailability of N-Shh or on the regulation of Ptc as described above, the end result may be to influence N-Shh-dependent survival and differentiation of neural epithelial cells (7,8,21). This hypothesis is supported by the megalin-deficient mouse phenotype, which demonstrates that megalin is required for normal viability and development of the neuroepithelium (19).