WO2009011892A2

WO2009011892A2 - Use of suif proteins and heparan sulfate in gdnf-dependent neural innervation and protection

Info

Publication number: WO2009011892A2
Application number: PCT/US2008/008754
Authority: WO
Inventors: Charles Emerson; Xingbin Ai
Original assignee: Boston Biomedical Research Institute
Priority date: 2007-07-18
Filing date: 2008-07-17
Publication date: 2009-01-22
Also published as: WO2009011892A3

Abstract

The present invention is based on the discovery that Sulf protein deficiencies lead to direct defects in GDNF-dependent neural innervation and fornnation of neural crest derived enteric glial cells. Sulfs regulate cellular Heparan Sulfate (HS) 6-O-sulfation, and thus reduce GDNF binding to HS, thereby enhancing GDNF signaling to induce neurite outgrowth for muscle innervation. Methods, including Parkinson's Disease, brain injury, and esophageal disorders. Included are methods wherein the increase in GDNF signaling is assayed by monitoring phosphorylation of downstream kinases and formation of glial cells.

Description

Use of SuIf Proteins and Heparan Sulfate in GDNF-dependent Neural Innervation and Protection

Background of the Invention

[0001] Proper function of the central and peripheral nervous systems requires the correct connections between innervating neurons and their target tissues. This is accomplished through a combinatorial mechanism involving the intrinsic developmental programs of innervating neurons and the extrinsic signals present in both the target tissues and the extracellular matrix. During innervation, target tissues provide neurotrophic factors that support the survival, differentiation, and neurite outgrowth of the innervating neurons (Zweifel et al., 2005). Neurotrophic factors that come from target tissues, as opposed to local sources, are called "target-derived." The extracellular matrix provides essential axonal guidance cues for innervating neurons and modulates neurotrophic factor signaling activity (Van Vector et al., 2006; Sariola and Saarma, 2003).

[0002] One family of neurotrophic factors is the neurotrophins, which include Nerve Growth Factor (NGF), Brain-derived neurotrophic factor (BDNF), Neurotrophin three (NT3), and Neurotrophin four (NT4). These proteins help stimulate and control neurogenesis. All four neurotrophins can bind to the low affinity neurotrophin receptor, p75.

[0003] Glial cell-line Derived Neurotrophic Factor (GDNF) is a small protein belonging to a second family of neurotrophic factors, the GDNF family of ligands (GFL). GDNF forms a multi-component complex with two receptors, the c-Ret receptor tyrosine kinase and the GDNF family receptor α-1 (GFRα-1) (Treanor 1996). GDNF is essential to the development and survival of many neurons, including neurons that use dopamine as a neurotransmitter. These dopaminergic neurons are the neurons which degenerate in the course of Parkinson's Disease. Parkinsonian animals treated with GDNF have undergone functional recovery (D. M. Gash et al., Nature 380: 252-255 (1996). Other studies have shown that GDNF protects dopaminergic neurons against the neurotoxic drugs methamphetamine and 6-hydroxydopamine (W. A. Cass, J. Neurosci 16: 8132- 8139 (1996); C. W. Shults, Neuroreport 7: 627-631 (1996). Because of GDNF's vital and neuroprotective role, it is a promising target for clinical studies of Parkinson's Disease and brain injury.

[0004] Previous studies in tissue culture have shown that GDNF signaling requires Heparan Sulfate (HS), a major component of the extracellular matrix. HS, covalently attached to a protein core of HS proteoglycans, is a linear polysaccharide of repeating disaccharide units of uronic acid linked to glucosamine (Esko and Lindahl, 2001). During HS biosynthesis in the Golgi body, disaccharide units can be sulfated at the 2-O position of uronic acid and the N-, 3-O, and 6-O positions of glucosamine by specific sulfotransferases. Sulfation along the HS chain is incomplete, thus creating highly sulfated domains separated by partially sulfated and non-sulfated domains (Esko and Lindahl, 2001). The highly sulfated domain is considered the major signaling domain of HS due to its interaction with a number of extracellular signaling ligands and their receptors. GDNF binding to HS requires 2-O- and 6-O-sulfate groups along the HS chain (Rickard et al., 2003), and inhibition of HS sulfation by chlorate blocks GDNF signaling in primary neurons (Barnett et al., 2002). The in vivo functions of HS in the GDNF signaling pathway have not been investigated.

[0005] Sulfs are newly discovered sulfatase family members with unique structural features, enzymatic activities, and signaling functions (Dhoot et al., 2001 ; Ai et al., 2005). Sulfs contain the essential enzymatic sequences conserved among all sulfatases. In addition, they contain distinct hydrophilic sequences that are required for Sulfs' enzymatic activities and to dock Sulfs on the surface of the cell (Dhoot et al., 2001; Ai et al., 2006). Vertebrates express two Sulfs, Sulfl and Sulf2, which have identical substrate specificity towards a subset of 6-O-sulfate groups in the highly sulfated domain of HS chains (Morimoto-Tomita et al., 2002; Ai et al., 2003). Sulfs enzymatically remodel the HS 6-O-sulfation pattern on the cell surface to regulate HS binding to signal ligands and receptors in a diversity of signaling pathways, including Wnt (Ai et al., 2003), FGF and HGF (Lai et al., 2003; Wang et al., 2004), BMP (Viviano et al., 2004), and Shh (Danesin et al., 2006). Sulfs cell surface localization and function in signaling regulation distinguishes them from other intracellular sulfatases, which function as metabolic enzymes to degrade sulfated intermediates (Kresse et al., 1980). Sulfs are dynamically expressed in the developing nervous system and myogenic progenitors, as well as in other tissue types (Dhoot et al., 2001; Ohto et al., 2002; Braquart-Varnier et al., 2004; Nagamine et al., 2005). Avian Sulfl controls Wnt-dependent myogenic specification and is implicated in Shh-regulated oligodendroglial specification (Dhoot et al., 2001; Danesin et al., 2006). SuIf regulation of the GDNF signaling pathway has not previously been studied.

[0006] Sulfs are ordinarily named after the species in which they are found: Hsulf (human); Msulf (mouse); and Qsulf (avian, quail). The present inventors investigated SuIf function in the nervous system and developmental signaling using MSuIf mutants. From their studies in the esophagus, the inventors describe herein the cooperative functions of MSulfs in GDNF signaling required for neuronal innervation and enteric glial cell formation.

Summary of the Invention

[0007] Heparan sulfate regulation of the neurotrophic factor-dependent organ innervation has not been investigated. In this study, genetic disruption was used to investigate the functions of the extracellular HS 6-O-endosulfatases, Sulfl and Sulf2, in esophageal innervation. MSulfl and MSulf2 are differentially expressed by GDNF- expressing esophageal muscle and innervating neurons. The present invention is based on the discovery that MSulfl and MSulf2 deficiencies lead to direct defects in GDNF- dependent neural innervation and formation of neural crest-derived enteric glial cells in the esophagus. MSulfs regulate cellular HS 6-O-sulfation in vivo, and thus reduce GDNF binding to HS, thereby enhancing GDNF signaling activity in neuroblastoma cells and embryonic esophageal explants to induce neurite outgrowth for muscle innervation. These findings reveal cooperative functions of MSulfs in controlling HS transmission and reception of GDNF from target muscle to innervating neurons, and establish Sulfs as new regulators of tissue interactions during neural development.

[0008] Methods are described in which SuIf proteins are used to regulate GDNF signaling, GDNF binding to HS, and GDNF-induced neurite outgrowth. The invention also includes methods wherein the increase in GDNF signaling is assayed by monitoring the phosphorylation of downstream kinases and the formation of glial cells. The invention provides methods of using Sulfs in GDNF-mediated neuroprotection in the treatment of pathologies characterized by neuronal innervation defects including Parkinson's Disease, brain injury, and esophageal disorders.

Brief Description of the Drawings

[0009] Figure 1. MSulfl and MSulf2 are differentially expressed in various embryonic tissues, including the esophagus.

(A) Characterization of mRNA and protein expression of MSulfl and MSulΩ at E 16.5 in the neural tube (NT), lung, dorsal root ganglion (DRG), bone, cartilage, and skeletal muscle (SK). In situ hybridization and immunohistochemistry were performed on cross- sections using specific MSuIfRNA in situ probes and antibodies against the hydrophilic domain of MSulfl (MSlHD) and the hydrophilic domain of MSulf2 (MS2HD), respectively. The expression patterns of MSuIf mRNA and protein completely overlapped. OL - oligodendr cyte; FP- floor plate of neural tube.

(B) Differential expression of MSulfl and MSulf2 in muscle and neural progenitors in the embryonic esophagus. Scale bars, 100 μm.

(a) MSulfl esophageal expression at E14.5. Outer and Inner refer to layers of the esophagus. Eso - esophagus; L - lumen; E - esophageal epithelium lining the lumen, (b-c) MSulfl mRNA (b) and protein (c) were detected at the outer layer of the esophagus and the protein was mostly on the cell surface (insert in c).

(d) The outer layer of the esophagus expresses SM22, a smooth muscle marker.

(e) Immunostaining with the rabbit anti-MSlHD antibody and the rabbit anti-SM22 antibody together identified MSulfl on the membrane of SM22-expressing cells. The insert shows the enlarged image of the outlined staining.

(f) MSulfl did not co-localize with the TuJl staining.

(g) GDNF w^s detected diffusively across the esophageal muscle layers and partially overlapped with MSulfl in the outer layer.

(h-j) MSulfl expression and co-localization with GDNF in the esophageal muscle layers at E16.5 by double staining. (k-n) MSulf2 esophageal expression at E 14.5. Cross-sections were double stained with the MSulf2 antibody and the TuJl antibody. MSulf2 tightly associated with TuJl staining. (C) Antibodies against MSulfl and MSuIG identify specific MSulfs. (a) The antisera against MSlHD and MS2HD detect selectively the corresponding antigens by Western blot assays. Purified MS lHD and MS2HD were resolved by SDS-PAGE and detected with the antisera against MS 1 HD and MS2HD, respectively, by Western blot assay, (b) Antibodies against MSlHD and MS2HD selectively label transfected cells using immunocytochemistry. The HB 12317 neuroblastoma cells were transfected with the expression vectors of MSulfl or MSulΩ. After 24 hr, the HB 12317 cells were fixed followed by immunocytochemistry using specific antibodies against MSulfl HD and MSulf2HD. Scale bar, 40 μm.

[0010] Figure 2. MSulfl-/-; MSulβ-/- mice have cellular HS 6-O-sulfation, postnatal growth, and dysfunctional esophageal phenotypes.

(A) Disaccharide analysis of HS. Radiolabeled HS was isolated from Mouse Embryonic Fibroblasts (MEFs) of wildtype, MSulfl-/-, MSulβ-/-, and MSulfl-/-; MSulβ-/- embryos at E14.5. After purification, [³⁵S]HS was chemically cleaved by nitrous acid deamination and the generated disaccharides were separated by HPLC. Individual disaccharides are represented as the percentage of the total radioactivity. Data presented are mean and standard deviation of a minimum of 2 independent samples of each genotype. The HS isolated from MSulfl-/- and the MSulfl-/-; MSulβ-/- MEFs is enriched for trisulfated disaccharide ISMS. M in disaccharide abbreviation stands for the 2,5-anhydromannitol deamination products of GIcNS residues. ** p<0.01, * p<0.05.

(B) Comparison of the body size and weight between the wildtype, MSulfl-/-, MSulβ-/-, and MSulfl-/-; MSulβ-/- female mice at postnatal day 12 (P 12) and after weaning (n=6 for each group).

(C) Histology of the adult esophagus and lung of the wildtype control and MSuIf 1-/- ; MSulβ-/- mice. MSuIf l-/-;MSulβ-/- mice have enlarged esophagi with food accumulated inside (compare al, a2 with bl, b2, respectively) and develop lung infections (n=14). Eso - esophagus; Out - outer skeletal muscle layer; Inn - inner smooth muscle layer; L - lumen. Scale bars, 100 μm. (D) Generation of MSuIJ 1 knockout Embryonic Stem (ES) cells.

(a) Sequence alignment of the enzymatic domain of MSulfl (MSuIfIE) and MSulf2 (MSulf2E) from amino acids 41-416. The sequences encoded by the second coding exon (exon 2) are shown in the black box. The essential Cys residue in exon 2 is marked by an asterisk star. The two sulfatase signature sequences are underlined. The homologous amino acids are in normal font. Sequences that are not conserved between MSulfl E and MSulf2E are in bold.

(b) Schematic gene targeting strategy for MSulfl. The MSulfl targeting vector contains a 4.5-kb 5' homologous arm and a 3.9-kb 3' homologous arm. A loxP site was inserted 3' to the MSulfl exon 2 by a Nhel site. Two external probes, located 5' and 3' to the homologous arms, respectively, were used for Southern blot. The probe shown is the 5' external probe. The MSulfl targeted allele was generated by homologous recombination in ES cells. The floxed exon 2 was subsequently removed by the Cre recombinase to generate the MSulfl mutant allele.

(c) Southern blot to identify the recombinant MSulfl ES cell clones. Genomic DNA was extracted from individual ES cell clones and cut by Kpnl restriction enzyme. Homologous recombination between the MSulfl genomic sequences and the targeting vector was detected by Southern blot using a radiolabeled 5' external probe. The MSulfl wildtype allele generates a 12.7-kb DNA fragment, whereas the targeted MSulfl allele generates a 10.3-kb DNA fragment. The Southern blot results using Sad restriction enzyme and the 3 ' external probe were not shown.

(d) Genotyping to detect the wildtype allele and the mutant allele of MSulfl.

(e) RT-PCR assay to detect the expression of MSulfl mRNA and MSulfl mRNA in MSulfl-/- embryos. RT-PCR using MSulfl primers located in exon 3 did not detect any MSulfl mRNA transcript in MSulfl-/- embryos, while a partial MSulfl mRNA transcript was detected in MSulfl-/- embryos using primers located in downstream exons 11 and 15. In contrast, MSulfl +/- embryos expressed full-length MSulfl mRNA and the expression of MSulfl mRNA was unaffected in MSulfl-/- embryos. The Glyceraldehyde 3- phosphate dehydrogenase (GAPDH) mRNA expression was used as the control of the total amount of mRNA in the assay. [0011] Figure 3. The esophagi of MSuIJ ϊ -/-; MSulfl-/- mice have normal skeletal muscle function, but impaired smooth muscle contractility. MSuIfI+/- ;MSulf2+/- is the control. (A) Postnatal development of the esophageal skeletal muscles in the control and MSuIf l-/-;MSulfl-/- pups. The cross-sections of the control and MSulfl-/-;MSulf2-/- mice were immuno-stained with antibodies against skeletal muscle markers, including (a) myosin heavy chain (MF20) and (b) Myf5, or incubated with Cy3 -conjugated alpha- bungarotoxin (alpha-BTx) to identify (d) acetylcholine receptors on muscle. The completion of the esophageal skeletal muscle formation at Pl 5 was assayed by immunostaining of the abdominal segments with an alkaline phosphatase-conjugated mouse antibody against (c) fast, skeletal myosin (sk-myosin). The antigen-antibody complex was visualized by the substrate BM purple. Arrowheads mark the junction between the esophagus and the stomach. The MSuIf 1 -/-;MSulf2-/- esophagi have completed the skeletal muscle formation in the esophagus at P 15. Sto - stomach. Scale bars, 100 μm.

(B) Physiological measurements of the esophageal skeletal muscles. The esophagi of the adult control and MSuIf 1 -/-;MSulf2-/- mice were stimulated by twitch and tetanus stimuli. Muscle contractility was measured and compared to conditions with the presence of selective ion channel blockers (n=3). The MSuIj 1 -/-;MSulf2-/- esophagi exhibited comparable skeletal muscle contractility in response to the electrical stimuli and ion channel blocks as the control esophagi.

(C) Physiological tests of the esophageal smooth muscle contractility. The smooth muscles of the control and the MSuIf 1 -/-;MSulfl-/- mutant esophagi were dissected and their contractile forces in response to various stimuli were measured (n=3). The MSulfl-/-; MSulf2-/- esophagi showed the diminished smooth muscle contractility induced by carbachol, but partially reduced in response to other chemicals.

(D) Quantification of the esophageal smooth muscle contractility induced by various stimuli shown in (C).

(E) Generation of MSulfl knockout ES cells, (a) Schematic gene targeting strategy for MSulfl. The MSulfl targeting vector contains a 2.5-kb 5' homologous arm and a 3.1-kb 3' homologous arm. A loxP site was inserted 3' to the MSulfl exon 2 by an AatII site. Two external probes, located 5' and 3' to the homologous arms, respectively, were generated for Southern blot. The probe shown is the 3' external probe. The MSulβ targeted allele was generated by homologous recombination in ES cells. The floxed exon 2 was removed by Cre recombinase to generate the MSulβ mutant allele, (b) Southern blot to identify the recombinant ES cell clones. Genomic DNA was extracted from the individual ES cell clones and cut with BamHI restriction enzyme. Homologous recombination between the MSulβ genomic sequences and the targeting vector was detected by Southern blot using a radiolabeled 3 ' external probe. The MSulβ wildtype allele generates a 12-kb DNA fragment, whereas the targeted MSulβ allele generates a 10.5-kb DNA fragment. The results of Southern blot using BgII restriction enzyme and the 5' external probe was not shown, (c) Genotyping to detect the wildtype allele and the mutant allele of MSulβ. (d) RT-PCR assays to detect the expression of MSuIj 7 mRNA and MSulβ mRNA in MSulβ-/- embryos. RT-PCR using primers located in targeted exon 2 of MSulβ did not detect any MSulβ mRNA transcript in MSulβ-/- embryos, while a partial MSulβ mRNA transcript was detected in MSulβ-/- embryos using primers located in downstream exons 6 and 8. MSulβ +/- embryos expressed full-length MSulβ mRNA and the expression of MSuIj 7 mRNA was unaffected in MSulβ-/- embryos. The GAPDHmRNA expression was used as the control of the total amount of mRNA in the assay.

[0012] Figure 4. The esophagi of MSuIj l-/-;MSulβ-/- mice have diminished neuronal innervation and enteric glial cells. Thoracic segments of the esophagus were immunolabelled either on cross-sections or by whole mounts with various antibodies to examine the neural innervation. Scale bars, 100 μm. Statistics were calculated by two- tailed student t-tests.

(A) Comparable TuJl staining and expression of GFRαl, p75, and GDNF between MSuIfI+/-; MSulβ '+/- control esophagi and MSulβ-/-; MSulβ '-/- esophagi at E14.5 (n=3). GDNFRl is GFRαl.

(B) Reduced neural innervation at esophageal smooth muscle layer in MSuIf l-/-;MSulβ-/- esophagi at El 8.5 (n=4). (a)The whole esophagus was mounted on the glass slide after whole-mount immunohistochemistry with the TuJl antibody. One side of the flattened esophagus was photographed. The number of neurons (cell body marked by *) on one side of a 1-mm longitudinal segment was counted and 3 non-overlapping segments were counted for each esophagus. Numbers presented were the averages of the neural number per 1-mm segment. (b)The innervation density at the smooth muscle layer was calculated by dividing the total number of innervating TuJl+ axons on cross-sections (shown by white arrowheads) by the length of smooth muscle. The length was not different between control and MSulfl -/-;MSulf2-/- esophagi at E18.5. A minimum of 5 serial cross-sections (200 μm apart) of each esophagus was quantified and the innervation density was normalized to the littermate control. (c)The smooth muscle innervation was also shown by P75 staining (white arrowheads). MSuIf 1-/-,MSuIp-/- esophagi had the same number of neurons as the control, while the smooth muscle innervation was reduced in MSulfl-/-;MSulf2-/- esophagi.

(C) Reduced esophageal innervation and enteric glial cells in MSulfl -/-;MSulf2-/- esophagi at P 15. Arrowheads indicated the TuJl+ axons innervating the smooth muscle layer. Arrows pointed to the enteric glial cells located between the skeletal muscle layers. The smooth muscle on cross sections of MSuIj l-/-;MSulf2-/- esophagi was -20% longer than that of the littermate controls. A minimum of 20 sections from 4 independent controls or MSulfl-/-; MSulfl-/- mice were counted. Data shown in the bar graphs represented the normalized innervation density and the average number of enteric glial cells per cross section.

(D) The esophagi of MSulfl -/-; MSulfl-/- adult mice have diminished neuronal innervation and enteric glial cells, (a) Detection of neuronal innervation and enteric glial cells in the esophagus of control and MSulfl-/-; MSulfl-/- mice. The cross-sections of the thoracic segment of the adult esophagi were immunolabelled with the TuJl antibody to detect neuronal innervation of the smooth muscle (indicated by arrowhead) and the antibody against GFAP to label enteric glial cells (pointed by arrows) in the esophagus (n=4). (b) Serial sections of 200 μm apart were used to quantify the total number of the TuJl+ axons innervating the smooth muscle layer and the GF AP+ enteric glial cells on each section. Innervation density was calculated by dividing the total number of innervating axons by the length the smooth muscle layer on cross sections which was ~85% longer in the MSulfl -/-;MSulfl-/- adult esophagi than in the littermate controls. A minimum of 20 sections from 4 independent controls or MSuIf 1-/-,MSuIp-/- mice were counted. Scale bars, 100 μm.

[0013] Figure 5. MSulfl and MSuIG regulate GDNF binding to heparin and GDNF signaling activity. Data presented were mean and standard deviation of a minimum of 3 independent experiments. ** PO.01, * PO.05 (two-tailed student-t test).

(A) MSulfl reduces GDNF binding to heparin. Heparin conjugated to agarose beads was digested by MSulfl or inactive QSuIfI(C-A) control. Various amounts of GDNF were incubated with enzyme-digested heparin beads (20 μl) to allow binding. The amount of GDNF bound to the beads was assayed by Western blot. The intensity of individual bands was quantified by Multi-analysis software (Bio-Rad). Numbers listed below the blot were normalized quantification of the individual bands from three independent experiments.

(B) MSulfl had no effect on GDNF binding to GFRαl . Heparin pre-digested either by MSulfl or inactive QSuIfI(C-A) was added to a mixture of GDNF (10 ng) and GFRαl - Fc (1 μg) to allow GDNF-heparin-GFRαl ternary complex formation. The complex was pulled down by protein A agarose beads. The amount of GDNF bound to GFRαl was assayed by Western blot and normalized to the amount of GFRα 1.

(C-E) MSulΩ enhanced the GDNF signaling activity in neuroblastoma cells. NG- 108- 15 cells that were stably transfected with the control (inactive QSuIf) vector or the MSulf2 expression vector were stimulated by GDNF at various concentrations for 5 min (C,D), or by GDNF (5 ng/ml) for various lengths of time (E). The activation of GDNF signaling pathway was analyzed by assaying the phosphorylation of the c-Ret receptor (p-Tyr) and the downstream AKT kinase using Western blot. Total c-Ret receptor or ERKs were used as loading control. Data shown were controlled for loading and then normalized to the basal level of control cells. GDNF activated the phosphorylation of c-Ret and AKT 3-10 fold above the basal level, and MSulf2 enhanced the phosphorylation levels up to 9-20 fold above the basal level. MSulf2 also prolonged the GDNF signaling activity in neuroblastoma cells. p-Ret - phosphorylated active receptor; p-AKT - phosphoylated AKT; total c-Ret - includes phosphorylated and non-phosphorylated forms of the c-Ret receptor. (F) MSulf2 had no effect on NGF signaling in cells. Serum-starved PC 12 cells that stably expressed MSuIf2 or inactive QSuIfI(C-A) were treated by NGF. The activation of NGF signaling was analyzed by assaying the phosphorylation of downstream ERKs. (G)(a-b) MSulf2 enhanced the GDNF signaling activity in the neuroblastoma cells. NG- 108-15 cells that were stably transfected with the control vector or the MSulΩ expression vector were stimulated by GDNF at various concentrations for 5 min, or by GDNF (5 ng/ml) for various lengths of time. The activation of GDNF signaling pathway was analyzed by assaying the phosphoryation of downstream Erk kinase using Western blot. Total ERK was used as the loading control. Data shown were controlled for the amount of cell lysates and then normalized to the basal level of the control cells. Data presented were mean and standard deviation of a minimum of 3 independent experiments. * P<0.05 (two-tailed student-t test).

Figure 6. MSuIf 1 -/-;MSulf2-/- esophagi had defective GDNF-dependent neurite outgrowth by El 1.5 esophageal explant assays. Esophagi were dissected from El 1.5 embryos and plated on collagen gel containing either BSA, GDNF, or neurotrophins at various concentrations. After 4-5 days, the whole explant was immunostained with the

TuJl antibody. Explants that failed to attach to collagen gel were not included in the assay. Scale bars, 1 mm in (A) and (B); 200 μm in (D). ** PO.01, * PO.05 (two-tailed student-t test).

(A-B) Neurite outgrowth of El 1.5 esophageal explants was selectively dependent on

GDNF, but not on neurotrophin. MSulfl-/-;MSulf2-/- esophagi failed to extend neurites at

10 ng/ml GDNF and showed reduced neurite outgrowth at 20 ng/ml GDNF and 50 ng/ml

GDNF.

(C) Quantification of the neurite outgrowth shown in (A)(right) and (B)(left). The proportional length of the extended neurite to the size of the explant was measured along

6 axes, 30° apart and the average was calculated to represent the neurite outgrowth of one explant. Data represented the mean and the standard deviation of a minimum of 3 individual cultures.

(D-E) Quantification of the total number of neurons in the explants. The neurons in the explants (dark cell body staining by the TuJ 1 antibody, pointed by arrows) were quantified using the bright field microscopy at low magnitude. Neurons were scattered, or had even migrated out of the control explants in the presence of 10 ng/ml GDNF. In control explants cultured in the presence of BSA or NGF and in GDNF-treated MSulfl-/-;MSulf2-/- explants, neurons tended to form clusters. The large clusters of neurons in MSuIf 1-/-; MSulf2-/- explants were quantified by adding the neuronal numbers at different focal planes.

Detailed Description of the Invention

[0014] The present invention involves a novel method of regulating GDNF signaling. Biochemical and cell biological studies described herein reveal that MSulfs, the primary regulators of cellular HS 6-O-desulfation in vivo, reduce GDNF binding to HS and selectively promote GDNF signaling without affecting Nerve Growth Factor (NGF) signaling.

[0015] To investigate the functions of SuIf genes during neural development and developmental signaling, mouse mutants were generated with targeted deletions which disrupt the function of MSuIf 1 and MSulf2 genes. The focus of this study was on the neuronal innervation defect of MSuIf double mutant mice. As is discussed in the Exemplification section which follows, defects observed during immuno-histological studies in the esophagus of MSuIj ^~l-/-;MSulf2-/~ mice suggest a role for MSulfs in the GDNF signaling pathway during the establishment of esophageal innervation (see Figures 3 and 4). The MSwZ/double mutant mice showed reduced levels of axon sprouting and innervation density at the smooth muscle layer. Smooth muscle innervation was defective in the MSuIf double mutant mice. The studies showed that MSu/fdeficiency leads to reduced esophageal innervation (Figure 4C) without affecting the number of intrinsic neurons, which suggests that MSulfs control GDNF signaling levels for differential aspects of GDNF-dependent neuronal development. The present invention includes using one or multiple SuIf proteins to promote GDNF signaling in neuronal cells during innervation. The invention includes methods for treating a pathology characterized by neuronal innervation defects comprising administering, to a patient, a solution comprising a SuIf protein in a physiologically compatible buffer in an amount, and for a period of time, sufficient to substantially increase GDNF signaling in neuronal cells.

[0016] SuIf proteins promote GDNF signaling likely by reducing GDNF-HS binding. When GDNF is released from the extracellular matrix, the GDNF ligand is mobilized and interaction with its receptors is promoted. To test whether Sulfs regulate GDNF signaling, heparin-agarose beads were used to show that MSulfl significantly decreases GDNF binding to heparin by up to 6 fold (Figure 5A). Thus the present invention includes administering a solution comprising Sulfl in a physiologically compatible buffer to substantially increase GDNF signaling in neuronal cells.

[0017] To test whether SuIG regulates signaling activity of GDNF in neurons, neuroblastoma cell lines that do not express Sulfs were transfected with a vector expressing MSul/2. Expression oϊMSulfi increased activation of the GDNF signaling pathway by increasing the phosphorylation levels of c-Ret and downstream kinases up to 2-3 fold (Figures 5C and 5D). These studies in the esophagus, show that MSulΩ enhances GDNF signaling in neural progenitors during innervation. Therefore, the present invention also includes administering a solution comprising Sulf2 in a physiologically compatible buffer sufficient to substantially increase GDNF signaling in neuronal cells. The invention includes methods wherein the GDNF signaling in neuronal cells is increased at least 2-fold.

[0018] MSuIf modulation of GDNF signaling is consistent with the essential roles of Sulfs in establishing functional neuron-target muscle interactions. MSulfs are essential developmental regulators of GDNF signaling during esophageal innervation. Loss of MSuIf function, which leads to a significant decrease in GDNF signaling activity, dramatically impacts innvervation.

[0019] As is discussed in the Background section, NGF is a neurotrophic factor that helps stimulate and control neurogenesis. The present study also shows that SuIf proteins promote the signaling activities of GDNF, but not NGF, in neuronal cell lines. This demonstrates the functional selectivity of Sulfs in neurotrophin signaling. As is discussed in the Exemplification and is shown in Figure 5(F), in assays of the phosphorylation of downstream kinases, MSulΩ had no effect on NGF signaling in PC 12 cells (neural cell line). On this ground, the present invention includes a method for treating a pathology characterized by neuronal innervation defects comprising administering, to a patient, a solution comprising a SuIf protein in a physiologically compatible buffer in an amount, and for a period of time, sufficient to substantially increase GDNF signaling in neuronal cells wherein NGF signaling in the patient is not substantially affected.

[0020] The current study has revealed a novel regulatory function for Sulfs and 6- O-sulfated HS sequences in the GDNF signaling pathway. As is detailed in the Exemplification section, an assay using heparin-agarose beads was used to demonstrate that MSulfl reduces GDNF binding to heparin. The assay showed that MSulfl activity significantly decreased GDNF binding to heparin by up to 6 fold at sub-saturating amounts of GDNF, and by -30% under conditions of excess GDNF (Figure 5A). While not wishing to be bound by theory, this assay indicates that MSuIf promotes GDNF signaling by reducing GDNF-HS binding to mobilize GDNF ligand and promote GDNF interaction with its receptors. MSulfl decreases GDNF binding to HS and potentially releases GDNF stored in the extracellular matrix of the target muscle for the innervating intrinsic neurons.

[0021] The results of this study reveal a novel function of Sulfs and HS 6-O- sulfated sequences in GDNF signaling and establish Sulfs as essential, new regulators of neuron-target tissue interactions. The finding that Sulfl 6-O-desulfation of heparin reduces GDNF binding to heparin is consistent with previous results using chemically desulfated heparin showing that the GDNF-heparin interaction is dependent on heparin 2- O- and 6-O-sulfation, but not on N-sulfation (Davies et al., 2003; Rickard et al., 2003).

[0022] In the experiments described herein, it was also determined whether Sulfl activity affected HS-regulated GDNF binding to GFRαl. The results showed that MSulfl activity had no effect on GDNF binding to GFRαl (Figure 5(B)). GDNF and GFRαl -Fc were incubated with heparin pre-digested with MSulfl and the resulting GDNF-heparin-GFRαl-Fc ternary complex was precipitated and analyzed. MSulfl desulfation of heparin did not affect heparin enhanced GDNF binding to GFRαl in vitro, indicating that MSulfl regulated HS 6-O-desulfation does not control GDNF binding to GFRαl. [0023] The present study shows that Sulfs quantitatively promote neurite outgrowth of intrinsic neurons to establish esophageal muscle innervation, but have no effect on the number of intrinsic neurons. The experiments herein further demonstrate, using MSulfs and esophageal explants, that neurite outgrowth is dependent on GDNF, but not other neurotrophins, and thus these observations establish functional specificity of Sulfs in the GDNF pathway during esophageal innervation. The present invention includes methods of promoting GDNF-induced neurite outgrowth by intrinsic neurons to innervate muscle.

[0024] MSulfl-/→;MSulf2-/- mice exhibit significant defects in multiple tissues. GDNF is an essential neurotrophic factor for enteric innervation. This report focuses on the analysis of the innervation in the esophagus. Immunohisto logical techniques showed that GDNF and MSulfl co-localize in esophageal muscle layers (Figure lB(g-j)). [0025] As is described in the Exemplification and shown in Figure 3, the esophageal defects in MSulfl -/-;MSulf2-/- mice result from impaired muscle contractility due to defects in neuronal innervation. In MSulfl -/-;MSulf2-/- mice, the smooth muscles showed a greatly diminished response to carbachol and a selective decrease in carbachol- induced smooth muscle contractility. These results indicate specific neuronal innervation defects. Immunohistological studies described herein also show that MSulfl -/-;MSulf2-/- esophagi have much reduced levels of axon sprouting and innervation density at the smooth muscle layer, (Figures 4(A), 4(B)). Figure 4(C) also shows reduced esophageal innervation in MSulfl -/-;MSulf2-/- esophagi providing additional evidence that smooth muscle innervation was defective in MSuIf double mutant mice. These observed esophageal defects show that Sulfs play a role in the GDNF signaling pathway during innervation and that SuIf deficiency leads to reduced innervation.

[0026] To further show that Sulfs are required for GDNF-induced neurite outgrowth, esophageal explant cultures were used. Experimental results established that GDNF is selectively required for neurite outgrowth of intrinsic neurons in embryonic esophagi (Figures 6A-6C). MSulfl -/-;MSulf2-/- explants failed to extend neurites at 10 ng/ml GDNF, and the neurite outgrowth was reduced to about one third of the control level at 20 ng/ml GDNF. Neurite outgrowth was also reduced at 50 ng/ml GDNF. Thus MSulfl and MSulΩ are required for GDNF-dependent neurite outgrowth in embryonic esophageal explant cultures. Results showed that the lack of neurite outgrowth by MSuIf 1 -/-;MSulf2-/- explants was due to disrupted GDNF signaling in intrinsic neurons. Figure 6A shows MSuIf 1 -/-; MSulft-/- esophageal explants with defective GDNF- dependent neurite outgrowth. Figure 6D shows that neurons within MSulfl -/-;MSulf2-/- explants formed large clusters, whereas neurons from control explants were scattered or migrating, indicating a defect in GDNF-induced neural migration in the absence of Sulfs. Therefore, embryonic MSulfl-/-; MSulfi-/- esophagi have defects in GDNF-dependent neural migration and neurite outgrowth for muscle innervation. MSw/^deficiency leads to defective neurite outgrowth in the esophagus both in vivo and in explant cultures, establishing MSuIf regulation of the GDNF pathway.

[0027] These neurite outgrowth studies, together with heparin binding and neural GDNF signaling studies, demonstrate that MSulfl and MSulf2 have dual, cooperative functions in GDNF signaling for establishing target-dependent esophageal innervation. MSulfl is actively expressed in esophageal muscle progenitors and mobilizes HS-bound GDNF from the extracellular matrix. MSulf2 is expressed by innervating neuronal progenitors and promotes reception of the GDNF signal to extend neurites to innervate esophageal muscle.

[0028] In the experiments with transfected neuroblastoma cell lines, discussed in the Exemplification section and above, cells transfected with an MSul/2 expression vector displayed increased phosphorylation levels of c-Ret and PKB/AKT by 2-3 fold and ERKs by up to 2 fold in response to GDNF (Figure 5C, 5D, 5G(a)), when compared to cells transfected with inactive or empty expression vectors. The MSulf2-expressing cells also showed sustained activation of downstream PKB/AKT and ERKs over 30 min after GDNF stimulation. In control cells, GDNF signaling activity rapidly declined 15 minutes after stimulation (Figure 5E, 5G(b)). However, MSulf2 had no effect on NGF-induced phosphorylation of AKT or ERK in PC 12 cells (Figure 5F), demonstrating the functional specificity of MSulf2 in the GDNF signaling pathway. Therefore, the present invention includes methods for treating a pathology characterized by neuronal innervation defects wherein the increase in GDNF signaling resulting from the treatment described herein is assayed by monitoring the phosphorylation of PKB/AKT or ERKs. An increase in phosphorylation following administration of a solution comprising a SuIf protein indicates an increase in GDNF signaling.

[0029] Previous studies have established that GDNF controls the proliferation and differentiation of enteric glial progenitors from neural crest progenitors. The present study shows that Sulfs are also required for the formation of glial cells. The cooperative functions of Sulfs in enteric glial cell formation in the esophagus is reported herein. Immunohistological techniques with antibodies against glial cell markers (GFAP and SlOO) were used to test whether glial cells were affected in MSulfl-/-;MSulf2-/- mice. The tests showed that the number of GFAP-expressing cells of the MSuIf 1 -/-;MSulf2-/- esophagi were reduced to one-third to one half of that in the control esophagi (Figure 4C, 4D). The present invention includes methods for treating a pathology characterized by neuronal innervation defects wherein the increase in GDNF signaling resulting from the treatment described herein is assayed by monitoring the formation of glial cells in the patient. An increase in the rate of glial cell formation following administration of a solution comprising a SuIf protein indicates an increase in GDNF signaling.

[0030] This study is the first in vivo evidence that HS and Sulfs play essential roles in neurotrophic factor signaling during the development of the nervous system. This regulatory role for Sulfs and HS 6-O-sulfation in GDNF signaling provides a basis for the development of new, HS-based therapeutic approaches to improve GDNF- mediated neuroprotection in treating Parkinson's Disease and brain injury. Since this study has demonstrated that Sulfs regulate GDNF binding to HS, the present invention includes methods of treating Parkinson's Disease, brain injury, or esophageal disorders to provide GDNF-mediated neuroprotection. The methods comprise administering, to a patient, a solution comprising a SuIf protein in a physiologically compatible buffer in an amount, and for a period of time, sufficient to substantially increase GDNF signaling in neuronal cells, wherein the administration of the SuIf protein solution has no substantial effect on NGF-signaling in the patient. The methods include treating brain injury caused by stroke.

[0031] As is discussed in the Background section, GDNF is essential to the development and survival of many neurons, including the dopaminergic neurons which degenerate in Parkinson's Disease. Prior studies have shown that Parkinsonian animals treated with GDNF have undergone functional recovery (Gash et al., 1996 supra). It is commonsensical to combine treatment with GDNF with Sulfs since they increase GDNF signaling. Therefore the present invention includes any of the above methods wherein the administered SuIf protein solution further comprises GDNF. In another embodiment, the invention includes a pharmaceutical solution comprising a SuIf protein and GDNF in a physiologically compatible buffer.

[0032] Previous studies in tissue culture, also discussed in the Background section, have shown that GDNF signaling requires HS. Therefore the present invention includes any of the above methods wherein the administered SuIf protein solution further comprises an HS compound. The HS compound may be heparin, hi another embodiment, the invention includes a pharmaceutical solution comprising a SuIf protein and an HS compound in a physiologically compatible buffer. The HS compound may be heparin.

Exemplification

[0033] In this study, gene targeting was used in combination with physiological, biochemical and cell biological approaches to investigate the in vivo functions of MSulfl and MSulf2 in the regulation of esophageal innervation.

[0034] To investigate SuIf function in the nervous system and developmental signaling, mouse mutants of MSulfl and MSul/2 were generated. The focus of this study is on the neuronal innervation defect of MSuIf double mutant mice. The cooperative functions of MSulfs in GDNF signaling required for neuronal innervation and enteric glial cell formation in the esophagus is reported. The roles of MSulfs in esophageal innervation are direct. MSulfl and MSuIEZ are dynamically and differentially expressed by the GDNF-expressing muscle progenitor and neuronal progenitors during the innervation of developing esophagus.

[0035] The neuronal innervation of muscles along the gastrointestinal tract is dependent on GDNF, a target-derived neurotrophic factor (Baloh et al., 2000). The esophagus contains both skeletal muscle and smooth muscle arranged in concentric outer and inner layers, respectively. Two groups of interacting neurons innervate esophageal muscle to coordinate muscle contractility during eating and breathing: "extrinsic" neurons whose cell bodies are located in ganglia outside of the esophagus (such as the vagal nerve, the nodose ganglion and the dorsal root ganglia), and resident neural crest- derived "intrinsic" neurons (Sang and Young, 1998; Neuhuber et al.₃ 2006). Intrinsic neurons, whose cell bodies reside within the longitudinal and circular skeletal muscles of the esophagus, innervate the smooth muscle layer and function as interneurons to communicate with extrinsic neurons. To establish innervation, muscle progenitors express GDNF beginning at embryonic day 10 (ElO), peaking between El l and El 6 and diminishing at E18 (Hellmich et al., 1996; Golden et al., 1999). GDNF not only promotes the proliferation of the enteric neural crest precursors and their differentiation into intrinsic enteric neurons and enteric glial cells (Heuckeroth et al., 1998; Gianino et al., 2003), but also acts as a target-derived chemoattractant for directed axonal outgrowth of extrinsic and intrinsic neurons (Young et al., 2001, 2004; Yan et al., 2004). Innervating neurons reach the target muscle during the initiation of GDNF expression in the embryonic esophagus. For example, neural crest progenitors migrate into the esophagus by ElO.5 (Durbec et al., 1996), and extrinsic nerve fibers from the nodose ganglion are identifiable in the esophagus around El 2 (Sang and Young, 1998). Formation of functional neuron-neuron and neuron-muscle connections and enteric glial cells initiates in the embryo and proceeds postnatally until completion around two weeks after birth (Sang and Young, 1997; Breuer et al., 2004), concurrent with the maturation of the outer esophageal muscle layer from smooth muscle to skeletal muscle (Patapoutian et al., 1995; Rishniw et al., 2003). Neuronal innervation and target muscle maturation are interdependent processes required to achieve the highly coordinated esophageal function (Neuhuber et al., 2006). Defects in either process can lead to neonatal death and a variety of esophageal disorders in animal models and in humans, such as achalasia and a motility disorder, congenital idiopathic megaesophagus (Longstretch and Walker, 1994; Baloh et al., 2000; Sumiyoshi et al., 2004; Park and Vaezi, 2005).

Histochemistry

[0036] Esophagi were dissected from control and MSulfl-/-;MSulf2-/- mice, rinsed gently in ice-cold PBS, and fixed in periodate-lysine-paraformaldehyde (PLP) for 2 hr at 4⁰C. Each esophagus was washed three times in PBS and then cryo-protected by equilibrating sequentially in PBS buffer containing 7%, 15%, and 25% sucrose. Tissues were embedded in Optimum Cutting Temperature compound (OCT) and 8 μm-thick serial cryo-sections were collected. For GDNF staining, fresh tissues were frozen in OCT and tissue sections were fixed in methanol for 15 min at 4⁰C. Sections were stained with hematoxylin and eosin (H&E). Adjacent, unstained sections were used for immunohistochemical studies (see below).

Isolation of the smooth muscle layer from the adult esophagi and whole mount immunostaining

[0037] The adult esophagai were dissected from control and MSuIf l-/-;MSulf2-/- mice, rinsed gently in ice-cold PBS, and fixed in 4% paraformaldehyde/PBS for 1 hr at 4⁰C. The smooth muscle layer was separated from the skeletal muscle layer by forceps and flattened on the filter paper followed by additional 1 hr PFA fixation before immunohistochemistry with the TuJl antibody.

Single and multi-label immunohistochemistry

[0038] Immunohistochemistry on tissue sections was performed as described previously (Ai et al., 2003), except for an additional 1-hr blocking with MOM (Mouse On Mouse) blocking reagent (Vector) for mouse monoclonal antibodies. Antigen- antibody complexes were detected either by fluorescence or by chromogenic substrate. For explant cultures or whole mount staining, the washes were extended to 30 min each. Primary antibodies include: (i) rabbit anti-GFAP (Dako; 1:100); (ii) rabbit anti-MSlHD (1: 100); (iii) rabbit anti-MS2HD (1: 100); (v) mouse anti-skeletal fast myosin (clone MY- 32) alkaline phosphatase conjugate (Sigma, 1 :300); (v) goat anti-GDNF (R&D Systems, 2 ug/ml); (vi) mouse anti-neuron-specific class III beta-tubulin antibody (clone TuJ-I) (R&D Systems, 1:500); (vii) rabbit anti-smooth muscle-specific SM22 (a gift from Dr. Mario Gimona, 1 :1000). (viii) rabbit anti-p75 (Upstate, 1:200), (ix) goat anti-GFRα-1 (R&D system, 1 μg/ml); (x) rabbit anti-Ret (Santa Cruz Biotechnology, 1 :50); (xi) mouse anti-S100 (Chemicon, 1 :100). For immunofluorescence staining, the secondary antibodies (1:300) were purchased from Molecular Probes, including goat anti-mouse 546 (1:300), goat anti-rabbit and donkey anti-goat 546. Slides were examined with a fluorescent microscope (Leica DMR) and images were taken using a digital camera (Leica DC300F). Immunostaining with chromogenic substrates was performed using an ABC kit (Vectastain) per manufacturer's instructions. For antibodies conjugated with alkaline phosphatase, the antigen-antibody complexes were detected by reacting with substrate BM purple (Roche Diagnostics).

Generation of antibodies against MSuIf 1 and MSulβ

[0039] The sequences of the hydrophilic domain (HD) oϊMSulfl and MSulβ were cloned into a bacteria expression vector pET-30 by PCR. The polypeptides (MSlHD and MS2HD) were expressed and purified under denaturing condition using Ni²⁺ column following manufacturer's protocols (Novagen). MSlHD and MS2HD were then injected into the rabbit to generate antiserum (Cocalico Biologicals, Reamstown, PA) followed by affinity purification using the antigen-conjugated sepharose column (Amersham Biosciences). The specificity of the antibodies was tested by Western blot and immunostaining on transfected cells and embryonic tissue sections.

MSulfl and MSulf2 are differentially expressed in multiple embryonic tissues, including the esophagus.

[0040] We characterized the expression oϊMSulfl and MSulβ by in situ hybridization and by immunostaining using specific antibodies generated against the hydrophilic domains of MSulfl (MSlHD) and MSulβ (MS2HD) (Figure 1). The tissue distribution of MSuIf mKN A completely overlaps with protein expression, establishing the specificity of MSuIf antibodies (Figure IA, IB). Both MSulfl and MSulf2 are associated with the cell membrane in various tissues, consistent with Sulfs as secreted and membrane-docking proteins (Dhoot et al., 2001). The expression levels of MSulfs are relatively low before E9.5 (data not shown). At E14.5 and E16.5, MSulfl and MSulf2 are expressed at higher levels in partially overlapping patterns by a variety of embryonic tissues (Figure IA) (Ohto et al, 2002; Nagamine et al., 2005). At E14.5, both MSulfl and MSulf2 are expressed in the floor plate of the neural tube, bone, and cartilage (Figure IA). MSulfl and MSulf2 also show differential expression in skeletal muscle and lung, respectively (Figure IA). In addition, MSulf2 is the major endosulfatase expressed in the nervous system, including in the oligodendrocyte precursors in the ventral neural tube, the dorsal neural tube, and neurons derived from neural crest progenitors in the dorsal root ganglion and the sympathetic ganglion (Figure IA; data not shown). Oligodendrocytes are neuroglia cells that form the myelin sheath.

[0041] MSulfl and MSulΩ are dynamically and differentially expressed in the muscle and innervating neurons in the embryonic esophagus (Figure 1 B). The expression of MSulfs is first detectable around El 1.5 in the esophagus (data not shown). Their expression levels peak around E14.5 (Figure IB), decrease dramatically by El 8.5, and are undetectable two weeks after birth and in the adult (data not shown). MSulfl expression does not co-localize with neuronal class III beta-tubulin labeled by the TuJ 1 antibody (Figure lB-f). Instead, MSulfl mRNA and protein were detected at the outer layer of the esophagus at E14.5 where the muscle forms (Figure lB-b, c). MSulfl exhibited a distinct punctuated pattern on the cell surface, similar to avian Sulfl (Dhoot et al., 2001). To confirm muscle expression of MSulfl, double-labeling with antibodies against MSulfl and a smooth muscle marker SM22 was performed. Although both antibodies were raised in rabbit, membrane-bound MSulfl was distinguished from intracellular SM22 by their differential subcellular localization (Figure IB-C, d, e). We found that MSulfl outlined the membrane of SM22-expressing cells at the outer layer of E14.5 esophagus, and was localized at both outer and inner muscle layers at E16.5 (Figure lB-h), establishing that MSulfl is expressed in embryonic esophageal muscles. Esophageal muscle progenitors also express GDNF mRNA, an essential neurotrophic factor for enteric innervation (Hellmich et al., 1996; Golden et al., 1999). To test whether MSulfl and GDNF co-localize, we tested four commercially available GDNF antibodies and found only one antibody from R&D Systems gave a weak signal on embryonic sections. Using this antibody, we characterized GDNF expression in the esophagus at E14.5 and E16.5 for comparison to MSulfl expression. GDNF was found on the cell surface and appeared diffusely across the esophageal muscle layers at E14.5 (Figure IB- g), including the MSulfl -expressing outer layer, and overlapped with MSulfl at E 16.5 (Figure lB-ij). In contrast to MSulfl expression by esophageal muscle progenitors, MSulf2 was detected within the esophageal muscle layers and tightly associated with the neuronal marker TuJl (Figure lB-k,l,m,n). As MSulΩ and neurotubulin have differential subcellular localization, the observed close association between these two proteins indicates that^'MSulf2 is expressed by innervating neuronal progenitors. The differential expression of MSulfl and MSulf2 by muscle progenitors and by neural progenitors that innervate the muscle, respectively, suggests that MSulfl and MSulf2 have distinct, cooperative functions in esophageal development.

HS preparation and disaccharide analysis

[0042] Mouse Embryonic Fibroblasts (MEFs) were isolated from the skin of E14.5 mouse embryos after dissociation with dispase II (Boehringer Mannheim, 2 mg/ml). MEFs were cultured in DMEM with 10% FBS. The radiolabeling, preparation, and structural analysis of the HS were performed as described previously (Ai et al., 2003).

Generation of MSulfl knockout mice

[0043] An MSulfl cDNA fragment was generated by RT-PCR from E9.5 mouse embryos using primers that generate a 171 -bp probe in the second coding exon (encoding amino acids 65-122). Using this probe, we identified the full-length MSulfl cDNA by screening an E8.5 mouse embryonic cDNA library (constructed by Dr. Brigid Hogan's lab) and two BAC clones by screening of a BAC Mouse II hybridization library (Incyte Genomics). The gene structure of MSulfl on Chromosome 1 was identified based on the sequencing database from the Mouse Genome Project (available via the Internet at "www" dot "ncbi" dot "nlm" dot "nih" dot "gov", wherein the word "dot" should be replaced by "." to form the address of a web page at the National Center for Biotechnology Information). To construct the targeting vector of MSulfl, we cloned the 4.5-kb Smal-BamHI fragment as the 5' flanking homologous sequences and the 3.9-kb Xbal fragment (containing the second coding exon) as the 3 ' homologous sequences into a vector containing floxed PGK-Neo cassette and PGK-DTA cassette. A loxP site was inserted at the Nhel site 3' to exon 2 of MSulfl gene. Maintenance, trans fection, and selection of 129SvEv Embryonic Stem (ES) cells, as well as generation of chimeric mice were performed as described previously (Tompers and Labosky, 2004). Southern blot analysis using Kpnl sites and probes 5' to exon 2 was performed to genotype the electroporated ES cells. Sequencing of the PCR products were also used to confirm the insertion oϊNeo cassette and the loxp site around the targeted exon 2. Two individual lines carrying the floxed allele were generated and crossed with a CMV-Cre transgenic line (The Jackson Laboratories) to remove the targeted exon 2 and generate a mutant MSuIf 1 allele. Sequencing of the RT-PCR products was performed to confirm the identity of the mutant MSuIf 1 transcript and the expression of MSulβ. MSuIf 1-1- mice were bred onto the C57BL/6 genetic background.

Generation of MSulβ knockout mice

[0044] The identification of MSulβ full-length cDNA and BAC clones (on Chromosome 2) containing the second coding exon was performed similarly as described above for MSuIf 1, using primers to generate a 189-bp probe (encoding amino acids 63- 125). The 3' flanking homologous sequences in the MSulβ targeting vector are a 2.5-kb BamHI-SacI fragment, while the right flanking homologous sequences including the targeted exon 2 are a 3.1-kb Sad fragment. A LoxP site was inserted into the AatII site. Mice carrying the floxed allele and the mutant allele were generated as described above. The insertion of the loxP site and the mutant transcript were confirmed by sequencing the PCR products. Two individual MSulβ floxed and mutant lines were bred onto the C57BL/6 genetic background. The MSuIJ ^"1-/- and the MSulβ-/- mice were mated to generate the double heterozygous mutants. MSuIf 1-/-; MSulβ-/- mice or embryos were generated by crossing MSuIfI+/- ; MSulβ +/- mice with MSuIf 1-/-; MSulβ '+/- mice or by mating between MSuIf l-/-;MSulβ+A mice. The MSuIf 1-/-; MSulβ-/- mice were in a mixed 129SvEv/C57BL/6 genetic background.

Genotyping

[0045]Genotyping was performed by DNA extraction from tail followed by PCR. Primer sequences and the annealing temperatures of the PCR reaction (35 cycles) follow:

MSuIf 1 null allele: (PCR product: 270 bp; annealing temperature: 54⁰C) Forward primer: 5' TCCCTC AT ATC AG A AAGTCTGG 3' Reverse primer: 5' GCCTCCTG AC AAGGTT ACT AGG 3' MSulfl wildtype allele: (PCR product: 150 bp; annealing temperature: 56⁰C) Forward primer: 5' GC ATAG AGTC AGTGGGTCAAAGTTG 3' Reverse primer: 5' GCCTCCTG AC AAGGTTACTAGG 3'

MSulf2 primers: ((PCR product: 880 bp for wildtype allele and 480 for null allele; annealing temperature: 54⁰C)

Forward primer: 5' GCAGGTTTGTACCCAACGC 3'

Reverse primer: 5' GGTTACTCCCACAATAAACTGGTG 3'

RT-PCR

[0046] Total RNAs were extracted from El 1.5 embryos with Trizol solution (Gibco) following manufacturer's protocol. 5 μg total RNAs were reverse transcribed (Promega) followed by PCR. Primer sequences and PCR conditions follow:

MSulfl #1 (annealing temperature 58⁰C, 30 cycles, 510-bp PCR product): Forward primer (in 3^rd exon): 5' TGTTTGTCGCAACGGCATC 3'; Reverse primer (in 6^th exon): 5' GGACCACGAATGAAGAAAGGC 3'.

MSulfl #2 (PCR product: 320 bp; annealing temperature 58⁰C; 30 cycles): Forward primer (in 11^th exon): 5' GCTGCTGGTGACATCAGGAATG 3'; Reverse primer (in 15^th exon): 5' AAGGGGTGAAGGTGACTCTTTAGC 3'.

MSulβ #1 (PCR product: 326 bp; annealing temperature 56⁰C; 30 cycles): Forward primer (in 2^nd exon): 5' ACACCAATGTGCTGTCCGTCTC 3'; Reverse primer (in 4^th exon): 5' CGTGAGGT AATCCGTGGAGTAGTC 3'.

MSulfi #2 (PCR product: 220 bp; annealing temperature 56⁰C; 30 cycles): Forward primer (in 6^th exon): 5' TCTGAACCCCCACATTGTCCTC 3'; Reverse primer (in 8^th exon): 5' CACTTTGTCACCCTCCCTCTTG 3'. GAPDH primers (PCR product: 452 bp; annealing temperature 56⁰C; 25 cycles): Forward primer: 5' ACCAC AGTCC ATGCC ATC AC 3'; Reverse primer: 5' TCCACCACCCTGTTGCTGTA 3'

MSulfl-/-;MSulf2-/- mice have an enlarged esophagus and show postnatal growth defects.

[0047] To investigate the functions of SuIf genes during neural development and embryo genesis, we generated mice with targeted deletions of the second coding exon (exon 2) of MSuIJ 1 and MSulft using Cre-LoxP technology (Figures 2D & 3E). Exon 2 encodes 80 highly conserved amino acids in the enzymatic domains of MSulfl and MSulf2, including a Cys residue that is essential for enzymatic activity (Ai et al., 2003). Although MSw/y mutant embryos express partial MSuIf transcripts and polypeptides recognizable by MSuIf antibodies (Figures 2D & 3E, data not shown), removal of exon 2 eliminates the full-length MSuIf transcripts (Figures 2D & 3E) and disrupts MSuIf function, as shown by the disaccharide analyses of ³⁵S-radiolabeled glycosaminoglycans (GAGs) (Figure 2A). GAGs were isolated from cultured primary MEFs of the El 4.5 skin which express both MSulfl and MSulf2 (Figures 2D & 3E). MSW/mutant MEFs of all genotypes have normal HS chain length, anionic properties and N-sulfation, and chondroitin sulfate contents (data not shown). However, loss of MSulfl significantly increased (-30%) the abundance of the trisulfated disaccharide, IdoA2S-GlcNS6S (ISMS), the known substrate of SuIf enzymes (Morimoto-Tomita et al., 2002; Ai et al., 2003), and proportionately decreased the disulfated IdoA2S-GlcNS (ISM) disaccharide (Figure 2 A, Table 1). In contrast, MSulfl-/- and MSulfl +/-;MSulf2+/- MEFs had unchanged levels of ISMS disaccharide, suggesting redundant MSuIf activity in MEF cells. The magnitude of the increase in ISMS in various MSuIf ^'mutant MEFs was generally proportional to the number of mutant MSuIf alleles (Table 1). Most significantly, HS isolated from MSulfl -/-;MSulf2-/- MEFs had double the ISMS content (63%), a level. comparable to that of heparin, a highly sulfated HS derivative. Taken together, the results of RT-PCR (Figures 2D & 3E) and disaccharide analyses (Figure 2 A, Table 1) establish that M^wZ/mutants are loss-of- function mutations, and that SuIf enzymes are the major regulators of HS 6-O-sulfation in vivo. Table 1. Summary of the general phenotypes of MSuIf mutant mice.

Genotype % ISMS % Survival Adult body fertility

IdoA2S-GIcNS6S weight

Wildtype 29 +/- 3.3 100% normal normal

MSulfl-/- 38 +/- 4.2** 100% normal normal

MSulβ-/- 33 +/- 0.7 100% -90% normal

MSuIf 1 +/-;MSulβ+/- 34 +/- 0.7 100% normal normal

MSuIf l-/-;MSulβ+/- 46 +/- 3.1** 100% ND normal

MSulfl+/-;MSulβ-/- 59** -90% 70-90% Reduced in both sexes'

MSuIj ϊ-/-;MSulβ-/- 61 +/- 5.1** 100% at birth, 40-70% Much reduced in 45% at P21 both sexes²

% ISMS represents % of total O-35S labeled disaccharides generated after deaminative cleavage (pH 1.5) of purified HS. M in ISMS stands fro the 2,5-anhydromannitol deamination products of 2-deoxy-2-sulfamido-α-D-glucopyranosyl (GIcNS) residues (** PO.01).

¹ The average litter size is 4-6, compared with the normal size 8-10. The average litter size is 2-4.

[0048] Single and double MSuIf ^'mutant mice are viable at birth with normal facial and body morphologies. Heterozygous and homozygous MSu/fsingle mutant mice are born at the expected Mendelian frequency. They are fertile and have a normal life span of over 2 years (Figure 2B, Table 1); however, MSulβ-/- mice are smaller and weigh -10% less than wildtype littermates. The MSuIj ^'1-/-; MSulβ-/- embryos and neonates are also morphologically indistinguishable from control littermates (data not shown), and fully viable based on observed Mendelian frequencies at El 1.5 (14/107, expected 1/8), at E14.5 (14/69, expected 1/4) and at birth (34/311, expected 1/8). [0049] MSuIf 1 -/-;MSulf2-/- mice exhibit significant defects in multiple tissues. This report focuses on the analysis of the esophageal innervation phenotype. A majority of MSuIJ l-/-;MSulf2-/- pups (41/76) exhibit severe growth defects, evident as early as postnatal day 3 (P3) (Figure 2B), and death around P 14. The gastrointestinal tract of dying pups was filled with air bubbles (data not shown), suggesting a dysfunction in contraction of the upper esophageal sphincter muscle during feeding. MSulfl-/-;MSulf2-/- pups that survive into adulthood (35/76 mice) are runted and have greatly reduced fertility (Figure 2B, Table 1). Approximately 75% of adult MSulfl-/-;MSulf2-/- mice develop megaesophagus phenotypes as early as two months of age, characterized by food accumulation in the esophagus, coughing, labored breathing, and lung infection (Figure 2C). Compared to the wildtype esophagus, MSuIj 1 -/-;MSulf2-/- esophagus had an enlarged lumen with no disruption of the arrangement of the outer skeletal muscle layers, the inner smooth muscle layer, or the esophageal epithelium by H&E staining and by immunostaining with specific markers for each cell type, although these tissue layers appear much thinner due to the dilation of the esophagus (Figure 2C; data not shown). We did not observe inflammatory cells in the MSulfl -/-;MSulf2-/- esophagi at P12 or in surviving adults (data not shown), ruling out the possibility that the esophagus dilation is caused by tissue damage. In addition, we did not detect ectopic expression of the intact MSuIf gene in single mutant esophagi (data not shown). As the megaesophagus phenotype was prevalent only in MSulfl-/-; MSulfl-/ - mice, but not in MSuIf 1-/- mice (0%) or MSulfl-/- mice (1~2%), we hypothesize that MSulfl and MSulf2 have distinct, but cooperative functions that contribute to the megaesophagus phenotype in MSulfl -/-;MSulf2-/- mice.

[0050] The distinct expression of MSulfl and MSulf2 in embryonic esophagi, yet a lack of the megaesophagus phenotype in MSwZ/single mutant mice, suggests a functional compensation between MSulfl and MSulf2. One likely mechanism is through their low levels of free secretion (data not shown, Morimoto-Tomita et al., 2002) such that the remaining MSuIf may compensate for the mutated MSuIf in nearby tissues. Measurement of contraction forces of the esophagus skeletal muscle and smooth muscle

[0051] Esophagus was dissected from P12 or adult mice. To measure contractions of skeletal muscles, the 5-mm longitudinal esophagus slices were cut out, tied with silk threads at both ends, and subsequently hooked up to tungsten needle tips. One of the needles was connected to a force transducer (AM801, SensoNor, Horten, Norway) and the other to a micromanipulator to stretch the muscle length to 1.1 times of the slack. Platinum electrode wires were placed near the tubes on both sides. The esophagus slices were stimulated by 1 msec duration square pulse (TS S20, Intracel, UK), one single pulse for twitch and 30 Hz for 1 minute for tetanus. To measure contractions of smooth muscle, the smooth muscle cross-rings were dissected from lmm esophagus tubes, cleaned from surrounding skeletal muscle and connective tissues, and hooked up to two needle tips. The rings were stretched to 1.2-1.3 times of the slack. The muscle preparations were then immersed into the normal external solution (150 mM NaCl, 4 mM KCl, 2 mM Ca- methanesulphonate (Ca-Ms), 2 mM Mg-Ms, 5.6 mM glucose, and 5 mM N-2- hydroxyethylρiperazine-N'-2-ethanesulfonic acid) in a well on a bubble plate to allow for a rapid solution exchange. Force records were amplified with a bridge-amplifier, which was connected to a data acquisition system (PowerLab; ADInstruments, Colorado Springs, CO) and displayed in a computer. Depolarizing external solution had K-Ms substituted equally for NaCl with other chemicals in the same concentrations. The solutions were neutralized by Tris to a pH of 7.4. Compounds used in the recoding were applied at the following final concentrations: 10 μM atropine, 2 μg/ml of alpha-BTx, 100 μM ATP, 124 mM K⁺, 30 μM histamine, and 30 μM carbachol. All experiments were carried out at 30°C.

Esophagi of MSulfl-/-;MSulf2-/- mice have normal skeletal muscle function, but impaired smooth muscle contractility.

[0052] The esophageal defects in MSuIf l-/-;MSulf2-/- mice may result from impaired muscle contractility due to defects in muscle maturation or in neuronal innervation. To distinguish between these possibilities, we first characterized the development and maturation of the outer esophageal skeletal muscle layer which expresses MStilfl (Figure IB). Immunohistological studies showed that the expression of myogenic genes, including myosin heavy chain (MHC), Myf5, MyoD or Myogenin, is unchanged in skeletal muscle at P 14, when MSuIj I -/-; MSul/2-/- pups start to die (Figure 3A; data not shown). Also, MSulfl-/-;MSulj2-/- esophagi normally complete skeletal muscle maturation at P15 (Patapoutian et al., 1995; Rishniw et al., 2003), as shown by the expression of fast skeletal myosin in the abdominal segment of the esophagus connecting to the stomach (Figure 3A). In addition, MSulfl-/-;MSulf2-/- mice that survive into adulthood have an apparently mature esophageal skeletal muscle layer that expresses fast skeletal muscle myosin and nicotinic acetylcholine receptor clusters indistinguishable from the MSuIf 1 +/-;MSulf2+/- controls (Figure 3 A; data not shown). Furthermore, isolated thoracic longitudinal segments of the adult MSulfl-/-;MSulf2-/- esophagi produced a comparable contractile force in response to twitch and tetanus electrical stimuli when compared with those of the MSuIf 1 +/-;MSulf2+/- controls (Figure 3B). Alpha-bungarotoxin (alpha-BTx, an inhibitor of the nicotinic acetylcholine receptor on skeletal muscle) partially but significantly inhibited both twitch and tetanus forces in both MSuIf 1 -/-; MSuIfI-/ - and control esophagi in a similar manner. The remaining forces after the treatment with alpha-BTx were resistant to atropine (an inhibitor of cholinergic receptors on smooth muscle) (Figure 3B), but were completely inhibited by further addition of tetrodotoxin in MSulfl-/-;MSulj2-/- and control esophagi (data not shown). These results demonstrate that the electrical stimuli-induced forces are produced largely by skeletal muscle in the esophagus and that the skeletal muscle contractility is not significantly different in the double mutant and control esophagi. Therefore, the development, maturation, and function of esophageal skeletal muscle is unaffected in MSuIj l-/-;MSulf2-/- mice.

[0053] By contrast, smooth muscle contractility is impaired in

MSuIf l-/-;MSulf2-/- esophagi. The contractile forces of the isolated smooth muscle ring at the thoracic segments of the esophagi were elicited by various stimuli such as high K⁺, carbachol, ATP and histamine (Woodsome et al., 2001). Compared to MSuIfI+/- ;MSulf2+/- controls, the MSulfl-/-;MSulf2-/- esophageal smooth muscles showed a greatly diminished response to carbachol, an agonist of the muscarinic receptor, but only partially reduced response to other stimuli (shown in Figure 3C; data quantified in Figure 3D). Similar decrease in carbachol-induced smooth muscle contractility was observed in MSuIj 7 V-; MSul/2-/- esophagi at P15 (data not shown). No smooth muscle contractility defects were observed in the abdominal segments of the MSuIf I -/-;MSulf2-/- esophagi (data not shown). Therefore, MSuIf 1 -/-; MSulf2-/- mice have an esophageal muscle defect similar to those observed in the motility disorder congenital idiopathic megaesophagus, but not in achalasia (Neuhuber et al., 2006). Carbachol-induced smooth muscle contractility correlates with the level of neuronal innervation in the esophagus, as shown in studies of Sprouty2 mutant mice (Taketomi et al., 2005). The selective impairment in carbachol-induced smooth muscle contractility of the MSuIJ l-/-;MSulf2-/- esophagi, therefore, indicates specific neuronal innervation defects rather than a general disruption of the smooth muscle structure.

Neuronal innervation and enteric glial cell numbers are diminished in the esophagus of MSulfl-/-;MSulf2-/- mice.

[0054] To test whether esophageal innervation is defective in MSuIf l-/-;MSulf2-/- mice, we performed immunohisto logical studies using the TuJl antibody to identify innervating nerve fibers of both intrinsic and extrinsic neurons. At E 14.5, when the esophageal inner smooth muscle is not innervated, the TuJ 1 staining was indistinguishable between MSuIfI+/- ;MSulβ+/- control and MSulfl-/-;MSulf2-/- esophagi (Figure 4A). In addition, no difference in the expression of GDNF, GDNF receptor α-1 (GFRα-1) c-Ret, or the low affinity neurotrophin receptor p75 was observed between control esophagi and MSulfl-/-;MSulβ-/- esophagi (Figure 4A, data not shown). However, at El 8.5, compared with control esophagi, MSuIf 1-/-; MSul/2-/- esophagi have much reduced levels of axon sprouting and innervation density at the smooth muscle layer though the number of intrinsic neurons and the size of the esophagi are unchanged (Figure 4B). We also used the antibody against p75 to label the endogenous neurons and their processes in the esophagus. We found no difference in p75 expression on the cell body or in the number of p75-expressing intrinsic neurons that are located between the longitudinal and circular layers of skeletal muscle between control and MSuIj ^rl-/-;MSulf2-/- esophagi (Figure 4B). However, p75 expression on axons innervating the smooth muscle layer is decreased -in MSuIj ^~l-/-;MSulf2-/- esophagi, providing additional evidence that smooth muscle innervation was defective. The reduced smooth muscle innervation in MSuIf I-/-; MSulβ-/- esophagi persists postnatally and in the adult (Figures 4C, 4D), which directly affects smooth muscle contractility (Figure 3C and 3D) and leads to gradual enlargement of the esophageal lumen from an -20% increase in the length of the smooth muscle at P15 to almost double the length in adult cross sections (Figure 2C).

[0055] " Intrinsic neurons and enteric glial cells are both derived from neural crest progenitors. Since neural innervation is defective in MSuIf 1-/-; MSulβ-/- esophagi, it is possible that enteric glial cells are also affected in these mutant esophagi. To test this hypothesis, we characterized the development of enteric glial cells in the esophagus using antibodies against the glial cell markers, glial fibrillary acidic protein (GFAP) and SlOO at P15 and in the adult when enteric glial cells form and mature (Figure 4C; data not shown). We found that the number of GFAP-expressing cells along the thoracic segments of the MSuIf 1-/-; MSulβ-/- esophagi are reduced to one-third and one half of that in the control esophagi at Pl 5 and in adult, respectively (Figures 4C, 4D). Esophageal innervation and formation of enteric glial cells from neural crest progenitors are controlled by GDNF. Therefore, the observed esophageal defects in MSuIf 1-/-; MSulβ '-/- mice suggest a role for MSulfs in the GDNF signaling pathway during the establishment of the esophageal innervation that is required for postnatal feeding and growth.

[0056] It was also shown that MSulfs are required for the formation of enteric glial cells in the esophagus. Although previous studies, mostly using neural crest cultures, have established that GDNF controls the proliferation and differentiation of enteric glial progenitors from neural crest progenitors, developmental mechanisms underlying the enteric glial formation in the esophagus remain largely unknown. The specification of the enteric glial progenitors in the esophagus likely happens before or around E12.5 (Chalazonitis et al., 1998). In our study, we also observed a small number of p75+TuJl- cells in the esophagus at E12.5, but not at E14.5. These cells may represent neural crest progenitors that have committed to an enteric glial fate. MSuIf in the esophagus is barely detectable before E12.5 and after El 8.5, a time when differentiated GFAP-expressing enteric glial cells start to form. Therefore, MSulfs regulate either the formation or proliferation of enteric glial progenitors. Discerning which will require identification of specific molecular markers to trace the origins of these progenitors. GDNF binding to heparin and GFRaI

[0057] The purification of MSulfl and the MSulfl digestion of heparin or heparin-conjugated agarose beads (Sigma) were performed as described previously (Wang et al., 2004). To allow GDNF binding to heparin, 20 μl of digested heparin- conjugated agarose beads were incubated with various amounts of GDNF in PBS (a total volume of 100 μl) at room temperature for 1 hr. The heparin-agarose beads were spun down and washed 2 times with PBS. The amount of GDNF bound to the heparin- conjugated agarose beads was assayed by immunoblot analysis. To allow GDNF binding to GFRαl, 10 ng GDNF, 1 μg GFRαl-Fc (the extracellular domain of GFRαl) and heparin were mixed in 50 μl PBS for 30 min at room temperature. The GDNF-heparin- GFRαl-Fc complex was purified with 10 μl protein A agarose beads. The amount of GDNF bound to GFRαl was assayed by immunoblot analysis.

MSulfl and MSuIf2 regulate GDNF signaling for esophageal function

[0058] MSulfs are hypothesized to have direct and distinct roles in GDNF signaling regulation, based on our finding that MSulfs are differentially expressed by embryonic muscle and neural progenitors in the esophagus (Figure 1). MSulfl and GDNF are co-expressed by the esophageal muscle progenitors and their expression levels are co-regulated to peak between E14 and E16 (Hellmich et al., 1996; Golden et al., 1999), suggesting their related functions. To test whether MSulfl regulates GDNF signaling, we first examined whether MSulfl activity modulates GDNF binding to HS. We compared GDNF binding to heparin-agarose beads that were pre-digested with either purified MSulfl or enzymatically inactive QSuIfI(C-A) (Figure 5A) (Ai et al., 2003). GDNF binds to heparin with very high affinity (Rickard et al., 2003), comparable to the level of Fibroblast Growth Factor 2 (FGF2) binding to heparin (data not shown). MSulfl activity significantly decreased GDNF binding to heparin by up to 6 fold at sub- saturating amounts of GDNF, and by -30% under conditions of excess GDNF (Figure 5A), demonstrating that MSulfl reduces GDNF binding to heparin. Secondly, we tested whether MSulfl activity affects HS-regulated GDNF binding to the GFRαl . GDNF and GFRαl-Fc were incubated with various amounts of heparin pre-digested by MSulfl or enzymatically inactive QSuIfI(C-A) to allow the formation of GDNF-heparin-GFRαl-Fc ternary complex. GDNF bound to GFRαl-Fc was precipitated with protein A beads and then analyzed by Western blotting assay. As reported previously, heparin enhanced GDNF binding to GFRαl by 2 fold (Figure 5B) (Rickard et al., 2003). However, MSulfl desulfation of heparin did not affect heparin activity in GDNF binding to GFRαl (Figure 5B), indicating that MSulfl regulated HS 6-O-desulfation does not control GDNF binding to GFRαl. These findings suggest that MSulfl functions in the embryonic esophagus to control the affinity of GDNF to HS and thus modulate GDNF release from the extracellular matrix to target GDNF-responding innervating neurons.

[0059] MSuIf modulation of GDNF signaling and the relatively late onset of MSuIf expression are consistent with their essential roles in establishing functional neuron-target muscle interactions rather than in regulating earlier GDNF-dependent events, such as the migration and proliferation of the neural crest progenitors and neural progenitors (Durbec et al., 1996; Chalazonitis et al., 1998). MSulfs are not obligatory components in the GDNF signaling pathway because neuroblastoma cells, which do not express MSulfs, are capable of transmitting the GDNF signal. In addition, high concentrations of GDNF rescued the neurite outgrowth defect of MSuIf mutant esophagi in our explant assays, demonstrating that MSw/^deficient neurons are competent to receive GDNF signaling. However, MSulfs are essential developmental regulators of GDNF signaling during esophageal innervation. Loss of MSuIf function, which leads to a significant decrease in GDNF signaling activity, dramatically impacts on the esophageal innvervation that is required for normal feeding, growth and postnatal survival. The regulatory roles of MSulfs in GDNF signaling are in contrast to the obligatory roles of GDNF or GDNF receptors, as demonstrated by the differences in neural survival between MSulfmutant mice and mice deficient in GDNF or GDNF receptors. Our observation that MSuIf deficiency leads to reduced esophageal innervation without affecting the number of intrinsic neurons also suggests that MSulfs control GDNF signaling levels for differential aspects of the GDNF-dependent enteric neuronal development. Based on our results, the smooth muscle innervation by intrinsic neurons whose cell bodies are located within esophageal skeletal muscle requires higher levels of GDNF signaling activity than the survival of these neurons. GDNF expression in the lower gastrointestinal tract is 3 to 5 fold higher than in the esophagus in both embryos and in the adult (data not shown; Peters et al., 1998), which likely explains why the innervation in the gut was much less affected than the esophageal innervation in MSuIf double mutant mice (data not shown), although functional studies need to be conducted to test the intestinal contractility.

[0060] The current study has revealed a novel regulatory function for Sulfs and 6- O-sulfated HS sequences in the GDNF signaling pathway. The finding that Sulfl 6-O- desulfation of heparin reduces GDNF binding to heparin is consistent with previous results using chemically desulfated heparin that the GDNF-heparin interaction is dependent on heparin 2-O- and 6-O-sulfation, but not on N-sulfation (Davies et al., 2003; Rickard et al., 2003). It was found that MSuIf activity has no effect on heparin-enhanced GDNF binding to GFRαl in vitro, indicating that the 6-O sulfates of HS are not required for GDNF-heparin-GDNF receptor ternary complex formation, which is different from the pivotal role of HS 6-O-sulfation in ternary complex formation in Fibroblast Growth Factor (FGF) signaling (Schlessinger et al., 2000). MSuIf promotes GDNF signaling likely by reducing GDNF-HS binding to mobilize GDNF ligand and promote GDNF interaction with its receptors. A similar mechanism has been proposed for SuIf regulation of Wnt signaling (Dhoot et al., 2001; Ai et al., 2003).

In situ hybridization

[0061] In situ hybridization was performed as described previously (Dhoot et al., 2001). DNA fragments for making the digoxigenin-labeled RNA probes were cloned into pTeasy vector (Promega) by PCR using primers at the 3' untranslated sequences of MSuIf genes. For the MSuIf 1 probe, the forward primer used was, 5' ATCAGCTCACCAGTCAGCAC 3', and the reverse primer was 5' CTAGCCATTTTGACTGGATAGA 3'. For the MSulfl probe, forward primer used was 5' TGTC AACC AC AC AGTCTTG A 3' and the reverse primer was 5' ACCAATAGCTAGACATTGGC 3'. Tissue culture

[0062] The NG 108-15 neuroblastoma cell line and PC 12 cells (American Type Culture Collection) were cultured in DMEM (Mediatech) with 10% fetal bovine serum (Mediatech) and 1% antibiotics (Gibco). Cells were transfected with pAG expression vectors containing MSuIf 1, Msulf, or QSuIfI(C-A) using Fugene-6 (Roche). Stably transfected cell lines were established by hygromycin selection (Sigma, 250 μg/ml). To activate signaling, NG 108- 15 cells or PC 12 cells (I X 10⁵/24-well) were serum-starved in DMEM for 6 hr before adding GDNF or NGF (R&D Systems), respectively, at various concentrations. After stimulation, cells were rinsed once with PBS before being lysed with 100 μl of IX Laemmli sample buffer (Sigma) for Western blot, or with RIPA buffer for co-immunoprecipitation with anti-Ret antibody.

Immunoblot analysis

[0063] Samples were separated on 12% SDS-PAGE gel followed by Western blot analysis (Dhoot et al., 2001). The intensity of the signal was quantified by Multi-analysis software (Bio-Rad). The primary antibodies for immunoblots include: rabbit anti-MAPK (1:4000, Sigma), mouse anti-phosphorylated MAPK (1:1000, Sigma), rabbit anti- phosphorylated PKB/ AKT (1 :1000, Cell Signaling), rabbit anti-GDNF (0.5 μg/ml), mouse anti-pTyr (Calbiochem, 1 :1000), rabbit anti-Ret (1 :500). The secondary antibodies include: HRP-conjugated goat anti-rabbit (1 :4000; Santa Cruz Biotechnology, Inc.) and HRP-conjugated goat anti-mouse (1:2000, Santa Cruz Biotechnology, Inc.).

[0064] MSulf2 is expressed by esophageal muscle-innervating neurons and thus may regulate the response of these neurons to GDNF. To test whether MSulf2 regulates signaling activity of GDNF in neurons, we established stably transfected neuroblastoma NG108-15 cell lines that expressed comparable levels of MSulf2 or inactive QSuIfI(C-A) (data not shown), or an empty expression vector. NG 108- 15 cells do not express detectable levels of endogenous MSulfs by immunostaining (data not shown). In response to GDNF, NG108-15 cells showed dose-dependent activation of the signaling pathway by phosphorylating the c-Ret receptor and the downstream kinases, Protein Kinase B (PKB/ AKT) and Extracellular signal-Regulated Kinases (ERKs), 3-10 fold above uninduced, basal levels (Figure 5C, 5D, 5G) (Lee et al., 2006). Expression of MSulf2, but not inactive QSuIfI(C-A) or empty vector, further increased the phosphorylation levels of c-Ret and PKB/AKT by 2-3 fold and ERKs by up to 2 fold in response to GDNF (Figures 5C, 5D, 5G(a)). Additionally, MSulΩ-expressing NG 108- 15 cells showed sustained activation of downstream PKB/Akt and Erks over 30 min after GDNF stimulation, compared to control cells in which GDNF signaling activity rapidly declined 15 min after stimulation (Figures 5E, 5G(b)). In contrast, MSulf2 had no effect on NGF-induced phosphorylation of AKT or Erk in PC 12 cells (Figure 5F), demonstrating the functional specificity of MSulf2 in the GDNF signaling pathway. This result is also consistent with previous findings that NGF signaling is independent of HS-sulfation (Barnett et al., 2002). The studies of neural cell lines provide evidence that MSulf2 enhances GDNF signaling in neural progenitors during esophageal innervation.

Esophagus explant cultures

[0065] The esophagus was dissected from El 1.5 mouse embryos and cultured in 500μl growth medium (DMEM (Mediatech) + 10% fetal bovine serum) on pre-solidified collagen gel in a 24-well plate. The collagen gel was made by mixing the acidic collagen (Purecol™, InaMed, 3mg/ml), 1OX DMEM and 0.1N NaOH to restore normal osmolality and pH according to manufacturer's protocol. GDNF, NGF, BDNF, NT3 or NT4 (R&D Systems) were added to the collagen/DMEM mixture immediately at various concentrations. The collagen/DMEM mixture was then aliquoted at 250 μl per well and allowed to solidify at 37⁰C for 1 hour. Esophagus explants were cultured on collagen gel for 4-5 days before fixation with 4% paraformaldehyde in PBS followed by immunohistochemistry as described above.

MSulfl and MSulf2 are required for GDNF-dependent neurite outgrowth in embryonic esophageal explant cultures

[0066] To further investigate whether MSulfl and MSulf2 are required for GDNF-induced neurite outgrowth of the endogenous neurons in the embryonic esophagus, we performed explant cultures by isolating the esophagi from El 1.5 embryos and culturing them on collagen gels in growth medium (Yan et al., 2004). After 4-5 days in culture, the explants adhered to the collagen gel and matured to spontaneously contract (Figure 6A). In the absence of GDNF, the esophagus explants grew without extending neurites onto the collagen gel, as shown by the lack of neurofilament immunoreactivity extending from the explants (Figure 6A). The neurite outgrowth from the explants was induced dose-dependently by the presence of GDNF in the collagen gel, but not by NGF, BDNF, NT3 or NT4 (Figure 6A, 6B, 6C, data not shown). This observation established that GDNF is selectively required for neurite outgrowth of intrinsic neurons in embryonic esophagi. The neurite outgrowth from control esophageal explants was induced by GDNF at 10 ng/ml, the lowest concentration tested, and the maximum induction was achieved at 50 ng/ml GDNF and higher (Figure 6A, 6C). In contrast, the MSulβ-/-;MSulβ-/- explants failed to extend neurites at 10 ng/ml GDNF, and the neurite outgrowth was reduced to about one third of the control level at 20 ng/ml GDNF. The defect of the neurite outgrowth of the MSuIj 1 -/-; MSul/2-/- esophageal explants was completely rescued by 100 ng/ml GDNF (Figure 6 A, 6C). Similar neurite outgrowth defects were also observed with the MSuIf I +/-;MSulf2-/- esophageal explants, but not with the MSuIf l-/-;MSulf2+/- explants (data not shown), consistent with the distinct MSulf2 expression by endogenous neurons. The lack of neurite outgrowth by MSuIf 1 -/-;MSulf2-/- esophageal explants at 10 ng/ml GDNF could be due to a disrupted GDNF signaling in intrinsic neurons, or a reduction in the number of intrinsic neurons. To distinguish between these two possibilities, we quantified the total number of TuJl -expressing intrinsic neurons within the explants after 4 days in culture. We found that GDNF is essential for the survival/growth of intrinsic neurons because control explants cultured without any neurotrophic factors or in the presence of NGF had only one third of the neurons compared with those treated with 10 ng/ml GDNF (Figure 6E). MSulfl-/-;MSulf2-/- esophageal explants, although they showed no significant neurite outgrowth in the presence of 10 ng/ml GDNF, had the same number of intrinsic neurons as the control explants (Figure 6E), consistent with the in vivo phenotype. In addition, neurons within the GDNF-treated control explants appeared scattered and a few neurons even migrated out of the explants (Figure 6D, data not shown). In contrast, neurons within MSulfl-/-;MSulf2-/- explants formed large clusters, suggesting a defect in GDNF- induced neural migration. Therefore, embryonic MSuIf 1 -/-; MSul/2-/- esophagi have defects in GDNF-dependent neural migration and neurite outgrowth for muscle innervation, but not in GDNF-dependent neuronal survival/growth.

[0067] Provided herein are several lines of evidence that regulation of esophageal innervation is directly controlled by selective MSuIf regulation of GDNF signaling. First, the temporal and spatial expression of GDNF and MSulfs is co-regulated in embryonic esophagi. MSulfl is co-expressed with GDNF in esophageal muscle and MSulf2 is expressed by innervating neural progenitors. Second, MSulfl decreases GDNF binding to heparan sulfate and potentially releases GDNF stored in the extracellular matrix of the target muscle for the innervating intrinsic neurons. Third, MSulf2 promotes the signaling activities of GDNF, but not NGF, in neuronal cell lines, demonstrating the functional selectivity of MSulfs in neurotrophin signaling. Fourth, neurite outgrowth by intrinsic neurons to innervate esophageal muscle is dependent on GDNF, but not on other neurotrophins, as shown by our explant assays and by previous genetic and culture studies (Baloh et al., 2000; Yan et al, 2004). MSΗ/^defϊciency leads to defective neurite outgrowth in the esophagus both in vivo and in explant cultures, establishing MSuIf regulation of GDNF pathway.

[0068] These neurite outgrowth studies, together with heparin binding and neural GDNF signaling studies, demonstrate that MSulfl and MSulf2 have dual, cooperative functions in GDNF signaling for establishing target-dependent esophageal innervation. MSulfl , which is dynamically expressed in esophageal muscle progenitors, mobilizes HS-bound GDNF from the extracellular matrix, while MSulf2, which is expressed by innervating neuronal progenitors, promotes reception of the GDNF signal to extend neurites to innervate esophageal muscle required for esophageal function after birth.

[0069] This study of SuIf regulation of GDNF signaling for esophageal innervation has provided the first in vivo evidence that HS plays essential roles in neurotrophic factor signaling during the development of the nervous system. Notably, the esophageal phenotype of M->u//~double mutant mice reflects an early embryonic signaling defect that is not revealed until birth and the onset of feeding behavior. Additionally, identification of a regulatory role for SuIf and HS 6-O-sulfation in GDNF signaling may provide a basis for the development of new, HS-based therapeutic approaches to improve GDNF-mediated neuroprotection in treating Parkinson's Disease and brain injury (Airaksinen and Saarma, 2002). Finally, in addition to the reported defects in GDNF signaling for esophageal development and function, MSuIf mutant mice exhibit defects in development and the regenerative capacity of stem cells in multiple tissues. Investigations are ongoing using MSuIf mutant mice to further understand the regulatory roles of SuIf genes in matrix signaling during development and disease (Hacker et al., 2005; Holt and Dickson, 2005).

Claims

What is claimed is:

1) A method for treating a pathology characterized by neuronal innervation defects, in a patient, the method comprising: a) providing a solution comprising a SuIf protein in a physiologically compatible buffer; b) administering the solution to the patient in an amount, and for a period of time, sufficient to substantially increase GDNF signaling in neuronal cells; the administration of the SuIf protein solution of step a) having no substantial effect on NGF-signaling in the patient.

2) The method of Claim 1 wherein the pathology is selected from the group consisting of Parkinson's Disease, brain injury, and an esophageal disorder.

3) The method of Claim 2 wherein the pathology is brain injury caused by stroke.

4) The method of Claim 1 wherein the SuIf protein is SuIf 1.

5) The method of Claim 1 wherein the SuIf protein is SuIf 2.

6) The method of Claim 1 wherein GDNF signaling is increased at least 2-fold.

7) The method of Claim 1 wherein the increase in GDNF signaling is assayed by monitoring phosphorylation of a kinase selected from the group consisting of PKB/AKT and ERKs, an increase in phosphorylation of the selected kinase following administration of the solution comprising the SuIf protein, relative to phosphorylation levels of the selected kinase in an otherwise identical assay in which the solution comprising the SuIf protein is not administered, being indicative of an increase in GDNF signaling. ) The method of Claim 1 wherein the increase in GDNF signaling is assayed by monitoring the formation of glial cells in the patient, an increase in the rate of glial cell formation following administration of the solution comprising the SuIf protein, relative to the rate of glial cell formation in an otherwise identical assay in which the solution comprising the SuIf protein is not administered, being indicative of an increase in GDNF signaling.

9) The method of claim 1 wherein the solution of step a) further comprises a heparan sulfate compound.

10) The method of claim 9 wherein the heparan sulfate compound is heparin.

1 l)The method of claim 1 wherein the solution of step a) further comprises GDNF.

12) A pharmaceutical solution comprising a SuIf protein and a heparan sulfate compound in a physiologically compatible buffer.

13) The solution of claim 12 wherein the heparan sulfate compound is heparin.

14) A pharmaceutical solution comprising a SuIf protein and GDNF in a physiologically compatible buffer.