A compartmentalized phosphoinositide signaling axis at cilia is regulated by INPP5E to maintain cilia and promote Sonic Hedgehog medulloblastoma.

Sonic Hedgehog (SHH) signaling at primary cilia drives the proliferation and progression of a subset of medulloblastomas, the most common malignant paediatric brain tumor. Severe side effects associated with conventional treatments and resistance to targeted therapies has led to the need for new strategies. SHH signaling is dependent on primary cilia for signal transduction suggesting the potential for cilia destabilizing mechanisms as a therapeutic target. INPP5E is an inositol polyphosphate 5-phosphatase that hydrolyses PtdIns(4,5)P2 and more potently, the phosphoinositide (PI) 3-kinase product PtdIns(3,4,5)P3. INPP5E promotes SHH signaling during embryonic development via PtdIns(4,5)P2 hydrolysis at cilia, that in turn regulates the cilia recruitment of the SHH suppressor GPR161. However, the role INPP5E plays in cancer is unknown and the contribution of PI3-kinase signaling to cilia function is little characterized. Here, we reveal INPP5E promotes SHH signaling in SHH medulloblastoma by negatively regulating a cilia-compartmentalized PI3-kinase signaling axis that maintains primary cilia on tumor cells. Conditional deletion of Inpp5e in a murine model of constitutively active Smoothened-driven medulloblastoma slowed tumor progression, suppressed cell proliferation, reduced SHH signaling and promoted tumor cell cilia loss. PtdIns(3,4,5)P3, its effector pAKT and the target pGSK3ß, which when non-phosphorylated promotes cilia assembly/stability, localized to tumor cell cilia. The number of PtdIns(3,4,5)P3/pAKT/pGSK3ß-positive cilia was increased in cultured Inpp5e-null tumor cells relative to controls. PI3-kinase inhibition or expression of wild-type, but not catalytically inactive HA-INPP5E partially rescued cilia loss in Inpp5e-null tumor cells in vitro. INPP5E mRNA and copy number were reduced in human SHH medulloblastoma compared to other molecular subtypes and consistent with the murine model, reduced INPP5E was associated with improved overall survival. Therefore our study identifies a compartmentalized PtdIns(3,4,5)P3/AKT/GSK3ß signaling axis at cilia in SHH-dependent medulloblastoma that is regulated by INPP5E to maintain tumor cell cilia, promote SHH signaling and thereby medulloblastoma progression.

Direct effects of Bmi1 on p53 protein stability inactivates oncoprotein stress responses in embryonal cancer precursor cells at tumor initiation.

Embryonal cancer can arise from postnatally persistent embryonal remnant or rest cells, which are uniquely characterized by the absence of p53 mutations. Perinatal overexpression of the MycN oncoprotein in embryonal cancer precursor cells causes postnatal rests, and later tumor formation through unknown mechanisms. However, overexpression of Myc in adult tissues normally activates apoptosis and/or senescence signals as an organismal defense mechanism against cancer. Here, we show that perinatal neuroblastoma precursor cells exhibited a transiently diminished p53 response to MycN oncoprotein stress and resistance to trophic factor withdrawal, compared with their adult counterpart cells from the TH-MYCN(+/+) transgenic mouse model of neuroblastoma. The adult stem cell maintenance factor and Polycomb group protein, Bmi1 (B-cell-specific Moloney murine leukemia virus integration site), had a critical role at neuroblastoma initiation in the model, by repressing p53 responses in precursor cells. We further show in neuroblastoma tumor cells that Bmi1 could directly bind p53 in a complex with other Polycomb complex proteins, Ring1A or Ring1B, leading to increased p53 ubiquitination and degradation. Repressed p53 signal responses were also seen in precursor cells for other embryonal cancer types, medulloblastoma and acute lymphoblastic leukemia. Collectively, these date indicate a general mechanism for p53 inactivation in some embryonal cell types and consequent susceptibility to MycN oncogenesis at the point of embryonal tumor initiation.

Canonical Notch signaling is not required for the growth of Hedgehog pathway-induced medulloblastoma.

Current treatment for medulloblastoma is successful in more than half of all cases but results in substantial disability in survivors. Accordingly, there is considerable interest in drugs that may target specific signaling pathways activated in the tumors, with inhibitors of both the Hedgehog and Notch pathways currently proposed as possible therapeutics. Here, we tested the hypothesis that Notch pathway inhibition in vivo may block the formation of Hedgehog-dependent medulloblastoma. We took the general approach of using a cre recombinase under the control of the GFAP promoter to generate medulloblastoma in mice carrying a conditional Ptc1 allele and introduced a conditional RBP-J allele to ablate canonical Notch signaling. Loss of RBP-J from the developing cerebellum led to a modest loss of stem cells and an overall developmental delay. These phenotypes could be partially compensated by activation of the Hedgehog pathway. Hedgehog-dependent medulloblastoma were not blocked by loss of RBP-J, indicating that canonical Notch signaling is not required for tumor initiation and growth in this model.

Sonic Hedgehog regulates Hes1 through a novel mechanism that is independent of canonical Notch pathway signalling.

Aberrant regulation of signalling mechanisms that normally orchestrate embryonic development, such as the Hedgehog, Wnt and Notch pathways, is a common feature of tumorigenesis. In order to better understand the neoplastic events mediated by Hedgehog signalling, we identified over 200 genes regulated by Sonic Hedgehog in multipotent mesodermal cells. Widespread crosstalk with other developmental signalling pathways is evident, suggesting a complex network of interactions that challenges the often over-simplistic representation of these pathways as simple linear entities. Hes1, a principal effector of the Notch pathway, was found to be a target of Sonic Hedgehog in both C3H/10T1/2 mesodermal and MNS70 neural cells. Desert Hedgehog also elicited a strong Hes1 response. While Smoothened function was found necessary for upregulation of Hes1 in response to Sonic Hedgehog, the mechanism does not require gamma-secretase-mediated cleavage of Notch receptors, and appears to involve transcription factors other than RBP-Jkappa. Thus, we have defined a novel mechanism for Hes1 regulation in stem-like cells that is independent of canonical Notch signalling.

GFP-tagged CFTR transgene is functional in the G551D cystic fibrosis mouse colon.

Trafficking of the cystic fibrosis transmembrane conductance regulator (CFTR) is central to its function, with the most common mutation, deltaF508, resulting in abnormal processing and trafficking. Therefore, there is a significant need to develop tools, which enable the trafficking of CFTR to be studied in vitro and in vivo. In previous studies it has been demonstrated that fusion of the green fluorescent protein (GFP) to the N-terminus of CFTR does lead to functional expression of CFTR chloride channels in epithelial cell lines. The aim of the present study was to examine whether it is possible to express GFP-tagged CFTR as a transgene in colonic and airway epithelial cells of cystic fibrosis (CF) mice and to correct the CF defect. Using the epithelial-specific human cytokeratin promoter K18, we generated bitransgenic mice cftr(G551D/G551) K18-GFP-CFTR(+/-), designated GFP mice. Transcripts for GFP-CFTR could be detected in bitransgenic mice by use of RT-PCR techniques. Expression of GFP-CFTR protein was detected specifically in the colonic epithelium by both direct GFP fluorescence and the use of an anti-GFP antibody. Ussing chamber studies showed that the ion transport defect in colon and airways observed in cftr(G551D/G551D) mice was partially corrected in the bitransgenic animals. Thus, K18-GFP-CFTR is functionally expressed in transgenic mice, which will be a valuable tool in studies on CFTR synthesis, processing and ion transport in native epithelial tissues.

G551D CF mice display an abnormal host response and have impaired clearance of Pseudomonas lung disease.

Several cystic fibrosis (CF) mouse models demonstrate an increased susceptibility to Pseudomonas aeruginosa lung infection, characterized by excessive inflammation and high rates of mortality. Here we developed a model of chronic P. aeruginosa lung disease in mice homozygous for the murine CF transmembrane conductance regulator G551D mutation that provides an excellent model for CF lung disease. After 3 days of infection with mucoid P. aeruginosa entrapped in agar beads, the G551D animals lost substantially more body weight than non-CF control animals and were less able to control the infection, harboring over 40-fold more bacteria in the lung. The airways of infected G551D animals contained altered concentrations of the inflammatory mediators tumor necrosis factor-alpha, KC/N51, and macrophage inflammatory protein-2 during the first 2 days of infection, suggesting that an ineffective inflammatory response is partly responsible for the clearance defect.

Murine CFTR channel and its role in regulatory volume decrease of small intestine crypts.

Cystic fibrosis (CF) is caused by mutations in the secretory Cl(-) channel CFTR (cystic fibrosis transmembrane conductance regulator). Variation in the severity of disease has been attributed to mutations in the CFTR gene that cause different degrees of dysfunction of the CFTR Cl(-) channel. However, studies of mouse models of CF indicate that the severity of intestinal pathology is not correlated with activity of the CFTR chloride channel. This observation suggests that other 'environmental' factors might be important in determining the severity of disease. In this respect, we have identified and characterised an additional cellular defect in intestinal epithelial cells of CF mice, the inability of these cells to regulate their volume after hypotonic challenge. Here, we review the function of murine CFTR as both a Cl(-) channel and as a regulator of volume-dependent homeostatic cell mechanisms.

G551D cystic fibrosis mice exhibit abnormal regulation of inflammation in lungs and macrophages.

The major cause of death in cystic fibrosis (CF) is chronic lung disease associated with persistent infection by the bacterium, Pseudomonas aeruginosa. S100A8, an S-100 calcium-binding protein with chemotactic activity, is constitutively expressed in the lungs and serum of CF patients. Levels of S100A8 mRNA were found to be three to four times higher in the lungs of mice carrying the G551D mutation in CF transmembrane conductance regulator compared with littermate controls. Intravenous injection of bacterial LPS induced S100A8 mRNA in the lung to a greater extent in G551D mice than in wild-type littermates. Localization of S100A8 mRNA and protein in the lung indicate that it is a marker for neutrophil accumulation. Bone marrow-derived macrophages from G551D mice were shown to also exhibit hypersensitivity to LPS, measured by induction of TNF-alpha. These results provide evidence that the pathology of CF relates to abnormal regulation of the immune system.

Isolation and characterization of human patched 2 (PTCH2), a putative tumour suppressor gene inbasal cell carcinoma and medulloblastoma on chromosome 1p32.

Mutations of the human Patched gene ( PTCH ) have been identified in individuals with the nevoid basal cell carcinoma syndrome (NBCCS) as well as in sporadic basal cell carcinomas and medulloblastomas. We have isolated a homologue of this tumour suppressor gene and localized it to the short arm of chromosome 1 (1p32.1-32.3). Patched 2 ( PTCH2 ) comprises 22 coding exons and spans approximately 15 kb of genomic DNA. The gene encodes a 1203 amino acid putative transmembrane protein which is highly homologous to the PTCH product. We have characterized the genomic structure of PTCH2 and have used single-stranded conformational polymorphism analysis to search for mutations in PTCH2 in NBCCS patients, basal cell carcinomas and in medulloblastomas. To date, we have identified one truncating mutation in a medulloblastoma and a change in a splice donor site in a basal cell carcinoma, suggesting that the gene plays a role in the development of some tumours.

Regulation of murine cystic fibrosis transmembrane conductance regulator Cl- channels expressed in Chinese hamster ovary cells.

1. We investigated the effect of protein kinases and phosphatases on murine cystic fibrosis transmembrane conductance regulator (CFTR) Cl- channels, expressed in Chinese hamster ovary (CHO) cells, using iodide efflux and the excised inside-out configuration of the patch-clamp technique. 2. The protein kinase C (PKC) activator, phorbol dibutyrate, enhanced cAMP-stimulated iodide efflux. However, PKC did not augment the single-channel activity of either human or murine CFTR Cl- channels that had previously been activated by protein kinase A. 3. Fluoride, a non-specific inhibitor of protein phosphatases, stimulated both human and murine CFTR Cl- channels. However, calyculin A, a potent inhibitor of protein phosphatases 1 and 2A, did not enhance cAMP-stimulated iodide efflux. 4. The alkaline phosphatase inhibitor, (-)-bromotetramisole augmented cAMP-stimulated iodide efflux and, by itself, stimulated a larger efflux than that evoked by cAMP agonists. However, (+)-bromotetramisole, the inactive enantiomer, had the same effect. For murine CFTR, neither enantiomer enhanced single-channel activity. In contrast, both enantiomers increased the open probability (Po) of human CFTR, suggesting that bromotetramisole may promote the opening of human CFTR. 5. As murine CFTR had a low Po and was refractory to stimulation by activators of human CFTR, we investigated whether murine CFTR may open to a subconductance state. When single-channel records were filtered at 50 Hz, a very small subconductance state of murine CFTR was observed that had a Po greater than that of human CFTR. The occupancy of this subconductance state may explain the differences in channel regulation observed between human and murine CFTR.

Comparison of the gating behaviour of human and murine cystic fibrosis transmembrane conductance regulator Cl- channels expressed in mammalian cells.

1. To investigate the function of the murine cystic fibrosis transmembrane conductance regulator (CFTR), a full-length cDNA encoding wild-type murine CFTR was assembled and stably expressed in Chinese hamster ovary (CHO) cells. 2. Like human CFTR, murine CFTR formed Cl- channels that were regulated by cAMP-dependent phosphorylation and intracellular ATP. However, murine CFTR Cl- channels had a reduced single-channel conductance and decreased open probability (Po) compared with those of human CFTR. 3. Analysis of the dwell time distributions of single channels suggested that the reduced Po of murine CFTR was caused by both decreased residence in the open state and transitions to a new closed state, described by an intermediate closed time constant. 4. For both human and murine CFTR, ATP and ADP regulated the rate of exit from the long-lived closed state. 5. 5'-Adenylylimidodiphosphate (AMP-PNP) and pyrophosphate, two compounds that disrupt cycles of ATP hydrolysis, stabilized the open state of human CFTR. However, neither agent locked murine CFTR Cl- channels open, although AMP-PNP increased the Po of murine CFTR. 6. The data indicate that although human and murine CFTR have many properties in common, some important differences in function are observed. These differences could be exploited in future studies to provide new understanding about CFTR.

Effect of IBMX and alkaline phosphatase inhibitors on Cl- secretion in G551D cystic fibrosis mutant mice.

Some cystic fibrosis transmembrane conductance regulator (CFTR) mutations, such as G551D, result in a correctly localized Cl- channel at the cell apical membrane, albeit with markedly reduced function. Patch-clamp studies have indicated that both phosphatase inhibitors and 3-isobutyl-1-methylxanthine (IBMX) can induce Cl- secretion through the G551D mutant protein. We have now assessed whether these agents can induce Cl- secretion in cftrG551D mutant mice. No induction of Cl- secretion was seen with the alkaline phosphatase inhibitors bromotetramisole or levamisole in either the respiratory or intestinal tracts of wild-type or cftrG551D mice. In contrast, in G551D intestinal tissues, IBMX was able to produce a small CFTR-related secretory response [means +/- SE: jejunum, 1.8 +/- 0.9 microA/cm2, n = 7; cecum, 3.7 +/- 0.8 microA/cm2, n = 7; rectum (in vivo), 1.9 +/- 0.9 mV, n = 5]. This was approximately one order of magnitude less than the wild-type response to this agent and, in the cecum, was significantly greater than that seen in null mice (cftrUNC). In the trachea, IBMX produced a transient Cl- secretory response (37.3 +/- 14.7 microA/cm2, n = 6) of a magnitude similar to that seen in wild-type mice (33.7 +/- 4.7 microA/cm2, n = 9). This response was also present in null mice and therefore is likely to be independent of CFTR. No effect of IBMX on Cl- secretion was seen in the nasal epithelium of cftrG551D mice. We conclude that IBMX is able to induce detectable levels of CFTR-related Cl- secretion in the intestinal tract but not the respiratory tract through the G551D mutant protein.

Characterisation of human patched germ line mutations in naevoid basal cell carcinoma syndrome.

Mutations in the human patched gene have recently been detected in patients with naevoid basal cell carcinoma syndrome. We have characterised a further 5 novel germ line mutations in patients presenting with multiple odontogenic keratocysts. Four mutations cause premature stop codons and one mutation results in an amino-acid substitution towards the carboxyl terminus of the predicted patched protein. No obvious genotype-phenotype correlations could be interpreted, consistent with previous studies.

Mapping the multiple self-healing squamous epithelioma (MSSE) gene and investigation of xeroderma pigmentosum group A (XPA) and PATCHED (PTCH) as candidate genes.

The MSSE gene predisposes to the development of multiple invasive but self-healing skin tumours (multiple self-healing squamous epitheliomata, MSSE). MSSE (previously named ESS1) was mapped to chromosome 9q by linkage analysis; haplotype analysis in families then suggested a common founder mutation and indicated that the gene lies in the interval D9S1-D9S29 (9q22-q31). Squamous cell carcinomata also develop as one of the complications of xeroderma pigmentosum, and one of the xeroderma pigmentosum genes (XPA) maps within the MSSE interval. We have investigated the hypothesis that a novel dominant mutation in XPA is responsible for MSSE. We screened the entire coding region, 3' untranslated region (UTR) and 5'UTR of XPA for germline mutations in MSSE families by single-stranded conformation polymorphism analysis and by direct DNA sequencing. No mutations were detected but a novel intragenic polymorphism was identified in the 5'UTR of XPA, in both MSSE-affected and unrelated normal individuals. This XPA polymorphism and nine new polymorphic markers that map in the MSSE region were typed in eleven MSSE families; XPA was excluded as the MSSE gene and the most likely location of MSSE was reduced to the interval between D9S197 and (D9S287, D9S1809). The Patched (PTCH) gene, which is mutated in naevoid basal cell carcinoma syndrome (NBCCS or Gorlin syndrome) lies in this interval and all MSSE families have been shown to share a common haplotype at three novel intragenic PTCH polymorphisms. Although no mutation has been detected in MSSE families, PTCH has not been excluded as the MSSE gene.

Targeted disruption of the Wnt2 gene results in placentation defects.

Wnt genes have been implicated in a range of developmental processes in the mouse including the patterning of the central nervous system and limbs. Reported here for the first time is the expression of Wnt2 in the early heart field of 7.5-8.5 dpc (days post-coitum) mouse embryos, making Wnt2 a potentially useful gene marker for the early stages of heart development. Expression was also detected in the allantois from 8.0 dpc and at later stages in the placenta and umbilicus. Mice deficient in Wnt2, generated by gene targeting, displayed runting and approximately 50% died perinatally. Histological analysis revealed alterations in the size and structure of placentas from these mice from 14.5 dpc. The placental defects were associated primarily with the labyrinthine zone and included oedema and tissue disruption and accumulation of maternal blood in large pools. There was also an apparent decrease in the number of foetal capillaries and an increase in the amount of fibrinoid material in the Wnt2 mutant placentas. These results suggest that Wnt2 is required for the proper vascularisation of the mouse placenta and the placental defects in Wnt2-deficient mice result in a reduction in birthweight and perinatal lethality.

A demonstration using mouse models that successful gene therapy for cystic fibrosis requires only partial gene correction.

Quantifying the level of transgene expression necessary for phenotypic effect is an important consideration in designing somatic gene therapy protocols. A nonlinear relationship between phenotype and gene activity is predicted by control analysis for any autosomal recessive condition. The unaffected phenotype of heterozygotes for autosomal recessive disorders demonstrates that 50% of the normal level of gene expression is sufficient to prevent disease. By extension, an exaggerated and positive effect on the mutant phenotype is predicted to arise from only a small addition of normal transgene expression delivered by gene therapy. We tested this expectation directly by intercrossing mice carrying different Cftr alleles which modulate Cftr gene expression from 0 to 100%. We demonstrate that 5% of the normal level of Cftr gene expression results in a disproportionately large correction of the chloride ion transport defect (50% of normal) and essentially complete rescue of the intestinal disease (100% survival). It follows that even modest levels of transgene expression and only partial correction of CFTR channel activity may have a significant clinical impact.

Murine WNT11 is a secreted glycoprotein that morphologically transforms mammary epithelial cells.

Wnt genes encode a set of structurally related cell surface glycoproteins that appear to have roles in cell-cell signalling. The ectopic expression of several murine Wnt genes has been implicated in the transformation of mammary epithelial and the onset of mammary tumours. Wnt11 is expressed in the developing embryo in a variety of structures including the dermatome/myotome junction of the somites, the truncus ateriosus region of the heart and limb mesenchyme. Here we report that Wnt11 encodes a glycoprotein that is secreted from expressing cells and becomes associated with the extracellular matrix. In addition, Rat2 fibroblasts expressing WNT11 (which are not morphologically altered themselves) are able to induce the transformation of adjacent C57MG mammary epithelial cells in co-culture experiments. These results suggest that WNT11 functions via a paracrine signalling mechanism to have a direct effect on the morphology and growth characteristics of mammary epithelial cells.

Cystic fibrosis mice carrying the missense mutation G551D replicate human genotype-phenotype correlations.

We have generated a mouse carrying the human G551D mutation in the cystic fibrosis transmembrane conductance regulator gene (CFTR) by a one-step gene targeting procedure. These mutant mice show cystic fibrosis pathology but have a reduced risk of fatal intestinal blockage compared with 'null' mutants, in keeping with the reduced incidence of meconium ileus in G551D patients. The G551D mutant mice show greatly reduced CFTR-related chloride transport, displaying activity intermediate between that of cftr(mlUNC) replacement ('null') and cftr(mlHGU) insertional (residual activity) mutants and equivalent to approximately 4% of wild-type CFTR activity. The long-term survival of these animals should provide an excellent model with which to study cystic fibrosis, and they illustrate the value of mouse models carrying relevant mutations for examining genotype-phenotype correlations.

+P5 (D1S3309E), a novel target binding site for the Wilms' tumour suppressor 1 (WT1) gene, maps to human chromosome 1q21-->22.

The Wilms' tumor suppressor 1 gene (WT1) encodes a zinc finger transcription factor critical for normal urogenital development. We have previously isolated a DNA fragment, +P5 (D1S3309E), to which all WT1 protein isoforms bind. Using PCR of a human x rodent somatic cell hybrid mapping panel, together with two-color fluorescence in situ hybridisation of +P5-containing cosmids and previously localised human chromosome 1q cosmids, we have mapped the +P5 fragment to chromosome 1q21-->q22.