| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
Departments of Genetics and Tumor Cell Biology [F. Z., L. M. N., L. N., C. M., C. J. S., M. F. R.], Developmental Neurobiology [R. J. S.], and Pathology [J. E. R.] and Howard Hughes Medical Institute [C. J. S.], St. Jude Children’s Research Hospital, Memphis, Tennessee 38105, and Department of Pathology, Johns Hopkins School of Medicine, Baltimore, Maryland 21205 [C. E.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
CKI proteins are differentially expressed during mouse embryonic development and in adult tissues. Whereas Ink4a and Ink4b are not normally expressed in embryos or in young animals, Ink4c and Ink4d are ubiquitously expressed in embryos and in many tissues of adult mice, usually in nonoverlapping patterns (9) . Both Ink4c and Ink4d are expressed in many regions of the developing central nervous system; Ink4c seems to be preferentially localized to proliferating neurons as they exit the cell cycle, whereas Ink4d is detected mainly in postmitotic neurons (10) . Expression of p19Ink4d and p27Kip1 is maintained in postmitotic cells of the adult brain, including Purkinje and granule neurons in the adult cerebellum (10) . Codeletion of both genes in the mouse allows differentiated neurons in various parts of the brain to undergo inappropriate cell divisions, leading to neuronal apoptosis and, ultimately, to gross neurological dysfunction and death between P14 and P24 (11) . This suggests that p19Ink4d and p27Kip1 normally collaborate to actively maintain certain neurons in a postmitotic state. In contrast, in the organ of Corti within the inner ear, p19Ink4d alone is required for maintenance of sensory hair cells in a quiescent state, whereas p27Kip1 keeps supporting cells from dividing. Loss of Kip1 leads to the appearance of supernumerary supporting cells, whereas disruption of Ink4d induces sensory hair cells to reenter the cell cycle and die, with loss of either CKI leading to deafness (12 , 13) . Like mice lacking Kip1 (14, 15, 16) , animals lacking Ink4c exhibit organomegally (17 , 18) . Both Kip1-null and Ink4c-null mice develop pituitary adenomas (14, 15, 16, 17, 18) , whose time of onset is greatly accelerated in animals lacking both genes (17 , 19) .
Loss of Ink4d does not accelerate tumor formation in Ink4c-null mice, but males lacking these two CKIs cannot produce viable sperm and are sterile (20) . We interbred the Ink4c/Ink4d-null mice with animals lacking p53, reasoning that apoptosis in developing male germ cells might be mitigated on a p53-null background and lead to testicular tumors. Mice lacking both Ink4c and p53 were highly tumor-prone, and the frequency of germ cell tumors was not increased by codeletion of Ink4d. Instead, we detected a high incidence of vascular tumors, as well as medulloblastomas. Studies of Ink4c and Ink4d expression in the developing cerebellum indicated that Ink4c is expressed primarily in cells of the pia mater and, between P5 and P10, in rapidly dividing cells of the EGL that are presumed to give rise to medulloblastoma. In contrast, Ink4d expression is restricted to Purkinje neurons. These data suggest that Ink4 proteins may play both cell-autonomous and non-cell-autonomous roles in tumor formation.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Tumor and Cell Transplants.
EOMA cells [a kind gift from R. Auerbach (University of Wisconsin); Ref. 22
] were transplanted by injection of 1 x 106 cells s.c. on each shoulder of NOD/SCID mice. Hemangiosarcomas were transplanted into NOD/SCID mice as described previously (23)
. Tumor growth was monitored daily.
Histopathology, Immunohistochemistry, and in Situ RNA Hybridization.
Moribund mice were sacrificed using CO2 inhalation. Tumors and organs from each animal were systematically collected for pathology and immunohistochemical analyses. Tissues were fixed in 10% neutral-buffered formalin and embedded in paraffin. Tissue sections of 5 µm were stained with H&E for histopathological analysis and characterized by immunohistochemistry (24)
. Heat-induced epitope retrieval was performed in citrate buffer (pH 6.0) or in target retrieval solution (pH 6.2; DAKO, Carpinteria, CA) as described previously (25)
. Vascular tumors were stained with antibodies to proteins that characteristically mark endothelial cells: CD31 and CD34 (PharMingen, San Diego, CA); VEGF-R2 (R&D Systems, Minneapolis, MN); and Factor 8 (DAKO). B- and T-cell lymphomas were characterized by staining with antibodies to CD3, CD45R/B220, CD138, terminal deoxynucleotidyl transferase, IgM, and 
light chain as described previously (25)
. Nonvascular tumors were stained with antibodies to determinants expressed on soft tissue mesenchymal and epithelial cells: desmin, muscle-specific actin, myogenin, smooth muscle actin, cytokeratin wide spectrum, and S-100 (all from DAKO); and cytokeratin AE1/AE3 (Chemicon International, Temecula, CA). Biotinylated second antibodies were from Vector Laboratories (Burlingame, CA) or Santa Cruz Biotechnology (Santa Cruz, CA). Negative controls were prepared by omitting primary antibody and substituting isotype-matched IgGs at equivalent concentrations.
For examination of brain tumors, formalin-fixed, paraffin-embedded mouse brain sections were deparaffinized, and antigen retrieval was performed for 1 min in a pressure cooker with antigen unmasking solution (Vector Laboratories). Endogenous peroxidase activity was quenched with 3% H2O2 for 5 min at room temperature. Sections were incubated for 45 min at room temperature with mouse monoclonal antibodies specific for nestin (1:300; BD PharMingen) or synaptophysin (1:100; DAKO) or with rabbit polyclonal antibodies for GFAP (1:1000; DAKO). Primary antibodies were visualized with either mouse or rabbit Vectastain Elite ABC kits (Vector Laboratories) following manufacturer’s instructions. In situ hybridizations were performed with antisense probes from Ink4d and Ink4c mouse cDNAs on sections of heads from E11.5 embryos to P5 and on sections of brain from adult mice, as described previously (10) .
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
70% incidence), leading to the death of afflicted animals within 6 months of birth (21
, 26) . Other tumors arising in p53-null animals include teratomas and sarcomas of various origins, including hemangiosarcomas at a significant but relatively low incidence (8–23%, depending on genetic background; Ref. 27
). The overall incidence of sarcomas (57% penetrance) is increased in p53+/- strains where the wild-type p53 allele is inactivated in the tumors, but the fraction of hemangiosarcomas is not correspondingly elevated (21)
. Ink4c-null mice, as well as those lacking both Ink4c and Ink4d, develop pituitary intermediate lobe tumors (75%) and T-cell lymphomas (10%) late in life, with mean latencies of 12 and 14 months, respectively (17
, 18
, 20)
.
Our initial goal was to try to restore male fertility in Ink4c/Ink4d-null mice by inhibiting p53-dependent apoptosis, so we bred these animals onto a p53-null or Arf-null background. Mice lacking p53 and one or two Ink4c alleles, regardless of Ink4d status, became moribund as early as 2 months of age, and by 30 weeks, all had succumbed to cancer (Fig. 1)
. In contrast, when crossed into an Arf-null background, mice lacking Ink4c did not develop any overt disease by 4 months of age, apart from incipient pituitary tumors. The survival curves (Fig. 1)
suggest that animals lacking p53 and only one copy of Ink4c develop tumors more slowly than mice lacking both Ink4c alleles, but these apparent differences were not statistically significant, given the sizes of the cohorts studied.
|
. Loss of one or two copies of Ink4c together with p53 did not significantly accelerate the onset of thymic lymphomas (mean latency, 16 weeks) that are associated with p53 loss alone. Animals lacking both Ink4c alleles (or, in one case, only one) developed mid-lobe pituitary adenomas, as expected (17
, 18)
. Mice lacking p53 and one (83%) or both (88%) Ink4c alleles also developed endothelial cell hyperplasias and vascular tumors of different degrees of severity, regardless of their Ink4d status. A significant number of animals developed brain tumors, most of which were medulloblastomas. Intriguingly, although medulloblastomas appeared in animals lacking one or both Ink4c alleles, no such tumors were observed in animals that also lacked both copies of Ink4d. Therefore, retention of Ink4d appears to facilitate medulloblastoma formation in the p53/Ink4c-null setting (see below).
|
. A gradient of pathologies was observed ranging from simple endothelial cell hyperplasia and benign angiomas to frank hemangiosarcomas (Fig. 2A)
. Histological staining revealed the presence of highly vascularized tissues with pleomorphic, dense nuclei and visible mitotic figures (Fig. 2B)
. Cells within the tumor masses expressed endothelial markers including CD31 (Fig. 2C)
, CD34 (Fig. 2D)
, VEGF-R2 (Fig. 2E)
, and von Willebrand factor (Factor 8; data not shown). The incidence of hemangiosarcomas was high in all p53-null mice lacking at least one Ink4c allele (Table 2)
.
|
|
were excised and divided, and pieces were injected s.c. into cohorts of three to six recipient animals. As a positive control, we used a transformed endothelial cell line, EOMA, which is vascular endothelial growth factor independent, can form capillaries when grown in Matrigel, and can be easily propagated by transplantation. Many of the primary tumors could be transplanted, reappearing and growing at s.c. sites of implantation within a month in at least some recipients (Fig. 2H)
. These transplanted tumors exhibited the same pathological characteristics as those seen in the primary tumors from Ink4c+/-/p53-/- or Ink4c-/-/p53-/- animals and were virtually indistinguishable from them (Fig. 2, H and I)
. These tumors can now be serially transplanted. To date, however, all attempts to propagate primary hemangiosarcomas in tissue culture have been unsuccessful, possibly because the tumors consist of multiple cell populations, of which endothelial cells represent only one component.
Medulloblastomas Develop in Mice Lacking Ink4c and p53 but Retaining Ink4d.
Unexpectedly, mice lacking p53 and Ink4c, but retaining one or two copies of Ink4d, developed brain tumors in addition to all other tumors listed in Table 1
. The frequency of brain tumor formation ranged from 8% to 24% in animals lacking Ink4c alleles (Table 1)
but dropped to 3.7% in mice also lacking both Ink4d alleles. The latter differences were significant (P < 0.02), implying that complete Ink4d loss counteracts these effects of Ink4c deletion. Moreover, whereas Ink4c loss predisposes animals to medulloblastoma formation (see below), the two brain tumors detected in animals lacking p53 and Ink4d but retaining one or both copies of Ink4c represented rare PNETs, perhaps reflecting only p53 deficiency, as described earlier by others (26)
. The similar frequency of tumor formation in Ink4c heterozygotes and Ink4c-null mice could reflect inactivation of the wild-type allele or haploinsufficiency at the Ink4c locus, which has been demonstrated in other settings (28)
. This has not been further studied.
Brain tumors in the Ink4c/p53-deficient animals, 15 in toto, included a majority (12)
that were predominantly cerebellar (Table 3)
. The tumors tended to spread along the surface of the cerebellar cortex and often grew outward as exophytic masses. Inward infiltration of the tumor through the molecular layer and into the granular cell layer and cerebellar white matter was also observed in the majority of cases (Fig. 3A)
. In one mouse, the cerebellar tumor extended into the brain stem; in another, separate tumor foci were seen in the olfactory bulb. Cerebellar tumors exhibited closely packed, round to oval embryonal cells with scant cytoplasm (Fig. 3B)
that closely resembled human medulloblastomas (Fig. 3C)
. The tumors often contained anaplastic regions with enlarged cells, cellular wrapping, and increased mitotic activity (Fig. 3D)
, features commonly found in human anaplastic medulloblastoma. Tumor nodules in the brain stem (Fig. 3E)
and cerebral cortex (Fig. 3F)
were also densely cellular and contained cells with prominent nucleoli and tumor giant cells. To confirm that the cerebellar tumors were medulloblastomas and not other primary central nervous system neoplasms or metastatic lesions, such as lymphomas, sections were stained for markers of glial (GFAP; Fig. 3G
) and neuronal (synaptophysin; Fig. 3H
) differentiation and for the intermediate filament nestin (Fig. 3I)
, which is expressed in primitive neuroepithelial cells and embryonal tumors. All cerebellar tumors stained positively for each of these markers in areas with little or no normal brain tissue, confirming that they exhibited divergent neuroectodermal differentiation and were bona fide medulloblastomas.
|
|
Examination of more than 30 cerebella from brain tumor-free animals of similar genotypes and ages (4–5 months) as the ones that developed medulloblastoma failed to reveal definitive precursor lesions. However, in several cases with smaller cerebellar lesions, the bulk of the tumor was clearly on the surface of the brain with only limited inward growth, suggesting that the tumors might arise at peripheral sites occupied by the proliferating EGL during cerebellar development. Brain stem lesions located adjacent to the fourth ventricle could have arisen from periventricular germinal matrix cells. An expanded and cytologically atypical ventricular zone subadjacent to the lateral ventricle in one mouse lacking an overt tumor could represent a precursor of supratentorial PNET. In animals with both cerebellar and cortical tumors, the cortical lesions could either have arisen independently or spread from the primary cerebellar site via the cerebral spinal fluid, as often occurs in humans.
To further assess the different contributions of Ink4c and Ink4d to tumor development, we evaluated their expression by in situ RNA hybridization in the cerebella of wild-type mice from E11.5 into adulthood (Fig. 4)
. At E11.5 (Fig. 4, A and K)
, the cerebellar anlage is a crescent-shaped structure consisting of a thick NE containing dividing neuroblasts and a DZ containing the earliest developing nuclear cells. At this time, an epithelial layer that covers the cerebellum will develop into the pial membrane (pia). Another epithelial layer, the medullary vellum (mv), will develop into the roof of the fourth ventricle. Expression of Ink4c RNA was detected in the developing pia (Fig. 4B)
, whereas Ink4d RNA was localized to the DZ (Fig. 4L)
. No expression of either gene was seen within the NE. By E13.5 (Fig. 4, C and M)
, the cerebellar anlage becomes a closed outpocketing of the hindbrain containing three distinct DZs. Recently born nuclear cells produced in the NE reside in DZ1 and move to their final position in DZ2, whereas early postmitotic Purkinje cells reside in DZ3. Ink4c and Ink4d were expressed in the pia (Fig. 4D)
and in DZ1 and DZ3 (Fig. 4N)
, respectively.
|
have migrated along the cerebellar surface to form the developing EGL, below which is a 3–4-cell-thick plate of Purkinje cells (PCPs; Fig. 4, E and P
). Ink4c was expressed in the pial membrane and in ossifying chondrocytes of the developing skull (Fig. 4F)
, and an increased intensity of Ink4d staining marked the PCPs (Fig. 4P)
. Importantly, at this time, cells in the GT and within the EGL do not express Ink4 RNAs. However, during the first 2 postnatal weeks of life, granule cells rapidly proliferate, leave the EGL, and migrate to form the IGL. At P5 (Fig. 4, G and Q)
, the cerebellar cortex, covered by the thickened Ink4c-expressing pia (Fig. 4H)
, consists of four layers: the rapidly proliferating granule cells of the EGL; an underlying molecular layer; a multilaminar PCP; and the nascent IGL. At this time, Ink4c expression was detected within the proliferative zone of the EGL. Other markers of proliferating GCPs, including the Math1 transcription factor and Ki67, are also expressed in this population, whereas Kip1 expression is confined to a deeper layer of cells that have exited the cell division cycle and to Purkinje cells (data not shown). Ink4d was expressed in Purkinje cells and in deeper cerebellar nuclear cells (Fig. 4R)
. Hence, Ink4c, but not Ink4d, is expressed at P5 in GCPs that are presumed to give rise to medulloblastomas (29)
.
In the adult cerebellum (Fig. 4, I and S)
, all cells formerly in the EGL have migrated to the IGL, so that a trilaminar cortex is established containing an outer molecular layer, a single layer of Purkinje cells (Fig. 4S
, pcl), and the IGL. High levels of Ink4c expression persisted only in the pial layer (Fig. 4J)
, whereas Ink4d expression was maintained in Purkinje cells and in deeper cerebellar nuclear cells (Fig. 4T)
.
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Vascular Tumorigenesis at High Penetrance.
Animals lacking both p53 and Ink4c presented with vascular tumors, the majority of which were invasive. Mice lacking p53 and Ink4a also develop such tumors with a high (56%) incidence (32)
, and it remains unclear whether loss of Ink4c has more profound effects than Ink4a in this regard, or whether the observed frequencies simply reflect strain differences. Ink4c+/-/p53-/- mice also developed hemangiosarcomas, but analysis of loss of heterozygosity in these tumors was not possible, due to the mixture of cell types that comprise the tumors and the ambiguity of which cell type (endothelial or mesenchymal) is the actual tumor cell. However, haploinsufficiency for tumor formation was recently revealed when Ink4c+/- mice were challenged with chemical carcinogens (28)
.
Primary hemangiosarcomas were serially transplantable by dividing tumors into pieces and introducing them s.c. into cohorts of immunocompromised mice. Not every animal receiving portions of the same primary tumor initially developed a secondary tumor at the injection site. Moreover, we were unable to establish tumor cell lines when we explanted primary tumors into culture. We therefore suspect that these tumors may be composed of interacting cell types that must be propagated in such a way that appropriate cell-cell interactions are maintained. In short, although the endothelial component of these tumors is prominent, there is no direct evidence that the proliferation of endothelial cells results from their oncogenic transformation, as opposed to their mitogenic stimulation by neighboring cells.
Roles of Ink4c and Ink4d in Medulloblastoma Formation.
Medulloblastomas, which are cerebellar embryonal neoplasms, do not arise in Ink4c-null or p53-null mice, but they appeared at a significant frequency in animals lacking both genes. Analysis of tumor histology, histochemistry, and patterns of invasion and metastasis revealed that these tumors shared many of the hallmarks of human anaplastic medulloblastomas. Medulloblastomas represent the most common malignant brain tumor in children and remain difficult to treat by combined surgical, chemo-, and radiotherapy (5-year survival, 50%). Similar lesions, known as PNETs, which can arise outside of the posterior fossa, albeit rarely, have been observed infrequently in p53-null animals (26)
.
During late embryogenesis, neural precursor cells migrate from a dorsal hindbrain structure, the rhombic lip, and advance along the cerebellar surface to form the EGL, which, in mice at birth, consists of only a single layer of cells. Postnatally, this population rapidly expands to generate a large pool of GCPs. Much of this expansion is regulated by diffusible factors, which, among others, include Shh released by neighboring Purkinje cells, Notch and Wnt ligands released by GCPs themselves, and the chemokine SDF-1
, which is released by pial cells on the cerebellar surface (29
, 33, 34, 35)
. Once expanded, GCPs exit the EGL, migrating through the Purkinje cell layer to form the IGL and maturing into granule neurons by the third postnatal week. By the time this process is completed, the major population of neurons in the entire adult brain comes to consist of cerebellar granule cells.
Most investigators believe that medulloblastomas arise from GCPs that sustain mutations and are unable to properly differentiate (29) . Consistent with this view, about 30% of human medulloblastomas sustain mutations in the Shh signaling pathway that involve genes such as Patched (Ptc1), Smoothened (Smo), and Suppressor of Fused (SuFu), leading to constitutive activation of Gli transcription factors (29 , 33) . Shh signaling also up-regulates the c- and N-Myc genes, which are amplified in 10% of medulloblastomas and highly expressed in many more (36 , 37) . Ptc1+/- mice develop medulloblastomas (29) , and the incidence of tumors approaches 100% on a p53-null background (although, here again, loss of Arf does not accelerate disease; Ref. 38 ). Animals lacking ligase IV die in utero but are rescued on a p53-null background (39) ; these mice also develop medulloblastomas, implying that DNA damage in the absence of p53 checkpoint control can initiate such tumors (40) . p53 and Rb loss can collaborate in this setting because their targeted codeletion in the cerebellum results in medulloblastomas with full penetrance (41) . In principle, Ink4c loss might mimic the effects of Rb loss and similarly contribute.
Oligonucleotide microchip arrays performed with RNAs taken from medulloblastomas arising in mice of different genotypes (including Ptc1+/-/p53-/-, LigIV-/-/p53-/-, and Ink4c-/-/p53-/-) compared with RNA from the cerebellum of normal 5-day-old mice revealed that the medulloblastomas exhibited strikingly similar patterns of gene expression. Many tumors overexpressed Gli1, N-Myc, c-Myc, cyclin D1, and the SDF-1 receptor CXCR4 by at least 5-fold (42)
. Therefore, they appeared to exhibit up-regulated Shh signaling, despite the different founding mutations that initiated tumorigenesis.
Ink4c expression was restricted to the developing postnatal EGL and otherwise was mostly confined to pial cells. We suspect that expression of p18Ink4c in GCPs acts together with p27Kip1 to regulate mitotic exit, after which only p27Kip1 expression is maintained in the postmitotic granule neurons that migrate from the EGL into the IGL. The brains of either Ink4c-null or Kip1-null mice are enlarged and contain extra neurons, but these cells seem to differentiate and migrate properly and make appropriate connections to other cerebellar neurons because no overt neurological deficits are observed in such mice (14, 15, 16, 17, 18, 19) . Although loss of Kip1 leads to increased proliferation in many parts of the brain, especially in regions of continuous proliferation such as the dentate gyrus, this is insufficient to induce brain tumors. Brain tumors have not been reported in any other mouse cohorts where Cip/Kip genes were deleted alone, in combination, or with p53. Therefore, defects in cell cycle regulation per se may not account for medulloblastoma development in Ink4c/p53-null mice.
Knoepfler and collaborators (43) recently found that Ink4c and Kip1 are highly expressed in the embryonic cerebellar primordium of mice in which N-Myc was conditionally deleted using a Nestin-Cre transgene. These mice develop a small cerebellum, presumably due to the abnormal expression of CKIs that suppress proliferation of the EGL. Hence, Ink4c and Kip1 are down-regulated by N-Myc, and this is likely necessary for proper granule cell development.
Ink4d was expressed in postmitotic Purkinje neurons, and its loss countered the development of medulloblastomas. Because these tumors are presumed to arise from GCPs, loss of Ink4c and Ink4d might affect the release of diffusible factors from pial and Purkinje cells, respectively, that sandwich the developing EGL. Because we saw no overt expansion of pial and Purkinje populations, there is no evidence that the role of the Ink4 proteins is to regulate their cell cycles. Instead, it may prove that the Ink4 proteins govern Rb family-dependent transcriptional programs that control cell-cell communication and the subsequent expansion and differentiation of GCPs. Evaluation of the levels of regulatory factors, such as Shh and SDF-1, produced by Purkinje and pial cells from Ink4-deficient animals should improve our understanding of the processes involved.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by NIH Grants CA 71907 (to M. F. R.), R25 CA23944 (to L. N.), Cancer Center Core Grant CA 21765, and by the American Lebanese Syrian Associated Charities of St. Jude Children’s Research Hospital. C. E. is the recipient of a Burroughs Wellcome Career Award in the Biomedical Sciences. C. J. S. is an Investigator of the Howard Hughes Medical Institute. 
2 Both of these authors contributed equally to this work.
3 To whom requests for reprints should be addressed, at St. Jude Children’s Research Hospital, Danny Thomas Research Tower, 5006C, Mail stop 350, 332 North Lauderdale, Memphis, TN, 38105. Phone: (901) 495-3481; E-mail: martine.roussel{at}stjude.org
4 The abbreviations used are: CDK, cyclin-dependent kinase; CKI, CDK-inhibitory protein; Rb, retinoblastoma; P, postnatal day; E, embryonic day; EGL, external granule layer; NOD, nonobese diabetic; SCID, severe combined immunodeficient; VEGF-R2, vascular endothelial growth factor receptor 2; GFAP, glial fibrillary acidic protein; PNET, primitive neuroectodermal tumor; DZ, differentiating zone; NE, neuroepithelial zone; GT, germinal trigone; PCP, Purkinje cell precursor; IGL, internal granular layer; GCP, granule cell precursor; Shh, Sonic hedgehog.
Received 4/14/03. Revised 6/ 5/03. Accepted 6/18/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  B. A. Hatton, E. H. Villavicencio, K. D. Tsuchiya, J. I. Pritchard, S. Ditzler, B. Pullar, S. Hansen, S. E. Knoblaugh, D. Lee, C. G. Eberhart, et al. The Smo/Smo Model: Hedgehog-Induced Medulloblastoma with 90% Incidence and Leptomeningeal Spread Cancer Res., March 15, 2008; 68(6): 1768 - 1776. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Hambardzumyan, O. J. Becher, M. K. Rosenblum, P. P. Pandolfi, K. Manova-Todorova, and E. C. Holland PI3K pathway regulates survival of cancer stem cells residing in the perivascular niche following radiation in medulloblastoma in vivo Genes & Dev., February 15, 2008; 22(4): 436 - 448. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Sasai, J. T. Romer, H. Kimura, D. E. Eberhart, D. S. Rice, and T. Curran Medulloblastomas Derived from Cxcr6 Mutant Mice Respond to Treatment with a Smoothened Inhibitor Cancer Res., April 15, 2007; 67(8): 3871 - 3877. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Zindy, T. Uziel, O. Ayrault, C. Calabrese, M. Valentine, J. E. Rehg, R. J. Gilbertson, C. J. Sherr, and M. F. Roussel Genetic Alterations in Mouse Medulloblastomas and Generation of Tumors De novo from Primary Cerebellar Granule Neuron Precursors Cancer Res., March 15, 2007; 67(6): 2676 - 2684. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Pogoriler, K. Millen, M. Utset, and W. Du Loss of cyclin D1 impairs cerebellar development and suppresses medulloblastoma formation Development, October 1, 2006; 133(19): 3929 - 3937. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Shakhova, C. Leung, E. van Montfort, A. Berns, and S. Marino Lack of Rb and p53 Delays Cerebellar Development and Predisposes to Large Cell Anaplastic Medulloblastoma through Amplification of N-Myc and Ptch2. Cancer Res., May 15, 2006; 66(10): 5190 - 5200. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Mendrzyk, B. Radlwimmer, S. Joos, F. Kokocinski, A. Benner, D. E. Stange, K. Neben, H. Fiegler, N. P. Carter, G. Reifenberger, et al. Genomic and Protein Expression Profiling Identifies CDK6 As Novel Independent Prognostic Marker in Medulloblastoma J. Clin. Oncol., December 1, 2005; 23(34): 8853 - 8862. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Uziel, F. Zindy, S. Xie, Y. Lee, A. Forget, S. Magdaleno, J. E. Rehg, C. Calabrese, D. Solecki, C. G. Eberhart, et al. The tumor suppressors Ink4c and p53 collaborate independently with Patched to suppress medulloblastoma formation Genes & Dev., November 15, 2005; 19(22): 2656 - 2667. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. A. Milne, C. M. Hughes, R. Lloyd, Z. Yang, O. Rozenblatt-Rosen, Y. Dou, R. W. Schnepp, C. Krankel, V. A. LiVolsi, D. Gibbs, et al. Menin and MLL cooperatively regulate expression of cyclin-dependent kinase inhibitors PNAS, January 18, 2005; 102(3): 749 - 754. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. R. Hallahan, J. I. Pritchard, S. Hansen, M. Benson, J. Stoeck, B. A. Hatton, T. L. Russell, R. G. Ellenbogen, I. D. Bernstein, P. A. Beachy, et al. The SmoA1 Mouse Model Reveals That Notch Signaling Is Critical for the Growth and Survival of Sonic Hedgehog-Induced Medulloblastomas Cancer Res., November 1, 2004; 64(21): 7794 - 7800. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Zindy, R. T. Williams, T. A. Baudino, J. E. Rehg, S. X. Skapek, J. L. Cleveland, M. F. Roussel, and C. J. Sherr Arf tumor suppressor promoter monitors latent oncogenic signals in vivo PNAS, December 23, 2003; 100(26): 15930 - 15935. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||
, Ink4d+/+, Ink4c+/+, p53-/- (n = 19); 
, Ink4d+/+, Ink4c+/-, p53-/- (n = 23); 
, Ink4d+/+, Ink4c-/-, p53-/- (n = 36); 
, Ink4d-/-, Ink4c-/-, p53-/- (n = 27); (
) Ink4d+/+, Ink4c-/-, Arf-/- (n = 62).
