Full Text (PDF)

Research Article

Patched2 Modulates Tumorigenesis in Patched1 Heterozygous Mice

Youngsoo Lee,1 Heather L. Miller,1 Helen R. Russell,1 Kelli Boyd,3 Tom Curran,2 and Peter J. McKinnon1

Departments of 1Genetics and Tumor Cell Biology and 2Developmental Neurobiology and 3Animal Resource Center, St Jude Children’s Research Hospital, Memphis, Tennessee

Abstract human patients was associated with poor prognosis (15). Many The sonic hedgehog (SHH) receptor Patched 1 (Ptch1) is features of PTCH2 make this a gene of interest and potentially relevant to tumorigenesis. These include the high similarity to critical for embryonic development, and its loss is linked to tumorigenesis. Germ line inactivation of one copy of Ptch1 PTCH1, the sites of expression, and the link to other tumor types predisposes to basal cell carcinoma and medulloblastoma in (17–22). mouse and man. In many cases, medulloblastoma arising PTC2 is located on chromosome 1p33-34, comprises 22 exons, from perturbations of Ptch1 function leads to a concomitant and shares about 73% amino acid similarity to PTC1 although with up-regulation of a highly similar gene, Patched2 (Ptch2). As significant sequence differences in the transmembrane domains 6 increased expression of Ptch2 is associated with medulloblas- and 7. Several alternatively spliced transcripts of the PTC2 gene toma and other tumors, we investigated the role of Ptch2 in have been identified in various human tissues, although the tumor suppression by generating Ptch2-deficient mice. In physiologic roles of PTC2 spliced forms remains to be determined striking contrast to Ptch1À/À mice, Ptch2À/À animals were (18, 22). Notably, PTC2 has been linked to familial/sporadic BCC and medulloblastoma (19, 22), tumors that are also associated with born alive and showed no obvious defects and were not cancer prone. However, loss of Ptch2 markedly affected tumor PTC1 mutation (8, 14). PTC2 was also suggested as a putative formation in combination with Ptch1 haploinsufficiency. tumor suppressor candidate whose loss was observed frequently Ptch1+/ÀPtch2À/À and Ptch1+/ÀPtch2+/À animals showed a in meningioma (20, 21), and decreased PTC2 expression was higher incidence of tumors and a broader spectrum of tumor associated with malignant peripheral nerve sheath tumors (17). types compared with Ptch1+/À animals. Therefore, Ptch2 Like PTCH1, PTCH2 is also a SHH target, although as the modulates tumorigenesis associated with Ptch1 haploinsuffi- expression pattern of PTCH1 and PTCH2 do not fully overlap, it is likely there are also separate functions for each (23–26). PTCH2 ciency. (Cancer Res 2006; 66(14): 6964-71) can bind HH proteins and can form a complex with SMOOTH- ENED (26). Although the physiologic function of PTC2 is still Introduction unclear, based on sequence and biochemical similarity to PTC1 and Signaling pathways that regulate development, such as the SHH aberrant expression in several tumors, it is possible that PTC2 is pathway, have been linked to tumorigenesis (1–3). SHH is a also important in the SHH signaling pathway and can modulate secreted molecule required for cell growth, patterning, and fate development and tumor formation. Therefore, to test if PTCH2 determination in many tissues (4–7). Germ line mutation of has functional overlap with PTCH1, including a role during tumori- PATCHED1 (PTCH1), a receptor for SHH, is responsible for the genesis, we generated Ptch2-deficient mice. In contrast to Ptch1À/À Gorlin syndrome, a familial cancer predisposition syndrome with a mice, which showed early embryonic lethality, Ptch2À/À animals high incidence of medulloblastoma and basal cell carcinoma (BCC; did not show any discernable abnormalities during development refs. 8, 9). Mutation of PTC1 also occurs in sporadic tumors, and or a propensity for tumorigenesis. However, when combined with germ line and somatic mutations of SUFU (suppressor of fused), a Ptch1 mutations, Ptch2 mutations promoted a dramatic increase in downstream negative regulator of the SHH pathway, were recently the incidence of tumorigenesis, suggesting a cooperative role of linked to medulloblastoma (10). Similar to Gorlin syndrome, Ptch2 with the tumor suppressor function of Ptch1. Thus, to our engineered mutations in Ptch1 in the mouse can lead to knowledge, these data are the first demonstration that whereas medulloblastoma, highlighting the direct relationship between Ptch2 is dispensable for development, it can influence the effect of SHH/PTC1 signaling and medulloblastoma (11–14). Ptch1 attenuation during tumorigenesis. Our previous studies examining the genesis of medulloblastoma identified a common cohort of gene expression changes that occurred in genetically distinct mouse models of medulloblastoma Materials and Methods (15, 16). In these mouse models, tumors occurred because of defects in either DNA repair, SHH signaling, or cell cycle regulation. Generation of Ptch2-deficient mice. A Mouse BAC genomic DNA Library (Invitrogen, Carlsbad, CA) was screened to identify clones One gene that was highly expressed in all mouse medulloblastomas containing the Ptch2 gene. To inactivate Ptch2, we deleted exons 5 to 17 studied was Ptch2. Furthermore, >30% of human medulloblastomas from the full-length 20 exon–containing murine Ptch2 gene. To do this, a showed increased expression of PATCHED2 (PTC2); this group of BclI and HindIII 12-kb genomic fragment was inserted into HindIII and BamHI sites of pBluscript II (Stratagene, La Jolla, CA), in which the SacII restriction site was inactivated. An oligomer containing a LoxP site was inserted in a PacI site present between a BclI site and exon 5 of the Ptch2 Requests for reprints: Peter J. McKinnon, Department of Genetics, St Jude gene. Then a Neo-tk cassette flanked by LoxP sites was inserted into a SacII Children’s Hospital, 332 North Lauderdale, Memphis, TN 38105. Phone: 901-495-2700; site between Ptch2 exons 17 and 18. Finally, W9.5 embryonic stem cells were Fax: 901-526-2907; E-mail: [email protected]. I2006 American Association for Cancer Research. electroporated with a NotI-linearized targeting construct. Embryonic stem doi:10.1158/0008-5472.CAN-06-0505 cells were selected by G418, and targeted integrations were identified using

Cancer Res 2006; 66: (14). July 15, 2006 6964 www.aacrjournals.org

Southern blot analysis (data not shown). pMC-Cre was then used to excise and antisense in situ hybridization probes were generated from a Ptch2 the floxed Neo-tk, and the embryonic stem cells were grown in gancyclovir IMAGE clone (clone 3972649, nucleotides 1891-3568, and 3¶ untranslated to ensure cre-mediated removal of the Neo-tk cassette. Targeted embryonic region), a Gli1 clone in pSPORT1 (nucleotides 701-1251, Genbank accession stem cells were microinjected into C57BL/6 blastocysts and implanted in no. AB025922), a Gli2 clone in pSPORT1 (nucleotides 1055-1835, Genbank pseudopregnant F1 B/CBA foster mothers and allowed to develop to term. accession no. X136212), and an Sfrp1 clone in pSPORT1 (nucleotides 247- Germ line transmission was confirmed by Southern blot and PCR analysis. 1375, Genbank accession no. U88566). All Ptch2 animals were maintained in a mixed genetic background of Quantitative real-time PCR. Total RNA samples were extracted from 129SvX1/SVJ and C57BL/6. The PCR condition for the genotyping snap-frozen tumor samples and control tissues using Trizol (Invitrogen). was 30 cycles of 95jC for 1 minute, 63jC for 1 minute, and 72jCfor real-time reverse transcriptase-PCR (RT-PCR) was done as described before 1 minute with primers Ptch2-1 (5¶ ACCGGCACAGCAGGCAGG), Ptch2-2 (15). A primer/Taqman probe set for Ptch2 was forward primer Ptch2-SDS-F (5¶ GTGAGCTCAGTGCCGTCTACACAGC), and Ptch2-3 (5¶ AGAGATGTG- (5¶-ATCCTAGCTGGGAGCCTGAAG), reverse primer Ptch2-SDS-R (5¶-TCCC- TCTGGCCACACAGAGC). These conditions gave a 755-bp PCR product for GCATCCCAGAGAGA), and Taqman probe Ptch2-SDS-FAM (5¶-TCCACTCT- a wild-type allele and a 663-bp product for the mutant allele. GGCTTCGTGCTTACTTCCA), which detected a portion of a missing part To examine the targeted Ptch2 allele, total RNA was extracted from P5 in the Ptch2 mutant allele (nucleotides 2366-2447, Genbank accession wild-type (WT) and Ptch2À/À brains using Trizol (Invitrogen), and RNA was no. NM_008958). Two different sets of primer/Taqman probe sets for Ptch1 separated on a 1% agarose gel in a MOPS/formaldehyde–containing buffer. were used: one set to detect mRNA from exon 21 and the other for exon 2 cDNA probes for either Ptch1 or Ptch2 were radiolabeled using a random (the exon disrupted in the mutant allele of Ptch1; ref. 14). To detect Ptch1 primed labeling system (Roche, Indianapolis, IN). exon 21, we used forward primer Ptch1-SDS-F (5¶-CAAGTGTCGTC- Mice. Ptch1 and Ptch2 double mutant animals were obtained by CGGTTTGC), reverse primer Ptch1-SDS-R (5¶-CTGTACTCCGAGTCGGAG- intercrossing of Ptch1+/ÀPtch2À/À or Ptch1+/ÀPtch2+/À. The details of Ptch1 GAA), and Taqman probe Ptch1-SDS-FAM (5¶-CCTCCTGGTCACACGAA- animal model were described before (13, 14). Ptch2À/Àp53À/À mutant CAATGGGTC). The primer set for Ptch1 exon 2 was forward primer animals were generated by intercrossing of either Ptch2+/Àp53+/À or Ptch1ex2-SDS-F (5¶-GGCTACTGGCCGGAAAGC), reverse primer Ptch1ex2- Ptch2À/Àp53+/À mice. All Ptch1-Ptch2 mice used in these studies were F2 SDS-R (5¶-GAATGTAACAACCCAGTTTAAATAAGAGTCT), and Taqman or later-generation littermates on a 129Svj Â C57BL/6 genetic background. probe Ptch1ex2-SDS-F (5¶-CCGCTGTGGCTGAGAGCGAAGTTTC). In addi- The presence of a vaginal plug was designated as embryonic day 0.5 (E0.5), tion, we used the following primer/probe sets for other genes: Gli1 forward and the day of birth as postnatal day 0 (P0). Animals were acclimated to (5¶-GCTTGGATGAAGGACCTTGTG), reverse (5¶-GCTGATCCAGCCTAA- controlled temperature and constant light/dark schedule with food and GGTTCTC), and Taqman probe (5¶-CCGGACTCTCCACGCTTCGCC); Sfrp1 water ad libitum. Full necropsy was done by the diagnostic Pathology forward (5¶-TCCTCCATGCGACAACGA), reverse (5¶-TGATTTTCATCCTCA- laboratory at St. Jude Children’s Research Hospital. All animals were housed GTGCAAACT), and Taqman probe (5¶ TGAAGTCAGAGGCCATCATTGAA- in an Association for Assessment and Accreditation of Laboratory Animal CATCTCTG); Math1 forward (5¶-ATGCACGGGCTGAACCA), reverse (5¶-TC- Care–accredited facility and were maintained in accordance with the NIH GTTGTTGAAGGACGGGATA), and Taqman probe (5¶-CCTTCGACCAG- Guide for the Care and Use of Laboratory Animals. The Institutional Animal CTGCGCAACG). The outcome of real-time PCR was analyzed with SDS Care and Use Committee at St. Jude Children’s Research Hospital approved ver2.0 software (ABI, Foster City, CA). 18S rRNA assay reagents (ABI) were all procedures for animal use. used as an internal control. Histology, immunohistochemistry, and in situ hybridization. For Spectral karyotyping. Medulloblastoma primary tumors were collected tissue fixation, embryos at E13.5 were submersed directly in buffered 4% 4 hours after colcemid treatment (10 Ag/mL i.p. injection). A total of six paraformaldehyde, and brains from P5 and 3-week-old animals were medulloblastoma (three Ptch1+/ÀPtch2+/À,twoPtch1+/ÀPtch2À/À, and one removed after trans-cardial perfusion with 4% buffered paraformaldehyde. Ptch1+/ÀPtch2+/À) were subject to spectral karyotyping analysis. A single- Fixed tissues were then cryoprotected in 25% sucrose in PBS and subjected cell suspension of medulloblastoma was subject to spectral karyotyping to immunohistochemistry. Tumors were extracted from euthanized animals analysis. Spectral karyotyping procedures were done according to the and fixed immediately in 10% neutral buffered formalin. The following SkyPaint hybridization and detection protocol (Applied Spectral Imaging, antibodies were used for immunohistochemistry: anti-Ki67 (1:1,000; Vista, CA), and a commercially prepared spectral karyotyping probe was Novocastra Laboratory, Newcastle upon Tyne, United Kingdom), anti-h- used to detect chromosomes. Sample pretreatment consisted of RNase A tubulin II (Tuj1; 1:1,000; BabCo, Richmond, CA), glial fibrillary acidic protein (100 Ag/mL) for 1 hour at 37jC and pepsin for 2 minutes at 20jC (50 Ag/mL (GFAP; 1:400; Sigma, St. Louis, MO), p27 (1:2,000; Santa Cruz Biotechnology, in 10 mmol/L HCl), with counterstaining by 4¶,6-diamidino-2-phenylindole. Santa Cruz, CA), and synaptophysin (1:200; Chemicon International, Sequence analysis of Ptch1 and Ptch2. To detect any mutations or Temecula, CA). Antigen retrieval was used for all immunohistochemistry. deletion of the Ptch1 and Ptch2 genes, cDNA was synthesized from Cryosections (10-Am thick) or paraffin sections (7-Am thick) were incubated medulloblastoma and rhabdomyosarcoma RNA by an oligo-dT primed RT with antibodies overnight at room temperature after quenching endoge- method using SuperScript II RNase HÀ Reverse Transcriptase (Invitrogen). nous peroxidase. Immunoreactivity was visualized with the vasoactive Using these cDNA samples as templates, full-length Ptch2 mRNA was intestinal peptide substrate kit (Vector Laboratories, Burlingame, CA) amplified with the following primer sets: Ptch2, a combination of forward according the manufacturer’s guide, or indocarbocyanine (Cy-3) conjugated primer Ptch2seq1F (5¶-AGCCTATGGCAGCGCTCAGATAACGCAGG) with secondary antibodies (Jackson ImmunoResearch Lab, West Grove, PA). For reverse primer Ptch2seq1R (5¶-CCTGCTGCCGCCCCAGCACAACCT- colorimetric detection, counterstaining with methyl green was done CAGTT), or for two overlapping fragments of Ptch2, a combination of followed by mounting slides with DPX (BDH Laboratory, Poole, United forward primer Ptch2seq1F with reverse primer Ptch2seqmt1 (5¶-CAG- Kingdom), and Gel/Mount (Biomedia Corp., Foster City, CA) was used for CATCTGAATGACCTGAGCGGAGCAGG) and a combination of forward fluorescence analysis. H&E staining was done by routine methods. Terminal primer Ptch2seq5F (5¶-ATTCCTGCGCTGCGGGCCTT) with reverse primer deoxynucleotidyl transferase–mediated nick-end labeling assays were done Ptch2seq1R. For Ptch1, two rounds of amplification were required as the using ApopTag (Fluorescein In situ Apoptosis Detection kit, Chemicon level of Ptch1 expression was very low. The first round of PCR was done International) to detect neuronal apoptosis in the developing brain 6 hours with forward primer Ptch1-14 (5¶-ACGCGCAATGTGGCAATGGAAGGC) and after irradiation (18 Gy at 403 rad/min from a cesium irradiator). In situ reverse primer Ptch1-R1 (5¶-GAAGCGGCCGCTTCAGATTTTAATTACCC) to hybridization was applied to detect Ptch1 and Ptch2 mRNA signals via amplify a full-length Ptch1. Subsequently, the PCR products were further exposure of emulsion-dipped slides in wild-type E16.5 and P7 brains and amplified with following five different sets of primers whose products Ptch2, Gli1, Gli2, and secreted frizzled related protein 1 (Sfrp1) mRNA signals overlapped each other and spanned a full-length of Ptch1. Set A: forward in Ptch1 and Ptch2 mutant medulloblastomas as described before (15) with primer Ptch1-F (5¶-ATGGCCTCGGCTGGTAACG) and reverse primer Ptch1- support from GENSAT at St. Jude Children’s Research Hospital (in situ 1(5¶-AAGGCCGGTCCATGTACCCATGGC); set B: forward primer Ptch1-2 methodology: http://www.stjudebgem.org/web/html/methods.php). Sense (5¶-GCTTAATCATTACACCTTTGGACTGC) and reverse primer Ptch1-6 www.aacrjournals.org 6965 Cancer Res 2006; 66: (14). July 15, 2006

(5¶-AAAGGAGCATAGTGCTTCTCTGC); set C: forward primer Ptch1-5 (5¶-TTGAGCCACAGGCCTACACAGAGC) and reverse primer Ptch1-8 (5¶-GTCTGAGGTGTCTCGTAGGCCG), set D; forward primer Ptch1-7 (5¶ TGGGAAACTGGGAGGATCATGC) and reverse primer Ptch1-10 (5¶ GCTC- AGGCGAAGGAGTGGGCAGTCG); and set E: forward primer Ptch1-9 (5¶-GTGGAGTTCACCGTCCACGTGGC) and reverse primer Ptch1-R2 (5¶-GAAGCGGCCGCTCAGTTGGAGCTGCTCCCCCACGGC). The PCR products were sequenced by a routine Big Dye Terminator (v.3) Chemistry approach on Applied Biosystem 3700 DNA analyzer.

Results Generation of Ptch2-deficient mice. Ptch1 is critical for development and also functions to prevent tumorigenesis. Given the importance of Ptch1 and its similarity to Ptch2 (73% similarity at the amino acid level), we sought to determine the potential relationship between these genes. Comparative expression analyses of Ptch1 and Ptch2 in the developing nervous system showed low and diffuse expression of Ptch2. In contrast, Ptch1 was expressed in distinct regions, such the ventricular zone in the ganglionic eminence and spinal cord during neurogenesis (Fig. 1A), suggesting distinct roles for these two genes during development. However, the developing cerebellum (P7) showed high expression of Ptch2 in the outer layer of the EGL, whereas Ptch1 was expressed more broadly in the EGL as well as granule cells in the internal granule layer (Fig. 1A). Therefore, we generated mice lacking Ptch2 in an attempt to further understand the relationship between these two proteins. To do this, we used gene targeting to remove exons 5 to 17 of the Ptch2 Figure 2. Normal development of Ptch2-deficient animals. A, wild-type and gene, a region that encodes the majority of transmembrane Ptch2 À/À embryonic brains at E13.5 (retina, Â200; other brain parts, Â400). domains (Fig. 1B). Consistent with the targeting strategy, Northern Tuj1 marks post-mitotic neurons and Ki67 identifies proliferating cells. Ptch2 À/À embryos underwent normal embryogenesis. B, wild-type and Ptch2 À/À cerebellum development (P5, Â400; 3 weeks old, Â200). Mitotic (Ki67) and post-mitotic (p27) populations were similar between wild-type control and Ptch2 À/À brains. Immunopositive signals were visualized with peroxidase substrate kit (purple staining). Methyl green was used for counterstaining (green staining).

blot analysis showed that Ptch2À/À animals produced a truncated Ptch2 message of 2.8 kb that lacks 2.3 kb of coding sequence (Fig. 1C). DNA sequencing of RT-PCR products of the Ptch2 message from Ptch2À/À cerebellum showed the only Ptch2 transcript was missing all sequences encoded by exons 5 to 17. Ptch2À/À animals were fertile and born at Mendelian frequency, and they aged without any discernible defects. Furthermore, analysis of embryogenesis showed no gross anatomic or histologic defects in Ptch2À/À embryos (Fig. 2A). We examined embryos at various ages and used immunohistochemical analysis to monitor cell proliferation and differentiation but found no differences between controls and Ptch2À/À animals. Additionally, the cerebellum of Ptch2À/À animals did not differ from the wild-type cerebellum during development and differentiation (Fig. 2A and B). We also found no differences in the response of Ptch2-null primary granule neurons to recombinant SHH compared with WT cells as determined by bromodeoxyuridine incorporation and Ki67 staining as both cell types were stimulated to the same degree after SHH (data not shown). We also used retroviral transduction of MEFs with a Gli1 promoter green fluorescent protein reporter construct, but whereas SHH stimulated Figure 1. Expression of Ptch1 and Ptch2 during development and the Gli-1 reporter expression, there were no differences between Ptch2- generation of Ptch2-deficient animals. A, in situ hybridization analysis of Ptch1 null and WT cells (data not shown). Therefore, the Ptch2 gene is and Ptch2 mRNA at E16.5 (Â200) and P7 (Â400) in the wild-type developing brain, showing the distinct expression patterns of these two genes. B, genomic dispensable during embryogenesis, in striking contrast to the early structure of the murine Ptch2 gene, which contains 20 exons. A LoxP sites lethality resulting from Ptch1 deficiency (14). flanked exons 5 and 17, which when excised by Cre recombinase deletes the Ptch2 null animals are not tumor prone. As Ptch1 majority of the transmembrane domains. C, Northern blot analysis showing the truncated Ptch2 mRNA in Ptch2 À/À animals. Ptch1 levels in the Ptch2 À/À heterozygous animals are tumor prone (14), we monitored samples together with a panel showing the ethidium bromide–stained RNA gel. Ptch2+/À and Ptch2À/À animals for tumors. Although Ptch2À/À

Cancer Res 2006; 66: (14). July 15, 2006 6966 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 2006 American Association for Cancer Research. Patched2 and Tumorigenesis animals occasionally developed spontaneous tumors, it was within mice to generate Ptch1+/ÀPtch2À/À and appropriate control the reference range and similar to matched WT animals (2 of 70 animals. Although mice harboring two mutant Ptch1 alleles were Ptch2À/À animals developed tumors by 18 months of age). Loss of embryonic lethal, all other allelic combinations were obtained at p53 rapidly accelerates tumorigenesis in Ptch1 heterozygous mice expected frequencies. Ptch1+/ÀPtch2À/À mice showed no overt (27); therefore, we introduced Ptch2 mutation onto a p53À/À phenotype, and anatomic analysis of embryos and adult brain background. Ptch2À/Àp53À/À animals showed a similar tumor showed normal histology (data not shown). However, we did find profile and latency to those of p53À/À animals (11 of 20 Ptch2À/À that many Ptch1+/ÀPtch2+/À and Ptch1+/ÀPtch2À/À mutant p53À/À animals developed thymoma or/and sarcoma by f4 animals suffered from intestinal serosal angiectasis (f19% and months of age). Ptch1+/À animals showed a high incidence of 18%, respectively; Fig. 3A), implying a problem with blood vessel medulloblastoma after irradiation at early postnatal times (28). formation compared with Ptch1+/À animals (3.7%; Table 1). Again, Ptch2À/À animals were similar to wild type, as low-dose Furthermore, we also found s.c. telangiectasia in 2 of 97 (2%) ionizing radiation (4 Gy) at P5 did not induce medulloblastoma in Ptch1+/ÀPtch2+/À animals. Ptch2À/À mice (data not shown). These data suggested that Ptch2 However, tumorigenesis in Ptch1+/À animals was markedly deficiency alone or in combination with p53 loss or ionizing affected by Ptch2 status. Ptch2À/À or Ptch2+/À in combination radiation is insufficient to induce tumorigenesis. Finally, whereas with Ptch1+/À haploinsufficiency significantly reduced tumor Ptch1+/À P5 cerebella are radiosensitive and show increased latency associated with Ptch1 loss (P = 0.006, log-rank test; Table ionizing radiation–induced cell death (15), Ptc2+/À and Ptc2À/À 1; Fig. 3B). By 8 months of age, 50% of Ptch1+/ÀPtch2À/À animals cerebellum was similar to controls at 6 hours after 18 Gy of had developed tumors, and 50% of Ptch1+/ÀPtch2+/À animals had ionizing radiation (results not shown). developed by 10 months of age, whereas only f15% Ptch1+/À Ptch1 and Ptch2 compound mutants have increased tumor Ptch2+/+ animals developed tumors by 12 months of age (Fig. 3B). susceptibility. To further assess Ptch2 function, we examined However, with increasing age (>12 months), Ptc1+/À animals were potential genetic interactions between Ptch1 and Ptch2 during less healthy and died with no obvious tumor burden at necropsy development and tumorigenesis. We intercrossed Ptch1+/ÀPtch2+/À (Fig. 3B). Ptch1+/ÀPtch2À/À compound mutant animals generally

Figure 3. Survival curve and histopathology of Ptch1 and Ptch2 combined mutations in tumorigenesis. A, two representative examples of intestinal serosal angiectasis (in one case, a rhabdomyosarcoma is also present). B, Kaplan-Meier survival analysis of animals. C, glossal rhabdomyosarcoma. Top, a Ptch1 +/ÀPtch2 À/À animal with tongue rhabdomyosarcoma. Bottom, H&E staining of a glossal rhabdomyosarcoma. D, histopathologic examination of rhabdomyosarcoma from the limbs and medulloblastoma. There were no apparent difference between Ptch1 +/ÀPtch2 +/À and Ptch1 +/ÀPtch2 À/À tumors. GFAP, glial fibrillary acidic protein.

www.aacrjournals.org 6967 Cancer Res 2006; 66: (14). July 15, 2006

Table 1. Tumor incidence in compound mutant mice

Genotype Case n Incidence (%) Period (mo) Mean (mo)

Ptch1+/ÀPtch2+/+ (n = 62) Medulloblastoma 4 6.4 4-6 4.7 Total sarcoma 14 22.6 Abdominal wall or intestine 7 11.3 10-19 13.3 Rhabdomyosarcoma (leg) 5 8.1 5-9 6.6 Glossal rhabdomyosarcoma 2 3.2 4-20 12 Angiectasis 2 3.7 9-11 10 BCC 1 1.6 9 9 Ptch1+/ÀPtch2+/À (n = 97) Medulloblastoma 15 15.4 1-8 4.1 Total sarcoma 40 41.2 Abdominal wall or intestine 18 18.5 6-16 10.4 Rhabdomyosarcoma (leg) 18 18.5 3-19 8.6 Glossal rhabdomyosarcoma 4 4.0 1-11 4 Angiectasis 18 18.5 2-19 9.8 BCC 5 5.1 7-14 10 Ptch1+/ÀPtch2À/À (n = 63) Medulloblastoma 11 17.4 2-10 4.9 Total sarcoma 25 39.6 Abdominal wall or intestine 15 23.8 2-14 8 Rhabdomyosarcoma (leg) 7 11.1 4-6 5.1 Glossal rhabdomyosarcoma 3 4.7 2-6 3.3 Angiectasis 12 19.0 5-14 9.7 BCC 3 4.7 10-14 12

developed a similar spectrum of tumors to those of Ptch1+/À, such sarcomas and angiectasis, a situation that was not observed in as medulloblastoma and rhabdomyosarcoma (14, 27, 29), although Ptch1+/À animals. Furthermore, 1 of 62 Ptch1+/À animals (1.6 %) with a much higher frequency (Fig. 3C; P = 0.05, medulloblastoma; had BCC, and Ptch2 mutation increased the frequency of BCC to P = 0.03, rhabdomyosarcoma; Table 1). To exclude genetic 5% (5 of 97 in Ptch1+/ÀPtch2+/À and3of63inPtch1+/ÀPtch2À/À). background effects, all data were acquired from later-generation In addition, 2 of 54 (3.7%) Ptch1+/ÀPtch2+/À males developed than second-generation backcrosses. Furthermore, our current testicular teratoma. These data suggest that Ptch2 mutations cohort is at the sixth generation, and increased tumorigenesis is enhance tumorigenesis, resulting from defective HH signaling apparent. Medulloblastomas from Ptch1+/ÀPtch2À/À (as well as pathway due to Ptch1 heterozygosity, implying a cooperation of Ptch1+/À+/ÀPtch2+/À) animals were histopathologically similar to Ptch2 loss with Ptch1 mutation. Ptch1+/À tumors and expressed typical neural markers, such as Ptch1 expression is lost in Ptch1+/ÀPtc2À/À tumors. We GFAP, Tuj1, and synaptophysin (Fig. 3D). Medulloblastoma checked if the decreased latency (Fig. 3B) observed in the Ptch1+/À signature genes (15), such as Gli1, Gli2, Math1, and Sfrp1, were Ptch2+/À tumors involved loss of the WT Ptch2 allele. We sequenced also overexpressed in Ptch1/Ptch2 tumors as detected by in situ Ptch2 cDNA derived from Ptch1+/ÀPtch2+/À tumor RNA samples hybridization (Fig. 4A), quantitative real-time PCR (Fig. 4B), and (two sarcomas and two medulloblastomas) using high-fidelity PCR microarray analysis (data not shown). Although Ptch1+/À medul- to amplify the Ptch2 open reading frame. We did not find any loblastoma expressed Ptch2 at high levels, the expression of Ptch2 mutations in Ptch2 mRNA in the tumors (data not shown), in Ptch1+/ÀPtch2+/À (or Ptch1+/ÀPtch2À/À) medulloblastoma was suggesting that inactivation of the WT Ptch2 allele does not occur much lower or absent (Fig. 4A and B), indicating that Ptc2 is not in Ptch1+/ÀPtch2+/À tumors. Furthermore, we amplified Ptch1 required per se for tumorigenesis. However, as the tumor incidence cDNA from the tumors and as previously reported (11, 13) did not was increased similarly when either Ptch2+/À or Ptch2À/À were find any mutations or deletions in the Ptch1 open reading frame present in combination with Ptch1+/À (Table 1), this indicates that (data not shown). However, using quantitative real-time PCR the relative levels of Ptch2 can influence tumorigenesis in Ptch1+/À analysis with two different probe sets that can discriminate the mice. We also generated Ptch1+/ÀPtch2À/À (or Ptch2+/À) mutant mutant and WT Ptch1 alleles, we found that expression of the Ptch1 animals on a p53À/À background, but in this situation, Ptch2 loss WT allele was absent or barely detectable in these tumors. In did not alter tumor latency, as Ptch1+/ÀPtch2À/À(Ptch2+/À)p53À/À contrast, the level of Ptch1 in Ptch1+/ÀPtch2À/À P5 and nonneo- tumor incidence was the same as Ptch1+/Àp53À/À (data not plastic adult cerebellum was comparable with those of WT samples shown). (Fig. 5A) with both sets of primer/probes. This strategy distin- Concomitant Ptch2 loss leads to an increase in rhabdomyosar- guishes the normal Ptch1 allele from the engineered Ptch1 mutation coma frequency in Ptch1+/À mutant animals (Table 1; Fig. 3C in which a LacZ gene interrupted Ptch1 exon2 (14), and the primer and D). This tumor type also showed more diversity in double set that identifies Ptc1 exon 21 quantifies Ptch1 expression from mutant animals, as it was found in the tongue (glossylpharyngial both alleles. Therefore, a lack of expression of exon 2 indicates that rhabdomyosarcoma; Fig. 3C), genitourinary track, trunk, and most the wild-type allele is silenced in medulloblastoma. We also frequently in the limbs. The combination of Ptch2 mutation with examined Ptch1+/ÀPtch2+/À, Ptch1+/ÀPtch2À/À, and Ptch1+/À me- Ptch1 haploinsufficiency also increased the frequency of multiple dulloblastoma samples using spectral karyotype to identify if different tumors within individual animals mostly involving enhanced genomic instability may occur when Ptch2 levels are

Cancer Res 2006; 66: (14). July 15, 2006 6968 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 2006 American Association for Cancer Research. Patched2 and Tumorigenesis reduced but found no evidence of chromosomal rearrangements particularly given the early embryonic lethality of Ptch1À/À resulting from additional inactivation of Ptch2. Interestingly, one animals. The in situ hybridization analysis showed spatiotemporal case of medulloblastoma (Ptch1+/ÀPtch2À/À) showed a loss of differences between Ptch1 and Ptch2 expression, and so the one chromosome 13 in which the Ptch1 gene is located (Fig. 5B), nonoverlapping expression between these two proteins suggests supporting the notion that Ptch1 loss, either by chromosomal loss that functional redundancy by Ptch1 does not account for or epigenetic silencing occurs in Ptc1+/-derived medulloblastoma, normal development in the absence of Ptch2. Thus, Ptch1 and consistent with previous reports in Ptc1+/À mice (30). Ptch2 function differently during development, perhaps with Ptch2 fulfilling more subtle roles during SHH signaling, or Ptch2 participating more in other HH signaling pathways, such as Discussion Desert HH whose role in development is less critical than SHH During development the HH pathway plays a crucial role in (26, 37, 38). regulation of morphogenesis and differentiation. SHH deficiency in It is evident that there is a direct link between medulloblas- the mouse results in defective patterning and multiple organ toma and genomic instability induced either endogenously or defects, including cyclopia, akin to holoprosencephaly, a human exogenously. Defects in DNA strand break repair in animals phenotype with SHH or PTCH1 mutation (31–35). Similarly, mice [e.g., DNA ligase IV or poly(ADP-ribose) polymerase deficiency] deficient for SHH signaling, such as inactivation of Ptch1 or Sufu,a can lead to genomic instability and transformation of neuro- negative regulator of the SHH pathway, die in utero due to progenitor cells in the developing cerebellum, resulting in developmental defects (14, 36). In this report, we have investigated medulloblastoma (39, 40). Furthermore, irradiation of neonatal the requirement for the HH binding protein Ptch2 during Ptch1+/À animals can also substantially increase medulloblasto- development and tumorigenesis. Our motivation for doing this ma occurrence (41) and can promote BCC precursor lesions to was the high similarity between Ptch1 and Ptch2 and the develop into nodular and infiltrative BCC (42). Therefore, we indispensability of hedgehog signaling for development and in tested if DNA damage after ionizing radiation showed any many cases the prevention of cancer. Therefore, we generated an differences in Ptch2À/À compared with WT animals. However, in animal model of Ptch2 deficiency. It was somewhat surprising that contrast to Ptc1+/À animals (15), g-radiation of Ptch2+/À or Ptch2-deficient animals were normal in all aspects we examined, Ptch2À/À P5 pups did not show any differences to controls as judged by radiation-induced apoptosis or increased tumor rate compared. These observations further confirm functional differences between Ptch1 and Ptch2. Given the relatedness between Ptch1 and Ptch2, we introduced Ptch2 mutations onto a Ptch1+/À background. In this situation, we found that Ptch2 dosage dramatically affected the occurrence of tumors in Ptch1+/À animals. This was reflected as reduced tumor latency and also as the occurrence of frequent multiple tumors in Ptch1/Ptch2 mutants compared with Ptch1+/À animals. The comparable effect of Ptch2À/À and Ptch2+/À toward tumorigenesis in the Ptch1+/À mice may reflect the fact that Ptch2 levels are substantially reduced in Ptch2+/À tissues. Similar to other murine models of medulloblastoma, we found that histopathologic analysis could not discriminate Ptch1+/ÀPtch2À/À mutant tumors from Ptch1+/À tumors. Additionally, double Ptch1/Ptch2 mutations did not alter the molecular fingerprint of the medulloblastoma, as overexpression of Gli1, Gli2, Sfrp1, and Math1 still occurred. Thus, although Ptch1+/À tumor dynamics are affected in Ptch2 mutations, the tumor identity is very similar to tumors arising in Ptch1+/À animals. Although our data describing tumorigenesis in the Ptch1+/À Ptch2À/À mice highlight the crosstalk between Ptch1 and Ptch2, other available data also support a role for PTCH2 for preventing tumorigenesis. PTCH2 is located on chromosome 1p33-34 in human, and a portion of chromosome 1 containing 1p33-44 was missing in >15% of various human tumors, including breast, colon, lung, ovarian cancer, melanoma, neuroblastoma, and testicular germ cell tumor (43). Independently, similar observations were made in neuroblastoma, testicular germ cell tumors, sporadic colorectal polyps, and meningiomas (20, 21, 44–46). In contrast, it Figure 4. Overexpression of medulloblastoma signature genes in has been reported that high PTCH2 (Ptch2) expression occurs in Ptch1 +/ÀPtch2 +/À and Ptch1 +/ÀPtch2 À/À tumors. A, expression of Sfrp1, Gli2, Gli1, and Ptch2 were detected using in situ hybridization in medulloblastomas familial and sporadic BCC (22) and in multiple models of murine (Â200). B, quantitative real-time RT-PCR data were used to measure mRNA medulloblastoma (15), although it is likely that in these cases, levels in rhabdomyosarcomas (Sar) and medulloblastomas (Med) in comparison Ptch2 overexpression may reflect activated SHH signaling (47) to wild-type P5 and adult cerebellums. The expression of Ptch2 was absent in Ptch2 À/À tumors and much lower in Ptch2 +/À tumors. However, rather than as a specific component of tumorigenesis. Therefore, Ptch1 +/ÀPtch2 +/+ medulloblastoma showed overexpression of Ptch2. the potential role of Ptch2 as a tumor suppressor in other www.aacrjournals.org 6969 Cancer Res 2006; 66: (14). July 15, 2006

Figure 5. Spectral karyotyping analysis and Ptch1 expression in tumors. A, quantitative real-time RT-PCR for Ptch1 using primers sets that discriminate the mutant and WT Ptc1 alleles. Two different sets of primer/probe were used to detect expression of exon 21 (in both mutant and wild-type alleles) and exon 2 (absent in the mutant allele). Wild-type Ptch1 expression was negligible in all Ptch1 +/À medulloblastomas examined. B, representative spectral karyotyping analysis of Ptch1 +/ÀPtch2 +/À medulloblastoma. Spectral karyotyping images show the spectral (RGB) color image on the left and the classified (pseudocolor) image on the right. In one case, a Ptch1 +/ÀPtch2 À/À medulloblastoma showed loss of one copy of chromosomes 13 and 14.

physiologic contexts besides Ptch1 mutations is clearly an area for Acknowledgments further investigation. Received 2/8/2006; revised 4/18/2006; accepted 4/26/2006. In summary, we have provided the first evidence that Ptch2 can Grant support: NIH grants NS-37956 and CA-21765, Cancer Center Support Grant modulate tumorigenesis in select settings. Our data likely reflect P30 CA21765, and the American Lebanese and Syrian Associated Charities of St Jude Children’s Research Hospital. the requirement for Ptch2 for subtle aspects of HH signaling, which The costs of publication of this article were defrayed in part by the payment of page can enhance tumorigenesis when suboptimal levels are present, charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. probably via collaboration with Ptch1 loss to maintain persistent We thank the Hartwell Center, the Cancer Center Cytogenetics Core, and the SHH signaling. Transgenic Core facility for their help with these studies.

References Mutations of the human homolog of Drosophila patched fingerprint for medulloblastoma. Cancer Res 2003;63: in the nevoid basal cell carcinoma syndrome. Cell 1996; 5428–37. 1. Hooper JE, Scott MP. Communicating with hedgehogs. 85:841–51. 16. Uziel T, Zindy F, Xie S, et al. The tumor suppressors Nat Rev Mol Cell Biol 2005;6:306–17. 9. Johnson RL, Rothman AL, Xie J, et al. Human homolog ink4c and p53 collaborate independently with patched 2. Ruiz i, Altaba A, Sanchez P, Dahmane N. Gli and of patched, a candidate gene for the basal cell nevus to suppress medulloblastoma formation. Genes Dev hedgehog in cancer: tumours, embryos and stem cells. syndrome. Science 1996;272:1668–71. 2005;19:2656–67. Nat Rev Cancer 2002;2:361–72. 10. Taylor MD, Liu L, Raffel C, et al. Mutations in sufu 17. Levy P, Vidaud D, Leroy K, et al. Molecular profiling 3. Wechsler-Reya R, Scott MP. The developmental predispose to medulloblastoma. Nat Genet 2002;31: of malignant peripheral nerve sheath tumors associated biology of brain tumors. Annu Rev Neurosci 2001;24: 306–10. with neurofibromatosis type 1, based on large-scale real- 385–428. 11. Zurawel RH, Allen C, Wechsler-Reya R, Scott MP, time RT-PCR. Mol Cancer 2004;3:20. 4. Berman DM, Karhadkar SS, Maitra A, et al. Wide- Raffel C. Evidence that haploinsufficiency of ptch leads 18. Rahnama F, Toftgard R, Zaphiropoulos PG. Distinct spread requirement for hedgehog ligand stimulation in to medulloblastoma in mice. Genes Chromosomes roles of ptch2 splice variants in hedgehog signalling. growth of digestive tract tumours. Nature 2003;425: Cancer 2000;28:77–81. Biochem J 2004;378:325–34. 846–51. 12. Zurawel RH, Allen C, Chiappa S, et al. Analysis of 19. Smyth I, Narang MA, Evans T, et al. Isolation and 5. Lum L, Beachy PA. The hedgehog response net- ptch/smo/shh pathway genes in medulloblastoma. characterization of human patched 2 (ptch2), a putative work: sensors, switches, and routers. Science 2004; Genes Chromosomes Cancer 2000;27:44–51. tumour suppressor gene in basal cell carcinoma and 304:1755–9. 13. Wetmore C, Eberhart DE, Curran T. The normal medulloblastoma on chromosome 1p32. Hum Mol 6. Ruiz i, Altaba A, Nguyen V, Palma V. The emergent patched allele is expressed in medulloblastomas from Genet 1999;8:291–7. design of the neural tube: prepattern, shh morphogen mice with heterozygous germ-line mutation of patched. 20. Sulman EP, Dumanski JP, White PS, et al. Identifica- and gli code. Curr Opin Genet Dev 2003;13:513–21. Cancer Res 2000;60:2239–46. tion of a consistent region of allelic loss on 1p32 in 7. Watkins DN, Berman DM, Burkholder SG, et al. 14. Goodrich LV, Milenkovic L, Higgins KM, Scott MP. meningiomas: correlation with increased morbidity. Hedgehog signalling within airway epithelial progeni- Altered neural cell fates and medulloblastoma in mouse Cancer Res 1998;58:3226–30. tors and in small-cell lung cancer. Nature 2003;422:313–7. patched mutants. Science 1997;277:1109–13. 21. Sulman EP, White PS, Brodeur GM. Genomic 8. Hahn H, Wicking C, Zaphiropoulous PG, et al. 15. Lee Y, Miller HL, Jensen P, et al. A molecular annotation of the meningioma tumor suppressor

Cancer Res 2006; 66: (14). July 15, 2006 6970 www.aacrjournals.org

locus on chromosome 1p34. Oncogene 2004;23: and disruption of granule cell development in a pre- polymerase causes medulloblastomas in p53(À/À) 1014–20. neoplastic stage of medulloblastoma. Development mice. Am J Pathol 2003;162:343–52. 22. Zaphiropoulos PG, Unden AB, Rahnama F, 2005;132:2425–39. 40. Lee Y, McKinnon PJ. DNA ligase iv suppresses Hollingsworth RE, Toftgard R. Ptch2, a novel human 31. Chiang C, Litingtung Y, Lee E, et al. Cyclopia and medulloblastoma formation. Cancer Res 2002;62:6395–9. patched gene, undergoing alternative splicing and up- defective axial patterning in mice lacking sonic 41. Pazzaglia S, Mancuso M, Atkinson MJ, et al. High regulated in basal cell carcinomas. Cancer Res 1999; hedgehog gene function. Nature 1996;383:407–13. incidence of medulloblastoma following X-ray-irradia- 59:787–92. 32. Litingtung Y, Lei L, Westphal H, Chiang C. Sonic tion of newborn ptc1 heterozygous mice. Oncogene 23. Frohlich L, Liu Z, Beier DR, Lanske B. Genomic hedgehog is essential to foregut development. Nat Genet 2002;21:7580–4. structure and refined chromosomal localization of the 1998;20:58–61. 42. Mancuso M, Pazzaglia S, Tanori M, et al. Basal cell mouse ptch2 gene. Cytogenet Genome Res 2002;97: 33. Belloni E, Muenke M, Roessler E, et al. Identification carcinoma and its development: insights from radiation- 106–10. of sonic hedgehog as a candidate gene responsible for induced tumors in ptch1-deficient mice. Cancer Res 24. Motoyama J, Heng H, Crackower MA, et al. Over- holoprosencephaly. Nat Genet 1996;14:353–6. 2004;64:934–41. lapping and non-overlapping ptch2 expression with shh 34. Roessler E, Belloni E, Gaudenz K, et al. Mutations in 43. Mertens F, Johansson B, Hoglund M, Mitelman F. during mouse embryogenesis. Mech Dev 1998;78:81–4. the human sonic hedgehog gene cause holoprosence- Chromosomal imbalance maps of malignant solid 25. Motoyama J, Takabatake T, Takeshima K, Hui C. phaly. Nat Genet 1996;14:357–60. tumors: a cytogenetic survey of 3185 neoplasms. Cancer Ptch2, a second mouse patched gene is co-expressed 35. Ming JE, Kaupas ME, Roessler E, et al. Mutations in Res 1997;57:2765–80. with sonic hedgehog. Nat Genet 1998;18:104–6. patched-1, the receptor for sonic hedgehog, are 44. Lothe RA, Andersen SN, Hofstad B, et al. Deletion of 26. Carpenter D, Stone DM, Brush J, et al. Characteriza- associated with holoprosencephaly. Hum Genet 2002; 1p loci and microsatellite instability in colorectal polyps. tion of two patched receptors for the vertebrate 110:297–301. Genes Chromosomes Cancer 1995;14:182–8. hedgehog protein family. Proc Natl Acad Sci U S A 36. Cooper AF, Yu KP, Brueckner M, et al. Cardiac and 45. Ohtsu K, Hiyama E, Ichikawa T, Matsuura Y, 1998;95:13630–4. cns defects in a mouse with targeted disruption of Yokoyama T. Clinical investigation of neuroblastoma 27. Wetmore C, Eberhart DE, Curran T. Loss of p53 but suppressor of fused. Development 2005;132:4407–17. with partial deletion in the short arm of chromosome 1. not arf accelerates medulloblastoma in mice heterozy- 37. Bitgood MJ, Shen L, McMahon AP. Sertoli cell Clin Cancer Res 1997;3:1221–8. gous for patched. Cancer Res 2001;61:513–6. signaling by desert hedgehog regulates the male germ- 46. Summersgill B, Goker H, Weber-Hall S, et al. 28. Pazzaglia S. Ptc1 heterozygous knockout mice as a line. Curr Biol 1996;6:298–304. Molecular cytogenetic analysis of adult testicular germ model of multi-organ tumorigenesis. Cancer Lett 2006; 38. Bajestan SN, Umehara F, Shirahama Y, et al. Desert cell tumours and identification of regions of consensus 234:124–34. hedgehog-patched 2 expression in peripheral nerves copy number change. Br J Cancer 1998;77:305–13. 29. Hahn H, Wojnowski L, Zimmer AM, et al. Rhabdo- during wallerian degeneration and regeneration. J 47. Ingram WJ, Wicking CA, Grimmond SM, Forrest AR, myosarcomas and radiation hypersensitivity in a mouse Neurobiol 2006;66:243–55. Wainwright BJ. Novel genes regulated by sonic hedgehog model of Gorlin syndrome. Nat Med 1998;4:619–22. 39. Tong WM, Ohgaki H, Huang H, et al. Null mutation of in pluripotent mesenchymal cells. Oncogene 2002;21: 30. Oliver TG, Read TA, Kessler JD, et al. Loss of patched DNA strand break-binding molecule poly(ADP-ribose) 8196–205.

www.aacrjournals.org 6971 Cancer Res 2006; 66: (14). July 15, 2006

Youngsoo Lee, Heather L. Miller, Helen R. Russell, et al.

Cancer Res 2006;66:6964-6971.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/66/14/6964

Cited articles This article cites 47 articles, 16 of which you can access for free at: http://cancerres.aacrjournals.org/content/66/14/6964.full#ref-list-1

Citing articles This article has been cited by 24 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/66/14/6964.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/66/14/6964. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.