Download - Th-MYCN Mice with Caspase-8 Deï¬ciency Develop Advanced

Tumor and Stem Cell Biology

Th-MYCN Mice with Caspase-8 Deficiency DevelopAdvanced Neuroblastoma with Bone Marrow Metastasis

Tal Teitz1, Madoka Inoue1, Marcus B. Valentine1, Kejin Zhu1, Jerold E. Rehg2, Wei Zhao3, David Finkelstein4,Yong-Dong Wang5, Melissa D. Johnson6, Christopher Calabrese6, Marcelo Rubinstein7, Razqallah Hakem8,William A. Weiss9, and Jill M. Lahti1,†

AbstractNeuroblastoma, the most common extracranial pediatric solid tumor, is responsible for 15% of all childhood

cancer deaths. Patients frequently present at diagnosis with metastatic disease, particularly to the bone marrow.Advances in therapy and understanding of the metastatic process have been limited due, in part, to the lack ofanimal models harboring bonemarrow disease. The widely used transgenicmodel, the Th-MYCNmouse, exhibitslimitedmetastasis to this site. Here, we establish the first genetic immunocompetent mousemodel for metastaticneuroblastoma with enhanced secondary tumors in the bone marrow. This model recapitulates 2 frequentalterations in metastatic neuroblasoma, overexpression of MYCN and loss of caspase-8 expression. Mousecaspase-8 gene was deleted in neural crest lineage cells by crossing a Th-Cre transgenic mouse with a caspase-8conditional knockout mouse. This mouse was then crossed with the neuroblastoma prone Th-MYCN mouse.Although overexpression of MYCN by itself rarely caused bone marrow metastasis, combining MYCN over-expression and caspase-8 deletion significantly enhanced bone marrow metastasis (37% incidence). Microarrayexpression studies of the primary tumorsmRNAs andmicroRNAs revealed extracellularmatrix structural changes,increased expression of genes involved in epithelial to mesenchymal transition, inflammation, and downregula-tion of miR-7a and miR-29b. These molecular changes have been shown to be associated with tumor progressionand activation of the cytokine TGF-b pathway in various tumor models. Cytokine TGF-b can preferentiallypromote single cell motility and blood-borne metastasis and therefore activation of this pathway may explain theenhanced bone marrow metastasis observed in this animal model. Cancer Res; 73(13); 4086–97. �2013 AACR.

IntroductionNeuroblastoma, a peripheral neural crest-derived child-

hood solid tumor, is a major medical challenge (1, 2). Half ofall patients with neuroblastoma have metastatic disease atdiagnosis, which carries a poor prognosis. Primary humanneuroblastoma tumors arise in the paraspinal ganglia or the

adrenal medulla, whereas disseminated disease appears inthe bone marrow (71% of patients), bones (56%), lymphnodes (31%), lungs (3%), and other internal organs (15%–45%). The new International Risk Group classification sys-tem of the disease divides the patients to 16 risk groups, withthe highest risk group being the one that presents withmetastasis to the bone marrow and has only 40%–50%survival rate (2, 3). The most commonly used preclinicaltransgenic mouse neuroblastoma model, the Th-MYCNmodel (4), exhibits a limited capacity for metastasis to thebone marrow (<5% incidence). In an attempt to establish animmunocompetent genetic metastatic model for neuroblas-toma, we crossed 2 genetically engineered mouse lines, eachwith a known molecular alteration, common to the aggres-sive disease, amplification of MYCN, and loss of expressionof caspase-8.

MYCN oncogene amplification is frequently seen in aggres-sive neuroblastoma and occurs in 25%–35% of human patients(5). Caspase-8 is a cysteine endoproteinase that cleaves peptidebonds after aspartic acids (6). In addition to its proapoptoticfunction as an initiator caspase in the extrinsic receptor-mediated death pathway (6), caspase-8 plays important rolesin mediating migration, adhesion, growth, immune response,differentiation, wound healing, fibrosis, and necroptosis incertain cell types (7–9). Suppression of caspase-8 expressionby epigenetic silencing occurs in approximately 70% of human

Authors' Affiliations: 1Departments of Tumor Cell Biology, 2Pathology,3Biostatistics, 4Computational Biology, 5Hartwell Center for Bioinformaticsand Biotechnology; and 6Animal Imaging Center, St. Jude Children'sResearch Hospital, Memphis, Tennessee; 7Instituto de Investigaciones enIngeniería Gen�etica y Biología Molecular, Consejo Nacional de Investiga-ciones Científicas y Tecnol�ogicas and Universidad de Buenos Aires,Buenos Aires, Argentina; 8Department of Medical Biophysics, OntarioCancer Institute, University of Toronto, Toronto, Canada; and 9Depart-ments of Neurology, Pediatrics and Neurological Surgery, University ofCalifornia, San Francisco, San Francisco, California

†Deceased.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

This article is dedicated to Dr. Jill M. Lahti who passed away on May 30,2012 after a long and courageous battle with cancer.

Corresponding Author: Tal Teitz, Department of Tumor Cell Biology, St.Jude Children's Research Hospital, Memphis, TN 38105. Phone: 901-595-2156; Fax: 901-595-2381; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-12-2681

�2013 American Association for Cancer Research.

CancerResearch

Cancer Res; 73(13) July 1, 20134086

on April 10, 2019. © 2013 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst March 27, 2013; DOI: 10.1158/0008-5472.CAN-12-2681

http://cancerres.aacrjournals.org/

neuroblastomas (10). Loss of caspase-8 has also been associ-ated with enhancement of the tumorigenic potential of SV40T-antigen transformed mouse embryonic fibroblasts (11) andwith providing an advantage in survival and metastasis ofengrafted neuroblastoma cell lines (12, 13). Nevertheless, therole of caspase-8 expression has not been tested thus far in vivoin an immunocompetent mouse model, which could havesignificance given the roles caspase-8 plays in the immunesystem.To circumvent the lethality in CASP8�/�mice, Salmena and

colleagues developed a conditional knockout mouse in whichLoxP siteswere introduced into theDNAflanking exons 3 and 4of the mouse caspase-8 gene (14). We mated these mice withTh-Cre transgenic mice, which express Cre recombinase onlyin the peripheral neural crest cells and in the brain catechol-aminergic neurons starting on day E9.5 (15), to selectivelydelete caspase-8 in the cells that give rise to neuroblastoma.Our preliminary results suggested that conditional knockout

of caspase-8 alone was insufficient to induce tumor formation.We studied 30 mice (129 � 1/SvJ background) harboring 2floxed caspase-8 alleles and Th-Cre for 6 to 9 months. Tumorsdid not develop in any of thesemice, suggesting that additionalgenetic alterations are needed. Thus, we tested the hypothesisthat the loss of caspase-8 facilitates MYCN-induced tumorformation or metastasis by using the Th-MYCN transgenicmouse neuroblastoma model (4). Because mouse geneticbackground influences tumor penetrance in this model (4),we backcrossed all mouse lines at least 6 generations (6–12) tothe 129� 1/SvJ background to ensure that differences in tumorformation were not due to strain variability.

Materials and MethodsMouse strainsTh-MYCN hemizygote mice were purchased from National

Cancer Institute (NCI; Bethesda, Maryland) mouse repository(strain #01XD2) on genetic background 129 � 1/SvJ and kepton this background. Floxed caspase-8 mice were received fromRazqallah Hakem (14) on 129� 1/SvJ, C57BL6 background andbackcrossed 6 to 12 generations to the 129�1/SvJ background.Th-Cre hemizygote mice were received from Dr. MarcelloRubinstein (15) on a B6.CBF2 background and backcrossed6 to 12 generations to the 129 � 1/SvJ background. This studywas carried out in strict accordancewith the instructions in theGuide to Care and Use of Laboratory Animals of the NIH(Bethesda, MD). The protocol was approved by the Institu-tional Animal Care and Use Committee at St. Jude Children'sResearch Hospital (IACUC protocol 420). All efforts were madeto minimize suffering.

Genotyping of mouse tissues and tumorsCaspase-8 primers:

Sense primer–50-CCAGGAAAAGATTTGTGTCTAGC-30

Antisense primer–50-GGCCTTCCTGAGTACTGTCACCTGT-30

PCR amplification of the wild-type caspase-8 allele gives a650-bp band; the caspase-8 floxed (exons 3 and 4) produces aband of 850 bp. Thus, the deleted cre recombinase digestedDNA band is 200 bp.

Th-Cre primers:

Sense primer–50-ATGTCCAATTTACTGACCTACAC-30

Antisense primer–50-CTAATCGCCATCTTCCAG-30

Th-MYCN primers:

Sense primer–50- CGACCACAAGGCCCTCAGTA-30

Antisense primer–50-CAGCCTTGGTGTTGGAGGAG-30

Quantitative RT-PCR for mouse caspase-8Total RNA was extracted from mouse tumors with Trizol

reagent (Life Technologies) and reverse transcribed withSuperscript II (Life Technologies). Primers for mouse cas-pase-8 were used to PCR amplify a 120-bp fragment from exon1 to exon 2 of the transcript, thus recognizing wild-typemessage and floxed caspase-8 message if stable. Primers used:sense primer–50–CCCTACAGGGTCATGCTCTT-30 antisenseprimer–50–CAGGCTCAAGTCATCTTCCA-30.

Antibodies and immunohistochemistryFor Western blots, we used antibodies: mouse caspase-8,

(Cell Signaling cat. no. 4927, Dil 1:1,000), MYCN (Cell Signalingcat. no. 9405, Dil 1:1,000), and Actin (Santa Cruz cat. no. 1616,Dil 1:2,000). For staining paraffin-embedded formalin-fixedtumors and tissues we used antibodies: caspase-8 1H11(Abcam cat. no. ab119892, dilution 1:200) and anti-caspase-3, Ki67, PGP9.5, synapthophysin, chromogranin A, and NFP asdescribed previously (16).

Ultrasound imagingUltrasound imaging of the mouse tumors was carried out as

described (16) using the VisualSonics VEVO-770 high frequen-cy ultrasound system (VisualSonics).

Microarray analysis of neuroblastoma tumor samplesTotal RNA was extracted from primary mouse tumors using

TRIzol reagent (Life Technologies). Samples were assayed withthe Affymetrix Mouse 430v2 GeneChip array and the Agilentmouse microRNA v18 array microRNA. Data was summarizedby the RMA protocol (17) using Partek Genomics suite 6.6.Outlier samples were detected and removed by principalcomponent analysis (PCA) and a batch correction correspond-ing to hybridization datewas applied. The datawere defined byclass and a series of unequal variance t tests were applied tocompare classes. Data was visualized and filtered by P value(0.05) and log ratio and submitted to Gene Set EnrichmentAnalysis (GSEA, www.broadinstitute.org/gsea) to assess GeneOntology (GO) enrichment. The mRNAs targeted by down-regulated microRNAs in the primary tumors samples withbone marrow metastasis were predicted using MirTarget2(18), and followed by an enrichment analysis using the Data-base for Annotation, Visualization and Integrated Discovery(DAVID) at the National Institute of Allergy and InfectiousDiseases at the National Institutes of Health (19). All array datahave been deposited in Gene Expression Omnibus at NationalCenter for Biotechnology Information, and are accessiblethrough GSE42548 (mRNA) and GSE42254 (microRNA). Select

Bone Marrow Metastasis in a Neuroblastoma Model

www.aacrjournals.org Cancer Res; 73(13) July 1, 2013 4087




array data were validated by quantitative real-time PCR (qRT-PCR) using TaqMan gene expression assays as described inSupplementary Fig. S1.

ResultsEstablishing a neuroblastoma mouse model with MYCNamplification and caspase-8 deficiency

Neural crest-specific deletion of caspase-8 in mice wasachieved by mating floxed and/or knockout caspase-8 alleles

mice with hemizygous Th-Cre mice and then with hemizygousTh-MYCN mice (Fig. 1A). Four genotypes of mice were estab-lished: caspase-8fl/fl (flox/flox), caspase-8fl/ko (flox/knockout),caspase-8wt/ko (wild-type/knockout), and the control cas-pase-8wt/wt (wild-type/wild-type). Cre-mediated caspase-8deletion in the caspase-8fl/fl primary tumors was assayedby genomic PCR (Fig. 1B, TU lanes). Caspase-8 was deletedin the primary tumors at varying efficiencies ranging from alow percentage of cells to complete deletion (Fig. 1B).

Figure 1. Characterization of the tumors formed in the Th-MYCN, Th-Cre, caspase-8 deleted mice (fl/fl, fl/ko, or wt/ko caspase-8). A, deficient caspase-8,Th-MYCN mice were obtained by mating floxed or heterozygous caspase-8 alleles mice with hemizygous Th-Cre mice and then with hemizygousTh-MYCN mice. B, validation of caspase-8 deletion status in the primary Th-MYCN, Th-Cre, caspase-8fl/fl neuroblastoma tumors using genomic PCR. TailDNA (TA) from mice with tumors and primay tumor DNA (TU) was amplified using primers flanking exons 3 and 4 of mouse caspase-8 gene. Only theCasp8fl bands were detected in the tail samples, whereas the Casp8fl and the deleted caspase-8 bands (Casp8del) were detected in varying levelsin the primary tumors. PCR amplification of the casp8fl3-4 allele (Casp8fl) generates a band of 850 bp and the deleted band (Casp8del) is 200 bp. C, Westernblot analyses of primary neuroblastoma tumors with an anti-mouse caspase-8 antibody. �/� mice are the Casp8 fl/fl and caspase8 fl/ko mice. D,qRT-PCR analysis for mouse caspase-8 was conducted on total RNA extracted from the primary tumors of the following Th-MYCN, Th-Cre mouse geneticgroups: wt/wt caspase-8, n ¼ 8; wt/ko caspase-8, n ¼ 5; �, P ¼ 0.007 compared with wt/wt, fl/ko caspase-8, n ¼ 6; �, P ¼ 0.003 compared with wt/wt;P values determined by t test. E, Kaplan–Meier survival curves of the Th-MYCN, Th-Cremicewith wt/wt caspase-8 (n¼ 85) or fl/fl caspase-8 (n¼ 90,P¼ 0.12compared with wt/wt), or fl/ko caspase-8 (n¼ 90, P¼ 0.30 compared with wt/wt), or caspase-8 heterozygous wt/ko (n¼ 43, P¼ 0.21 compared with wt/wt).P values were determined by Gehan–Breslow–Wilcoxon log-rank test.

Teitz et al.

Cancer Res; 73(13) July 1, 2013 Cancer Research4088




Caspase-8 expression was also assessed in the primarytumors at the mRNA and protein levels. Most Th-MYCN,Th-Cre, caspase-8wt/wt (labeled þ/þ) primary abdominaltumors (19/20) expressed endogenous caspase-8, as deter-mined by RNA protection (data not shown) and immuno-blot assays (Fig. 1C). Caspase-8 expression in the wt/ko(þ/�) tumors was reduced to approximately 50% of wildtype, whereas the Th-MYCN, fl/fl, and fl/ko caspase-8 allelestumors (labeled �/�) had lower caspase-8 expression,around 20% average, suggesting that the floxed allele ishypomorphic (Fig. 1C and D). Levels of caspase-8 protein inthe primary tumors was also determined by immunohis-tochemistry with anti-caspase-8 antibody and found to below or undetectable (Supplementary Fig. S2). The primary

deficient caspase-8 tumors stained for neuronal markerstypical to neuroblastoma as synapthophysin and PGP9.5(Supplementary Fig. S3) and were indistinguishable fromthe wild-type tumors in their incidence, latency, mass (Figs.1E, 2, 3A and Supplementary Fig. S4). In addition, nostatistical significant differences were found in the numberof proliferating cells in the primary tumors as determinedby Ki-67 staining (both groups had >95% positive cells, datanot shown), or in the number of primary tumor cellsundergoing apoptosis as determined by immunostainingwith cleaved caspase-3 (Supplementary Fig. S5). MYCNexpression comparing primary tumors with and withoutcaspase-8 was also not significantly changed (Supplemen-tary Fig. S6).

Figure 2. Caspase-8 deficiencydoes not change the incidence,latency, location, or growth rate ofthe mouse primary neuroblastomatumors. A, primary tumors could beidentified by ultrasound imagingstarting at mouse age 6 to 7 weeks.Tumors (red) arose in Th-MYCN,Th-Cre, wt/wt caspase-8, orTh-MYCN, Th-Cre, caspase-8–deficientmice (fl/ko and fl/fl caspase-8) in the paravertebral ganglia, eithersurrounding or near the aorta (purple;81 � 7% of the cases) or near thekidney (green) and adrenal gland (19� 7% of the cases); n¼ 28 for the wtcaspase-8 group and n ¼ 41 for thecaspase-8–deficient group. B,typical examples of preneoplasticislets observed in mice withcaspase-8 wt/wt, Th-MYCN (n ¼ 10)or deficient caspase-8, Th-MYCN(n¼9) at ages 10, 21, and49days areshown (H&E). No preneoplastic isletswere observed in mice withoutMYCN. Arrows point to LN, lymphnode; PN, preneoplastic lesions;BM, Bone marrow; A, aorta; VE,vertebra; GN, ganglia. Scale bar,50 mm.






Figure 3. Deficiency of caspase-8 in the Th-MYCNmousemodel enhancesmetastasis preferentially to theBM. A, no significant difference in themeanmass ofthe primary neuroblastoma tumors of the Casp8wt/wt, Th-MYCN, mouse group, and the caspase-8 deficient (Casp8fl/ko and Casp8fl/fl mice, labeled �/�),Th-MYCNmice at the time interval metastasis was assayed (14� 6-week-oldmice). Data are expressed asmean�SEM (n¼ 12 for bothmouse groups). B, asignificantly higher incidence of metastatic neuroblastoma was found in the BM of the caspase-8–deficient mouse group compared with the wtcaspase-8 group (n¼ 21, wt caspase-8mice and n¼ 28, caspase-8–deficent; �,P¼ 0.014; P determined by Fisher exact test). C, metastatic incidence to theovaries did not differ statistically between thewt caspase-8 group and the deficient caspase-8 group (n¼ 7,wt caspase-8;n¼ 9, caspase-8–deleted;P¼0.10by Fisher exact test). D, typical examples of H&E staining of the secondary tumors in the BM of caspase-8–deficient mice in long bones, such asthe femur, tibia, and humerus. The BM metastatic tumor cells stained positive for the neuronal markers synapthophysin and tyrosine hydroxylase. Areas oftumor cells in the BM are marked with arrows and asterisks indicate bones. Th-MYCN mouse BM with no metastasis served as control. Scale bar is50 mm except in images 1, 3, 4, and 6 from top left it is 200 mm. BM, bone marrow.

Teitz et al.





Growth characterization of the primary neuroblastomatumors in the genetically engineered caspase-8–deficientTh-MYCN mouse modelComparison of overall survival and primary tumor onset

between the differentmouse genetic groups, all on the 129� 1/SvJ background, andharboringwild-type or deficient caspase-8yielded no significant statistical difference (Fig. 1E and Sup-plementary Fig. S4) with the limitation that all mice had to besacrificed after 16 to 17 weeks due to high primary tumorburden. To determine whether there were any changes in thegrowth or location of the primary tumors before this point, wemonitored weekly the tumors in both caspase-8–expressingand caspase-8–deleted Th-MYCN mice by ultrasound imaging(Fig. 2A). No significant differenceswere found in the frequencyor initial location of the tumors. Primary tumors in all micegroups were located in areas surrounding the aorta (81� 7% inboth groups) or near the adrenal gland or kidney (19 � 7% inboth groups; Fig. 2A). In addition, we examined mice at earliertime points (ages 10, 21, and 49 days) to determine whetherthere were variations in the number of initiating preneoplastichyperplasia cells arising during development in the paraspinalganglia (Fig. 2B). These studies were based on previous workthat showed higher frequencies of hyperplastic cells in theparaspinal ganglia of heterozygous Th-MYCN mice comparedwith 129 � 1/SvJ wild-type mice littermates 1 to 5 weeks afterbirth (20, 21). These experiments as well did not reveal anydifferences in the appearance or incidence of hyperplasticneuroblasts between Th-MYCN mice with and without cas-pase-8 expression (Fig. 2B), Thus, we conclude that caspase-8deficiency does not significantly contribute to initial primarytumor formation.

Caspase-8–deficient Th-MYCN mice have preferentiallyenhancedneuroblastomametastasis to the bonemarrow

We then screened Th-MYCN mice harboring advanced-stage primary tumors (ages 9–17 weeks, Fig. 3A) from thewild-type and caspase-8–deficient groups for secondarymetastatic tumors to determine whether caspase-8 plays arole inmetastasis in vivo. Detailed necropsy of all major organswas carried out (Fig. 3B–D). In agreement with the datadescribed above, no statistical significant difference wasfound in the average size of the initial primary tumors (Fig.3A). We did detect, however, a significant difference in thefrequency of secondary tumors in the bone marrow (Fig. 3Band Table 1). Eighty sections were cut from various bones ofthe mice, including the sternum, the long bones, and thevertebra. Tumor cells were identified and the number ofbone marrow tumor foci was determined by hematoxylin andeosin (H&E) staining and immunohistochemistry with theneuroblastomamarkers synapthophysin, tyrosine hydroxylase,and/or PGP9.5 (Table 1). From 27 mice with deficiency incaspase-8 (16fl/ko and 11fl/fl, labeled�/�), 10mice (37%) hadmetastatic foci in their bonemarrowwithin 10 to 17weeks. Thesize of the foci ranged from clusters of 5 to 10 tumor cells tolarge sheets compromising up to a quarter of the bonemarrowcells' population. Twenty one mice with Caspase-8wt/wt werescreened for bone marrow metastasis and only one mouse(4.5% incidence) had bonemarrowmetastasis at weeks 10 to 17(P¼ 0.014, Fig. 3B andD andTable 1). Bonemarrow tumor cellswere negative for caspase-8 expression as determined byimmunostaining (Supplementary Fig. S2) and had very lowapoptosis levels (about 1%) measured by immunostaining forcleaved caspase-3 protein (Supplementary Fig. S5).

Table 1. Metastasis to the bone marrow in the Th-MYCN–deleted caspase-8 mice

Casp8 genotype Mouse number(s) Age (weeks) Sternum Long bones Vertebra

wt/wt 1–20 10–17 N N Nwt/wt 21 10 N 2a 1fl/ko 1–10 10–17 N N Nfl/ko 11 10 N N 1fl/ko 12 9 N N 1fl/ko 13 10 N N 1fl/ko 14 17 N 1a Nfl/ko 15 16 N 1 1fl/ko 16 10 N 1 Nfl/fl 1–7 10–17 N N Nfl/fl 8 10 N 1a Nfl/fl 9 11 N N 1fl/fl 10 10 1 6a 1a

fl/fl 11 11 N 1 N

NOTE: Mice were generated from at least 3 to 4 mating combinations during a 2-year period.n ¼ 21 for wt/wt casp-8 mice; n ¼ 16 for fl/ko casp-8 mice; and n ¼ 11 for fl/fl casp-8 miceAbbreviation: N, No tumor cells detected.aNeuroblastoma cells were confirmed by immunohistochemistry with synapthopysin, tyrosine hydroxylase, and/or PGP9.5 antibodiesandbymorphology after H&E staining. Numbers in table indicate independent foci ranging in size froma cluster of 5 to 10 tumor cells tolarge sheets of cells.






The incidence of metastatic dissemination to other organsof the mice, other than the bone marrow (Supplementary Fig.S7) was not significantly enhanced in the caspase-8–deficientgroup (shown for ovaries, Fig. 3C). The secondary metastaticlesions stained positive for neuronal markers typical for neu-roblastoma (Supplementary Fig. S7 shown for PGP9.5). Met-astatic ovaries were stained for cleaved caspase-3 and had 1%to 2% of cells undergoing apoptosis, equal to the percentage ofcells undergoing apoptosis in the matching primary tumors(Supplementary Fig. S5).

Array expression analysis of the caspase-8–deficientprimary tumors reveals changes in extracellularmatrix proteins and tumor cell–ECM interactingproteins

We compared the mRNA and microRNA expression profilesof a group of 6 primary tumors with wild type caspase-8 levelsin animals with no bone marrow metastasis to a group of 7primary tumors with deficient caspase-8 levels in animals thathad bone marrow metastasis. Heat map of the top 20 statis-tically significant genes that differ in expression between the 2groups are shown in Fig. 4A. The list included genes known tohave a role inmetastasis, epithelial-to-mesenchymal transition(EMT), cell detachment from the extracellular matrix (ECM),fibrosis, wound healing, and inflammation as Tfpi2 (tissuefactor pathway inhibitor-2, a serine proteinase inhibitor (22),Snai2 (snail homolog 2 known also as SLUG, a neural cresttranscription factor (23), Myct1, myc target 1, a direct c-myctarget gene (24), Serpinh1, a serine or cysteine peptidaseinhibitor known also as Hsp47, a collagen-specific molecularchaperone (25), Emcn, amucin-like sialglycoprotein that inter-feres with the assembly of focal adhesion complexes andinhibits interaction between cells and the ECM (26) and Fos(27, 28). The complete list of genes that are different betweenthe groups and had statistical significance above P¼ 0.05 is inSupplementary Table S1. The list included MMP-15, matrixmetallopeptidase 15 (29), 1.3-fold increase in the primarytumors with bone marrow metastasis. The complete list ofmRNAs was submitted to GSEA and the top enriched geneset was the ECM structural constituents with a nominalP value of 0.008 (Fig. 5). This gene set included upregulationin expression of Tfpi2, LAMA4 (laminin alpha 4; refs. 30, 31),FBLN2 (fibulin 2; ref. 32), PRELP (ECM protein that func-tions to anchor basement membranes to the underlyingconnective tissues; ref. 33), COL4A2 (collagen type 4 alpha2, (refs. 34, 35; Fig. 5). mRNA analysis was done also on agroup of 10 deficient caspase-8 mice that did not showmetastasis to bone marrow by histology (Fig. 4B). The topgenes that came up are in Fig. 4B and all the gene changeswith P < 0.05 are included in Supplementary Table S2. Thistumor group had 4 genes that overlapped with the meta-static group: Fos (downregulation), Lancl1 (downregula-tion), Emcn (upregulation), and Myct1 (upregulation), sug-gesting these gene expression changes occur before metas-tasis is observed in the bone marrow.

Analysis of microRNA expression was done on primarytumors with wild-type caspase-8 and no bone marrow metas-tasis compared with caspase-8–deficient primary tumors with

detected bone marrow metastasis (Supplementary Table S3).The top microRNAs changes (fold changes >30%) are shownin Fig. 6, and included downregulation of miR-29b (1.86 fold)and miR-7a (1.43-fold) expression in the deficient caspase-8group. Suppression of miR-29 by TGF-b1/Smad3 signaling hasbeen shown to promote collagen and other ECM componentsexpression and to promote renal fibrosis (36, 37). MiR-7 wasshown to be suppressed in human neuroblastoma (38), breastcancer, and glioblastoma and its downregulation was associ-ated with tumor metastasis (38, 39). Its forced expression intumor cells inhibited EMT transition and metastasis of breastcancer cells via targeting focal adhesion kinase expression (39).The mRNAs targeted by the downregulated microRNAs in theprimary tumors with bone marrow metastasis in this studywere predicted byMirTarget2 and analyzed byGO enrichment.Interestingly, this analysis showed highly enriched expression(P < 10�6) for the identical gene set that came up in the mRNAanalysis, the gene set of the ECM structural components ascollagens and laminins.

DiscussionTumor suppressor genes are defined as genes whose loss-

of-function in tumor cells contribute to the formation and/or maintenance of the tumor phenotype. The findings pre-sented in this report provide proof that caspase-8 functionas a metastatic bone marrow tumor suppressor gene inneuroblastoma. Loss of caspase-8 expression does not affectprimary tumor formation in the Th-MYCN mouse but it doespromote selective metastasis formation and maintenance inthe bone marrow. These results are in accordance withprevious studies in mice that showed that a deficiency ofTRAIL-R, a protein at the apex of the caspase-8–signalingpathway, enhances metastasis of squamous cell carcinomato lymph nodes without affecting primary tumor develop-ment (40). It also supports the recent clinical findings that alack of caspase-8 correlates with relapse in human neuro-blastoma patients evident by bone marrow metastatic dis-ease (10).

Caspase-8 suppression by epigenetic silencing has alsobeen reported in other human tumors including small celllung carcinoma, primitive neuroectodermal tumors, alveolarrhabdomyosarcoma, medulloblastoma, and retinoblastoma(41). It will be of great interest to investigate whethercaspase-8 loss can also enhance bone marrow metastasisin these tumors.

Deletion of caspase-8 in the mouse primary tumor in thisanimal model was a driver genetic event that led to approx-imately 7-fold increase in Th-MYCN-induced preferentialmetastasis incidence to the bone marrow (from 5% averageincidence to 37%). Interestingly, metastasis to other organsincluding the abdominal and pulmonary lymph nodes orlungs was not significantly different. This is in contrast towhat we previously found by engrafting human cells in thechick embryo or injecting human neuroblastoma tumorcells deficient in caspase-8 expression directly to the bloodstream of immunodeficient mice (12). In these experiments,metastasis incidence of cells with decreased caspase-8expression was increased to both the bone marrow and to

Teitz et al.





the lungs at the same frequency. In the animal model studiedin this work, we concentrated on the effect of the developingprimary tumors on metastasis and determined the transcrip-tional changes occurring at the primary tumor before andafter metastasis are detected in the bone marrow. Although

we did not find statistical significant changes in the percent-age of cells undergoing apoptosis in the primary tumors, wefound changes in the ECM structure of the caspase-8–defi-cient primary tumors as upregulated expression of collagen4A2 and laminin a4 once metastasis is detected in the bone

Figure 4. Heat map of the top20 genes that are differentiallyexpressed in the primary tumors(P < 0.05).






marrow. These ECM changes are likely to cause increasedstiffness of the primary tumor and changes in mechano-transduction properties, which have been shown in differenttumor types to correlate with advanced stage of disease (42–44). In addition, we see transcriptional changes that wouldsuggest increased motility and migration of the caspase-8–deficient tumor cells by upregulated expression of genesinvolved in EMT (as Snai2, Twist1, and TfpI2), enhanceddetachment of the tumor cells from the ECM (effected byEmcn, PRELP, miR-7), and increased fibrosis (accumulationof ECM proteins and downregulation of miR-29b). Interest-ingly, EMT changes have been observed recently in vivo in abreast cancer animal model specifically when the oncogenemyc was amplified (45). ECM constituents changes as accu-mulation of collagens and laminins has been described infibrosis of tissues and tumors (9, 34, 35, 37) and in woundresponse processes (8). Caspase-8 downregulation has beenlinked to wound processes in vivo in which accumulation ofcollagen and other ECM structural proteins occur (8) andfibrosis is seen in mice that have deficient caspase-8 in theirepidermal tissues (9). In addition, caspase-8 has direct inter-action with the ECM proteins by being in complexes withintegrins (12, 46) and as a part of the focal adhesion complex(7). Deficiency of caspase-8 in the primary tumors thus couldcause changes in the ECM structure and/or possible activa-

tion of a wound-like process that triggers deposition of ECMproteins by fibroblasts in the stroma and can activate pro-duction of various cytokines. The cytokine TGF-b is one ofthe major cytokines to be activated in response to wound/injury processes (47), fibrosis (36, 37), or as direct changes inthe stiffness and mechanotransduction properties of theECM (48). Importantly for the bone marrow preferentialmetastasis, TGF-b has been shown by intravital imagingexperiments to be transiently and locally activated in breastcancer motile cells and switch the cells from cohesive tosingle cell motility (49). Cells restricted to collective invasionwere capable of lymphatic invasion but not blood bornemetastasis (49). Thus transient activation of TGF-b prefer-entially in the caspase-8 primary tumors as result of ECMremodeling and/or fibrosis can potentially promote singlecell motility, increase invasion to the blood vessels, andenhance bone marrow metastasis. Interestingly, we see upre-gulated expression in the caspase-8–deficient primarytumors of genes known to be induced by TGF-b as Tgif2and Tgfb1i1 (Supplementary Table S1) and downregulation ofmiR-29b associated with TGF-b activation (36, 37). Intravitalimaging experiments in our caspase-8–deficient mouse mod-el could shed light whether indeed increased blood bornemetastasis of single cells contributes to the preferred metas-tasis to the bone marrow.

Figure 5. GSEA enrichment plot of GO term ECM structural constituent. Genes in the GO term ECM constituent showed significant enrichment in BMmetastasis versus no BMM samples. The top portion of the figure plots the enrichment score for each gene, and the bottom portion shows the valuesof the ranking metric moving down the list of the ranked genes. The table indicates that the majority of the genes in the term were significantly enriched andupregulated in the caspase-8–deficient BMM samples.

Teitz et al.





In this work, we observed downregulation of microRNA-7expression in the caspase-8–deficient mouse primary tumors,which has been described recently in human neuroblastomaand was associated with metastatic advanced stage (38).Finally, our finding that the loss of caspase-8 in the mouse

primary tumor cells significantly promotes metastasis tothe bone marrow, the most common site for metastasisin human neuroblastoma (71% of patients at diagnosis;refs. 1–3), indicates that this Th-MYCN/caspase-8–deletedanimal model should be useful for testing therapies formetastatic neuroblastoma. Ongoing experiments in ourlaboratory are aimed at purifying the metastatic bone mar-row cells for gene expression analysis and surgically resect-ing or debulking the primary tumors to allow even furtherprogression of the metastatic process in the bone marrow.Our preliminary experiments also indicate the feasibility ofestablishing orthotopic allograft models using this geneti-cally engineered model by passaging the primary tumor cellsfrom mice to mice and thus establishing uniform animalcohorts suitable for drug screenings (16). Labeling theprimary and secondary metastatic tumor cells in vivo inthis animal system by breeding to a fluorescence and/or aluciferase mouse reporter line in which expression of thereporter gene is cre-recombinase mediated (50) is currentlyin progress and should facilitate the monitoring of tumorcell growth and responses to therapeutic modalities.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: T. Teitz, M. B.Valentine, J.E. Rehg, J.M. LahtiDevelopment ofmethodology:T. Teitz, M. Inoue, M. B. Valentine, J.E. Rehg, M.D. Johnson, C. Calabrese, J.M. LahtiAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): T. Teitz, M. Inoue, M. B.Valentine, Y.-D. Wang, M.D.Johnson, C. Calabrese, M. Rubinstein, R. Hakem, W.A. WeissAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): T. Teitz, M. Inoue, M. B. Valentine, J.E.Rehg, W. Zhao, D. Finkelstein, Y.-D. Wang, C. Calabrese, J.M. LahtiWriting, review, and/or revision of the manuscript: T. Teitz, J.E. Rehg, W.Zhao, Y.-D. Wang, C. Calabrese, R. Hakem, W.A. Weiss, J.M. LahtiAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): K. Zhu, M.D. Johnson, Y.-D. Wang,C. CalabreseStudy supervision: T. Teitz, J.M. Lahti

AcknowledgmentsThe authors thank all members of Dr. Jill Lahti's laboratory for discussion,

Joshua Stokes from St. Jude BMC and Judith Hyle for help with preparing thefigures, the staff at the St. Jude Veterinary Pathology Core including Sean Savage,Patricia J. Varner, Dorothy Bush, and Pamela Johnson, the staff at the St. JudeHartwell Center for Bioinformatics and Biotechnology including Dr. GeoffreyNeale, Ridout Granger, Deanna Naeve, Melanie Loyd, and John Morris, and thestaff at St. Jude Animal Imaging Shared Resource for assistance.

Grant SupportThe work was funded by NIH grant RO1 CA067938 (J.M. Lahti), Comprehen-

sive Cancer Center Support Grant P30 CA021765, the American LebaneseAssociate Charities (ALSAC), and NCI grants PO1 CA81403 and RO1 CA102321(W.A. Weiss.)

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicate thisfact.

Received July 6, 2012; revised March 13, 2013; accepted March 14, 2013;published OnlineFirst March 27, 2013.

Figure 6. microRNAs that aredifferentially expressed in thecaspase-8–deficient primary tumors.






References1. Maris JM. Recent advances in neuroblastoma. N Engl J Med

2010;362:2202–11.2. Owens C, Irwin M. Neuroblastoma: the impact of biology and coop-

eration leading to personalized treatments. Crit Rev Clin Lab Sci2012;49:85–115.

3. Cohn SL, Pearson AD, London WB, Monclair T, Ambros PF, BrodeurGM, et al. The International RiskGroup (INRG) classification system: anINRG Task Force report. J Clin Oncol 2009;27:289–97.

4. Weiss WA, Aldape K, Mohapatra G, Feuerstein BG, Bishop JM.Targeted expression of MYCN causes neuroblastoma in transgenicmice. EMBO J 1997;16:2985–95.

5. Brodeur GM.Neuroblastoma: biological insights into a clinical enigma.Nat Rev Cancer 2003;3:203–16.

6. Muzio M, Chinnaiyan AM, Kischkel FC, O'Rourke K, Shevchenko A, NiJ, et al. FLICE, a novel FADD-homologous ICE/CED-3-like protease, isrecruited to the CD95 (Fas/Apo-1) death-inducing signaling complex.Cell 1996;85:817–27.

7. Teitz T, Jiang M, Twitchell D, Lahti JM. Caspase-8 roles in cancerand development—Lessons in multitasking. In Kevin V. Urbano, ed.Advances in Genetic Research, vol. 9. Hauppauge, NY: Nova Publish-ers; 2012; pp. 55–82.

8. Lee P, Lee D, Chan C, Chen S, Ch'en I, Jamora C. Dynamic expressionof epidermal caspase 8 simulates a wound healing response. Nature2009;458:519–24.

9. Li C, Lasse S, Lee P, Nakasaki M, Chen S, Yamasaki K, et al.Development of atopic dermatitis-like skin disease from the chronicloss of epidermal caspase-8. Proc Natl Acad Sci U S A 2010;107:22249–54.

10. Grau E, Martinez F, Orellana C, Canete A, Ya~nez Y, Oltra S, et al.Hypermethylation of apoptotic genes as independent prognosticfactor in neuroblastoma disease. Mol Carcinog 2011;50:153–62.

11. Krelin Y, Zhang L, Kang TB, Appel E, Kovalenko A, Wallach D.Caspase-8 deficiency facilitates cellular transformation in vitro. CellDeath Differ 2008;15:1350–5.

12. Stupack DG, Teitz T, Potter MD, Mikolon D, Houghton PJ, Kidd VJ,et al. Potentiation of neuroblastoma metastasis by loss of caspase-8.Nature 2006;439:95–9.

13. Ragu _enez G, M€uhlerthaler-Mottet A, Meier R, Duros C, B�enard J,Gross N. Fenretinide-induced caspase-8 activation and apoptosis inan established model of metastatic neuroblastoma. BMC Cancer2009;9:97.

14. Salmena L, Lemmers B, Hakem A, Matysiak-Zablocki E, Murakami K,Au PY, et al. Essential role for caspase 8 in T-cell homeostasis and T-cell mediated immunity. Genes Dev 2003;17:883–95.

15. Gelman DM, Noain D, Avale ME, Otero V, Low MJ, Rubinstein M.Transgenic mice engineered to target Cre/LoxP-mediated DNArecombination into catecholaminergic neurons. Genesis 2003;36:196–202.

16. Teitz T, Stanke JJ, Federico S, Bradley CL, Brennan R, Zhang J, et al.Preclinical models for neuroblastoma: establishing a baseline fortreatment. PLoS ONE 2011;6:e19133.

17. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ,Scherf U, et al. Exploration, normalization, and summaries of highdensity oligonucleotide array probe level data. Biostatistics 2003;4:249–64.

18. Wang X. miRDB: a microRNA target prediction and functional anno-tation database with a wiki interface. RNA 2008;14:1012–7.

19. Dennis G, Sherman BT, Hosack DA, Yang J, Gao W, Lane HC, et al.DAVID: database for annotation, visualization, and integrated discov-ery. Genome Biol 2003;4:3.

20. Hansford LM, Thomas WD, Keating JM, Burkhart CA, Peaston AE,Norris MD, et al. Mechanisms of embryonal tumor initiation: distinctroles for MycN expression and MYCN amplification. Proc Natl AcadSci U S A 2004;101:12664–9.

21. AlamG, Cui H, Shi H, Yang L, Ding J,Mao L, et al. MYCN promotes theexpansion of Phox2B-positive neuronal progenitors to drive neuro-blastoma development. Am J Pathol 2009;175:856–66.

22. NeaudV,HisakaT,Monvoisin A,BedinC,BalabaudC, Foster DC, et al.Paradoxical pro-invasive effect of the serine proteinase inhibitor tissue

factor pathway inhibitor-2 on human hepatocellular carcinoma cells.J Biol Chem 2000;275:35565–9.

23. Vitali R, Mancini C, Cesi V, Tanno B, Mancuso M, Bossi G, et al. Slug(SNAI2) down-regulation byRNA interference facilitates apoptosis andinhibits invasive growth in neuroblastoma preclinical models. ClinCancer Res 2008;14:4622–30.

24. Rogulski KR, Cohen DE, Corcoran DL, Benos PV, Prochownik EV.Deregulation of common genes by c-Myc and its direct target, MT-MC1. Proc Natl Acad Sci U S A 2005;102:18968–73.

25. Yang Q, Liu S, Tian Y, Hasan C, Kersey D, Salwen HR. Methylation-associated silencing of the heat shock Protein 47 gene in humanneuroblastoma. Cancer Res 2004;64:4531–8.

26. Kinoshita M, Nakamura T, Ihara M, Haraguchi T, Hiraoka Y, Tashiro K.Identification of human endomucin-1 and -2 as membrane-bound O-sialoglycoproteins with anti-adhesive activity. FEBS Lett 2001;499:121–6.

27. Ordway JM, Williams K, Curran T. Transcription repression in onco-genic transformation: common targets of epigenetic repression incells transformed by Fos, Ras or Dnmt1. Oncogene 2004;23:3737–48.

28. Ostler KR,YangQ, LooneyTJ, ZhangL, VasanthakumarA, TianY, et al.Truncated DNMT3B isoform DNMT3B7 supresses growth, inducesdifferentiation, and alters DNA methylation in human neuroblastoma.Cancer Res 2012;72:4714–23.

29. TaoG, Levay AK,Gridley T, Lincoln J.Mmp15 is a direct target of Snai1during endothelial to mesenchymal transformation and endocardialcushion development. Dev Biol 2011;359:209–21.

30. Sprenger CCT, Drivdahl RH, Woodke LB, Eyman D, Reed MJ, CarterWG, et al. Senescence-induced alterations of laminin chain expressionmodulate tumorigenicity of prostate cancer cells. Neoplasia 2008;10:1350–61.

31. Huang X, Ji G, Wu Y, Wan B, Yu L. LAMA4, highly expressed in humanhepatocellular carcinoma from Chinese patients, is a novel marker oftumor invasion and metastasis. J Cancer Res Clin Oncol 2008;134:705–14.

32. LawEWL,CheungAKL, KashubaVI, PavlovaTV, ZabarovskyER, LungHL, et al. Anti-angiogenic and tumor-suppressive roles of candidatetumor-suppressor gene, Fibulin-2 in nasopharyngeal carcinoma.Oncogene 2012;31:728–38.

33. Happonen KE, Furst CM, Saxne T, Heinegard D, Blom AM. PRELPprotein inhibits the formation of the complement membrane attackcomplex. J Biol Chem 2012;287:8092–100.

34. Bianchini G, Qi Y, Alvarez RH, Iwamoto T, Coutant C, IbrahimNK, et al.Molecular anatomy of breast cancer stroma and its prognostic value inestrogen-receptor positive and –negative cancers. J Clin Oncol 2010;28:4316–23.

35. Stewart DA, Cooper CR, Sikes RA. Changes in extracellular matrix(ECM) and ECM-associated proteins in the metastatic progression ofprostate cancer. Reprod Biol Endocrinol 2004;7:2–15.

36. Qin W, Chung ACK, Huang XR, Meng X, Hui DSC, Yu C, et al. TGF-b/Smad3 signaling promotes renal fibrosis by inhibiting miR-29. J AmSoc Nephrol 2011;22:1462–74.

37. Wang B, Komers R, Carew R, Winbanks CE, Xu B, Herman-EdelsteinM, et al. Suppression of microRNA-29 expression by TGFb-1 pro-motes collagen expression and renal fibrosis. J Am Soc Nephrol2012;23:252–65.

38. Guo J, Dong Q, Fang Z, Chen X, Lu H, Wang K, et al. Identification ofmicroRNAs that are associated with tumor metastasis in neuroblas-toma. Cancer Biol Ther 2010;9:446–52.

39. Kong X, Li G, Yuan Y, He Y, Wu X, Zhang W, et al. MicroRNA-7inhibits epithelial-to-mesenchymal transition and metastasis ofbreast cancer cells via targeting FAK expression. PLoS ONE2012;7:e41523.

40. Grosse-WildeA, VoloshanenkoO,BaileySL, LongtonGM,SchaeferU,Csernok AI, et al. TRAIL-R deficiency in mice enhances lymph nodemetastasis without affecting primary tumor development. J Clin Invest2008;118:100–10.

41. Fulda S. Caspase-8 in cancer biology and therapy. Cancer Lett2009;281:128–33.

Teitz et al.





42. Lu P, Weaver VM, Werb Z. The extracellular matrix: a dynamic niche incancer progression. J Cell Biol 2012;196:395–406.

43. Butcher DT, Alliston T, Weaver VM. A tense situation: forcing tumourprogression. Nat Rev Cancer 2009;9:108–22.

44. JaaloukDE, Lammerding J.Mechanotransduction gone awry. Nat RevMol Cell Biol 2009;10:63–73.

45. Trimboli AJ, Fukino K, Bruin A, Wei G, Shen L, Tanner SM, et al. Directevidence for epithelial-mesenchymal transitions in breast cancer.Cancer Res 2008;68:937–45.

46. Lahti JM, Teitz T, Stupack DG. Does integrin-mediated celldeath confer tissue tropism in metastasis? Cancer Res 2006;66:5981–4.

47. Anscher MS. Targeting the TGF-b1 pathway to prevent normal tissueinjury after cancer therapy. Oncologist 2010;15:350–9.

48. Maeda T, Sakabe T, Sunaga A, Sakai K, Rivera AL, Keene DR, et al.Conversion of mechanical force into TGF-b-mediated biochemicalsignals. Curr Biol 2011;21:933–41.

49. Giampieri S,ManningC,Hooper S, Jones L, Hill CS, Sahai E. Localizedand reversible TGFbeta signaling switches breast cancer cells fromcohesive to single cell motility. Nat Cell Biol 2009;11:1287–96.

50. Madisen L, Zwingman TA, Sunkin SM, Oh SW, Zariwala HA, Gu H,et al. A robust and high-throughout Cre reporting and characteri-zation system for the whole mouse brain. Nat Neurosci 2010;13:133–40.






2013;73:4086-4097. Published OnlineFirst March 27, 2013.Cancer Res Tal Teitz, Madoka Inoue, Marcus B. Valentine, et al. Neuroblastoma with Bone Marrow MetastasisTh-MYCN Mice with Caspase-8 Deficiency Develop Advanced

Updated version

10.1158/0008-5472.CAN-12-2681doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2013/03/22/0008-5472.CAN-12-2681.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/73/13/4086.full#ref-list-1

This article cites 49 articles, 19 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/73/13/4086.full#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/73/13/4086To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-12-2681

http://cancerres.aacrjournals.org/content/suppl/2013/03/22/0008-5472.CAN-12-2681.DC1

http://cancerres.aacrjournals.org/content/73/13/4086.full#ref-list-1

http://cancerres.aacrjournals.org/content/73/13/4086.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/73/13/4086


Download - Th-MYCN Mice with Caspase-8 Deï¬ciency Develop Advanced

Top Related