A Cre-conditional MYCN-driven neuroblastoma mouse model as an improved tool for preclinical studies

Abstract

Neuroblastoma, a childhood cancer that originates from neural crest-derived cells, is the most common deadly solid tumor of infancy. Amplification of the MYCN oncogene, which occurs in approximately 20-25% of human neuroblastomas, is the most prominent genetic marker of high-stage disease. The availability of valid preclinical in vivo models is a prerequisite to develop novel targeted therapies. We here report on the generation of transgenic mice with Cre-conditional induction of MYCN in dopamine β-hydroxylase-expressing cells, termed LSL-MYCN;Dbh-iCre. These mice develop neuroblastic tumors with an incidence of >75%, regardless of strain background. Molecular profiling of tumors revealed upregulation of the MYCN-dependent miR-17-92 cluster as well as expression of neuroblastoma marker genes, including tyrosine hydroxylase and the neural cell adhesion molecule 1. Gene set enrichment analyses demonstrated significant correlation with MYC-associated expression patterns. Array comparative genome hybridization showed that chromosomal aberrations in LSL-MYCN;Dbh-iCre tumors were syntenic to those observed in human neuroblastomas. Treatment of a cell line established from a tumor derived from a LSL-MYCN;Dbh-iCre mouse with JQ1 or MLN8237 reduced cell viability and demonstrated oncogene addiction to MYCN. Here we report establishment of the first Cre-conditional human MYCN-driven mouse model for neuroblastoma that closely recapitulates the human disease with respect to tumor localization, histology, marker expression and genomic make up. This mouse model is a valuable tool for further functional studies and to assess the effect of targeted therapies.

Figure 1

Generation of transgenic LSL-MYCN mice. (a) Graphical representation of the ROSA26 locus with recombinase-mediated cassette exchange (RMCE) sites used to introduce the RMCE exchange vector containing the MYCN transgene. The Rosa26 locus is displayed before (top) and after (center) insertion of MYCN by RMCE, and after cre-recombinase-mediated removal of the transcription termination site 5′ to the MYCN allele (bottom). Localizations of primers used for genotyping (A1 and A2) and the PCR-based validation of floxing out the transcriptional site 5′ of the MYCN allele (B1 and B2) are displayed. Splice acceptor site (SA), polyadenylation signal (pA), internal ribosome entry site (IRES), chicken actin gene promotor (CAG), transcriptional STOP cassette made of the human Growth Hormone polyadenylation signal (hGHpA), human MYCN open reading frame (MYCN). (b) Representative genotyping PCR validating the MYCN knock-in allele in heterozygous and homozygous LSL-MYCN mice (primers used: A1 and A2); wild type (wt), heterozygous LSL-MYCN (+/−), homozygous LSL-MYCN (+/+). (c) Representative PCR validating absence or presence of the transgene inserted into the ROSA26 locus in wt, heterozygous (+/−) and homozygous (+/+) LSL-MYCN mice.

Figure 2

Double-transgenic LSL-MYCN;Dbh-iCre mice develop tumors derived from the neural crest. (a) Kaplan–Meier analysis indicating the presence and the time to detection of palpable tumors in mice (that is, tumor-free survival) heterozygous for LSL-MYCN and mice double transgenic for LSL-MYCN and Dbh-iCre (Log-rank test). (b) Representative result of PCR validating the removal or presence of the transcriptional termination site 5′ to the MYCN transgene in tumor and control tissues, respectively. Wild type (wt), double-transgenic LSL-MYCN;Dbh-iCre (+/−). (c) Bioluminescence imaging of three representative LSL-MYCN;Dbh-iCre mice carrying palpable tumors at the superior cervical ganglion (I), adrenals (I, II, III) or celiac ganglion (III). Color code indicates luciferase activity (low=blue; high=red). (d) High frequency ultrasound images of palpable tumors arising from superior cervical ganglion (left) and adrenal (middle), and three-dimensional reconstruction of adrenal tumor (right). (e) Growth curves of tumors, as detected by high-frequency ultrasound. (f) Macroscopic images during autopsy of mice carrying palpable tumors arising from both adrenals and the celiac ganglion (left) and from the superior cervical ganglion (right).

Figure 3

Tumors of LSL-MYCN;Dbh-iCre mice resemble human neuroblastoma in terms of histology and molecular expression patterns. (a) MYCN expression (qPCR) in four representative tumors from LSL-MYCN;Dbh-iCre mice compared with control tissues. Expression was normalized to normal adrenal glands. Student's t-test: ***P<0.001. (b) Western blot analysis confirms MYCN expression in tumors (tu) compared with heart (he) tissue collected from four representative double-transgenic mice. (c) Hematoxylin and eosin (H&E) staining shows small, round blue cells typical for neuroectodermal tumors. Scale bars=100 μm (left) and 50 μm (right). (d) Electron micrographs show neuronal structures, including neurosecretory vesicles (red arrows). Scale bar=500 nm. (e) Reverse transcription-qPCR confirms significantly increased expression of the murine orthologs of the human neuroblastoma marker genes dopamine β-hydroxylase (Dbh), tyrosine hydroxylase (Th) and paired-like homeobox 2b (Phox2b) in tumors compared with normal control tissues (Student's t-test; Dbh: P=0.005, Th: P=0.02, Phox2b: P=0.0006). (f, g) Immunohistochemistry confirms expression of neuroblastoma markers, Ncam1 and Th. Scale bars=200 μm.

Figure 4

Murine neuroblastomas recapitulate genomic aberrations of human neuroblastomas. (a) Mouse karyotype overview of all genomic imbalances detected in 13 murine neuroblastomas (green bars: gained regions, red bars: lost regions). (b) Partial ratio plot for the mouse chromosome 6 region encompassing the ROSA26 amplicon in tumor 9 (right) and copy number of the MYCN transgene in 13 tumors from heterozygous LSL-MYCN;Dbh-iCre mice (+/−), as assessed by qPCR (insert left). Bars represent mice with normal chromosome 6 copy number (white), with whole chromosome 6 gain (black) and with focal chromosome 6 amplification (striped). (c) Tumor/control ratio plot for mouse chromosome 11 in tumor 9 showing partial chromosome 11q gain, corresponding to gain of almost the entire human chromosome 17.

Figure 5

Tumors from heterozygous LSL-MYCN;Dbh-iCre mice recapitulate human neuroblastoma at transcriptional level. (a) Table of selected gene sets from the MSigDB C2 collection, enriched among genes upregulated in the tumors from heterozygous LSL-MYCN;Dbh-iCre mice based on GSEA. (Rank of gene set in overall list of gene sets, ranked according to decreasing normalized enrichment score (NES; rank), number of genes in each set (size), NES). (b) GSEA enrichment plots showing upregulation of a gene set representing cell cycle (I) and markers downregulated during neuronal differentiation (II) in the transcriptional profiles of neuroblastoma tumors from heterozygous LSL-MYCN;Dbh-iCre mice. Depicted is the plot of the running sum for the MSigDB gene set within the LSL-MYCN;Dbh-iCre neuroblastoma data set, including the maximum enrichment score and the leading edge subset of enriched genes. FDR = false discovery rate. (c) The LSL-MYCN;Dbh-iCre signature score in the 967 cell lines in the Cancer Cell Line Encyclopedia showing the highest signature score in neuroblastoma cell lines, followed by medulloblastoma cell lines. (d) The MYCN mRNA (I) and miRNA (II) gene signature in normal adrenal medulla and MYCN- and ALK^F1174L-driven tumors. P<0.05 (*) was considered significant. (e) The cumulative distribution of the significance score [-10log(pfp)] associated with differential expression in tumors from heterozygous LSL-MYCN;Dbh-iCre mice (+/−) versus normal adrenals, for genes in the human non-MYCN-amplified neuroblastoma signature (black) and all other genes (gray). Genes in the human non-MYCN-amplified neuroblastoma signature show more significant differential expression compared with all remaining genes (Kolmogorov–Shmirnov test, P<0.001).

Figure 6

(a) Macroscopic images of cells cultivated after explantation of LSL-MYCN;Dbh-iCre tumors. Scale bars=100 μm. (b) Representative genotyping PCR validating the MYCN knock-in allele (left) and the Dbh-iCre transgene (right) in cells cultured from murine neuroblastoma. Heart (he) from wild-type (wt) and heart (he) and tumor (tu) from heterozygous LSL-MYCN;Dbh-iCre mice (+/-) as controls. (c) PCR validating the removal of the transcriptional termination site 5′ of the MYCN allele in cells cultivated after explantation of LSL-MYCN;Dbh-iCre tumors. Wild-type (wt), heterozygous LSL-MYCN;Dbh-iCre (+/−), heart (he), tumor (tu). (d) Bioluminescence imaging of mNB-A1 cells. Luciferase activity: low=blue; high=red. luciferin (luc). (e) MYCN expression (qPCR) in mNB-A1 cells compared with various control tissues and to a representative LSL-MYCN;Dbh-iCre tumor. Expression was normalized to that in normal adrenal glands. Student's t-test: ***=P<0.001; NS=not significant. (f) Western blot analysis confirms MYCN expression in mNB-A1 cells compared with heart and LSL-MYCN;Dbh-iCre tumor. (g) Tumor growth after engraftment of 10⁷ mNB-A1 cells into three nude mice at day 0. (h) Bioluminescence imaging of mNB-A1 cells growing in nude mice. Luciferase activity: low=blue, high=red.

Figure 7

Treatment of mNB-A1 and re-grafted tumors. (a) JQ1 treatment of mNB-A1 cells significantly reduced cell viability in MTT assays compared with untreated or DMSO-treated cells. (b) Western blot analysis confirmed no MYCN protein regulation in mNB-A1 cells treated with JQ1 compared with DMSO-treated cells. (c) MLN8237 treatment of mNB-A1 cells significantly reduced cell viability in MTT assays compared with untreated or DMSO-treated cells. (d) Western blot analysis confirmed MYCN downregulation in mNB-A1 cells treated with MLN8237 compared with DMSO-treated cells. (e, f) A mRNA signature score established from genes differentially expressed after JQ1 treatment of neuroblastoma cell lines (e), and a MYCN mRNA signature score (f) for mNB-A1 cells treated with either DMSO, JQ1 or MLN8237. (g) Western blot analyses of Brd4, MYCN, Myc, E2f1 and Cyclin D1 expression in re-grafted tumors from LSL-MYCN;Dbh-iCre mice treated with JQ1 or DMSO. Actin and Gapdh were used as loading controls. (h) Re-grafted tumors from JQ1- or DMSO-treated mice were examined histologically after hematoxylin/eosin (H&E) staining or immunostaining for cleaved caspase 3 (apoptotic cells) or Ki-67 (actively proliferating cells). Representative images are shown. (i) Bar graphs show the mean relative apoptosis (left) and proliferation (right) calculated from three representative images from each re-grafted tumor from groups of mice treated with either JQ1 or DMSO. Significance was calculated by Student's t-test: *P<0.05, **P<0.01, ***P<0.001.