Liver metastasis is one of the major causes of death in colorectal cancer (CRC) patients. To understand this process, we investigated whether the gene expression profiling of matched colorectal carcinomas and liver metastases could reveal key molecular events involved in tumor progression and metastasis. We performed experiments using a cDNA microarray containing 17,104 genes with the following tissue samples: paired tissues of 25 normal colorectal mucosa, 27 primary colorectal tumors, 13 normal liver and 27 liver metastasis, and 20 primary colorectal tumors without liver metastasis. To remove the effect of normal cell contamination, we selected 4,583 organ‐specific genes with a false discovery rate (FDR) of 0.0067% by comparing normal colon and liver tissues using significant analysis of microarray, and these genes were excluded from further analysis. We then identified and validated 46 liver metastasis‐specific genes with an accuracy of 83.3% by comparing the expression of paired primary colorectal tumors and liver metastases using prediction analysis of microarray. The 46 selected genes contained several known oncogenes and 2 ESTs. To confirm that the results correlated with the microarray expression patterns, we performed RT‐PCR with WNT5A and carbonic anhydrase II. Additionally, we observed that 21 of the 46 genes were differentially expressed (FDR = 2.27%) in primary tumors with synchronous liver metastasis compared with primary tumors without liver metastasis. We scanned the human genome using a cDNA microarray and identified 46 genes that may play an important role in the progression of liver metastasis in CRC.

Malignant peripheral nerve sheath tumors (MPNSTs) are aggressive, frequently metastatic sarcomas that are associated with neurofibromatosis type 1 (NF1), a prominent inherited genetic disease in humans. Although loss of the NF1 gene predisposes to MPNST induction, relatively long tumor latency in NF1 patients suggests that additional genetic or epigenetic abnormalities are needed for the development of these nerve sheath malignancies. To study the molecular pathways contributing to the formation of MPNSTs in NF1 patients, we used a zebrafish tumor model defined by nf1 loss in a p53-deficient background together with the overexpression of either wild-type or constitutively activated PDGFRA (platelet-derived growth factor receptor-α) under control of the sox10 neural crest-specific promoter. Here we demonstrate the accelerated onset and increased penetrance of MPNST formation in fish overexpressing both the wild-type and the mutant PDGFRA transgenes in cells of neural crest origin. Interestingly, overexpression of the wild-type PDGFRA was even more potent in promoting transformation than the mutant PDGFRA, which is important because ~78% of human MPNSTs have expression of wild-type PDGFRA, whereas only 5% harbor activating mutations of the gene encoding this receptor. Further analysis revealed the induction of cellular senescence in zebrafish embryos overexpressing mutant, but not wild-type, PDGFRA, suggesting a mechanism through which the oncogenic activity of the mutant receptor is tempered by the activation of premature cellular senescence in an NF1-deficient background. Taken together, our study suggests a model in which overexpression of wild-type PDGFRA associated with NF1 deficiency leads to aberrant activation of downstream RAS signaling and thus contributes importantly to MPNST development—a prediction supported by the ability of the kinase inhibitor sunitinib alone and in combination with the MEK inhibitor trametinib to retard MPNST progression in transgenic fish overexpressing the wild-type receptor.

Earlier reports showed that hyperplasia of sympathoadrenal cell precursors during embryogenesis in Nf1-deficient mice is independent of Nf1’s role in down-modulating RAS-MAPK signaling. We demonstrate in zebrafish that nf1 loss leads to aberrant activation of RAS signaling in MYCN-induced neuroblastomas that arise in these precursors, and that the GTPase-activating protein (GAP)-related domain (GRD) is sufficient to suppress the acceleration of neuroblastoma in nf1-deficient fish, but not the hypertrophy of sympathoadrenal cells in nf1 mutant embryos. Thus, even though neuroblastoma is a classical “developmental tumor”, NF1 relies on a very different mechanism to suppress malignant transformation than it does to modulate normal neural crest cell growth. We also show marked synergy in tumor cell killing between MEK inhibitors (trametinib) and retinoids (isotretinoin) in primary nf1a-/- zebrafish neuroblastomas. Thus, our model system has considerable translational potential for investigating new strategies to improve the treatment of very high-risk neuroblastomas with aberrant RAS-MAPK activation.

Malignant peripheral nerve sheath tumors (MPNSTs) are soft-tissue sarcomas that frequently arise in patients with neurofibromatosis type 1 (NF1). Most of these tumors are unresectable at diagnosis and minimally responsive to conventional treatment, lending urgency to the identification of new pathway dependencies and drugs with potent antitumor activities. We therefore examined a series of candidate agents for their ability to induce apoptosis in MPNST cells arising in nf1/tp53-deficient zebrafish. In this study, we found that DNA topoisomerase I-targeted drugs and mTOR kinase inhibitors were the most effective single agents in eliminating MPNST cells without prohibitive toxicity. In addition, three members of these classes of drugs, either AZD2014 or INK128 in combination with irinotecan, acted synergistically to induce apoptosis both in vitro and in vivo. In mechanistic studies, irinotecan not only induces apoptosis by eliciting a DNA damage response, but also acts synergistically with AZD2014 to potentiate the hypophosphorylation of 4E-BP1, a downstream target of mTORC1. Profound hypophosphorylation of 4E-BP1 induced by this drug combination causes an arrest of protein synthesis, which potently induces tumor cell apoptosis. Our findings provide a compelling rationale for further in vivo evaluation of the combination of DNA topoisomerase I-targeted drugs and mTOR kinase inhibitors against these aggressive nerve sheath tumors.

Background: The information from different data sets experimented under different conditions may be inconsistent even though they are performed with the same research objectives. More than that, even when the data sets were generated from the same platform, the data agreement may be affected by the technical variation among the laboratories. In this case, it is necessary to use the combined data set after adjusting the differences between such data sets, for detecting the more reliable information. Results: The proposed method combines data sets posterior to the discretization of data sets based on the ranks of the gene expression ratios, and the statistical method is applied to the combined data set for predictive gene selection. The efficiency of the proposed method was evaluated using five colon cancer related data sets, which were experimented using cDNA microarrays with different RNA sources, and one experiment utilized oligonucleotide arrays. NCI-60 cell lines data sets were used, which were performed with two different platforms of cDNA microarrays and Affymetrix HU6800 oligonucleotide arrays. The combined data set by the proposed method predicted the test data sets more accurately than the separated data sets did. The biological significant genes were detected from the combined data set, which were missed on the separated data sets. Conclusion: By transforming gene expressions using ranks, the proposed method is not influenced by systematic bias among chips and normalization method. The method may be especially more useful to find predictive genes from data sets which have different scale in gene expressions.