Search

Although proteasome inhibitors such as bortezomib had significant therapeutic effects in multiple myeloma and mantel cell lymphoma, they exhibited minimal clinical activity as a monotherapy for solid tumors, including colorectal cancer. We found in this study that proteasome inhibition induced a remarkable nuclear exportation of ubiquitinated proteins. Inhibition of CRM1, the nuclear export carrier protein, hampered protein export and synergistically enhanced the cytotoxic action of bortezomib on colon cancer cells containing wild-type p53, which underwent G2–M cell-cycle block and apoptosis. Further analysis indicated that tumor suppressor p53 was one of the proteins exported from nuclei upon proteasome inhibition, and in the presence of CRM1 inhibitor KPT330, nuclear p53, and expression of its target genes were increased markedly. Moreover, knockdown of p53 significantly reduced the synergistic cytotoxic action of bortezomib and KPT330 on p53+/+ HCT116 cells. In mice, KPT330 markedly augmented the antitumor action of bortezomib against HCT116 xenografts as well as patient-derived xenografts that harbored functional p53. These results indicate that nuclear p53 is a major mediator in the synergistic antitumor effect of bortezomib and KPT330, and provides a rationale for the use of proteasome inhibitor together with nuclear export blocker in the treatment of colorectal cancer. It is conceivable that targeting nuclear exportation may serve as a novel strategy to overcome resistance and raise chemotherapeutic efficacy, especially for the drugs that activate the p53 system. Mol Cancer Ther; 16(4); 717–28. ©2016 AACR.

NSCLC is a common cancer with the highest mortality and few effective treatment options, except for the patients with EGFR (Erlotinib, Iressa) and ALK (Crizotinib) activating mutations. Recent studies indicated that some patients with the activating EGFR mutations but without pre-exposure to EGFR tyrosine kinase inhibitors (TKIs) are resistant to TKIs. The proposed mechanisms (1) include: T790M gatekeeper mutation preventing accessing the kinase activation site, Kras mutation, pTEN loss (2), or c-met amplification (3) that activates ERBB3 signaling (4). Patient Derived Xenograft (PDX) models have been suggested to more closely represent patient tumors (5, 6). We established several PDXs, including HuPrime® LU858, LU859, LU357, LU387, LU1901 and LU2503, using tumor tissues from individual NSCLC patients. LU858/859 contain the “classic” activating EGFR mutation, L858R; LU1868 contain L858R/T790M double mutations; LU357/LUF387 contain the common exon 20 insertion (2319-2320 AACCCCCAC); and LU1901/2503 have wild type EGFR. To our surprise, the two models with single L858R mutation, just like the model with double mutations, respond to the erlotinib poorly, as seen in some of clinic cases (1). Since they do not have T790M and Kras mutation, we examined them for c-met status, finding that their c-met gene are significantly amplified with also increased c-met/HGF expression as measured by at mRNA and protein levels (IHC), consistent with those seen in clinic (3). Furthermore, LU357 is sensitive to erlotinib while LU387 is not, corresponding to no c-met amplification in LU357 but significant amplification in LU387. All these suggest that the observed erlotinib resistance in these models could potentially be attributed to c-met signaling activation via c-met/HGF autocrine. To this end, we treated LU858, LU859 and LU2503 (also with c-met amplification) with c-met inhibitor, Crizotinib. Indeed, Crizotinib caused significant tumor response as expected in all these models. In particular, the combination treatment of LU858 and LU859 with erlotinib caused complete tumor regression by overcoming the resistance, suggesting a new potential effective treatment option. The results using PDX models can potentially help to design a prospective human study to demonstrate improved efficacy by the combination treatment via patient stratification to include the naïve NSCLC patient subjects with c-met amplification/EGFR mutation. 1. A. F. Gazdar, Oncogene 28 Suppl 1, S24, 2009 2. M. L. Sos et al., Cancer Res 69, 3256, 2009 3. T. Kubo et al., Int J Cancer 124, 1778, 2009 4. A. B. Turke et al., Cancer Cell 17, 77, 2010. 5. L. Ding et al., Nature 464, 999, 2010 6. E. Marangoni et al., Clin Cancer Res 13, 3989, 2007 Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 2344. doi:1538-7445.AM2012-2344

c-Met is a receptor tyrosine kinase that interacts with its ligand, HGF, secreted from surrounding stroma (paracrine signaling). This interaction is involved in cell growth, invasion, metastases and angiogenesis of a number of solid tumors. c-met gene amplification and increased gene expression have been frequently associated with these tumors. c-met is thus considered as an important cancer target for therapy of the cancers that their growth depends on activated c-met/HGF signaling1, including many HCC. As for any investigational c-met inhibitors, a highly relevant c-met/HGF experimental cancer model is critical to evaluate its efficacy and associated changes in biomarkers for predicting the likely responders. While patient derived xenograft (PDX) models, without any in vitro manipulation, mirror patients’ histopathological and genetic profiles2,3, the successfully engrafted PDX models usually lost its human stroma in mice and thus its c-met dependency. This is due to the lack of paracrine c-met/HGF mechanism in mice (no cross reactivity between mouse and human), and the tumors with such c-met/HGF paracrine dependency are selected against. The engrafted models might have changed its properties from the original human tumors by relying on other signaling pathways for survival, and thus not suitable for evaluating c-met inhibitors. Apparently, neither of these scenarios can be adequately used to evaluate c-met inhibitors. One alternative to this is engraftment of primary human tumor into HGF-transgenic immunocompromised mice, which would support the growth of c-met dependent PDX via paracrine mechanism. We recently screened a number of PDX HCC models using a c-met inhibitor and identified two responders, indicating their reliance on c-met/HGF signaling for growth, among many non-responders. The non-responders likely grow by relying on signaling pathway other than c-met/HGF. The two responders, on the other hand, likely grow through c-met/HGF autocrine mechanism. To further investigate this, we performed genomic profiling and IHC analysis of c-met and HGF on these two responders and other non-responders. Our results demonstrated that these two responders have amplification of HGF gene and express high levels of both c-met and HGF in both mRNA and protein (IHC). This is in contrast to no gene amplification and low level expression of both genes for non-responders. All these are consistent with that these two responders grow in mice via c-met/HGF autocrine mechanism. In conclusion, these two HCC models could be very useful models for evaluating c-met inhibitors. 1. Nat Rev Drug Discov 7, 504-516 (2008). 2. Nature 464, 999-1005 (2010). 3. Clin Cancer Res 13, 3989-3998 (2007). Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 1322. doi:1538-7445.AM2012-1322