In This Issue

See article, p. 685.

Activating KRAS mutations occur in the vast majority of pancreatic ductal adenocarcinomas (PDA), but attempts to pharmacologically inhibit RAS have been unsuccessful. Because several small-molecule inhibitors of RAS effectors are already in clinical development, defining the pathways required for KRAS-driven PDA initiation and progression could lead to readily testable therapeutic approaches. Collisson and colleagues evaluated the effects of constitutive activation of 2 downstream RAS effectors, BRAF and PI3K, in the pancreas. Mutationally activated BRAF^V600E, but not PIK3CA^H1047R, led to rapid formation of pancreatic intraepithelial neoplasia (PanIN) lesions that were morphologically indistinguishable from those caused by Kras mutation. Also, like oncogenic KRAS^G12D, BRAF^V600E cooperated with dominant-negative p53 to rapidly induce PDA in 100% of mice. Taken together, these data suggest that RAF-mediated signaling accounts for most, if not all, of the effects of mutant KRAS in the initiation and progression of PDA, and that downstream blockade of RAF signaling by MEK1/2 inhibition may be an effective therapeutic approach. Indeed, a MEK1/2 inhibitor had a cytostatic effect on established PDA orthotopic xenografts that led to a significant survival advantage. MEK suppression also inhibited the growth of multiple human PDA cell lines, though some activated AKT in response to MEK inhibition. Combined MEK and AKT inhibition synergized to inhibit the growth of resistant PDA cell lines, further providing a rationale for the development of MEK and PI3K/AKT inhibitors in PDA.

See article, p. 694.

Oligonucleotide decoys have been developed as a method to selectively inhibit transcription factor activity through competitive inhibition of binding to cis-regulatory elements. Sen and colleagues conducted an exploratory, first-in-human phase 0 trial to determine the pharmacodynamic activity of an oligonucleotide decoy for STAT3, which is constitutively active in head and neck squamous cell carcinomas (HNSCC) and many other cancers. A pretreatment biopsy was obtained prior to intratumoral injection of a STAT3 decoy or saline control solution during tumor resection surgery, and a post-treatment biopsy was collected from the site of injection after tumor resection. No adverse events were reported, and the trial met its primary endpoint of demonstrating that STAT3 decoy treatment led to decreased levels of STAT3 target gene products compared with the control. Because the linear double-stranded form of oligonucleotide decoys makes them sensitive to nuclease degradation if administered systemically, the authors generated modified STAT3 decoys with either a nucleotide or hexaethylene glycol linker connecting the 2 DNA strands at one or both ends. These modified STAT3 decoys had longer serum half-lives and higher melting temperatures than the parental decoy, and intravenous injection of the cyclic STAT3 decoy into tumor-bearing mice significantly suppressed tumor growth and STAT3 target gene expression and induced tumor regression in 2 of 10 mice. These findings will facilitate further clinical development of the STAT3 decoy and may provide a generalizable framework for inhibition of other transcription factors.

See article, p. 706.

To successfully colonize metastatic sites, tumor cells must adapt to distinct tissue microenvironments after extravasation. Shibue and colleagues describe a process by which disseminated cancer cells interact with the extracellular matrix to facilitate proliferation. Focal adhesion kinase (FAK) activation was required to induce the proliferation of metastatic tumor cells in 3-dimensional cultures and within the lung parenchyma in vivo. Phosphorylated FAK colocalized with integrin β₁, an essential factor for FAK activation, at elongated adhesion plaques that also correlated with areas of increased proliferation. Prior to the formation of these adhesion plaques, filopodium-like protrusions (FLP) containing actin and integrin β₁ were observed and served as nucleation sites for assembly of the mature, elongated plaques in multiple metastatic breast cancer cell lines. The formation of FLPs was dependent on the expression of cytoskeleton regulatory proteins, most notably mDIA2, which promotes actin polymerization, and the Rho family GTPase RIF. In contrast to metastatic cells, nonaggressive tumor cells rarely formed FLPs or elongated plaques, suggesting that FLPs enhance metastatic potential. Consistent with this idea, inhibition of FLP formation resulted in reduced FAK activation and diminished metastatic cell proliferation. Moreover, whereas FLPs did not affect primary tumor size, FLPs were necessary for the generation of macroscopic metastases in various target organs. These results highlight a distinct cell biologic process and provide insight into the mechanisms underlying the establishment and selective outgrowth of metastatic colonies.

See article, p. 722.

Indoleamine 2,3-dioxygenase (IDO) is an immunomodulatory enzyme that inhibits T-cell growth by depleting cellular tryptophan levels and has become an attractive therapeutic target for restoration of antitumor immune responses. Given the lack of genetic evidence for a role of IDO in tumorigenesis, Smith and colleagues bred homozygous Ido1-null alleles into a Kras-mutant inducible lung cancer mouse model and found that IDO-deficient, Kras-mutant mice had a reduced tumor burden and pulmonary vascular density compared with IDO-expressing mice. Additionally, Ido1-null mice developed fewer pulmonary metastases following engraftment with a highly metastatic breast cancer cell line. As inflammation is known to play an important role in the development of Kras-induced lung tumors, the authors evaluated inflammatory cytokine levels in the lungs of tumor-bearing mice and found that Ido1-null mice had markedly lower levels of interleukin-6 (IL-6) compared with Ido1–wild-type mice. Given the known role of IL-6 in the development of myeloid-derived suppressor cells (MDSC), which inhibit antitumor immune responses, these findings suggested that IDO may be required for IL-6–mediated MDSC activity. Indeed, MDSC expansion was delayed and MDSC immunosuppressive activity impaired in tumor-bearing Ido1-null mice, and constitutive expression of IL-6 in tumor cells rescued the immunosuppressive activity of MDSCs and led to a significant increase in the metastatic burden of IDO-deficient mice. Collectively, these results suggest that IDO orchestrates vascularization, inflammation, and immunosuppression to establish a favorable environment for tumor outgrowth and metastasis.

See article, p. 736.

Glioblastoma can be classified into unique molecular subtypes based on genetic heterogeneity and clinical responses. To investigate the role of microRNAs (miR) in glioblastoma, Genovese and colleagues used a systems biology approach that integrated computational modeling of miR–mRNA networks and functional analysis of these relationships in glioblastoma pathogenesis. The Context Likelihood of Relatedness (CLR) algorithm was used to predict miR–mRNA nodes in glioblastoma, of which only a small percentage represented putative direct interactions. However, many of the nodes showed differential expression among glioblastoma subtypes, indicating that these networks may contribute to subtype-specific expression signatures. Clinically, miR-34a expression was prognostic, as low miR-34a levels defined a subset of proneural (PN) glioblastoma patients with overall improved survival. Functionally, miR-34a overexpression in a mouse model of PN-like glioblastoma inhibited tumor formation and decreased tumor spheroid self-renewal, whereas decoy-inhibition of miR-34a expression enhanced tumorigenicity, supporting a tumor suppressive role for miR-34a in PN glioblastoma. This effect is in part mediated via negative regulation of a known miR-34a target, PDGFRA. In addition, transcription factor binding site enrichment analysis identified SMAD4, a mediator of TGF-β signaling, as a potential intermediate regulator linking miR-34a–mRNA nodes. Indeed, SMAD4 was shown to be directly regulated by miR-34a, and SMAD4 expression was sufficient to reverse miR-34a–mediated tumor suppression. These results uncover an important signaling axis in glioblastoma and demonstrate the value of network modeling in cancer biology.

Note: In This Issue is written by Cancer Discovery Science Writers. Readers are encouraged to consult the original articles for full details.