Vaccinia Virus: Applications for Eradicating Human Tumor Vasculature
Targeting the tumor vasculature for treatment of solid tumors has been an objective for decades. Several inhibitors, including vascular disrupting agents (VDA) for manipulating VEGF/VEGF receptor (KDR) pathways, have been developed. Unfortunately, the outcome in clinical trials has been disappointing due to the acquired resistance to the therapeutic agent and/or activation of alternative pathways of progression by tumors. Breitbach and colleagues examined the therapeutic efficacy of a genetically engineered Vaccinia virus in destroying tumor vasculature in xenografts and human hepatocellular carcinoma (HCC) patients in a phase I clinical trial. They used JX-594, a Vaccinia virus developed by Wyeth that lacks the active thymidine kinase (TK) gene and expresses human granulocyte-monocyte colony stimulation factor (hGM-CSF, CSF2) and β-galactosidase (GLB1) transgenes. To monitor replication in vitro, they developed a GFP-expressing JX-594 construct (JX-594-GFP+/β-gal). In vitro JX-594 selectively infected VEGF and FGF-activated normal human umbilical vein endothelial cells or dermal microvascular endothelial cells and expressed the transgenes. In vivo, a murine version of the engineered virus lacking the TK gene efficiently disrupted the vasculature surrounding subcutaneous xenografts of 4T1 breast cancer cells in BALB/c mice. Decreased intratumoral vessel formation and blood flow, together with increased hypoxia, were evident in the tumors. The normal vasculature remained undamaged when the compound was tested in rabbits (a more sensitive animal than mice or rats). In a subsequent phase I clinical trial to evaluate the efficacy of JX-594 in 18 HCC patients, intravenous injection of JX-594 virus effectively infected tumor-associated endothelial cells in the patients and, remarkably, resulted in disruption of the vasculature along with decreased perfusion, which persisted for at least 8 weeks without substantial damage to the surrounding normal vasculature or wound healing. These intriguing observations suggest the potential therapeutic efficacy of a Vaccinia virus approach in disrupting tumor vasculature, which, based on potential safety in humans, warrants further testing and development.
Clinical Chemotherapeutic Selectively Targets Tumor-Associated Macrophages
Overwhelming evidence shows that macrophages abundant in experimental and human tumor microenvironments can promote cancer growth and spread. In cancer models, depleting or blocking the influx and tumor-promoting activity of macrophages leads to enhanced response to chemotherapy. Early-phase clinical trials attempting to translate this research to patient benefit are under way. Related to these observations, the killing of macrophages was recently shown to be a key mechanism of action for a recently approved DNA-binding chemotherapeutic agent, trabectedin. Trabectedin, originally isolated from the sea squirt Ecteinascidia turbinata during a screen of plant and marine organism extracts carried out by the National Cancer Institute over 50 years ago, is now an approved treatment for advanced soft tissue sarcoma and ovarian cancer. Germano and colleagues show in 4 mouse cancer models that trabectedin selectively depleted monocytes in blood, spleen, and tumor, leading to a reduction in tumor angiogenesis. Through their use of trabectedin-resistant malignant cell lines or depletion of myeloid cells in mice, the authors show that targeting monocytic lineage cells is a key component in the anticancer activity of trabectedin. Blood monocytes (but not granulocytes or lymphocytes) were decreased in trabectedin-treated patients and in posttherapy biopsies from trabectedin-treated sarcoma patients. With regard to the question of how trabectedin kills cells of the mononuclear phagocyte lineage, it appears that the agent activated the extrinsic apoptotic pathway downstream of TRAIL receptors, and its selectivity was due to differential expression of death signaling and decoy receptors in cells of the monocyte lineage. Another interesting question is why macrophages express death receptors that are not “protected” by decoy receptors. The authors suggest this relates to a checkpoint for preventing overactivity of macrophages, which can be long-term tissue residents. Germano and colleagues provide proof of concept for macrophage targeting in humans, opening new perspectives for exploiting this unique property of trabectedin in combination cancer treatments.
Endothelial Cells Mediate a Stem Cell Phenotype via a Paracrine Effect in Colorectal Cancer
Colorectal cancer (CRC) is a leading cause of cancer death. The median survival of patients with metastatic disease is less than 2 years despite recent advances in therapy. To identify new treatment avenues, these authors investigated the biology of cancer stem cells (CSC) in CRC. Recognizing that tumor progenitor cells localize to the perivascular niche, they investigated the role of endothelial cells in promoting the CSC phenotype. They first isolated primary cultures from CRC specimens and cocultured these with liver parenchyma endothelial cells. After 3 days of coculture, they examined cancer cells for enrichment of CSC characteristics and observed an increase in Aldefluor-positive and CD133 (PROM1)-positive cells, both of which are markers of CSC. To evaluate whether direct cell–cell contact was required, they treated tumor cells with endothelial cell–conditioned medium and found similar CSC enrichment. Conditioned medium from established colorectal cancer tumors or fibroblasts could not induce this phenotype, indicating specificity to endothelial cells. Furthermore, endothelial cell–conditioned medium increased the tumorigenicity and metastatic potential of CRC in serial dilution experiments. To establish relevance, the authors showed that CD133-positive CRC cells were in proximity to endothelial cells, as measured by CD31 (PECAM1) staining. Because the stem cell phenotype also involves resistance to chemotherapy, they showed that endothelial cell–conditioned media promoted resistance to oxaliplatin and 5-fluorouracil in CRC cells. To determine the mechanism of the endothelial cell–dependent conversion of colorectal cancer cells, they tested reporter constructs for Notch, Wnt/β-catenin, and Sonic hedgehog pathways and observed Notch pathway activation in response to endothelial cell-conditioned medium. Importantly, the effects of the conditioned medium were blocked by treatment with a γ-secretase inhibitor, which inhibits the Notch pathway. Furthermore, in human specimens, expression of Notch intracellular domain, an effector of Notch pathway activation, colocalized with CD133 positivity. Next, they showed that endothelial cells secrete a soluble form of the Notch ligand Jagged-1 (JAG1), resulting in a paracrine effect in CRC and Notch pathway activation. In this system, Jagged-1 was secreted as a C-terminally truncated protein, truncated at amino acid residue 1054 by ADAM17, enabling secretion of Jagged-1 as a soluble factor. Finally, blockade of ADAM17 activity prevented the endothelial cell–dependent promotion of the CSC phenotype, implicating its key role in this process. These results show that endothelial cells are more than simply conduits for blood flow; these cells also contribute soluble factors to promote the CSC phenotype and chemoresistance. Future studies may test whether Notch pathway activation could be targeted in patients with metastatic CRC.
MYCN as a Biomarker of BET Bromodomain Inhibitors
Epigenetic “readers” recognize and bind to covalent modifications of chromatin, with bromodomains in particular recognizing acetylated lysine residues on histone tails. The bromodomain and extraterminal domain (BET) family, composed of BRD2-4 and BRDT, regulates transcription, epigenetic memory, and cell growth. The inhibitor JQ1 is a thienotriazolo-1,4-diazepine that displaces BET bromodomains from chromatin by binding competitively to the acetyl lysine recognition pocket. To identify predictors of sensitivity to JQ1, Puissant and colleagues screened approximately 700 genetically characterized tumor cell lines, identifying amplification of MYCN as a top predictive marker of response. Their analysis of genome-wide expression showed downregulation of both MYCN and the MYCN transcriptional program. In MYCN-driven neuroblastoma cell lines, genetically engineered models, and patient-derived xenograft mouse models, treatment with JQ1 led to blockade of MYCN and target genes, increased apoptosis, and improved survival. Thus, amplification of MYCN appears to be a biomarker of sensitivity to BET bromodomain inhibitors, consistent with results from earlier studies showing that inhibition of BET bromodomains blocked expression and function of MYC in hematologic malignancies. Collectively, these observations support the testing of clinical BET-bromodomain inhibitors in MYC- and MYCN-driven cancers.
A Glowing Response to Early Tumorigenesis
The study of very early events in tumorigenesis and the modeling of slow-growing, less aggressive neoplasms remain challenging. Burd and colleagues developed a luciferase knockin mouse model that generates a reporter system that visualizes the expression of the tumor suppressor p16 (Cdkn2a; p16LUC). They use this mouse to report on very early tumorigenesis in multiple tumor models and to explore whether induction of p16 indicates increased risk of cancer development. Interestingly, they also identify early induction of p16 in tumor-associated stroma, including bone marrow–derived cells. The authors first validated their approach by showing that stimuli known to induce p16INK4a also activate luciferase in vivo. For example, they detected increased luciferase activity in p16+/LUC mice during mammary gland involution and in healing skin wounds. Having shown that luciferase activity is a faithful read-out for induction of p16INK4a, the authors examined luciferase activity across a large cohort of aging mice. Mirroring the known increase in p16INK4a expression with age, luciferase activity increased on average by 6.9-fold in the mice from 16 to 80 weeks of age. Tremendous heterogeneity was also apparent in the luciferase activity in mice of a similar age. Using this heterogeneity, the authors tested the hypothesis that increased induction of p16, suggesting increased levels of cell stress and cell senescence, predicts cancer-related mortality. Surprisingly, total body luciferase activity and the rate of change in luciferase activity did not predict development of a spontaneous malignancy in mice, suggesting that the age-dependent development of spontaneous cancer may not be associated with the global accumulation of senescent cells. Using 14 murine tumor models involving multiple driver mutations and cell types, including well-characterized genetically engineered mouse models, the authors showed the utility of the p16LUC allele to detect very early neoplasms and tumor dissemination to other organs. By transplanting tumor cells without the p16LUC allele into p16+/LUC mice the authors also identified robust activation of p16LUC in the tumor-associated stroma. They further demonstrated that bone marrow–derived cells contribute to the non–cell autonomous activation of p16LUC by performing bone marrow transplant of p16+/LUC cells into mice without the p16LUC allele. These findings indicate that in early neoplastic lesions, both tumor cells and tumor-associated stromal cells are induced to express p16INK4a. The p16LUC knockin mouse has potential as a tool to detect early neoplasia in vivo and to study the response of the stroma to neoplasia.
Note: Breaking Advances are written by Cancer Research Editors. Readers are encouraged to consult the articles referred to in each item for full details on the findings described.
RSS Feeds