|
|
|
 |
||
|
Cancer Redox Biology Faculty Calls a Workshop
The CRB Faculty recently assessed the state of the science in a workshop on “Redox-Based Non-Steroidal Anti-Inflammatory Drugs (NSAIDs) in Cancer Treatment, Prevention and Angiogenesis: A Novel Solution to an Old Problem.” It was, in part, a response to an extensive hearing by the U.S. Food and Drug Administration (FDA) in 2005 that suggested the entire class of NSAIDs might have problematic side effects.
Dr. David Wink of CCR’s Radiation Biology Branch (right) presents Nobel Laureate Dr. Louis Ignarro (UCLA) with a lifetime achievement award at the Grand Rounds Lecture at the workshop. A few years before, researchers had claimed that NSAIDs such as aspirin could inhibit cyclooxygenase (COX) enzymes and help prevent colorectal cancer. The oncology community soon discovered that aspirin causes side effects such as stomach ulcers. Renewed optimism followed with the discovery of selective COX-2 inhibitors, such as celecoxib or rofecoxib, which targeted COX-2 exclusively, reducing prostaglandin (PG) E2 levels in cells without causing stomach toxicity. However, these new agents also brought new side effects. The workshop assessed NSAIDs from a CRB perspective. The day opened with Dr. Louis Ignarro, Nobel Laureate in Medicine from the David Geffen School of Medicine at the University of California, Los Angeles, delivering the Grand Rounds Lecture. He presented his exceptional discoveries that clarified the role of NO as a unique signaling molecule beneficial for patients who have heart disease. The CRB Faculty and invited speakers then shared some of the translational research aimed at the COX-2 target. NO-aspirin, a potential new drug to inhibit COX-2 enzymes with NO covalently attached, drew much attention from the scientists. Data presented by Basil Rigas, MD, DSc, of the Cancer Prevention Division of the State University of New York at Stony Brook showed that NO-aspirin was over 1,000-fold more potent than aspirin at inhibiting colon cancer cellular proliferation. And in animal models of colon and pancreatic cancer, it reduced tumor proliferation with little toxicity. Novel NSAIDs such as this may eventually provide more treatment options to patients with adenocarcinomas in breast and lung tissues.
Other researchers showed that the selective COX-2 inhibitor celecoxib is proving effective in lung cancer treatment. In a phase II clinical trial, celecoxib increased the survival of lung cancer patients compared with those who received chemotherapy alone. In a phase I study, again treating lung cancer, this agent used in combination with radiotherapy increased patient response more than radiation alone. Sulfur (S)-NSAIDs were evaluated as well. Piero del Soldato, PhD, of CTG Pharma in Milan, Italy, reported that the S-NSAIDs showed little gastrointestinal toxicity. Grace Yeh, PhD, and David Roberts, PhD, described their important properties with respect to cancer prevention and treatment. Larry Keefer, PhD, presented chemical delivery systems that can target NO to specific sites within the body. The CRB Faculty is an excellent example of successful multidisciplinary teaming within the CCR that leverages existing resources to support translational research. The success of this Faculty stems from a well-defined mission, strong leadership, and excellent communication among the members. Much preclinical work on NO remains to be done, but based on the pace of the CRB Faculty to date, this work will be accomplished soon and used to guide researchers in their selection of lead compounds to take forward to clinical trials.
Nitric Oxide, a Mediator of Inflammation, Regulates Tumorigenesis
We used a genetic strategy to determine if NO· and p53 cooperatively regulate tumorigenesis in p53-deficient mice, an animal model of cancer-prone Li-Fraumeni syndrome. We generated mice that were deficient in both p53 and NOS2 and also eight other combinations. To make a valid comparison, these mice were crossbred to be more than 99% C57BL6. Lymphomas developed more rapidly in TP53–/–NOS2–/– or TP53–/–NOS2+/– mice than in TP53–/–NOS2+/+ mice. Likewise, sarcoma and lymphomas developed faster in TP53+/–NOS2–/– or TP53+/–NOS2+/– than in TP53+/–NOS2+/+ mice. These observations suggest that NO· suppresses tumorigenesis in this model, but how? The balance of cellular proliferation and cell death can influence tumor development, and NO· is involved in both of these processes. Therefore, we asked whether there are differences in apoptosis and proliferation in mice with different NOS2 status. When we compared double-knockout mice, we found that TP53–/–NOS2+/+ mice showed a higher apoptotic index and a decreased proliferation index with an increased expression of the death receptor ligands, CD95 and TRAIL, and the cell cycle checkpoint protein, p21waf1, in the spleen and thymus prior to tumor development. The role of NO· in immune modulation argues in favor of a different immune profile in mice deficient in both p53 and NOS2 when compared with p53-deficient mice. We found that TP53–/–NOS2–/– mice produced a four-fold higher level of anti-inflammatory interleukin 10 (IL-10), when compared with TP53–/– mice. The increased level of IL-10 may also contribute to the rapid tumor development in the p53 and NOS2 double-knockout mice. IL-10 favors the type 2 T-helper cell (TH2) response and induces antigen-activated CD4+ cells to become CD4+/CD25+ T-regulatory cells. T-regulatory cells can also produce IL-10 and are potently immunosuppressive. IL-10 also inhibits the maturation of dendritic cells. Therefore, the loss of NO·-related TH1-antitumor response and IL-10–mediated inhibition of specific immune recognition by impairment of dendritic cell function and suppression of immune function through T-regulatory cells may enhance tumorigenesis. Therefore, we propose two models: a cell cycle arrest and increased apoptosis model, and an immune suppression model for NO·-mediated suppression of tumorigenesis (Figure 1). We are currently testing the hypothesis that IL-10 accelerates tumor development in TP53 knockout mice by generating TP53/IL-10/NOS2 triple-knockout mice. Depending on their intestinal microbial flora, IL-10 mice can develop inflammatory colitis and colon cancer. TP53/IL-10/NOS2 triple-knockout mice may have increased tumor latency of lymphoma, leukemia, and sarcoma, but decreased tumor latency and increased incidence of colon cancer as compared with TP53–/–NOS2–/– mice. However, the question remains of how does NO· regulate IL-10 production? We are currently investigating the possible mechanisms that might be responsible for NO·-mediated reduction of IL-10 in TP53-deficient mice.
Figure 1. Mechanistic models of nitric oxide (NO·)–mediated
inhibition of tumorigenesis in p53-deficient mice. Is NO·-mediated regulation of tumorigenesis in p53-deficient mice altered under an inflammatory microenvironment? We have preliminary evidence that higher levels of NO·, produced by treating these mice with heat-inactivated Corynebacterium parvum, a potent inflammatory agent, increase spontaneous tumorigenesis. Because the p53-deficient mice primarily develop spontaneous lymphoma, leukemia, sarcoma, and rarely carcinoma, future studies to investigate the role of NO· in the modulation of carcinoma are warranted.
Generic Features of Tertiary Chromatin Structure
Various studies have attempted to visualize tertiary chromatin structure, but virtually all of them have examined the structures in cells that were subjected to various extraction procedures and/or chemical fixation. A prevailing concern about these studies has, therefore, been whether the structures detected are artifacts of the preparation procedure. To address this concern, we and others have examined tertiary chromatin structures in live cells. Although the resolution limit of the light microscope (approximately 0.2 μm) does not permit visualization of the folding of the 10- or even 30-nm chromatin fibers, live cell imaging completely eliminates the question of fixation artifacts. We achieved live cell imaging of tertiary structures by using large tandem arrays of transcription factor binding sites integrated into a chromosome and then visualized by a green fluorescent protein (GFP) tag. When transcriptionally activated, these tandem arrays unfolded from a single punctum (about 0.5 μm in diameter) into a series of adjacent puncta that we referred to as “beads.” These beaded structures rapidly refolded into a single bead when transcription was inhibited. Although tandem arrays occur naturally—for example, the ribosomal gene cluster—most normal chromatin is composed of different genes with different promoters that are often widely interspersed with non-coding DNA. Thus, despite the advantage of imaging live cells, it is not clear whether the results from tandem array systems can be extrapolated to most chromatin. Consequently, we identified completely natural systems in which to investigate tertiary chromatin structure. As a start, we hypothesized that the tandem arrays would be good models for regions of natural chromatin where a high percentage of genes were active. We identified two such regions. One was the MHCII cluster on human chromosome 6 in which at least 13 genes spanning a nearly 750-kb domain are simultaneously activated by interferon. The other was a nearly 400-kb region on human chromosome 22, which we identified by a bioinformatics analysis of microarray data. We found that this region was much more transcriptionally active in Jurkat cells than in Raji cells. At present there is no live cell imaging technique to visualize tertiary chromatin structure in these transcriptionally active domains, so we resorted to a fixation and hybridization procedure known as DNA fluorescent in situ hybridization (FISH). To validate that this DNA FISH procedure did not disrupt the tertiary structures observed by light microscopy, we compared DNA FISH images of the tandem array with live cell images and showed that the two yielded comparable structures. With confidence that tertiary structure would be preserved, we then used DNA FISH to examine these structures in the transcriptionally active domains on human chromosomes 6 and 22. Remarkably, we found structures very similar to those detected in the live cell tandem arrays. The prominent feature was again a series of adjacent puncta or beads approximately 0.5 μm in diameter that refolded into a single bead when transcription was inhibited. Thus, despite vast differences in gene density and levels of transcriptional activation, these disparate chromosomal domains shared common beaded features of tertiary chromatin structure. Consequently, we propose that beads may be generic features of higher-order chromatin organization. The presence of beads in light microscope images indicates that higher-order chromatin must be clustered into largely distinct subdomains. The ability of these beads to split into a series of adjacent beads indicates that the folding of chromatin within a cluster or bead must be sufficiently organized to readily enable this bifurcation process. These results, therefore, impose constraints on how tertiary chromatin must be organized. The determination of the actual folding patterns of a 30-nm chromatin fiber within a bead requires higher resolution imaging than afforded by light microscopy. We are now pursuing methods for correlative fluorescence and X-ray microscopy that will permit three-dimensional reconstruction of cryo-preserved intact nuclei at 20-nm resolution and, thereby, enable us to assess 30-nm chromatin fiber organization within the beaded structures detected in our GFP-tagged tandem array.
Chromatin Epigenetics: Nucleosome-binding Proteins Modulate the Levels of Histone Posttranslational Modifications in Chromatin
We explored the possibility that chromatin-binding structural proteins, such as HMGNs, which are devoid of enzymatic activity, are part of the mechanism that regulates the levels of chromatin modifications. HMGNs are a family of proteins that bind to nucleosomes without specificity for the underlying DNA sequence and induce structural and functional changes in chromatin. We reasoned that the binding of these proteins to nucleosomes may induce local or global changes in chromatin and shift the equilibrium between the activities of histone-modifying and -demodifying enzymes. To test this hypothesis, we used specific antibodies to examine the levels of several modifications in the tail of histone H3, isolated from either wild-type, or from Hmgn1–/– mouse embryonic fibroblasts (MEFs). We found that loss of HMGN1 protein elevated the levels of phosphorylation at serine 10 (H3S10p) and serine 28 (H3S28p) but decreased the levels of acetylation at lysine 14 (H3K14ac). Reexpression of the HMGN1 in the Hmgn1–/– cells, by induction of stably integrated vectors, elevated the levels of H3K14ac and reduced the levels of H3S10p, proof that the levels of these modifications are indeed regulated by HMGN1. Significantly, reexpression of a mutant HMGN1 that does not bind to chromatin did not alter the levels of these modifications, an indication that HMGN1 modulates histone modifications by binding to chromatin. Further in vitro studies and analysis of cells treated with HDAC inhibitors revealed that HMGN1 elevates the levels of H3K14ac by enhancing the activity of a specific HAT, rather than by reducing the demodifying activity of an HDAC. Similar studies on the mechanisms whereby HMGN1 reduces the levels of H3S10p indicate that the binding of the protein to nucleosomes hinders the ability of specific kinases to access and modify their targets. The in vivo studies are fully supported by in vitro analyses indicating that HMGN1 affects these H3 modifications only in the context of chromatin. The modification of purified H3 protein is not affected by HMGN1. Thus, HMGN1 can either enhance or reduce the level of a specific modification in chromatin. Interestingly, a close homolog of HMGN1, named HMGN2, enhanced the acetylation of H3K14 more efficiently than HMGN1 but did not inhibit the phosphorylation of H3S10, suggesting HMGN variant–specific effects on histone modification. Current analysis of a series of mutants in which distinct domains of HMGN1 were swapped with domains of HMGN2 indicate that distinct domains of the proteins are involved in enhancement of acetylation and reduction of phosphorylation, further proof for HMGN-specific effects on histone modifications. We have previously demonstrated that the linker histone H1 inhibits the acetylation of H3K14 (Herrera JE et al. Mol Cell Biol 20: 523–9, 2000). Taken together, these studies establish that chromatin-binding structural proteins modulate the levels of chromatin modifications, most likely by altering the ability of chromatin modifiers to access and modify their targets. HMGNs and H1 may affect the accessibility of sites in chromatin either by inducing global changes in the “compaction” of the chromatin fiber, or by inducing steric changes at the local levels of the nucleosome. Our studies demonstrate that structural proteins alter the equilibrium generated by the activities of the enzymes that determine the levels of chromatin modifications, and point to an additional mechanism that regulates these epigenetic markers (Figure 1). Figure 1. Model illustrating the effects of the structural chromatin-binding proteins H1, HMGN1, and HMGN2 on the levels of specific histone modifications. These proteins affect the levels of histone modifications by altering the dynamic equilibrium between the enzymes that continuously add or remove chemical tags, such as acetyl or phosphate groups, to histone tails. N2, HMGN2; S10, serine 10; K14, lysine 14; S10p, phosphorylated S10; H1, histone 1; K14ac, acetylated K14; HAT, histone acetylase; HDAC, histone deacetylase.
Axon Guidance Cues in Tumor and Developmental Angiogenesis
Hints for such targets have unexpectedly come from the developmental patterning mechanism for the nervous system. Increasing evidence is emerging that similar patterning mechanisms are used during the development of neural and vascular networks in vertebrates. Molecules that guide axons to their targets appear to have counterparts that guide the patterning of vessels. Four families of guidance molecules, namely netrins, semaphorins, ephrins, and slits, and their cognate receptors, mediate axon guidance. Each of these guidance families has recently been shown to contain at least one ligand-receptor pair that plays a functional role in patterning of the vasculature. Our study focuses on the slit-robo signaling system in vascular development and investigates one member of the Roundabout (Robo) family, robo4, in vessel guidance. Robos were originally found in Drosophila to mediate repulsive cues for slit ligands by preventing the re-crossing of axons once they have crossed the midline. Four members of the Robo family have been identified. Robo1, robo2, and robo3 are primarily expressed in the central nervous system (CNS) and help guide axons. Outside the nervous system, robos have been thought to play a role in leukocyte trafficking and kidney branching morphogenesis. However, until the discovery of the fourth member of this family, robo4, and its expression patterns in developing mouse vasculature, robos were unappreciated in vascular development. Besides developmental angiogenesis, robo4 is expressed in sites of active angiogenesis in tumor vessels. Because robo4 is both selective to tumor vasculature and is a cell-surface receptor, it is an attractive target for tumor endothelium. To investigate robo4 function, we studied its ortholog in zebrafish during embryonic development. We characterized robo4 gene expression and pursued functional studies by gene knockdown approaches using morpholinos (antisense oligonucleotides). Robo4’s in situ expression showed three interesting features. First, its expression in the embryonic zebrafish vasculature was transient and was seen in both angioblasts (Figure 1, parts A and B, white asterisk) and intersomitic vessels (ISVs) (Figure 1, parts C and D, black asterisks). Second, robo4 expression in notochord and ISVs of the trunk region overlapped, such that a concomitant decrease in robo4 expression in notochord in a rostral-to-caudal manner was immediately followed by expression in ISVs, suggesting a mechanistic role for robo4 in the ISV sprouting process (Figure 1, parts B and C). Third, robo4 expression in rostral sprouts ceased first, prior to its expression in caudal sprouts, suggesting a temporal regulation for the ISV sprouting process. To assess phenotype, we performed in situ studies with both endothelial (flk) and neuronal (acetylated tubulin) markers in robo4 knockdown embryos. Compared with wild-type embryos (Figure 1, part E), robo4 knockdown embryos displayed either a lack of ISV sprouts (Figure 1, part F) or linear ISV sprouts, albeit weak ones in the zebrafish trunk region. Figure 1. Robo4 expression patterns in embryonic zebrafish vascular development. Panels A through D show robo4 in situ (blue) across 20 (A) and 22 (D) somites in embryos. Panels B and C are high-power images of the trunk region of (A) and (D), respectively. (E) Wild-type embryo and (F) morpholino-injected embryo. Panels E and F depict trunk regions of 22-somite embryos double stained for flk RNA (blue, endothelial marker) and antibody to acetylated tubulin (brown, neuronal marker). N, notochord; ISV, intersomitic vessels. Black asterisks depict the location of ISVs, and the white asterisk depicts angioblasts. To visualize the ISV vessel growth in a live embryo, we pursued time-lapse imaging of the trunk vasculature in embryos injected with morpholinos. Using live time-lapse imaging in robo4 knockdown vascular-specific transgenic embryos, we demonstrated that when vessel sprouting occurs incorrectly, vessel formation is often aborted. Also, embryos often displayed a loss of temporal and spatial regulation of ISV sprouting from the dorsal aorta (DA). The surprising finding was that prior to our gene knockdown study, it was widely assumed that robos were primarily negative regulators of guidance; however, our study suggests that removal of a presumptive negative guidance cue resulted in the collapse of vessels as opposed to the expected increase in sprouts. This result suggests that in the absence of guidance molecules, vessels normally have a default mechanism that prevents them from growing into incorrect tissue sites. This helps clarify why robo4 would be upregulated in tumor vasculature since, from a tumor standpoint, a cancer cell would subvert or override normal developmental guidance checkpoints and utilize this mechanism to form a chaotic vascular network, which is seen routinely in tumor vasculature. Another interpretation of our knockdown result is that besides traditional repulsive cues, robos mediate attractive cues as well. In fact, preliminary unpublished evidence from our laboratory suggests that this cannot be excluded, and perhaps, tumors use axon guidance molecules, which are known to have bi-functional properties, to fulfill their growing needs.
Scientific Advisory CommitteeIf you have scientific news of interest to the CCR research community, please contact one of the scientific advisors (below) responsible for your areas of research.
CCR Frontiers in Science—StaffCenter for Cancer Research Robert H. Wiltrout, PhD, Director Deputy Directors Douglas R. Lowy, MD Editorial Staff Sue Fox, BA/BSW, Senior Editor * Palladian Partners, Inc. FOR INTERNAL USE ONLY |
||
 |
everal
oxidation-reduction mechanisms are important in cancer etiology and treatment.
There is so much to learn about nitric oxide (NO) and its oxidation-reduction
reaction, called redox, that David Wink, PhD, of the NCI CCR Radiation Biology
Branch, formed the Cancer Redox Biology (CRB) Faculty to do just that. The Faculty
has successfully enhanced communication and promoted collaboration among biochemists,
chemists, clinical oncologists, epidemiologists, and others interested in the
molecular mechanisms by which redox stress alters cancer development and tumor
spread.
itric
oxide (NO
ll
processes involving DNA, including replication, transcription, and repair, must
occur within the context of chromatin. However, this packaging of the DNA molecule
is understood with certainty at only its lowest level of organization, namely
the wrapping of DNA around nucleosomes yielding a fiber 10 nm in diameter. This
nucleosomal fiber is further wound in some way—the details are still controversial—to
yield a thicker fiber of approximately 30 nm. Yet these first two levels of
packaging are just the beginning, as much more compaction is required to fit
our genome into the nucleus. These unknown higher levels of chromatin folding
are referred to as higher-order or “tertiary” chromatin structure.
Our recent work has suggested that generic, conserved features are within such
tertiary chromatin structure.
osttranslational
modifications in the tails of the nucleosomal histones are important epigenetic
markers and serve as potential targets for cancer therapy. In living cells,
the levels of histone modifications are not fixed. They are in a continuous
state of flux and reflect the equilibrium reached between the activities of
enzymes that continuously modify, and those that continuously demodify specific
histone residues. For example, the equilibrium between the opposing activities
of histone acetylases (HATs) and deacetylases (HDACs) or kinases and phosphatases
determines the levels of acetylation and phosphorylation in chromatin. Recruitment
of histone-modifying or -demodifying enzyme complexes to specific sites is part
of the mechanism that regulates gene expression. 
hree
decades of angiogenesis research has culminated with the first U.S. Food and
Drug Administration (FDA)–approved anti-angiogenic drug, Avastin, which
is showing promise in the clinic. Avastin, a monoclonal antibody to vascular
endothelial growth factor (VEGF), sops up VEGF secreted by tumors and thus denies
them the ability to grow new blood vessels. However, VEGF has other physiological
functions, which are compromised in Avastin-treated patients. Because of this,
discovering novel targets of vascular endothelium has taken on a new sense of
urgency and prompted the search for the next generation of anti-angiogenic drug
targets that are tumor-vasculature specific. Cell-surface moieties are of particular
interest; historic evidence suggests that they might be suitable as targets.
