FAQ

Study Sections

Results of FARE 2005

Results of FARE 2004

Results of FARE 2003

Results of FARE 2002

Results of FARE 2001

Winning Abstracts from FARE2002

Bishow B. Adhikari, NIAMS
Modulation of skeletal muscle activation by diffusible calcium sensor proteins
In striated muscle, force is generated by the ATP-powered cyclic interaction of myosin motors in the thick filaments with the actins in the thin filaments in the contractile machinery, sarcomere. Regulatory proteins provide a precise control of the timing, duration and power of these motors. While it is well known that the motor activity is triggered by calcium via the troponin-tropomyosin complex in the thin filament, details remain obscure and the potential roles of other regulatory proteins are unexplored.
Two recent observations provide strong motivation for our current research. First, in vitro biochemical studies indicate that the thin filament protein, nebulin, along with a calcium sensor protein (S100 or calmodulin) is capable of modulating the interaction between actin and myosin. Second, two S100 isoforms, S100A1 and S100B, are present in significant amounts in skeletal and cardiac muscles and interact with sarcomeric proteins. They are also implicated in the normal development and functioning of the heart. To explore their potential role in contractility, we measured the effect of S100A1, S100B and calmodulin on the calcium activation of isometric ATPase, tension and stiffness in skinned, single rabbit psoas muscle fibers. At a concentration of 5-80 micromolar these proteins significantly reduce active tension, stiffness and ATPase at sub-maximal (pCa ~6.1-5.6) levels of activation. The effectiveness of inhibition follows the order: 100A1>S100B>calmodulin. Hill analysis of the tension-pCa curves and ATPase-pCa curves reveals a significant reduction of pKs and an appreciable increase in cooperativity as reflected by the Hill coefficients. The S100s exhibited an effect at sub-maximal, but not at maximal level of activation (pCa 4.0), whereas calmodulin reduced tension and stiffness at all activation levels.
These findings reveal that diffusible calcium sensor proteins are capable of altering the actin-myosin interaction in skeletal muscle fibers. The significantly higher potency of S100A1 compared to S100B or calmodulin, combined with the higher abundance of S100A1, suggest a strong physiological relevance for S100A1 in skeletal muscle contraction. A similar role for S100B and calmodulin is obscured by their reported low abundance and their lower potency and needs to be further examined. We propose that the S100A1, and perhaps S100B and calmodulin are viable candidates for physiological modulation of muscle contraction.

Hirofumi Akari, NIAID
The HIV-1 Accessory Protein Vpu Induces Apoptosis by Suppressing the NF-kB dependent Expression of Anti-Apoptotic Factors
The human immunodeficiency viruses type 1 (HIV-1) and type 2 (HIV-2) are the etiologic agents of AIDS. Despite overall genetic similarity, epidemiological studies have shown that HIV-2-infected individuals develop AIDS at a much lower frequency than HIV-1-infected people, suggesting that differences in the viral genomes account for their different pathogenic properties. The main structural difference between HIV-1 and HIV-2 is the presence of the vpu gene in HIV-1 but not in HIV-2.The possible contribution of this gene to pathogenesis in HIV-1-infected individuals is unclear. However, there is evidence from the macaque monkey model supporting a role for Vpu in viral pathogenesis.
HIV-1 Vpu is an integral membrane protein with a unique affinity to beta-TrCP (TrCP), a key member of the SCF (SkpI-Cullin-F-box) E3 ubiquitin ligase complex that is involved in the regulated degradation of cellular proteins, including IkB. Interestingly, we recently found that Vpu competitively inhibits TrCP-dependent degradation of IkB resulting in the suppression of NF-kB activity in Vpu-expressing cells. NF-kB has a central role in the expression of genes involved in the regulation of apoptosis. It is therefore possible that Vpu could contribute to the virus-induced cytopathic effects by modulating NF-kB-dependent expression of anti-apoptotic genes. Based on this hypothesis, we have explored the possible involvement of Vpu in HIV-1-induced apoptosis.
We found that in HIV-1-infected CD4+ T cells, Vpu contributed significantly to the induction of apoptosis. Using an inducible expression system, we showed that the effect of Vpu on apoptosis was specific and did not require the coexpression of other viral proteins. Of note, mutation of a TrCP-binding motif in Vpu abolished its apoptogenic property, indicating the requirement of TrCP interaction. Analysis of cellular factors involved in the induction of apoptosis demonstrated that Vpu indeed efficiently down-modulated the NF-kB-dependent expression of anti-apoptotic proteins such as Bcl-xL and A1/Bfl-1.Concomitantly, Vpu expression resulted in increase in the level of active caspase-3.Thus, Vpu induces apoptosis through activation of the caspase pathway by way of inhibiting the NF-kB-dependent expression of anti-apoptotic genes. These findings may in part explain why HIV-1 generally exhibits more severe in vivo pathogenicity than HIV-2 which does not encode a vpu gene.

Salvatore Alesci, NICHD
E1-/E4+ ADENOVIRAL VECTORS MAY IMPAIR ADRENAL RESPONSE TO STRESS
Recombinant adenoviral vectors (Ad) are effective in transferring foreign genes to a variety of cells and tissue types, both in vitro and in vivo. However, during the gene transfer, these vectors may alter the cell function and its local environment. The adrenal gland may be a target for adenovirus. In fact, morphological changes in the adrenal cortices have been described in mice and humans after adenoviral infection. However, the effect of adenoviral vectors on adrenal steroidogenesis is not known. To test this hypothesis, we transduced bovine adrenal cortical cells in primary culture with two E1-/E4+ (AdGFP and LacZ) and two E1+/E4- (dl 1011 and dl 1014) adenoviral mutants. Transduction by the vector AdGFP, expressing green fluorescent protein (eGFP) was close to 100% in cells examined by fluorescent microscopy. Electron microscopy showed nuclear segmentation and mitochondrial alterations. Basal cortisol secretion was significantly increased by both AdGFP (165 ± 52% of control, m ± SE, p = 0.016) and LacZ (263 ± 37%, p = 0.016), but not by dl 1011 (37 ± 11%, p = 0.190) and dl 1014 (15 ± 14%, p = 0.619). Additionally, ACTH (10-9 M) stimulated cortisol secretion was significantly decreased by both AdGFP (55 ± 11% of control, p = 0.039) and LacZ (77± 2%, p < 0.001), but not by dl 1011 (19 ± 5%, p = 0.193) or dl 1014 (0.5 ± 5%, p = 0.967).
Our data suggests that E1-/E4+ adenoviral vectors may interfere with the hypothalamic-pituitary-adrenal axis function whereas E1+/E4- mutants show no discernible effects in vitro. A normal cortisol secretion is important for an adequate response to stress, infection and inflammation. On this basis, E1+/E4- vectors may represent a safer tool than E1-/E4+ vectors for in vivo gene therapy. The high level of transduction in adrenocortical cells suggests that adenoviral vectors may be useful for suicide gene therapy of adrenocortical carcinomas with poor prognosis and for corrective gene therapy in congenital adrenal hyperplasia.

Julio C. Aliberti, NIAID
Lipoxin-mediated inhibition of IL-12 production by dendritic cells: A novel regulatory pathway for the innate immune response
The induction of IL-12 from dendritic cells (DC) is thought to be a key event in the initiation of cell mediated immunity. We have previously shown that microbial stimulation with a soluble extract (STAg) of the protozoan parasite Toxoplasma gondii results in a potent DC IL-12 response both in vitro and in vivo and that this cytokine production depends on the expression of the chemokine receptor CCR5 on the responding DC. Nevertheless, after initial STAg inoculation mice undergo a transient state, termed paralysis, in which DC cannot be restimulated in vivo by further injection of STAg. We now show that DC paralysis is associated with the down-regulation of CCR5 on DC freshly isolated from STAg-injected animals. Interestingly, Lipoxin A-4 (LXA-4), an anti-inflammatory eicosanoid derived from the catalysis of arachidonic acid by lypoxygenases (LO) also known to induce down-modulation of CCR5 as well as other chemokine receptors, is induced after microbial stimulation in vivo, peaking by 6 hr in serum and splenic homogenates. To assess the possible role of lipoxins in the down-regulation of CCR5 expression as well as IL-12 production by DC in vivo, mice were treated with LO inhibitors during the induction of paralysis and at the time of STAg challenge. This procedure resulted in the restoration of normal IL-12 responsiveness. To confirm the involvement of LO in the induction of DC paralysis we used 5- and 15-LO knockout animals. Importantly, mice deficient in 5-LO failed to display normal paralysis following STAg injection while 15-LO deficient animals were as readily paralyzed as wild-type mice. The defect in paralysis induction in the 5-LO mice was associated with a more rapid recovery of CCR5 expression in comparison to both the wild-type control and 15-LO deficient animals. Finally, in support of the above in vivo observations, DC cultured overnight in the presence of LXA-4 showed markedly suppressed IL-12 responses to STAg, but not to LPS in vitro. These findings demonstrate a novel in vivo pathway for the regulation of DC IL-12 production mediated by 5-LO derived LXA-4.This mechanism may have evolved to prevent the immunopathologic consequences of uncontrolled cytokine production by these cells.

Maria G. Ferrari, NIDDK
Role of Edg1 receptor during embryonic development
Sphingolipid signaling pathways have been implicated in many cellular events. Sphingosine-1-phosphate (SPP), a sphingolipid metabolite found in high concentrations in platelets and blood, stimulates members of the endothelial differentiation gene (Edg) family of G protein-coupled receptors and triggers cell growth, survival, migration, and morphogenesis. During embryonic development in mice, Edg1 receptor is highly expressed in forebrain, heart and endothelial cells of arteries and capillaries and weakly in smooth muscle cells (SMC) surrounding the aorta and vessels. Disruption of the Edg-1 gene in mice (Edg1 KO) leads to intrauterine death at embryonic day (E) 12.5-E14.5 due to hemorrhage. Embryos have underdeveloped limbs and craniofacial defects. Vasculogenesis and angiogenesis appear normal. However, vascular maturation is incomplete because SMC are clustered to the ventral side of vessels and fail to surround them uniformly. Results suggest that SPP regulates migration of SMC around vessels via Edg1.To address how Edg1 functions in a particular tissue, we are using the Cre/loxP system to produce tissue-specific knockouts. We have generated knockout mice in which Edg1 is deleted in ectodermal-derived tissues by using mice expressing cre under the nestin promoter. Ectoderm generates tissues such as brain, craniofacial structures (skull and cartilage), the outflow tract and wall of aortic arches. Mice expressing the Edg1 gene flanked by loxP sequences (Edg1-lox) were crossed with nestin-cre mice. Edg1-lox/nestin-cre embryos die at E12.5-E14.5.Unlike the Edg1 KO, conditional mutant embryos do not show bleeding at E12.5; the aorta and brain vessels are covered by SMC, although pericardial edema is observed. The limbs are well developed. Mutant embryos have similar craniofacial defects as the Edg1 KO. Some of the Edg1-lox/nestin-cre embryos have defects in neural tube closure. The phenotype of the Edg1-lox/nestin-cre embryos suggests that Edg1 is important in neural tube closure and neural crest development and/or migration. Currently we are analyzing the causes that might involve proliferation, cell death, migration or differentiation of the ectodermal-derived cells.

Kristina A. Archer, NIAID
Analysis of the Mechanism of Decoy Receptor Regulation of TRAIL Cytotoxicity in Lymphocytes
The tumor necrosis factor (TNF) related apoptosis inducing ligand (TRAIL) receptor subfamily consists of four receptors, TR1, TR2, TR3, and TR4 that coordinately regulate the cytotoxic function TRAIL. It has been observed that TRAIL exhibits selective cytotoxicity against cancer cells. The extracellular region of the four receptors share two conserved cysteine-rich domains (CRD) and a partial CRD. When TR1 and TR2 bind TRAIL ligand, the cytoplasmic death domains of these receptors aggregate to transduce the death signal, causing caspase-mediated apoptosis. However, both TR3 and TR4, which lack functional cytoplasmic death domains and therefore are known as decoy receptors, are not able to activate apoptosis. It has yet to be determined how these decoy receptors regulate the signaling of TR1 and TR2.Previously, it had been thought TR3 and TR4 trimers compete with TR1 and TR2 for ligand binding. However, it was found that the decoy receptors have lower affinity for TRAIL than the functional receptors, indicating that ligand competition would not be an efficient method of regulation. Due to the high homology among the TRAIL receptors, it is possible that TR3 and TR4 pre-associate to form mixed complexes with TR1 and TR2.To investigate whether decoy receptors can cross-associate, we have performed co-immunoprecipitation assays using differentially-tagged receptors as well as fluorescence resonance energy transfer on cells transfected with fluorophore-tagged receptors. Cross-association occurring between the TRAIL receptors has been observed. In order to determine if TR3 and TR4 can protect cell from TRAIL-induced death, we performed a dominant interference assay in which we took cells that endogenously express TR2 and transfected them with TR3 or TR4.We found that the presence of the decoy receptors protected the cells from death in the presence of TRAIL. Finally, to determine whether heterogeneous receptor complexes regulate TRAIL function, mutant receptors, that were unable to bind ligand but retain the ability to pre-associate, were transfected into cells lacking endogenous TRAIL receptors. The occurrence of cell protection in the absence of ligand binding would indicate that ligand competition is not a method of regulation. Thus far, these experiments have shown that cross-association between decoy and functional TRAIL receptors are possible.

Gerald S Baron, NIAID
CELL-FREE CONVERSION OF RAFT-ASSOCIATED PRION PROTEIN TO THE PROTEASE-RESISTANT STATE
Transmissible spongiform encephalopathies (TSEs) are infectious neurodegenerative diseases. The infectious agent, or prion, may be composed of an isoform of a glycosylphosphatidylinositiol (GPI)-anchored protein called PrP. Like other GPI-anchored proteins, PrP is enriched in sphingolipid-cholesterol-rich membrane microdomains, or rafts. In a cell culture model using persistently infected neuroblastoma cells, localization to rafts is required for the conversion of the normal, protease-sensitive isoform (PrP-sen) to the protease-resistant TSE-associated isoform (PrP-res). However, the effect of membrane association on the interaction between PrP-sen and extracellular PrP-res molecules is unknown. The ability of raft-associated PrP-sen to interact with exogenous PrP-res molecules was directly assessed in an attempt to model the molecular processes that occur during the initial infection of cells with exogenous PrP-res-associated TSE agents. A cell-free conversion reaction approximating physiological conditions was developed containing purified rafts as a source of PrP-sen and crude brain microsomes from either normal animals or TSE-infected mice, as a source of PrP-res. Raft-associated PrP-sen was not converted to PrP-res unless the PrP-sen was released from rafts by treatment with phosphatidylinositol-specific phospholipase C (PI-PLC), which does not release PrP-res, or the combined membrane fractions were treated with a fusogenic concentration of polyethylene glycol (PEG). The exact co-migration during SDS-PAGE of the PI-PLC-induced conversion product with that formed using purified PI-PLC-treated PrP-sen suggested the former molecule lacked a GPI-anchor. Conversely, the PEG-induced conversion product possessed a GPI-anchor as determined by PI-PLC digestion after conversion. PEG pretreatment of PrP-res microsomes alone did not enhance their inherent conversion efficiency of purified PrP-sen. Similar results were obtained using rafts from three different cell lines and prepared by detergent-based and detergent-free methods, indicating the results are neither cell clone-specific nor artifacts of the raft preparation procedure. These observations raise the possibility that as TSE infections spread along neural pathways in the host, cell-to-cell transfer across synapses requires exchange of membranes or insertion (i. e. , 'painting') of infecting PrP-res into the raft domains of recipient cells.

Brigitte E. Beer, NIAID
Characterization of Novel Simian Immunodeficiency Virus from Redcapped Mangabeys (Cercocebus torquatus) - Implications for HIV and the Origin of the Vpx Gene
Primate lentiviruses are presently classified into six lineages, SIVcpz from chimpanzees, SIVsm from sooty mangabeys, SIVagm from African green monkeys, SIVlhoest from L'Hoest monkeys, SIVsyk from Sykes' monkeys, and SIVcol from Colobus monkeys. These six lineages are approximately equidistant and share a common genome structure (LTR-gag-pol-vif-vpx/r-tat-rev-env-nef-LTR). Two of those lineages, SIVcpz and SIVsm, are the source of HIV-1 and HIV-2, respectively. So far, only the SIVsm/HIV-2 lineage has been shown to encode for a vpx gene. On a recent journey to Nigeria, we obtained samples from 12 feral redcapped mangabeys (Cercocebus torquatus) in order to characterize novel SIVs and determine the implications of these viruses for humans. Two redcapped mangabeys were seropositive for SIV, and the new virus strains were named SIVrcm. Molecular characterization showed that SIVrcm is recombinant between different lentiviruses that circulate in the primate habitat (primarily SIVcpz and SIVagm) and therefore cannot be classified into a single lineage. This finding is comparable to HIV-1 strains that are recombinant between different clades except that the parental strains of SIVrcm are more divergent. SIVrcm also encoded for a vpx gene in addition to vpr, suggesting a common ancestor with SIVsm and HIV-2.Surprisingly, SIVrcm and SIVsm/HIV-2 were highly divergent across the entire genome, including vpx (42% identity). It has been discussed that vpx originated from a nonhomologus recombination with vpr from SIVagm from the sympatric sabaeus monkey and we created phylogenetic trees of vpx and vpr including SIVrcm that supported this theory. Therefore, the second hypothesis for the origin of vpx - that it was derived from a duplication of the vpr gene - seems unlikely. In terms of coreceptor usage, SIVrcm utilized CCR2B instead of CCR5.The lack of CCR5 usage was possibly due to a unique 24 bp deletion in the CCR5 gene that occurred at an allelic frequency of 77 percent in Nigerian redcapped mangabeys. Therefore, SIVrcm is an example of a primate lentivirus that adapted its coreceptor usage to allow productive replication in its natural host. Similarly, HIV-1 is able to use alternative coreceptors in order to broaden its host range, for example CCR8 in human thymocytes. Finally, SIVrcm replicated efficiently in primary human PBL blasts to an extent similar to HIV-1 and HIV-2.This demonstrates that SIVrcm could result in a HIV-3 epidemic if transmitted to humans.

John M. Beggs, NIMH
INFORMATION IS STORED OVER MANY HOURS BY SEQUENCES OF NEURAL ACTIVITY IN SMALL NETWORKS OF CORTICAL NEURONS
While tremendous effort has been directed toward understanding the neural mechanisms of memory at the synaptic and systems levels, almost no experimental work has been done to investigate how small networks of neurons could store information. We hypothesized that small groups of neurons could store information over many hours, possibly in patterns of neural activity.
To address this issue, we cultured slices of rat cortex for 4 weeks and then recorded their spontaneous activity for 48 hrs with an array of 60 microelectrodes. Slice cultures were chosen to preserve the cortical circuitry that is commonly destroyed in dissociated cultures. Other labs using dissociated cultures reported that spontaneous activity was "random" and "synchronously bursting". At long time scales, the activity in our slice cultures also seemed to fit this pattern, but when local field potentials were binned at 4 milliseconds, a dramatically different view of activity emerged. The cortical cultures produced spatial patterns of activity that appeared in distinct sequences that would later repeat themselves. Analysis programs were written to identify an activity sequence and group it with other sequences of similar structure. Similarity between sequences was defined as the fraction of electrodes showing overlapping activity. For statistical significance, similarity values from actual data were compared to similarity values from 20 shuffled control data sets. Information theory was used to quantify the information shared between distributions of activity sequences recorded during the first hour and subsequent hours.
The cultures (n = 4) produced 1,500 ± 200 (mean ± s.d.) sequences per hour that could be clustered into 110 ± 23 distinct groups. In contrast, shuffled data produced significantly fewer (20 ± 15) distinct groups. To investigate whether these structured sequences were robust over time, recordings made in later hours were compared to the activity recorded in the first hour. Remarkably, many of the activity sequences were repeated with nearly exact similarity even 37 hrs later. In fact, the distributions of activity 37 hrs later retained 94 ± 5 % of the maximum information that was theoretically possible, while shuffled data retained significantly much less (5 ± 4%). The diversity and persistence of robust activity sequences in cortical slice cultures strongly suggests that we have revealed basic elements of memory at the small network level.

Igor M. Belyakov, NCI-CCR
The importance of mucosal CTL vs systemic CTL in reducing viral load after intrarectal challenge of rhesus macaques with pathogenic SHIV
The mucosal surface is a major natural route of HIV-1 entry and the gut is a major site of HIV replication. Thus one likely characteristic of a successful HIV vaccine is the ability to induce an immune barrier at mucosal sites. Recent studies from our lab and others of the mucosal transmission of virus in mouse models, as well as in monkeys and humans, indicate that HIV-specific cytotoxic T lymphocytes (CTL) are an important component of the mucosal immune barrier. These considerations led us to hypothesize that CTL generated in the gut mucosa by mucosal immunization may be more effective at clearing the virus than CTL generated elsewhere by systemic immunization. Here we demonstrate that mucosal (colon and mesenteric lymph node) SIV-specific CTL responses can be induced by intrarectal immunization of Mamu A*01-positive monkeys with synthetic HIV/SIV peptide vaccines in the presence of a mucosal adjuvant LT(R192G). After intrarectal challenge with pathogenic SHIV-Ku, monkeys immunized intrarectally with the HIV/SIV peptide vaccine showed a rapid decline of viral titer to undetectable levels, both in blood and in the gut reservoir, a significantly lower set point compared with subcutaneously (SC) immunized animals and adjuvant only controls, perhaps because the CTL were focused on the major site of SHIV replication. The numbers of CD4+ and CD8+ cells of IR immunized animals were also better preserved in intrarectally immunized monkeys, and the colonic CTL directly ex vivo were higher 200 days after challenge than in the other groups. Moreover, reduction of viremia correlated with post-infection CTL activity in the gut, and with clearance of SHIV from the gut mucosal tissues which were probably the major site seeding virus into the blood stream. Thus, mucosal peptide immunization can induce mucosal CTL in primates and impact the course of pathogenic retroviral infection. These data thus for the first time provide important support for the concept in primates that the induction of SIV or HIV-specific CTL responses in the mucosa, the main site of SIV/HIV replication, is at least as important as inducing high level systemic CTL responses in protection against HIV/SIV infection, not just because of mucosal transmission, but because of the gut as a major reservoir of virus. This fundamental concept will be important for the optimal design of effective AIDS vaccine.

Julie M. Bork, NIDCD
MUTATIONS IN CDH23 CAUSE CONGENITAL DEAFNESS (DFNB12) AND A DEAF/BLINDNESS SYNDROME (USH1D)
Previous genetic mapping studies had defined broad overlapping regions on human chromosome 10q for loci causing a form of nonsyndromic congenital deafness (DFNB12) and congenital deafness associated with progressive blindness (Usher syndrome type 1D, USH1D). In this study, we mapped nine consanguineous families segregating recessive deafness to chromosome 10q21; one family allowed us to refine the DFNB12 locus to 0.5 cM. We identified candidate genes from raw genomic DNA sequence using a positional cloning strategy. One great candidate, a novel cadherin-like gene, eventually named CDH23, was initially detected using gene prediction programs and BLAST homology searches against public and private databases. PCR amplification and northern blot analysis allowed us to demonstrate CDH23 expression in the cochlea and retina of humans. We sequenced probands from each family linked to DFNB12 and USH1D, since their overlapping regions suggested they might be allelic. We found missense mutations in five families with congenital deafness, and nonsense and splice site mutations in four families with congenital deafness and progressive blindness. Thus missense mutations of CDH23 result in isolated congenital deafness, while nonsense and splice site mutations yield congenital deafness associated with progressive blindness. Cadherin-23 is predicted to be a transmembrane protein with a large extracellular domain containing multiple cadherin repeats. Similarity to known cadherin proteins suggests that these repeats may function in a calcium-dependent manner in cellular adhesion. The cytoplasmic domain shows no similarity to any known proteins or protein motifs in the database. Therefore we are attempting to identify interacting protein partners of the unique cytoplasmic domain by yeast two-hybrid and phage display analyses. An understanding of the expression of cadherin-23 in the retina is especially important since CDH23 mutations cause progressive blindness, as opposed to congenital deafness. Preliminary immunohistochemical experiments with antibodies generated against the cadherin-23 cytoplasmic domain suggest it is localized to the pigmented epithelium, outer limiting membrane, inner nuclear layer and ganglion cell layer of the mouse retina. Our ongoing investigation of CDH23 expression and cadherin-23 localization should provide insight into the pathology of the human retina in Usher syndrome patients.

Cesario V. Borlongan, NIDA
BRADYKININ B2 RECEPTOR AGONIST, CEREPORT, FACILITATES NEUROPROTECTIVE EFFECTS OF CYCLOPSORINE-A IN AN ANIMAL MODEL OF PARKINSONISM
While mainly used as an immunosuppressant, newly identified properties of Cyclosporine-A (CsA) suggest its potential as a therapeutic agent for neurological disorders. Accumulating evidence reveals that CsA may be considered a multiple site of action therapeutic drug since it can modulate brain functions by blocking calcium channels, restoring mitochondrial energy metabolism, providing neurotrophic factors and scavenging excess free radicals. However, CsA does not easily cross the blood brain barrier (BBB), thus requiring high doses (>10 mg/kg) and chronic administration of the drug to produce protective effects. Inherent with this regimen are deleterious immunosuppressant side effects, including nephrotoxicity and hepatotoxicity. If a means of facilitating CsA entry into the brain could be identified, then it might be possible to use lower, but still therapeutically active doses of CsA, while avoiding the side effects associated with high dose. Because the bradykinin B2 receptor agonist, Cereport, has been shown to produce a transient increase in BBB permeability, we examined the effects of combined CsA and Cereport treatment in the unilateral 6-OHDA model of Parkinson's disease. Adult rats were intravenously injected daily (for 7 consecutive days pre-lesion) with vehicle, CsA alone (1 mg/kg, a low dose without either immunosuppressive or neuroprotective effects, or 10 mg/kg, a high dose that produces both immunosuppression and neuroprotection), or CsA (1 mg/kg) in combination with Cereport (9 µg/kg, a dose that allows entry of 1kDalton-size molecules for 15 mins). Established motor asymmetry tests, at 1 month post-lesion, revealed that a low dose of CsA was neuroprotective when combined with Cereport, but not when given alone. Tyrosine hydroxylase (TH) immunohistochemistry demonstrated that while near complete (>90%) depletions of nigral TH+ neurons were noted in lesioned rats that received vehicle infusion or low dose CsA alone, lesioned rats that received low dose CsA + Cereport exhibited a significant protection of nigral TH+ neurons and a marked reduction in the loss of striatal TH+ fibers. These histological data support the sparing of dopamine nigrostriatal pathways as the mechanism underlying attenuation of 6-OHDA-induced behavioral abnormalities. The safe and effective administration of low dose CsA with Cereport-enhanced BBB permeability offers a novel way of producing protective effects in the CNS without the toxic liabilities of high dose CsA.

Emmanuel Bourdon, NICHD
Selective targeting and degradation of oxidized proteins: IRP2 as a model protein.
Oxidative stress is a key element in the genesis of many diseases including atherosclerosis, aging, diabetes and Parkinson's disease. Iron and oxidative stress are intimately associated and mammalian cells have developed sophisticated regulatory iron uptake and distribution systems. These latter are accurately regulated through the Iron Regulatory Proteins, IRP1 and IRP2, which differs from each other in their mode of regulation. In iron replete cells, IRP1 gains an iron-sulfur cluster and serves as a cytosol aconitase, IRP2, on which our attention has been focused, is oxidized and degraded. Our results showed that a 73 amino acids degradation domain allows in IRP2 to sense iron, be oxidized, ubiquitinated and degraded in presence of this metal. We determined the site and nature of IRP2 oxidation and calculated an initial Kd of IRP2 for iron. In vivo studies have shown that manganese, cobalt, and copper, ions known to compete for iron binding sites, can protect IRP2 from degradation.
These data constitute the first example of an oxidation event acting as a signal for proteosomal degradation. To identify other proteins subject to iron-dependent oxidation and degradation, cellular lysates from RD4 cells grown with or without iron and lactacystin, a proteasome inhibitor, were assayed for oxidative damage by the carbony assay. We found that over 100 proteins are oxidatively modified and then selectively degraded whereas the non-oxidized forms of these proteins are relatively stable. In vivo and in vitro studies also revealed that iron-treatment dramatically increased ubiquitination implying that many proteins are ubiquitinated in response to iron-induced oxidative damage. Using proteomic techniques, we have identified 10 oxidized proteins and are currently demonstrating iron-dependent oxidation and ubiquitination for each. While the proteasome has been previously shown to be involved in degradation of some oxidatively damaged proteins, our observations constitute the first demonstration that ubiquitination and proteasomal degradation is a general pathway for the turnover of oxidatively damaged proteins. The selective targeting and removal of oxidatively modified proteins such as IRP2 represents an important and under-appreciated component of the cellular defense against oxidative stress and the diseases associated with it.

Jay H Bream, NCI-CCR
Identification of an IL-2 Responsive Element in the Human IFN-gamma Promoter
IFN-gamma is a multifunctional cytokine that is involved in the development of Th1 cells and in cellular responses to a variety of pathogens. IL-2 and IL-15 are known to induce the expression of IFN-gamma from NK and T cells. Despite concentrated efforts to identify the transcriptional control mechanism(s) by which IL-2 or IL-15 induces IFN-gamma mRNA expression, no such genomic regulatory regions have been described. Utilizing DNAse I hypersensitivity analysis, we have identified a potentially new regulatory region distal to previously known elements. We have further characterized this region and found a site that has sequence homology to a consensus STAT 5 binding site. To test for the ability of this site to bind STATs, we performed electrophoretic mobility shift analysis (EMSA) with nuclear extracts from IL-2, IL-12, or IL-15 treated PHA blasted primary human T cells. A strong, unique complex was formed with extracts from IL-2 or IL-15 stimulated cells but not with IL-12 treated cells. Disruption of the potential STAT site eliminated binding of the IL-2- or IL-15-induced complex. Supershift analysis identified both STAT 5A and STAT 5B in this complex suggesting the formation of STAT 5A/STAT 5B heterodimers. To investigate further the ability of STAT 5 proteins to interact with this site, we employed chromatin immunoprecipitation analysis (CHIP) and show in vivo binding of IL-2-induced STAT 5A and STAT 5B to this region. Moreover, in preliminary experiments, this distal STAT site was linked 5' to an existing luciferase construct containing the core 2.9 kb IFN-gamma promoter. When trasfected into a T cell line, the -2.9/STAT construct was responsive to IL-2 stimulation while the -2.9 kb vector was not. This report is the first evidence of IL-2- or IL-15-induced STAT proteins interacting with the human IFN-gamma promoter.

Fraser D. Brown, NHLBI
Phosphoinositides and Arf6-regulated membrane traffic
ADP-ribosylation factor (Arf) proteins are a family of Ras-related GTPases that regulate membrane trafficking. Arf6 regulates the movement of membrane between the plasma membrane (PM) and a novel endosomal compartment and influences actin structure at the PM. Phosphoinositides, such as phosphatidylinositol-4,5-bisphosphate (PIP2) are membrane-associated lipids implicated in membrane trafficking and actin remodelling. Arf6 was recently shown, in vitro, to activate phosphatidylinositol-4-phosphate 5-kinase (PIP5K), an enzyme that catalyses PIP2 production. This led us to determine whether Arf6 regulates membrane traffic through PIP5K and its product, PIP2.To localize PIP2 we used a fusion protein composed of the PH domain from phospholipase C fused to green fluorescent protein. PIP2 co-localized with Arf6 at the PM and on endosomal structures. Persistent activation of Arf6, achieved by expressing Arf6Q67L, a constitutively active mutant, induced some membrane ruffling and the formation of PIP2-positive, actin-coated vacuoles that accumulated in the cell. These vacuoles trapped PM proteins that traffic through the Arf6 endosomal membrane system. PIP5K overexpression mimicked the effects seen with Arf6 Q67L. Examination of the dynamics of PIP2 trafficking in live COS cells showed some movement of tubulo-vesicular structures in resting cells. Stimulation of Arf6 activity by expression of EFA6, an Arf6 exchange factor, resulted in continuous membrane ruffling and stimulated the formation of PIP2 positive macropinosomes that were dynamic and capable of recycling membrane back to the PM. In contrast, expression of Arf6Q67L or PIP5K induced the formation of vesicular structures that fused with each other and resulted in the accumulation of the enlarged vacuoles. Failure to recycle this membrane to the PM halted membrane ruffling. Collectively, these results confirm that PIP5K is a major in vivo target of Arf6 and suggest that recycling of membrane into and out of the Arf6 endosome requires repeated cycles of Arf6 GDP/GTP exchange and synthesis and turnover of PIP2.Understanding how Arf6-regulated PIP2 turnover controls the formation of protrusions and vacuoles will clarify the relationship between membrane traffic and actin rearrangements that underlie the changes in cell shape during dynamic processes such as cell migration and metastasis.

Dmitry V. Bulavin, NCI-CCR
Initiation of a G2/M checkpoint after UV radiation requires p38 kinase
The response to genotoxic stress can be considered a multistage process involving initiation of cell cycle arrest as well as maintenance of arrest during DNA repair. While maintenance of G2/M checkpoint(s) involves Chk1, Chk2/Rad53 and upstream components, the mechanisms involved in its initiation are less well defined. Here we report that p38 kinase has a critical role in the initiation of a G2 delay after UV radiation. Inhibition of p38 blocked the rapid initiation of this checkpoint in both human and murine cells after UV radiation. p38 was found to bind and phosphorylate in vitro Cdc25B at Ser309 and Ser361 as well as Cdc25C at Ser216, and these were required for binding to 14-3-3 proteins. Inhibition of p38 prevented in vivo phosphorylation of Cdc25B at Ser309 and 14-3-3 binding after UV radiation, and mutation of the site was sufficient to inhibit the checkpoint initiation. In contrast, in vivo Cdc25C binding with 14-3-3 was not affected by p38 inhibition after UV radiation. We propose that regulation of Cdc25B phosphorylation by p38 is a critical event for initiation of the G2/M checkpoint after UV radiation.

Stephen C. Bunnell, NCI-CCR
The Visualization of Dynamic Changes in T Cell Morphology and the Distribution of Signaling Molecules During T Cell Activation.
The engagement of the T cell receptor (TCR) induces changes in T cell shape and cytoskeletal architecture that are concurrent with the biochemical events triggered by the TCR. Physiologically, these changes result in the apposition of the responding T cell to a stimulatory antigen-presenting cell, and the subsequent development of a complex structure referred to as the "immunological synapse". However, efforts to study the earliest events in this process in real time have been frustrated by the complex topology of this interaction. To facilitate the collection of confocal images of activated T cells at high temporal and spatial resolution, we plated Jurkat T cells on coverslips coated with stimulatory antibodies to the TCR and dynamically visualized both the developing contacts and the distribution of EGFP-tagged molecules within these contacts. By tracking the distribution of EGFP-tagged actin over time we established that the inital contact between the T cell and the stimulatory coverslip triggers the directed extension of lamellipodia toward the coverslip and the growth of the contact region. Within 2 minutes the expanding contact becomes bounded by a continuously remodeled circumferential actin-rich ring. These processes are strictly dependent on actin polymerization and the activation of the Src-family tyrosine kinases Lck and Fyn. Cytoplasmic calcium increases also play a role, as intracellular calcium chelators impair both the recruitment of EGFP-actin to the coverslip and the expansion of the contact. In addition, LAT, a crucial hematopoietic adaptor, plays an important role in these processes, initially enabling more extensive and uniformly directed cytoskeletal rearrangements, and subsequently facilitating the maintenance of T cell-coverslip contacts. Studies in which the degree of morphological rearrangement induced by TCR ligation was carefully quantitated revealed that both the PLC-gamma1 and Grb2/Gads binding sites in LAT are required in order to promote the transient and sustained morphological changes described above. Having established a mophological context for the biochemical events triggered by TCR ligation, we have begun to track the distribution of essential signaling molecules during process of contact formation. Our recent observations suggest that crucial signaling adaptors are recruited to sites of T cell-coverslip contact within seconds of their formation.

Kevin D. Burroughs, NIEHS
Insulin-like growth factor-I (IGF-I) stimulates vascular endothelial growth factor (VEGF) synthesis in prostate epithelial cells: requirement of multiple signaling pathways
Growth of solid tumors beyond 1-2 millimeters in diameter requires recruitment of a blood supply by stimulating the growth of new capillaries from existing host vasculature, a process known as angiogenesis. VEGF is a potent angiogenic protein that is overexpressed in a subset of prostatic intraepithelial neoplasia (PIN) foci, indicating that it may promote the transition of latent, precancerous lesions to prostatic carcinoma. In light of the fact that high serum IGF-I levels represent a putative risk factor for prostate cancer, we investigated IGF-I signaling pathways that influence VEGF synthesis in primary cultures of normal, human prostate epithelial cells (PrEC). Treatment of confluent PrEC with Long-R3 IGF-I increased the concentration of VEGF165 in culture medium in a dose-dependent manner. A phosphatidylinositol 3 (PI3)-kinase inhibitor LY294002 or MAP kinase kinase (Mek1) inhibitor PD98059 each blocked IGF-I stimulation of VEGF synthesis, demonstrating a dependence on both pathways for this response. IGF-I increased nuclear levels of hypoxia-inducible factor-1 (HIF-1), a transcription factor elevated by hypoxia and important for hypoxic stimulation of VEGF production. LY294002 blocked IGF-I-induced elevation of HIF-1; however, PD98059 had no effect. In PrEC, IGF-I also induced the expression of c-fos, a transcription factor that is a downstream target of MAP kinase signaling and may regulate VEGF expression. PD98059, but not LY294002, inhibited IGF-I induction of c-fos. Treatment of an immortalized prostate epithelial cell line (PZ-HPV-7) with IGF-I also increased VEGF165 in culture medium. However, production of VEGF was inhibited by LY294002, but not by PD98059 in the PZ-HPV-7 cells, which unlike PrEC constitutively express c-fos. Reporter assays using a 1.5 kb region of the human VEGF promoter indicated that IGF-I increases transcription of the VEGF gene and that both PI3-kinase and MAP kinase pathways participate in regulating VEGF transcription. A dominant negative c-fos blunted VEGF reporter activity in response to IGF-I in PZ-HPV-7 cells. These data indicate that IGF-I stimulation of VEGF synthesis in prostate epithelial cells requires both PI3-kinase and MAP kinase pathways, which increase HIF-1 and c-fos, respectively. Further characterization of the pathways that regulate VEGF production will likely provide new pharmacological targets to prevent the growth and metastasis of solid tumors.

Stefano Campaner, NCI-CCR
A mitotic role for SIL suggested by cell cycle regulation and interaction with PIN1

SIL is an immediate early gene whose expression correlates with proliferation in a wide array of cell lines and human tissues.
The SIL gene was originally discovered in a chromosomal rearrangement in which the SCL hematopoietic transcription factor was placed under the control of the SIL promoter. This rearrangement is found in 25% of childhood T-cell acute lymphoblastic leukemias. SIL is required for maintenance of the midline during embryonic development. Mouse embryos homozygous for a null mutation in the SIL gene display apoptosis of the midline structures, abnormal left/right asymmetry and blocked Sonic Hedgehog signaling. In an attempt to understand the molecular details of SIL function we set out to analyze the regulation of the SIL protein during the cell cycle. While the SIL mRNA remains stable throughout the cell cycle, the amount of protein increases as cell progress form G1 to G2/M. As cells transit mitosis, SIL disappears, becoming undetectable in early G1.The pattern of down regulation matches the degradation of cyclin B1, suggesting that SIL disappears during anaphase. In addition, SIL is hyperphosphorylated in cells blocked at prometaphase. PIN1 is a recently discovered peptidyl-prolyl isomerase that regulates progression through mitosis. PIN1 is able to bind a subset of mitotic phosphorylated proteins and thereby modulate their biological activity through the isomerization of prolyl phospho-Ser/Thr peptide bonds. We therefore set out to test whether SIL could be a substrate of PIN1 using a GST-pull down assay. The results show that SIL is specifically able to bind PIN1.The association is cell cycle regulated since only SIL from mitotic extract is pulled down in a GST based assay. Furthermore, the interaction depends on the SIL phosphorylation state and can be abolished by phosphatase treatment. SIL binding is mediated by the WW domain of PIN1.We also mapped the interaction site on SIL. The SIL sequence displays three clusters of putative phosphorylation sites. Using a combination of deletions and point mutations we mapped the binding site to the second of the three clusters. This spans a region containing 8 phosphorylation sites. SIL point mutated in cluster II is not phosphorylated during mitosis and no longer binds to PIN1.
Overall, our data identify SIL as a target of the PIN1 pathway, suggesting a role for the SIL in the control and/or execution of mitotic progression.

Liu Cao, NIDDK
Premature cell senescence induced by loss of Brca1 (breast tumor suppressor gene 1) can be overcome by the inactivation of p53 at increased risk of malignant transformation
Breast cancers occur at high frequency and affect 1 in 9 American woman. Germline mutations of Brca1 result in 90% of familial cases of breast cancer and 50% cases of combined breast and ovarian cancers. Previous investigations indicated that targeted disputation of tumor suppressor Brca1 in mouse mammary epithelium results in tumorigenesis, however, its germline mutations cause embryonic lethality with diminished growth defect. How can the same gene be required for embryonic development and tumor suppression? What is the relationship between Brca1 and p53 in embryonic development and tumorigenesis? To gain understanding into these contradictory issues, we characterized Brca1 mutant embryos and embryonic fibroblast (MEF) cells. We found that late stage Brca1 mutant embryos and MEF cells derived mutant embryos exhibited senescence. Starting from early passages, mutant cells displayed flattened and enlarged morphology, irreversible permanent cell cycle arrest, and exhibited spontaneously apoptosis. The senescent Brca1 mutant embryos and cells were accompanied by expression of senescence-associated b-galactosidase, upregulation of p53 and p53-dependent accumulation of p21.Moreover, the Brca1 mutant mice also exhibited aging-related changes in their skeletal and skin, including osteoarthritis and abnormally thickened epidermis. To investigate whether the senescence was due to p53 activation, we introduced a p53 null mutation into Brca1 mutant cells and found that lack of p53 prevented senescence-associated phenotypes and spontaneous apoptosis. Notably, some Brca1 mutant cells could overcome senescence and become immortalized through spontaneous mutation of p53.Interestingly, the immortalized cells showed increased proliferation capacity and higher expression of cyclins A and D1.Although occurring at low frequency, immortalized cells could undergo malignant transformation as manifested by the foci formation in soft agar and tumorigenesis in nude mice while such events were not observed in immortalized wild type control cells. These observations suggested that loss of Brca1 induced p53-dependent senescence, which may act as a physiologic and/or pathologic response, result in embryonic lethality. In contrast, loss of p53 allowed Brca1 mutant cell survival, accumulation of further genetic alterations, and ultimately leads to tumorigenesis. This study provides new insights into a mechanistic basis of Brca1 function in embryonic development and tumorigenesis.

Kimberley T. Cesen-Cummings, NIEHS
Pregnancy-related signaling pathways as therapeutic targets for uterine leiomyoma
Uterine leiomyomas, benign smooth muscle tumors of the uterus more commonly called fibroids, are the most prevalent gynecological neoplasm in women of reproductive age, yet currently there is no adequate medicinal therapy for this disease. Hysterectomy, the most commonly recommended treatment for symptomatic leiomyomas, is unacceptable for a majority of women, especially those who wish to retain their fertility. Pharmacological interventions such as hormone manipulation using GnRH agonists or androgens have limited and varying success and are accompanied by serious side effects. Our goal is to better understand the etiology and pathobiology of uterine leiomyomas so that we may develop and implement more effective therapeutics for the treatment of these tumors. Leiomyomas have phenotypes similar to differentiated myometrial cells of pregnancy, but unlike the myometrial cells found during pregnancy, the tumor cells have lost or cannot respond to the key signals for dedifferentiation or cell death. Moreover, if these signals could be introduced or restored in leiomyomas, the tumor cells may display a non-neoplastic phenotype. We questioned whether a phenotype characteristic of the myometrium at parturition is also an attribute of leiomyomas that arise in the Eker rat model for leiomyomas. We found that Eker rat leiomyomas and tumor-derived cells (ELT-3 cell line) express genes associated with parturition including Cx43, COX-1, COX-2, and oxytocin receptors (OTR). Moreover, in the ELT-3 cells, 17beta-estradiol increases the expression of OTR, and oxytocin binding to OTR increases intracellular cytosolic calcium demonstrating that this signaling pathway is functional. Using the ELT-3 cell line, pregnancy-related signaling pathways were investigated as potential targets for therapeutic intervention. COX-2-mediated prostaglandin synthesis was evaluated as a possible therapeutic avenue for decreasing leiomyoma growth, however, the prostaglandins PGE2 and PGF2alpha were found to stimulate, rather than inhibit, proliferation of ELT-3 leiomyoma cells. Conversely, oxytocin (50 nM, 100 nM) inhibited 17beta-estradiol-induced growth of ELT-3 cells, indicating that signaling via the OTR was antiproliferative in leiomyoma cells. These data suggest that signaling pathways such as those mediated via OTR may be useful therapeutic targets for the treatment of uterine leiomyomas.

Francis K. Chan, NIAID
Analysis of the mechanisms of TNF-induced death in T cells
Tumor Necrosis Factor alpha (TNFa) can elicit many cellular responses including programmed cell death or apoptosis through binding to its two cell surface receptors, p60 TNFR-1 and p80 TNFR-2.TNFa-induced apoptosis has been largely attributed to p60 TNFR-1.However, numerous reports have also suggested a role for p80 TNFR-2 in apoptosis induction in T cells. The mechanism by which p80 mediates apoptosis is largely unknown. To investigate the role of p80 TNFR-2 in mediating TNFa-induced apoptosis, I generated stable transfectants of the T cell leukemia cell line Jurkat that over-expresses p80 TNFR-2.Unlike parental Jurkat cells that express only the p60 TNFR-1, the p80-expressing Jurkat cells undergo programmed cell death in response to TNFa. Interestingly, TNFa-induced death in the p80-expressing cells requires signal transduction from both the p60 and p80 TNFRs. Stimulation of either receptor alone is not sufficient to trigger the response. Moreover, this TNFa-induced death can proceed in the presence of strong NF-kB induction, which confers pro-survival signals against TNF-induced cytotoxicity. Activation of p80 TNFR-2 results in the degradation of TRAF2, which binds to both the p60 and p80 signaling complexes. Two other adapter proteins, FADD and RIP, are also recruited to the p60 TNFR-1 signaling complex. Importantly, cells that are deficient in RIP are minimally sensitive to TNF-induced death that can be completely blocked by the pan-caspase inhibitor zVAD. In contrast, Jurkat cells that are deficient in FADD are exquisitely sensitive to TNF-induced death that is not blocked by the caspase inhibition. Strikingly, treatment of FADD-deficient cells with geldanamycin, which results in the abrogation of RIP, results in complete protection against TNF-induced death. Taken together, these data suggest a model of TNFR-1 signaling in which the adapter proteins FADD, RIP and TRAF2 compete for binding to the p60 TNFR-1 signaling complex. The recruitment of FADD leads to caspase-dependent death, whereas the recruitment of RIP results in caspase-independent death. Signaling through p80 TNFR-2 is required to switch on these previously dormant death machineries through degradation of TRAF2, which normally binds to the p60 signaling complex and prevents death.

Lin Chen, NIDDK
Generation of the mouse model for human Apert syndrome
Occuring at an incidence of around 1 in 2500 infants, craniosynostosis (CS) is a relatively common birth defect characterized by the premature fusion of one or more of the cranial sutures. Apert syndrome (AS) is a clinically distinct condition characterized by CS, craniofacial anomalies, and symmetric syndactyly of the digits. Recently, AS was found to be caused by mutation at a specific amino acids (Ser252) of FGFR2.To generate mouse model for AS, we introduced a Fgfr2-Ser250Trp mutation, which is equivalent to human Ser252Trp mutation, into mouse genome using homologous recombination. The Heterozygous mice carrying the Ser250Trp mutation (Fgfr2 250/+) were slightly abnormal at birth and displayed AS like phenotypes. They exhibited craniofacial malformations including mild malocclusion, shorten anterior-posterior axis, increased bottom-top height, and widely spaced eyes (hypertelorism). The mutant mice have normal limbs, however their body size is smaller than that of wild type littermate controls. A long time controversy hypothesis suggests that cranial base is a primary site of defects in CS. Our analysis of this mutant mice does not support this hypothesis as all mutant mice examined showed no sign of premature fusion of cranial bases. Histological examination revealed significantly more overlaping between osteogenic fronts of frontal and parietal bones. Notably, the overlapping osteogenic fronts in mutant skull are usually thinner than wild type. The mutant also showed significantly decreased number of the mesenchymal cells that are located between the overlapping frontal and parietal bones. Theoretically, the decrease of mesenchymal cells could be caused by alterations in cell proliferation, differentiation or apoptosis. We found no obvious changes in proliferation, differentiation of suture osteoblast cells. Then we checked apoptosis of sutures and found that the mutant coronal sutures showed increased apoptosis in the mesenchymal cells and suture bones, which is accompanied by upregulation of Bax, a pro-apoptotic protein. In summary, we generated the first real mouse model for human Apert syndrome. Our data indicated that the increased apoptosis occurred at coronal suture may be the primary reason for the AS. This mouse strain will benefit studies on the pathogenesis and therapies for Cs and provide a valuable in vivo model for the studies of FGFRs and their roles in bone development

Qiufang Cheng
, NEI
LOVASTATIN-INDUCED CATARACT CAN BE PREVENTED BY GERANYLGERANYL PYROPHOSPHATE
The cholesterol lowering agent lovastatin causes cataracts in vitro. Lovastatin not only blocks the synthesis of cholesterol, but also of the isoprenoids farnesyl pyrophosphate (FP) and geranylgeranyl pyrophosphate (GGP) which are required by certain proteins for membrane localization and activity. We have tested the hypothesis that the cataracts result from decreased levels of isoprenoids. Lenses from 75-100 g Sprague-Dawley rats were exposed in organ culture to 100 uM lovastatin +/- FP or GGP for up to 8 days. After 3-4 days lenses exposed to lovastatin alone developed opacity, starting in the equatorial region and becoming complete by 6 days. Histologically, the lovastatin-induced cataracts had markedly swollen epithelial cells and abundant vacuoles in the lens posterior cortex. Supplementation of the medium with GGP (7.5 uM) virtually eliminated these effects while FP did not. Neither isoprenoid adversely affected the lenses in the absence of lovastatin. GGP also prevented TUNEL positive apoptotic epithelial cells which were common in lovastatin-exposed lenses. Metabolic labeling with 3H-geranylgeraniol or 3H-farnesol followed by SDS-PAGE and fluorography was used to determine the complement of prenylated proteins in the lenses following 24 hour culture +/- lovastatin. Labeling was stronger in the epithelium than in lens fiber cells and was increased by pre-incubation with lovastatin, indicating depletion of intracellular isoprenoid pools. Proteins labeled by 3H-geranylgeraniol were largely in the 20-30 kDa range characteristic of the Ras superfamily small GTPases. These enzymes are the major proteins known to be prenylated and act as molecular switches regulating numerous critical cellular functions. Farnesylated proteins included some in the size range of small GTPases, but also several at higher molecular masses. The large majority of prenylated species were found in the insoluble protein fraction, consistent with membrane association.
These data support the concept that cataracts in this system result from decreased prenylation, and more specifically, the decreased geranylgeranylation of one or more proteins. It is likely that these proteins are small GTPases of the Ras superfamily. Studies using 2D electrophoresis and mass spectroscopy are underway to identify and characterize these species.

Uimook Choi, NIAID
Persistent Low Level Retrovirus Gene Marking of Hematopoietic Stem Cells Following Non-Myeloablative Chemotherapy Conditioning in a Rhesus Macaque Model
In the absence of in vivo selection, clinically successful autologous hematopoietic stem cell gene therapy will likely require marrow conditioning to reduce the number of resident stem cells. For application of human gene therapy studies, a chemotherapy-based marrow conditioning regimen may be safer than one which involves ionizing radiation. We used rhesus macaques to determine whether a non-myeloablative chemotherapy regimen is comparable to non-myeloablative total body irradiation (TBI) facilitating engraftment of autologous gene-marked stem cells. The chemotherapy regimen is similar to clinical trial for allogeneic stem cell transplantation. Cytokine-mobilized rhesus peripheral blood stem cells from 3 animals were enriched by immunoadsorption for CD34+ cells and transduced with a clinical-grade retrovirus (MFGS-gp91). The cells were cultured in X-VIVO 10 media containing 1% HSA, Pixykine (interleukin 3/GM-CSF fusion protein) 10ng/ml, G-CSF 10ng/ml, FLT3L 100ng/ml, and SCF 50ng/ml for 96 hrs in a gas-permeable container coated with RetroNectin. The cells were exposed to fresh viral supernatant for 8 hours each day. Two animals were conditioned with cyclophosphamide 50mg/kg for 2 days and fludarabine 25mg/square miter for 3 days prior to infusion of transduced cells. One animal was conditioned with 600Gy of TBI. Lineage-specific gene marking was carefully analyzed by real time PCR using primers that targeted the unique sequence in the retroviral backbone from peripheral blood samples. Myelosuppression was mild in two animals conditioned with chemotherapy to compare with the irradiated animal. The first of two animals conditioned with chemotherapy showed more than 1% gene marking in granulocytes 2 weeks after infusion, which decreased to 0.02% at 2 months. This level of gene marking has remained stable for 10 months. The second animal displayed 0.02% granulocytes marking after 2 weeks which dropped to 0.001% after 8 weeks. The lower gene marking in the second animal was expected due to the poor ex-vivo transduction. Stable marking is approximately 15 fold higher in the animal conditioned with TBI. These data suggest that persistent low level gene marking is possible in non-human primates following mild chemotherapy-based bone marrow conditioning. Ongoing studies are underway to determine whether dose escalation of chemotherapy will allow for engrafment comparable to that which is achieved by TBI.

Anand P. Chokkalingam, NCI-DCEG
Vitamin D receptor gene polymorphisms, insulin-like growth factors, and prostate cancer: a population-based case-control study in China
Operating through the vitamin D receptor (VDR), vitamin D shows strong inhibitory effects on prostate cancer growth in both laboratory and human studies. Similarly, through modulation of insulin-like growth factor (IGF) bioactivity, IGF binding proteins (IGFBPs) inhibit growth of prostate cells; we and others have shown that circulating IGFBP levels are inversely associated with prostate cancer risk. Importantly, recent laboratory evidence indicates that vitamin D-induced prostate cancer inhibition is accompanied by increased IGFBP expression, suggesting that the vitamin D and IGF regulatory systems may operate together to affect prostate cancer. Based on this evidence, we examined the a priori hypothesis that VDR gene polymorphisms influence prostate cancer risk - either independently or in conjunction with plasma IGF or IGFBP levels - in a population-based case-control study in Shanghai, China. Histologically-confirmed cases of primary prostate cancer newly diagnosed between 1993 and 1995 (N=191) and randomly selected age-matched population controls (N=304) submitted to in-person interviews and blood draws. Genomic DNA was used to determine VDR FokI and BsmI polymorphism genotypes; IGF and IGFBP concentrations were determined from plasma. No significant independent association of either the BsmI or the FokI VDR markers with prostate cancer was observed, though a small effect of BsmI cannot be ruled out owing to the low observed prevalence of its B allele. However, among men with the ff FokI genotype, those with the highest tertile of IGFBP-3 had a significantly decreased risk versus those with the lowest tertile (OR=0.14; 95% CI=0.04-0.56; p for trend<0.01), while among men with the FF and Ff FokI genotypes IGFBP-3 was not associated with risk. Similarly, IGFBP-1 was significantly inversely associated with prostate cancer risk among men with the ff FokI genotype (OR=0.25; 95% CI=0.07-0.85; p for trend=0.02), but not among men with the FF and Ff genotypes. No such FokI specific effects were noted for IGF-I or IGF-II. As variants of the FokI VDR marker have been shown to be functionally different, these novel findings suggest a functional interaction between two primary research areas within prostate cancer epidemiology: the vitamin D regulatory system and the IGF axis. Further studies are needed to confirm these findings and to elucidate biological mechanisms for the observed interactions.

Anthony D. Cristillo, NHLBI
Identification of novel targets of immunosuppressive agents by cDNA-based microarray analysis
T cell activation follows TcR-CD3 engagement. The immunosuppressive agents, cyclosporine (CsA) and tacrolimus (FK506) have been instrumental in TcR-CD3 dependent intracellular signaling pathways. CsA and FK506 each bind to families of intracellular immunophilin receptors, termed the cyclophilins (CyP) and FK506-binding proteins (FKBP), respectively. The complexes of CsA-CyP and of FK506-FKBP both bind to and inhibit the activity of the serine/threonine phosphatase calcineurin, thereby inhibiting T-cell activation. We used cDNA-based microarray analysis to identify early, CsA- and FK506-sensitive genes in purified human peripheral blood T lymphocytes, hybridizing simultaneously two different cDNA samples each labeled with a different fluorophore. We reasoned that changes in gene expression common to both CsA- and FK506-treatment would likely be secondary to calcineurin inactivation. No transcripts specifically regulated by only one immunosuppressive agent were identified, suggesting that immunophilin-dependent, calcineurin-independent gene expression was below the limits of detection of our analysis. We did identify, however, a number of CsA-and FK506-sensitive (and calcineurin-dependent) genes induced or inhibited following anti-CD3 mAb ligation + PMA relative to resting cells. Hierarchy profiles and cluster analysis of target genes suggested relative specificity. While some transcripts are well characterized (e. g. IL-2 and lymphotactin), the array data suggest a number of novel calcineurin-dependent substrates. Analysis of mRNA transcription and protein in both CD4+ and CD8+ T-cell subpopulations, used to verify and extend these cDNA microarray results, demonstrated both induction of Stat5a and inhibition of IL-16, genetic elements not previously appreciated to be affected by either CsA or FK506.Moreover, expression of several expressed sequence tags (ESTs) were modulated by T-cell activation and regulated by drug. We cloned full length cDNAs from human PBL mRNA. Bioinformatic analysis of one isolated target suggests a transmembrane protein with strong homology to a Nedd4 WW domain-binding protein. Transient overexpression of this protein appears to enhance NFAT transcriptional activation by a mechanism currently under investigation. Further analysis may identify novel, functional calcineurin (and immunophilin)-dependent T-cell signaling pathways that activate transcription factors that in-turn coordinately control gene expression.

Xinle Cui, NHLBI
Identification of ARTS-1: A Novel Aminopeptidase that Selectively Binds to TNFR1, not TNFR2, and Up-regulates TNFR1 Ectodomain Shedding
Tumor necrosis factor (TNF) is a multifunctional cytokine that mediates inflammation, immunoregulation, cytotoxicity, and antiviral effects. Both the type 1 (TNFR1) and type 2 (TNFR2) TNF receptors can be proteolytically cleaved and shed as soluble TNF binding proteins that modulate TNF activity. Utilizing a yeast two-hybrid approach, we identified a novel aminopeptidase that directly binds to and upregulates TNFR1 shedding, which we have named ARTS-1 (Aminopeptidase Regulator of TNFR1 Shedding). The full-length human ARTS-1 cDNA contains 4845 nucleotides, and encodes a 941-amino acid protein. ARTS-1 is a type II integral membrane protein with a single hydrophobic transmembrane a-helical domain, a very short hydrophobic intracellular amino-terminal domain and a large 913-amino acid extracellular domain that contains a consensus zinc metalloprotease catalytic motif (HEXXH(Y)18E). Northern blot analysis demonstrated that human ARTS-1 is widely expressed as a 4.8 kb transcript in multiple tissues. Immunoblot analysis demonstrated ARTS-1 as a 100 kDa membrane-associated protein in multiple human epithelial cells, endothelial cells and fibroblasts. ARTS-1 selectively binds to TNFR1, but not TNFR2 in vivo as demonstrated by coimmunoprecipitation performed with both epithelial and endothelial cell membranes. Cell lines overexpressing full length ARTS-1 mRNA demonstrated a 3-fold increase of sTNFR1 in culture supernatants and decreased membrane-associated TNFR1, while cell lines expressing anti-sense ARTS-1 mRNA had a 70% decrease of sTNFR1 in culture supernatants and increased membrane-associated TNFR1.TNFR2 shedding was not altered in either the ARTS-1 overexpression or antisense cell lines. To characterize the aminopeptidase activity of ARTS-1, the extracellular domain of ARTS-1 was expressed as a GST fusion protein in E. coli and purified utilizing a glutathione Sepharose 4B affinity column. Kinetic constants were determined by Lineweaver-Burk analysis, GST-ARTS-1 aminopeptidase activity was selective for non-polar amino acid substrates over a four-fold range of enzyme activity based upon kcat / Km values. No activity against either acidic or basic amino acid p-nitroanilides was identified. In conclusion, we have cloned and characterized a novel aminopeptidase that binds to and up-regulates TNFR1 shedding, but not TNFR2, and also possesses aminopeptidase activity towards non-polar amino acid residues.

Edna Cukierman, NIDCR
Taking Cell-Matrix Adhesions to the Third Dimension
Cells secrete and organize extracellular matrix (ECM), which provides structural support for their adhesion, migration, and tissue organization, besides regulating cellular functions such as growth and survival. Matrix interactions are vital for vertebrate development; disorders in these processes have been associated with fibrosis, cancer, and other diseases. Cell-matrix adhesions consist of transmembrane receptors (integrins) that anchor cytoskeletal and signal transduction molecules to their extracellular environment. Integrins operate both as recipients and generators of signals that trigger multiple processes. The choice of integrins differs depending on both the cell type and the substrate. Focal and fibrillar adhesions have been studied intensively in fibroblasts.
Although fibroblast-to-extracellular matrix adhesions have been studied extensively in vitro, little is known about their in vivo counterparts. We have overcome this limitation by developing a new approach in which we mimic natural substrates by utilizing cell-derived 3D matrices. We characterized the composition and functions of adhesions in three-dimensional tissues, in natural 3D matrices from tissue samples, or in 3D matrices generated from cell cultures. We described the process of maturation of 3D matrix adhesions compared to plating cells on artificial 2D and 3D substrates. By applying new technologies for digital imaging and morphological measurements, we identified a new adhesive structure - the three-dimensional-matrix adhesion - which is charachteristic of cell-matrix adhesions in vivo.
3D-matrix adhesions in vitro and in vivo differed in molecular composition from focal and fibrillar adhesions in their content of alpha-5 beta-1 integrin, paxillin, and other cytoskeletal components, and in tyrosine phosphorylation of the focal adhesion kinase. Cells with 3D-matrix interactions showed substantially enhanced biological cell activities (e. g. adhesion, elongated morphology, migration, and proliferation) and markedly constricted integrin utilization to alpha-5 beta-1 integrin. Our studies indicate the importance of these novel 3D-matrix adhesions, which differ in structure, localization, and function from classical focal and fibrillar adhesions. We suggest that future studies in this research area should focus on these 3D-matrix adhesions, which are more representative of adhesions in vivo.

Raya Dana, NIAID
Roles for ADP-Ribosylation Factor 6 (ARF6) and its Effector, Phospholipase D1, in the Respiratory Burst in Phagocytic Cells.
Phagocytic cells respond to a variety of stimuli to produce superoxide and related antimicrobial reactive oxygen species in a process referred to as the respiratory burst. Recently, we presented evidence linking formyl-peptide (fMLF)-induced activation of NADPH oxidase to activation of the small GTPase, ARF6, and its downstream effector, phospholipase D (PLD). We have now compared the fMLF-elicited responses in neutrophil-like PLB-985 cells to those of two other agonists, PMA and opsonized yeast particles (OZ), and have explored the subcellular distribution of ARF6 and related signaling intermediates in response to all three agonists. As was observed with fMLF, the OZ-stimulated respiratory burst appears to involve both ARF6 and PLD; this pathway was affected by both dominant-negative and active mutant forms of ARF6 and by the PLD inhibitor 2,3-diphosphoglycerate. Furthermore, involvement of the PLD1 isozyme was indicated with all three agonists, since transfection of the inactive PLD1 mutant, PLD1-K898R inhibited the burst in response to all three stimuli. In contrast, PMA stimulates the burst in a PLD-dependent, but ARF6-independent manner, as the ARF6 mutants had no effect. Both physiological agonists, fMLF and OZ, cause a rapid redistribution of ARF6 from a diffuse, punctate staining pattern to a more condensed staining of larger, vesicular structures within the cell peripheri. In OZ-stimulated cells, ARF6 accumulated near newly formed phagosomes after 15-30 minutes, while in fMLP-stimulated cells the ARF6 assumed the more dispersed appearance of unstimlated cells within this later time frame. Finally, confocal fluorescence staining of ARF6 in fMLP-stimulated cells resembled, both spatially and temporally, the staining of two essential oxidase components, Rac2 and p47phox, suggesting a direct role of ARF6 in the assembly or activation of NADPH oxidase. These findings provide evidence that both physiological agonists signal activation of the respiratory burst through ARF6 and PLD1-dependent pathways, which might serve as targets for affecting the oxidative output of phagocytic cells and related inflammatory processes.

Sunit Das, NINDS
Syntaphilin: An Inhibitory Regulator Coupling Synaptic Vesicle Exocytosis and Endocytosis
Synaptic vesicle (SV) exocytosis is mediated by a series of interactions between the membranes of SVs and the presynaptic terminal, culminating in the Ca++-dependent fusion of the two membranes and release of the vesicle contents into the synaptic cleft. This process depends upon the assembly of a stable SNARE complex, a trans-helical protein structure consisting of the SV protein VAMP/synaptobrevin and the plasmalemmal proteins syntaxin and SNAP-25 that likely facilitates mixing of the two lipid bilayers and provides the energy that drives membrane fusion. Following exocytosis, synaptic vesicles and SNARE proteins are recycled; as with SV exocytosis, the process of endocytosis depends upon the interaction of a set of proteins that constitutes the basic molecular machinery of endocytosis, most prominently, the proteins clathrin, dynamin and amphiphysin. Identifying proteins that modulate SV exo- and endocytosis is critical to elucidating the molecular mechanisms that control neurotransmitter release at central synapses. Here we characterize a functionally significant interaction between syntaphilin, a protein known to control SNARE assembly via its binding with syntaxin, and dynamin, a protein required for SV endocytosis. Immunohistochemical staining of cultured hippocampal cells demonstrates co-localization of dynamin and syntaphilin in a punctate pattern along dendritic processes. Dynamin immunoprecipitates with syntaphilin in brain homogenate and in lysate from co-transfected HEK 293 cells; a direct interaction between these two proteins was confirmed by a dynamin pull-down experiment with recombinant His-tagged syntaphilin. Biochemical assays demonstrate that syntaphilin competes with amphiphysin for binding to dynamin and thereby inhibits the formation of the dynamin-amphiphysin complex required for SV endocytosis. Functional studies examining the effect of syntaphilin overexpression on endocytosis in COS-7 cells are now underway. Our findings suggest that syntaphilin may function as a negative regulator of clathrin and dynamin-dependent endocytosis. As shown previously, overexpression of syntaphilin in cultured hippocampal neurons significantly reduced neurotransmitter release. Given its inhibitory effect on both synaptic vesicle exocytosis and endocytosis, we suggest that syntaphilin may function as a molecular clamp of syntaxin and dynamin, thereby acting as a general inhibitory modulator of synaptic vesicle release and recycling.

Angelo Del Parigi, NIDDK
Changes in neuronal activity in the caudolateral orbitofrontal cortex after a satiating meal in lean, obese, and post-obese people
Sensory-specific satiety is the decrease in the pleasantness of sensory inputs that occurs after prolonged exposure to the same sensory stimulation. This mechanism is thought to contribute to the termination of a meal and its attenuation/absence might promote hyperphagia. Human subjects fed to satiety report that the perceived taste intensity and sensory quality of the food is unmodified, but the pleasure is reduced. In macaques, neurons of the caudolateral orbitofrontal cortex (CLOF) stop responding to the taste of a food when the animals are fed to satiety. It is not known whether there is a functionally homologous region in the human brain. We hypothesized that a decrease in neuronal activity in the CLOF in response to a satiating meal would be attenuated/absent in obese compared to lean individuals. Post-obese subjects (who achieved near-normal body weight through diet and exercise) were also studied to understand the possible importance of the CLOF activity to the natural history of obesity. Using positron emission tomography and measures of regional cerebral blood flow (rCBF, a marker of neuronal activity), we investigated the activity of the CLOF in response to hunger (36-h fast), taste stimulation (2 mL of a liquid formula meal) and administration of a satiating meal (the same formula used as the tastant) in 21 lean (never-obese) (M/F=11/10, 33+/-9y, 22+/-7% body fat, mean+/-SD), 23 obese (M/F=11/12, 29+/-6y, 39+/-5%), and 11 post-obese (M/F=3/8, 40+/-6y, 26+/-4%) subjects. In lean individuals, no significant changes in rCBF in the CLOF were observed under any of the experimental conditions. In obese subjects, a significant (p

Dimiter G. Demirov, NIAID
THE E2-LIKE DOMAIN OF TSG101 SUPRESSES HIV-1 BUDDING IN A p6-DEPENDENT FASHION
A 6 kDa, Pro-rich protein, p6, is encoded by the C-terminus of the HIV-1 gag gene. Previously, we demonstrated that a highly conserved Pro-Thr-Ala-Pro (PTAP) sequence near the N-terminus of p6 is essential for efficient virus budding. Mutations in this motif result in a dramatic accumulation of tethered virus particles at the plasma membrane. Analogous motifs, collectively termed "late" or "L" domains to reflect their role in virus budding, have been identified in a number of other retroviral Gag proteins.
Several lines of evidence suggest that retroviral L domains may function by interacting with the cellular ubiquitination machinery: the late domain of Rous sarcoma virus reportedly interacts with a ubiquitin ligase (E3) family member, and proteasome inhibitors, which disrupt the ubiquitination pathway, impair retrovirus particle release. The product of the cellular tumor suppressor gene TSG101 contains a domain related to E2 ubiquitin-conjugating enzymes, which binds HIV-1 Gag in a p6-dependent fashion. As TSG101 is reportedly a multifunctional protein, we used a truncated construct, containing the E2-like domain and the adjacent, putative substrate-recognition Pro-rich domain. We refer to this truncated construct as the TSG101-5'. We examined the impact of overexpressing this construct on HIV-1 assembly. Intriguingly, we observed that TSG101-5' potently inhibits virus production in a dose-dependent manner. Examination of cells coexpressing HIV-1 Gag and TSG101-5' by electron microscopy reveals a defect in virus budding very reminiscent of that observed with p6 mutants. In addition, the effect of TSG101-5' is dependent upon an intact p6 L domain; the assembly and release of virus-like particles produced by a p6 Gag truncation mutant, or by mutants lacking a functional p6 PTAP motif, is not affected by TSG101-5' overexpression. Similarly, assembly and release of murine leukemia virus is insensitive to TSG101-5', indicating that it selectively inhibits the release of HIV-1 particles bearing a functional p6 L domain. These data suggest a link between the E2-like domain of TSG101 and HIV-1 L domain function. Ongoing experiments will further elucidate the mechanism responsible for the virus release block, and will examine the role of TSG101 interacting proteins in HIV-1 release. These data indicate that the TSG101-5' construct is a useful tool to probe the molecular mechanisms of HIV-1 particles release.

Alexandrine J. Derrien, NIAID
IMPORTANCE OF CONSERVED TYROSINE RESIDUES AND EGFR-INDUCED TYROSINE PHOSPHORYLATION OF RGS16 IN THE REGULATION OF Gi-MEDIATED SIGNAL TRANSDUCTION
Regulators of G protein signaling (RGS) proteins have emerged as important determinants of the precision kinetics associated with signaling by heterotrimeric G proteins. RGS proteins have been shown in some cases to accelerate turn-on rates (inward rectifying potassium channels) or inhibit downstream effector output (e. g. MAP kinase) induced by activation of Gi-coupled receptors. With few exceptions, virtually nothing is known of the physiological roles of RGS proteins nor of the regulation of their own catalytic activity. In this study, we show that RGS16 GTPase accelerating (GAP) activity and its ability to modulate Gi-coupled signaling is affected by two conserve tyrosine residues located near the C-terminus. Both purified src and EGFR tyrosine kinases phosphorylated RGS16 but not RGS2 in vitro. In HEK293T cells, RGS16 underwent tyrosine phosphorylation after stimulation by both EGF and carbachol via the m2R. EGF stimulation leads primarily to the phosphorylation of residues Y168 in 293T cells whereas src kinase preferentially phosphorylated Y177 in vitro. The src kinase inhibitor radicicol or dominant negative src diminished RGS16 phosphorylation provoked by EGF or carbachol but only to the extent that EGFR phosphorylation was inhibited. In contrast, a specific inhibitor of EGFR kinase (coumpound 32) completely abolished m2R- or EGFR-induced RGS16 phosphorylation, suggesting that, in these cells, EGFR is the kinase primarily responsible for RGS16 phosphorylation. Recombinant purified RGS16 mutant proteins Y168F and Y168/177F, but not the RGS16Y177F, demonstrated diminished GAP activity on Gai1.However, all three mutant proteins were unable to inhibit m2R-induced MAP kinase activation in HEK 293T cells. Purified EGFR kinase enhanced the catalytic activity of RGS16 in vitro. These results suggest that EGFR and/or src tyrosine phosphorylation of both residues may regulate the activity of RGS16 in mammalian cells. This study also provides the first evidence that receptor tyrosine kinase pathways may negatively regulate activity of G protein-coupled signaling pathways.

Constantine Dimitrakakis, NICHD
Testosterone inhibits estradiol's mitogenic effects upon primate mammary epithelium in vivo
In normal cycling women, the ovary produces balanced amounts of estrogen, androgen and progesterone. Years ago, it was found that 'replacing' estradiol alone in menopausal or oophorectomized women led to uterine hyperplasia and cancer, and that this effect could be prevented by giving progesterone, which opposes estrogen's proliferative effect upon uterine cells. Unfortunately, progesterone does not oppose and may even augment estrogen's stimulatory effect upon mammary epithelium. Estrogen is thought to increase the risk of breast cancer primarily by increasing mammary epithelial proliferation (MEP). Observing the anti-mammogenic effects of androgens in many clinical situations, we hypothesized that androgens may oppose estrogen's mitogenic effects upon the mammary epithelium. To test this hypothesis, we treated ovariectomized rhesus monkeys with sc pellets containing vehicle (7), estradiol (E2, n=4), E2 & progesterone (E2/P, n=5) or E2 plus testosterone (E2/T, n=5). The MEP index, assayed by Ki67 immunoreactivity, was increased 7-fold in both E2 and E2/P groups (P<0.0001). The E2/T group, however, showed only a 3.5-fold increase in MEPI, despite estradiol levels that were significantly higher than the other active treatment groups. Thus, addition of testosterone to the hormone replacement treatment (HRT) significantly attenuated E2's mitogenic effects. We repeated the study using physiological rather than pharmacological doses of ovarian steroids. In this experiment, E2 alone and E2/P increased MEPI compared to placebo controls by ~4-fold, while the E2/T combination did not significantly increase the MEPI above control values (E2 levels were similar in the 3 active treatment groups). To determine if endogenous testosterone normally inhibits endogenous estradiol's proliferative effects, intact, cycling monkeys were treated with flutamide, an androgen receptor antagonist or placebo. The MEPI was increased by ~2-fold in the flutamide group (P = 0.03). These data suggest that testosterone may serve as a natural, endogenous protector of the breast and limit estrogen's mitogenic and cancer promoting effects on mammary epithelium. Unfortunately, in giving menopausal women oral estrogen for HRT, we may inadvertanly aggravate estrogen-induced risk, since oral estrogens decrease testosterone production and availability. Thus, we propose clinical trials evaluating testosterone as a component of menopausal hormone replacement

Federica Dipalma, NIDCD
Identification of the Varitint-Waddler gene: a link between deafness and pigmentation defects.
Positional cloning of mouse deafness mutations has proven an important means for identifying new genes involved in human deafness, and for elucidating the pathological and molecular processes underlying auditory system defects. Varitint-Waddler (VaJ) is a semidominant mouse mutation causing deafness and coat color defects due to the lack of pigment producing melanocytes. In VaJ mutant mice, the lack of strial pigmentation in the inner ear correlates with an abnormal endochlear potential and subsequent hearing loss. Inner ear abnormalities are also found in the organ of Corti and include degeneration of hair cells (HC). We find that early signs of disorganization of the HC stereocilia are present at P2 and affect both inner and outer hair cells. Published data and the early onset of such disorganization suggest the hair cells defects occur independently of the reduced strial pigmentation.
To identify the responsible gene, intercrosses were used to generate a precise genetic map of markers flanking the VaJ interval (0.15 cM) on chr 3.We then assembled a 350Kb contig of overlapping BAC clones across the critical region. We sequenced two BAC clones constituting a minimum tiling path across the VaJ interval, and identified two genes using GENESCAN and BLAST search algorithms. The two genes are located 50Kb apart from each other and encode two unknown transmembrane proteins, UNP and FD. Both proteins have 73% sequence similarity to mucolipins, a new family of ion channel proteins and 43% similarity to polycystins. We show by RT PCR that UNP is expressed in the kidneys, spleen, brain and in the organ of Corti. Sequence analysis of PCR products from genomic DNA and cDNA of normal and VaJ mutant mice revealed two point mutations in the coding sequence of UNP resulting into two amino acid changes. The mutations segregated with hearing loss and coat color phenotype in our colonies and were absent in parental and control strains. One of such mutations was also found in the original Va allele from which VaJ arose manifesting a less severe phenotype. Expression of UNP and FD in Xenopus oocytes is ongoing to evaluate their function as ion channels, and to explore the possibility of an interaction between the two genes. Further characterization of the VaJ gene by in situ hybridization and immunohistochemistry will contribute to elucidate the mechanism of hearing loss and the pathway linking pigmentation defects and inner ear abnormalities in this unique mouse model.

Anna Fogdell-Hahn, NCI-CCR
Molecular genetics of neuronal degeneration: studies of reaper, grim, and hid in mutant rhodopsin induced apoptosis
Certain dominant mutations in rhodopsin induce apoptosis of the photoreceptor (PR) neurons of humans and Drosophila, leading to blindness in both organisms. In humans, the disease caused by these mutations, retinitis pigmentosa (RP), occurs in about 1 in 4000 adults. Our goal is to understand the mechanism by which these mutations induce apoptosis. In Drosophila, a chromosomal interval defined by the H99 deficiency is required for all embryonic cell death. Three genes in the interval, hid, grim, and reaper, are important regulators of developmental apoptosis. However, their role in terminally differentiated cells is unknown. We asked whether hid, grim, and/or reaper are involved in PR cell death. Detailed time courses of apoptosis were determined in three different lines of flies carrying dominant mutations in the Drosophila rhodopsin gene, ninaE. All three alleles, which are equivalent to those that cause RP in humans, induced apoptosis with similar kinetics. Interestingly, degeneration in the ninaE/+ heterozygotes required exposure to visible-wavelength light. Therefore, all investigations were carried out under carefully controlled environmental conditions. At 500 lux constant light exposure, the loss of one copy of the H99 interval strongly suppressed PR cell death. Therefore, one or more genes in the interval is important for the activation of neuronal apoptosis in response to mutant rhodopsin. We are now determining which of these genes is responsible. A second question is whether genes in the interval are not only involved but required. Therefore, H99/H99 mitotic clones were generated in ninaE mutant-background eyes using FLP-ase mediated recombination. Importantly, PR neurons homozygous for H99 were found to have normal morphology suggesting that genes in the H99 interval are not required for normal retinal development. Studies are underway to determine whether cell death is completely blocked in the H99 homozygous clones. It is expected that homologues of the responsible genes will be found to act in mutant rhodopsin induced apoptosis in humans as well as flies.

Oksana I. Dukhanina, NCI-CCR
Prolonged expression of a soluble TGF-beta antagonist in vivo protects against metastasis without any adverse side-effects.
TGF-betas play key roles in many biological processes. In tumorigenesis they are proposed to play a dual role, functioning as tumor suppressors early in the process, but as pro-oncogenic factors in late-stage metastatic disease. To probe this complexity in vivo we have developed a transgenic mouse overexpressing a soluble TGF-beta antagonist (SR2F), consisting of the extracellular domain of the type II TGF-beta receptor fused to the Fc domain of human IgG1.Using the MMTV-LTR promoter, high level expression of SR2F was obtained in the mammary glands, seminal vesicle and serum. To determine the effect of the antagonist on metastatic efficiency, an isogenic metastatic melanoma cell line was injected into the tail vein of transgenic and control mice. Transgenic mice showed a decrease in the number of metastases in liver, lung, spleen and pancreas when compared with wildtype controls. In the liver, where most metastases formed, a 3-fold reduction in the number of metastases was observed (p< 0.03). This proves that TGF-beta plays a pro-metastatic role in late-stage cancer. In order to determine if the SR2F was also effective at preventing metastasis from a primary tumor, the MMTV-neu mouse model of breast cancer was crossed with MMTV-SR2F mice or with FVB/N control mice. Tumor latency, tumor multiplicity and total tumor burden were unaffected by the presence of SR2F. However, the neu/SR2F bigenic mice showed a 3.3-fold decrease (p = 0.04) in the number of mice with lung metastases when compared with mice expressing neu alone. This suggests the surprising conclusion that, in vivo, the SR2F antagonist can prevent metastasis without accelerating primary tumor formation. To determine whether the SR2F interferes with any other normal functions of TGF-beta, we have analyzed cohorts of aged SR2F and FVB/N mice for evidence of any other pathologies. None were observed. In particular there was no evidence for the multi-focal inflammation, auto-immune disease and enhanced tumorigenesis that are characteristic of the TGF-beta1 null mouse. The conceptual problem with the clinical use of TGF-beta antagonists has been the potential interference with the many important roles of TGF-beta in normal tissues. Our data show the unexpected finding that prolonged exposure to the SR2F soluble antagonist protects against experimental metastasis without any undesirable side-effects. This makes the SR2F antagonist a very promising candidate for clinical use against metastasis.

Michelle L. Dykstra, NIAID
Epstein-Barr Virus co-opts lipid rafts disrupting BCR signaling and trafficking of antigen
The B cell antigen receptor (BCR) plays a dual role in B cell activation. First, upon crosslinking by antigen, the BCR initiates a signal transduction cascade leading to the transcription of various genes associated with B cell activation. In addition, the BCR is internalized to deliver its bound antigen to intracellular compartments for processing and presentation by MHC class II molecules. Recent evidence indicates that BCR signaling influences BCR trafficking resulting in an accelerated targeting of antigen. In collaboration with a colleague, I recently showed that upon crosslinking by antigen the BCR translocates into cholesterol- and sphingolipid-rich membrane microdomains, or lipid rafts, that concentrate the Src-family kinase Lyn and exclude the phosphatase CD45R. The BCR is phosphorylated in the rafts and subsequently trafficks to antigen processing compartments. If translocation of the BCR into lipid rafts represents an integral step for B cell activation, then it might be anticipated that lipid rafts would be targets of pathogens to disarm B cell function. The Epstein Barr Virus (EBV) establishes a persistent latent infection in human B cells, and it has been shown previously that the EBV latent membrane protein 2A (LMP2A) associates with Lyn and Syk and both blocks BCR signaling upon receptor crosslinking and generates its own signal for B cell survival. The location of LMP2A in B cell membranes and its influence on BCR raft association were determined. Lipid rafts from human EBV latently infected B cells were isolated by sucrose gradient fractionation following membrane solubilization in 1% Triton X-100 at 4 deg C. LMP2A was found to be constitutively present in lipid rafts and its isolation in rafts was dependent on membrane cholesterol. LMP2A residing in the rafts excluded the BCR from entering rafts, thus blocking both BCR signaling in the rafts and accelerated internalization of the BCR. Mutant LMP2A that does not bind Lyn permits BCR signaling and raft translocation but still blocks antigen trafficking, indicating independent control of these BCR functions. Thus, EBV co-opts the lipid rafts to signal itself and to disarm both the signaling and antigen processing functions of the BCR by independent mechanisms. Studies in progress to delineate the mechanisms by which LMP2A blocks BCR function should provide significant detail concerning the mechanisms underlying the dual roles of BCR signaling and antigen targeting.

Nadukkudy V. Eldho, NIAAA
DOCOSAHEXAENOIC VS. DOCOSAPENTAENOIC ACID-THE DIFFERENCE THAT THE LOSS OF A SINGLE DOUBLE BOND MAKES
Synaptosomal membranes of neurons may have docosahexaenoic acid (22:6w-3, DHA) content up to 50 mol% of all lipid fatty acids. DHA is a polyunsaturated fatty acid with 22 carbons and 6 double bonds. Insufficient supply of DHA or its w-3 fatty acid precursors through the diet, in particular during brain growth in infancy, results in its replacement by docosapentaenoic acid (22:5w-6, DPA) with just one double bond less. This replacement is related to impaired neural function. Considering the high concentration of DHA, an influence on integral membrane protein function appears likely. My hypothesis is that the DHA-DPA replacement can alter lateral pressure density profiles of neural membranes thereby shifting activation of receptor proteins. To confirm this hypothesis, I investigated properties of two polyunsaturated lipids, 18:0(d35)-22:6w-3 phosphatidylcholine (PC) and 18:0(d35)-22:5w-6 PC through solid-state NMR in the biologically relevant lamellar state. 2D NOESY in combination with magic angle spinning, relaxation rate analysis, and classical 2H NMR order parameter studies provided more than 100 parameters, sampling conformations with spatial resolution on the angstrom-scale and temporal resolution from pico- to milliseconds. I was taken by surprise by the apparent low potential barriers for structural transitions of DHA and DPA since they contain six or five rigid cis-double bonds. Both DHA and DPA chains are very motionally disordered. This must be the consequence of lower potential barrier of rotation between methylene and ethylenic carbons. But, the DHA chain appeared to be more flexible at the methyl end with significantly shorter motional correlation times. Furthermore, the 18:0(d35) chain paired with DPA has significantly higher order parameters compared to that of DHA, which corresponds to a decrease in average bilayer thickness and an increase in area per lipid molecule for the DHA-containing lipid. My NMR parameters have been used to set up MD simulations that revealed rapid DHA transitions between extended and looped conformations. The density of DHA near the lipid-water interface is higher, providing extra freedom of movement for the paired sn-1 chain. This redistribution of sn-2 chain density towards the lipid-water interface is partially reversed when going from DHA to DPA. We will calculate lateral pressure density profiles of these membranes, to predict differences in free energy between normal- and activated states of receptor proteins.

Michael Elkin, NIDCR
Estrogen regulates expression of heparanase, an enzyme important in angiogenesis and tumor progression.
Breast cancer is the most common malignancy and the second leading cause of cancer-related deaths among women in the US. Although numerous epidemiological studies clearly suggest that estrogen is one of the main driving forces in breast tumorigenesis, the precise mechanisms by which estrogen acts in cancer promotion remain poorly understood. Even less is known about molecular pathways underlying the more recently described pro-angiogenic activity of estrogen.
Heparanase is an enzyme that degrades heparan sulfate, the main polysaccharide component of extracellular matrix (ECM). The ECM, in particular subendothelial and subepithelial basement membranes, represent major barriers to cell invasion. Heparan sulfate is a key factor in the self-assembly and integrity of the ECM multimolecular architecture, and, therefore, cleavage of heparan sulfate in the ECM by heparanase is a prerequisite for cell invasion. In addition, heparanase liberates heparan sulfate-binding growth factors from ECM depots and makes them available for growth factor dependent processes. We have previously demonstrated the decisive role of heparanase in tumor progression, metastasis, and angiogenesis. Here, we now have identified 3 estrogen response element sequences in the heparanase promoter and studied the regulation of heparanase gene expression by estrogen. Using a reporter construct we found that the transcriptional activity of heparanase promoter was significantly increased in the presence of physiological concentrations of estrogen. Estrogen treatment induced a ~300% increase in heparanase mRNA expression in estrogen receptor (ER) positive MCF-7 human breast carcinoma cells, confirming the promoter study data. This effect was abolished in the presence of the antiestrogen tamoxifen. In the ER negative MDA-231 breast carcinoma cell line, similar estrogen treatment did not result in any increase in heparanase expression, indicating that the classic ER pathway is involved in transcriptional activation of heparanase, rather than numerous alternative mechanisms of estrogen-regulated cell responses. Taken together, these findings indicate that heparanase might be an important downstream estrogen effector, mediating the angio- and tumorigenic activities of estrogen. Understanding the role of steroid hormones in the control of tumorigenesis and neovascularization could be highly relevant for the development of combination anti-cancer therapies in breast and in other types of malignancies.

Chad A. Ellis, NCI-CCR
Identification and Characterization of Rig, A Novel Farnesylated Ras-Like GTPase Exhibiting Tumor Suppressor Properties.
Ras superfamily proteins are structurally similar yet functionally diverse proteins, having roles in regulation of cell growth, differentiation and survival. Ras oncoproteins serve as molecular switches, shuttling between active GTP-bound and inactive GDP-bound states. To function as oncogenes, Ras proteins must be processed at their carboxyl-terminus by a farnesyl lipid moeity. This relatively rare, post-translational modification has been successfully targeted for clinical drug development. As the precise cellular target of these compounds, Farnesyl Transferase Inhibitors, remains unidentified an attempt was undertaken to discover novel farnesylated monomeric GTPases and characterize their role in tumorigenesis. As a result of our bioinformatic search, a novel Ras-related protein was identified and designated Rig (Ras-related inhibitor of cell growth). Amino acid sequence analysis revealed that Rig shares > 60% overall similarity with Ras, Rap and Noey2, near identity in the guanine-nucleotide and effector binding domains and contains a C-terminal CAAX motif. Based on primary amino acid sequence comparison of all Ras-like GTPases, Rig and the Noey2 tumor suppressor belong to a distinct Ras protein sub-family. Furthermore, biochemical examination of recombinant proteins demonstrated that Rig could be post-translationally farnesylated and geranylgeranylated in vitro. Rig expression is limited to cardiac and neural tissue. Rig protein expression is down regulated in primary human astrocytomas and glioblastoma multiforme (GBM). Reintroduction of Rig, via induced expression, inhibited neural-tumor cell growth. In NIH3T3 fibroblasts, Rig inhibited Ras-mediated cellular transformation by 50% and Ras-mediated MAPK activation by >70%. Also, Rig interacts with Raf-1 and RPIP8 in vivo, effector molecules for Ras and Rap2A, respectively. Therefore, Rig is a novel Ras superfamily member that inhibits normal and tumor cell growth as well as oncogenic transformation by interacting with known and possibly novel effector molecules. Furthermore, Rig protein expression is suppressed in early and late stage (GBM) brain tumors prompting us to hypothesize that Rig functions as a tumor suppressor. We are further characterizing the tumor suppressor potential of Rig as well as the role of post-translational isoprenylation in Rig-mediated growth inhibition.

Christopher E. Ellis, NHGRI
Src family protein tyrosine kinase's phosphorylate alpha-synuclein on tyrosine 125 and affect its ability to interact with various brain proteins
Alpha-Synuclein (AS) is a 140 amino acid protein implicated genetically in the pathogenesis of early onset Parkinson's Disease (PD) through missense mutations identified in families with autosomal dominant PD. Additionally, AS immunoreactivity is detected in Lewy Bodies, the pathological hallmark of PD. The recent observation that mice and flies expressing a human AS transgene recapitulate characteristics of PD suggest that AS may be involved directly in the development of this disease. Thus, it is of interest to determine the specific function(s) of AS and its role in PD. A comparison of the protein sequence between Synuclein family members revealed that all four tyrosine residues of AS are conserved in all orthologs and beta-Synuclein paralogs, suggesting that these residues may be important functionally. In addition, sequences flanking tyrosine 125 form a phosphorylation consensus site. For these reasons, experiments were performed to determine whether AS can be phosphorylated on tyrosine residue(s) in human cells. Indeed, AS is phosphorylated within 2 minutes of pervanadate treatment in AS transfected cells. Tyrosine phosphorylation occurs primarily on tyrosine 125 and was inhibited by PP2, a selective inhibitor of Src protein tyrosine kinases (PTKs), at concentrations consistent with inhibition of Src function. Additionally, AS was phosphorylated directly in co-transfection experiments using c-Src and Fyn expression vectors and in in-vitro kinase assays with purified kinases. These data suggest that AS can be a target for phosphorylation by Src family PTKs, which is intriguing due to the overlap in brain region specific expression and subcellular localization of c-Src and AS. To elucidate functional consequences of this modification we initiated experiments to identify proteins that interact with the C-terminus of AS (residues 100-140) in a phosphorylation dependent manner. Biotinylated C-terminal peptides, phosphorylated at tyrosine 125 or non-phosphorylated, were incubated with brain lysate, and interacting proteins isolated on a streptavidin column, separated by SDS-PAGE and visualized by protein staining. In addition to several proteins isolated with both phosphorylated and non-phosphorylated peptides, 6 proteins (~8,10,28,45,48, & 70 kD) copurify only with the non-phosphorylated peptide and thus bind specifically to the C-terminus of AS in a phosphorylation dependent manner. Currently, we are identifying these proteins by mass spectroscopy.

Maria L. Allende, NCI-CCR
HVMNE, a Novel Lymphocryptovirus Related to Epstein-Barr Virus, Induces Lymphoma in New Zealand White Rabbits
HVMNE is a novel Epstein-Barr (EBV)-like virus recently isolated from a simian T-cell leukemia/lymphoma virus (STLV) type I/II-negative Macaca nemestrina with a CD8+ T-cell mycosis fungoides/cutaneous T-cell lymphoma. Here we demonstrate that intravenous inoculation of irradiated HVMNE-infected T-cells or a cell-free virus pellet from the J94356PBMC cell line in New Zealand white rabbits results in seroconversion to the viral capsid antigen (VCA) of EBV in most of the exposed animals. In addition, all animals that seroconverted to VCA developed malignant lymphoma within months from inoculation. Four control rabbits, inoculated with heat-inactivated culture supernatants from the same cell line failed to seroconvert to VCA and did not develop disease. Disseminated lymphoma cells of mixed origin were detected in most vital organs, including the spleen, liver, lungs, kidneys, and heart of the affected rabbits. Neoplastic infiltrates were also observed in lymph nodes, the thymus, the skin and in subcutaneous tissues. HVMNE DNA and EBV-like RNA (EBER) expression was demonstrated in the lymphomatous organs as well as in two transformed T-cell lines, one established from the lymph node and the other from the blood from the two lymphomatous animals. Detailed analysis of one of these T-cell line revealed the persistence of HVMNE DNA, expression of an LMP1-like protein and the acquisition of IL-2-independence coupled with a constitutive activation of the Jak/STAT pathway.
The demonstration of HVMNE's high oncogenicity in New Zealand white rabbits provides a valuable animal model for human T-cell lymphoma whereby genetic determinants for T-cell transformation by this EBV-like animal virus can be studied.

Steven E. Finkelstein, NCI-CCR
Virtual Bronchoscopy in the Evaluation of Thoracic Malignancies
Objective: Virtual bronchoscopy (VB) is a novel technique employing 3-D reconstruction of 2-D helical CT scan images for the noninvasive evaluation of the tracheobronchial tree. This study assesses the diagnostic potential of VB by comparing virtual bronchoscopic images with the gold standard of fiberoptic bronchoscopy (FB) in the evaluation of patients diagnosed with thoracic malignancies.
Experimental Design: Thirty-two consecutive patients who were diagnosed with thoracic malignancy, and underwent VB for evaluation of suspected tracheobronchial lesions, were prospectively studied. For each VB, 200 to 300 contiguous 1.25mm images of the thorax were obtained in one or two 17-second breath holds using a multislice CT scanner. VB images were 3-D reconstructed and interpreted blind to the actual endoscopic findings. The lesions present on VB were compared with actual FB findings.
Results: Anatomic computer simulation of the bronchial tree was successfully created in all patients. As of January 2001, 20 patients who had undergone VB had had correlative FB within 2 weeks. Seven patients (35%) were found to have FB within normal limits. In all patients with normal anatomy, VB accurately correlated with the FB findings. Thirteen patients (65%) had a total of 22 abnormal FB findings. VB detected 18/22 abnormal FB findings: 13/13 obstructive lesions; 5/6 endoluminal lesions; and 0/3 mucosal lesions. VB yielded additional information by detecting seven obstructive lesions that were not visualized by FB due to size limitation of the bronchoscope and lesions beyond stenoses. The sensitivity of VB was 100% for obstructive lesions, 83% for endoluminal lesions, 0% for mucosal lesions, and 82% for all abnormalities; the specificity of VB was 100%.
Conclusions: This study suggests that VB is an accurate and noninvasive method for identifying obstructions and endoluminal lesions, and for visualizing the tracheobronchial tree beyond stenoses. Although FB remains the gold standard for identifying airway patency and mucosal lesions, VB yields additional information that may be useful in the management of airway malignancies.

André Fisahn, NICHD
Neuronal gamma oscillations and the connection between the M1 muscarinic receptor and the hyperpolarization-activated current (Ih) - A journey from network phenomenon to ion channel
Much of the electrical activity in the brain is oscillatory in nature. One prominent rhythm are gamma oscillations (30-80Hz). They represent the synchronous discharge of action potentials by networks of pyramidal neurons and are widely accepted to play an important role in higher brain functions such as perception, learning and memory. Interestingly, they are disrupted in Alzheimer's, a disease that is also connected to cholinergic-muscarinic receptor dysfunction. We have investigated the role of the M1 muscarinic receptor in gamma oscillations in area CA3 of hippocampal slices of both wild-type (WT) and M1 knockout mice (KO).
Using extracellular field recordings we have shown that the muscarinic receptor agonist muscarine induces gamma oscillations in WT but not in KO slices. Therefore, activation of the M1 receptor is necessary for the induction of cholinergic-muscarinic gamma oscillations. Patch clamp experiments revealed that muscarine causes a substantial inward current in WT pyramidal neurons that contributes to a pronounced membrane depolarization. In contrast, in KO neurons the inward current is small and causes only a modest depolarization. This indicates that a major component of muscarine modulation is absent in KO pyramidal neurons. Muscarine has been shown to modulate two major voltage-gated conductances in pyramidal neurons, IM and Ih. Selective antagonists and different voltage dependent characteristics allowed us to study them in isolation. Our data show that the muscarinic modulation of IM is unaltered in KO slices, indicating that activation of the M1 receptor is not linked to IM as previously assumed. In contrast, the increase of Ih by muscarine in WT was absent in KO pyramidal neurons. We could demonstrate that muscarine increases Ih not by shifting its activation curve towards more positive membrane potentials but by increasing its maximal conductance. Finally we tested our results in extracellular field experiments in WT slices. Confirming the patch clamp data an IM antagonist was unable to induce rhythmic activity whereas an Ih antagonist abolished all muscarine-induced gamma oscillations.
Our findings bring together several lines of research and bridge the gap between cellular currents and neuronal network activity. For the first time we show a crucial role for Ih, and rule out a role for IM, in gamma oscillations. Concomitantly, our data delineate a novel pathway that links the increase of Ih to M1 muscarinic receptor activation.

Ross Fitzgerald, NIAID
Evolutionary genomics of Staphylococcus aureus: insights into the origin of methicillin-resistant strains and the toxic shock syndrome epidemic
An emerging theme in medical microbiology is that extensive variation exists in gene content among strains of many pathogenic bacterial species. However, this topic has not been investigated on a genome scale with strains recovered from patients with well-defined clinical conditions. Staphylococcus aureus causes several life-threatening diseases in humans such as septicemia, endocarditis and toxic shock syndrome, and also causes economically important infections in cows and sheep. A DNA microarray representing >90% of the S. aureus genome was used to characterize genomic diversity, evolutionary relationships, and virulence gene distribution among 36 strains of divergent clonal lineages, including methicillin-resistant strains and organisms causing toxic shock syndrome. Genetic variation in S. aureus is very extensive, with approximately 22% of the genome comprised of dispensible genetic material. Eighteen large chromosomal regions of difference (RDs) were identified, and 10 of these regions have genes that encode putative virulence factors or proteins mediating antibiotic resistance. We find that lateral gene transfer has played a fundamental role in the evolution of S. aureus. Phylogenetic analysis indicates that the mec gene has been horizontally transferred into distinct S. aureus chromosomal backgrounds at least 5 times, demonstrating that methicillin-resistant strains have evolved multiple independent times, rather than from a single ancestral strain. This finding resolves a long-standing controversy in S. aureus research. The DNA microarray data confirmed that the majority of strains causing toxic shock syndrome are genetically related and have shared a common ancestor. However, clearly they are not genetically identical, and the last ancestor has not been very recent in evolutionary time. This indicates that the epidemic of toxic shock syndrome that occurred in the 1970s was caused by a change in the host environment, rather than rapid geographic dissemination of a new hypervirulent strain. DNA microarray analysis of large samples of clinically-characterized strains provides broad new insights into evolution, pathogenesis, and disease emergence.

Anna M. Dipietrantonio, NINDS
Incorporation of host cell proteins into viral particles - a possible mechanism for autoimmune disease?
Multiple sclerosis (MS) is a chronic neurological disease, characterized by destruction of the white matter in the brain by autoreactive infiltrating T and B lymphocytes. Genetic factors increase the MS susceptibility, but environmental factors are generally believed to be important triggers of the disease. Current data from our group supports an association between human herpes virus 6 (HHV-6) infection and MS. These are increased levels of anti-HHV-6 IgM in serum and CSF, higher lymphoproliferative responses, a higher T cell precursor frequency against the A variant of HHV-6, and detection of viral DNA in serum, urine and white matter plaques from MS patients.
A possible mechanism whereby viruses could induce autoimmunity is by incorporating host cells proteins into the viral envelope during the budding process. If viruses that can replicate in the myelin producing oligodendrocytes in the brain, incorporate myelin proteins, and then drain out from the brain, an immune response against both the virus and the myelin proteins might be elicited. This represent a true autoimmune response directing the activated T and B cells to specifically destroy myelin in the brain, as seen in MS. The aims of this project are to investigate the possible incorporation of host cell proteins into the HHV-6 virion, if these are different depending on what type of cell the virus is cultivated in, and if this is relevant in MS.
Incorporation of membrane cofactor protein, CD46, the cellular receptor for HHV-6, into the HIV, SIV and CMV virions protects these viruses against complement virolysis. With an immunoaffinity column we purified soluble CD46 from the serum of MS patients and controls and found a higher level of soluble CD46 in serum of MS patients. DNA was prepared from the anti-CD46 eluates and amplified by PCR for three different HHV-6 genes. Two of four MS patients had positive PCR for all three genes tested, whereas none of the 16 controls were positive for any of the genes. The detection of HHV-6 genome co-eluted with soluble CD46 might be interpreted as an incorporation of the CD46 molecule into the HHV-6 virion. Higher levels of anti-CD46 antibodies have been reported in MS patients, which possibly could be due to that HHV-6 renders CD46 immunogentic by incorporating CD46 into the virion. We are currently infecting human oligodendrocytes with HHV-6 to determine possible incorporation of myelin proteins into the virus.

Stephan Frank, NINDS
Dynamin-related protein mediates apoptosis
The physiological cell suicide program of apoptosis ensures the controlled elimination of unwanted or seriously damaged cells and is therefore of fundamental importance for embryogenesis, maturation and function of the immune system, as well as for the normal cellular turnover. On the intracellular level, many apoptosis signaling pathways are regulated by mitochondrial organelles.
In healthy cells mitochondrial morphology is maintained by balanced fusion and fission events. We made the discovery that, during apoptosis, mitochondria disintegrate into multiple fragments. This led us to examine the role of Drp1, a dynamin-related protein that mediates outer mitochondrial membrane fission, in apoptotic cell death. Upon induction of apoptosis, Drp1 translocates from the cytosol to mitochondria, where it preferentially localizes to potential sites of organelle division. Inhibition of Drp1 by expression of a dominant negative mutant counteracts mitochondrial fragmentation, prevents the loss of the mitochondrial membrane potential and the release of cytochrome c, and reveals a reproducible swelling of the organelles. Remarkably, inhibition of Drp1 (including inhibition experiments on endogenous level by microinjection of Drp1-blocking antibodies) blocks cell death.
These results, which were obtained using multiple inducers of various mitochondria-dependent pathways in various cell lines, warrant the conclusion that dynamin-related protein 1 generally functions in the mitochondrial pathways of apoptosis.

Patrick S. Frost Bellgowan, NIMH
Voxel-wise Estimation of Hemodynamic Onset Delays During a Lexical Decision Task.
Introduction: Attempts to derive temporal information from FMRI data have been limited to ROI analyses that increase delay estimation error due to between voxel hemodynamic variability. The present study uses a voxel-wise estimation of hemodynamic delays associated with perceptual and cognitive task manipulations.
Task: Three male subjects performed a lexical decision task comprised of words and pronounceable non-words rotated at either 0o (NR) or 120o (R). Forty stimuli of each type were presented for 750 ms followed by a fixation crosshair with an average ISI of 3000 ms.
FMRI Methods: Four sets of 384 echo-planar images using a TR of 1000 ms and fourteen 6mm non-contiguous sagittal slices were collected in a randomized event-related design. General linear contrasts derived from the deconvolved impulse response functions (IRF's) comparing words vs. non-words, and rotated vs. non-rotated stimuli were used to restrict the analyses to active voxels within the left fusiform and superior frontal gyrus. Delays were calculated by correlating a shifted gamma-variate function with the derived IRF's and determining which shifted function resulted in the strongest correlation. Voxel delays from each stimulus type and ROI were submitted to an ANOVA to determine regions sensitive to rotation and lexical delays.
Results: The subjects performed 92% correct on the lexical decision task with reaction times (RT's) of 879, 1591, 1174, and 1895 ms for the non-rotated words, rotated words, non-rotated non-words, and rotated non-word stimuli, respectively. Active voxels in the left fusiform gyrus were found to have significant hemodynamic onset delays due to the rotation manipulation. The left superior frontal gyrus demonstrated significant delays due to the lexical decision with non-words showing longer onset delays than words.
Discussion: Results demonstrate that voxel-wise delay estimation can successfully dissociate regions associated with RT delays due to perceptual and lexical manipulations. However, the temporal resolution used in the present study (100 ms) is not sufficient for discerning RT differences observed in most cognitive tasks. Currently we are developing analysis techniques to more accurately estimate hemodynamic delays that also incorporate the width of the IRF. In combination these to estimates may allow for detection of both the locus of the delay and regions affected by delayed processing.

Theresa M. Geiman, NCI-CCR
Identification of novel DNMT3B complexes containing HDAC1, HDAC2, the SNF2 family member SNF2H, and the KAP-1/Tif1-beta corepressor
Cytosine methylation is a covalent modification of DNA that can lead to repression of chromatin. This repression can occur through the methylation and the recruitment of methyl-CpG-binding proteins, corepressors, histone deacetylases, and chromatin remodeling proteins. In this way, there are several possible layers of regulation leading to the alteration of chromatin structure, and therefore the modification of cellular processes such as transcription. The exact mechanisms of inducing repression through DNA methylation remain uncertain, but the study of enzymes that perform DNA methylation, DNA methyltransferases (DNMTs), and their interacting proteins should provide valuable clues in this process. In an attempt to elucidate the function of one of the DNMTs, DNMT3B, we have purified protein complexes from both murine (P19) and human (HeLa) cell lines. DNMT3B is a de novo DNA methyltransferase that is particularly interesting because it is overexpressed in human tumors and mutated in human ICF syndrome (Immunodeficiency, Chromosome instability, and Facial anomalies). This mutation results in decreased DNA methylation and increased genomic instability. Enzymatic assays reveal that the DNMT3B complexes exhibit DNA methyltransferase, histone deacetylase, and ATP/nucleosomal/histone-dependent ATPase activities. This is the first ex