P E O P L E

RECENTLY TENURED

Cynthia Dunbar  received her B.A. and M.D. degrees at Harvard University, then completed internal medicine residency training at Boston City Hospital. She joined the Clinical Hematology Branch of the NHLBI in 1987 as a medical staff fellow. Aside from a year at the University of California at San Francisco, where she received further clinical training in hematology-oncology, she has been at the NHLBI ever since. She is currently director of the Hematology Fellowship Training Program and active in clinical branch activities in bone marrow transplantation, as well as her laboratory activities.

Cynthia Dunbar with photo of daughters, Anna and Alexa Cerf

The current research in my laboratory and clinical research program focuses on improving our understanding of functional characteristics of pluripotent hematopoietic stem cells (PHSCs) and using that knowledge to improve the transfer of exogenous genes into these cells in order to treat a wide variety of congenital and acquired human diseases. The PHSC has been an obvious prime target for gene therapy applications because of the ease of collecting these cells (not just from bone marrow but, more recently, also from peripheral blood and cord blood sources); because it is now possible to maintain and manipulate the cells ex vivo with hematopoietic growth factors; and because retroviral vectors can be used to transfer genes to murine PHSCs at relatively high efficiencies.

During my postdoctoral studies in Arthur Nienhuis' laboratory from 1987 to 1990, we used retroviral gene transfer techniques in the murine model to study the consequences of overexpression of cytokine genes such as IL-3 and IL-6. Besides helping us refine techniques to improve gene transfer efficiency into murine stem cells, these studies also provided insight into the possible roles of autocrine growth factor production in myeloproliferative diseases and leukemias, and further work defined a possible intracellular autocrine pathway for IL-3 signaling.

The Clinical Hematology branch has one of the few nonhuman primate hematopoietic transplantation facilities in the world, and in 1991, we began to use this model to work toward human trials of retroviral gene transfer directed at PHSCs. The primate experiments showed that the efficiencies of gene transfer to repopulating stem cells appeared to be much lower than the mouse model, and important safety data were obtained when we found that the presence of recombinant, replication-competent helper virus could lead to aggressive T-cell lymphomas in monkeys heavily immunosuppressed by the transplantation procedure. In the laboratory, we began to explore the optimal conditions for safe, practical, and efficient gene transfer to human progenitor cells. We found that vector transduction of purified CD34+ progenitor cell populations over the course of 72 to 96 hours in the presence of IL-3, IL-6, and stem cell factor (SCF) produced high (30-90%) transduction efficiencies of either bone marrow or peripheral blood progenitor cells, assayed either by standard semi-solid media colony assays or by long-term culture assays designed to study more primitive cells.

Thus, in 1992-1993, we initiated the first genetic marking trial of hematopoietic progenitor and stem cells in adult patients who were undergoing autologous transplantation for multiple myeloma or breast cancer, in collaboration with our NHLBI clinical service, investigators in the Medicine Branch of NCI, and the Department of Transfusion Medicine of the Clinical Center. The goals of the study were to assess the feasibility of using these techniques in humans before attempting trials with potentially therapeutic genes, and to compare bone marrow and "mobilized" peripheral blood cells as sources for gene transfer targets. If marking was successful, it would also begin to answer important questions about transplantation biology, including the relative kinetics and durability of reconstitution after transplantation with cells from the bone marrow compared with peripheral blood.

The first phase of the trial has been completed, and the good news is that we found long-term marking of all lineages with cells derived from PHSCs in some patients but, unfortunately, at levels (less than 0.1-1%) that are unlikely to be clinically useful for most therapeutic applications. Cells derived from both the marrow and peripheral blood grafts contributed to the short- and long-term marking. No marked relapses of breast cancer or myeloma were seen, and no toxicity or helper virus generation was detected.

We have continued the clinical trials with modifications of the transduction conditions to try and improve these results and have also examined several factors in the laboratory to improve the efficiency of transfer. One area of interest has been a potentially negative role of other cytokines or regulators of hematopoiesis during ex vivo culture and transduction. Transforming growth factor-b (TGF-ß) and macrophage inflammatory protein-1' (MIP-1') had previously been shown to inhibit the growth of certain progenitor populations ex vivo, but their effect on true PHSCs had not been studied. In a murine competitive repopulation model, we found that TGF-ß but not MIP-1' significantly inhibited PHSCs cultured ex vivo under conditions commonly used for retroviral transduction. More interestingly, we found that adding an antibody that neutralizes TGF-ß in an ex vivo culture improved PHSC activity, suggesting that autocrine or paracrine production of this mediator may be part of the difficulty in keeping PHSCs alive ex vivo and in stimulating them to divide so that they can be transduced by retroviral vectors.

Ongoing studies will examine the effect of this manipulation on gene transfer efficiency in the murine and primate models. We are also using the primate model and gene marking techniques to study ex vivo culture conditions that have been reported to allow actual expansion of primitive human cells. These conditions clearly produce an increase in total cell numbers and progenitor numbers, but the effect on true PHSCs is not known, and may even be negative, as these culture conditions may push true stem cells to terminally differentiate. We have also been developing new vector systems that will allow immediate selection of transduced cells without the need for prolonged culture in selective media. Another project is to develop vectors that would transduce nondividing cells.

We are also continuing clinical studies to improve assays of gene transfer. Current in vitro assays for human stem cells are inadequate and nonpredictive, and the animal models, especially the mouse, have not allowed quantitative predictions. Ongoing small-scale trials of the transfer of the glucocerebrosidase gene into hematopoietic cells of patients with Gaucher Disease (in collaboration with investigators at NINDS) and transfer of the multidrug resistance gene to patients undergoing autologous transplantation with breast cancer (in collaboration with NCI researchers) may not yet benefit these specific patients, but they should prove invaluable in learning more about the behavior of stem and progenitor cells after transplantation in these different situations and help direct further improvements of gene transfer techniques.

Thomas Leto received his Ph.D. from the University of Virginia in 1980 and did postdoctoral work at Yale University before joining the Laboratory of Clinical Investigation of NIAID in 1988. He is now a senior investigator in the Laboratory of Host Defenses, NIAID.

Thomas Leto

My interests are in the area of protein structure as it relates to function, with a focus on the formation of intracellular protein complexes. My research at NIAID has applied these interests to clinically relevant host defense systems in phagocytic blood cells. We have been studying an enzyme, NADPH oxidase, that generates reactive oxidants that kill invading microbes. This multi subunit enzyme assembles from membrane and cytosolic components during cellular activation; its importance in host defense is evident in patients with chronic granulomatous disease (CGD), whose inherited deficiency in oxidant production renders them susceptible to bacterial and fungal infections. CGD is the result of a defect in any one of four essential components of the oxidase.

In my early work at NIH, my colleagues and I identified the genes that encode two cytosolic oxidase components (p47-phox and p67-phox) affected in a majority of autosomal-recessive CGD patients. Much of our progress in understanding the NADPH oxidase relied on the development of gene expression systems that we used to restore the defective enzyme from CGD patients. These advances provided the impetus for ongoing CGD gene therapy studies in the Laboratory of Host Defenses.

In addition to its direct clinical relevance, this system has provided fertile ground for investigators interested in signal transduction and cellular activation. We recognized that within the deduced structures of p47-phox and p67-phox there are duplicated sequence motifs of about 60 amino acids, called Src homology 3 (SH3) domains. These motifs are found in a variety of intracellular proteins that participate in diverse signaling cascades in organisms ranging from yeast to humans. On the basis of clues from these other systems, we deduced the role of the oxidase SH3 domains in activation of this enzyme. SH3 domains are recognition modules for proline-rich target sequences and, in the case of the oxidase, we identified several specific SH3 targets within other oxidase components. Our work demonstrated that the SH3 domains are major links that bring this enzyme complex together during cell activation.

We have since found several ways to interfere with oxidase assembly based on disruption of SH3 interactions. In one case, we showed that a single proline mutation can cause CGD by blocking movement of cytosolic components to the membrane. We also showed that another SH3 domain-containing protein (p40-phox) inhibits the enzyme, and we have recently discovered a proline- rich natural peptide called PR-39 that can inhibit the oxidase when added to whole cells.

We hope that these molecular details will be used some day to design drugs that block undesirable effects of inflammation on "innocent bystander" host tissues which occur as phagocytes produce oxidants intended to kill pathogens. Many disease states, such as septic shock, arthritis, neurodegenerative diseases, atherosclerosis, cancers, and aging, are thought to result, in part, from inappropriate over-production of reactive oxidants during chronic or acute inflammatory processes. The oxygen- dependent defense system we are studying has apparently been around for a long time, evolutionarily speaking, since we have evidence for conserved oxidase counterparts in the plant kingdom. It will be interesting to see whether the plant enzyme is regulated by conserved signal-transduction systems. Future work will also explore other possible roles of phagocyte oxidants, such as intercellular communication during wound healing.

James Dee Higley received his Ph.D. from the University of Wisconsin, Madison, in 1985 and did postdoctoral work at the university and at NIH before joining the Laboratory of Clinical Studies of NIAAA and the Laboratory of Comparative Ethology, NICHD. He became a fellow in the Laboratory of Clinical Studies, NIAAA, in 1991 and is now a research psychologist in that laboratory.

James Higley

Violence and alcoholism are endemic public health problems affecting Americans and their families. My principle research has focused on the neurobiology of these two problems in a nonhuman primate model. The research subjects are rhesus macaques that are selectively bred for high or low central nervous system (CNS) serotonin functioning. This research had its genesis in our discovery that CNS serotonin turnover, as measured by low concentrations of the major metabolite of serotonin, 5- hydroxyindoleacetic acid (5-HIAA), in cerebrospinal fluid (CSF), is highly heritable. Subsequent studies showed that early experiences also play a major role in CNS serotonin functioning, with parental neglect reducing CSF 5-HIAA concentrations to levels lower than those in subjects receiving normative parental care. These early rearing experiences appear to cause reductions in CNS serotonin that persist from the neonatal period into adulthood.

One of the more important findings from my laboratory is that differences in CSF 5-HIAA concentrations between individuals are highly stable, showing consistent stability beginning as early as day 15. Interindividual differences in CSF 5-HIAA concentrations in infancy were found to predict interindividual differences in adulthood. Such findings suggest that two individuals with differing CNS serotonin reactivity may respond quite differently to the same stimuli. Furthermore, this propensity may be trait-like, with long-term, predictable differences.

Reduced CNS serotonin functioning has deleterious effects on the acquisition of social competence in the developing and adult primate. Human studies have shown that men with low CSF 5-HIAA concentrations are more likely to engage in violent behaviors, abuse alcohol, and engage in other antisocial behaviors. Paralleling these findings in humans, rhesus macaques with low CSF 5- HIAA concentrations engage in impulsive violent behavior, consume alcohol at high rates, have few social partners, and engage in high risk behavior. They are more likely to be ostracized from their social groups and die at an early age, often from violence. In a recent study, we found that CNS testosterone augments the effect of reduced CNS serotonin functioning on violence, although subjects with high CSF 5-HIAA concentrations were unlikely to engage in violence even if they had high testosterone, a finding that suggests that testosterone may produce competitive and aggressive motivations, but the intensity and timing of the aggressive or competitive response may be under serotonin control. Indeed, pharmacological treatment using serotonin reuptake inhibitors reduces aggression and other impulsive behaviors. Interestingly, subjects with high CSF 5-HIAA concentrations are more likely to become socially dominant, a measure of competent social behavior in nonhuman primates. Unlike subjects with low CSF 5-HIAA concentrations, subjects with high CSF 5-HIAA concentrations have little trouble falling asleep at night.

Early rearing experiences, such as parental neglect, not only reduce the responsiveness of the serotonin system, they also affect behaviors that are under serotonin control, such as aggression and alcohol consumption. For example, monkeys reared in social settings with other age-mates but no adults present (peer-reared monkeys) are more likely to be removed from their social groups for aggressive injuries requiring medical treatment. They consume alcohol at rates twice as high as subjects reared with their parents. Female peer-reared monkeys often abuse or neglect their infants, a behavior hardly ever seen in females reared by their parents. Peer-reared monkeys seldom become socially dominant.

Studies of human sons of alcoholic fathers show that they have a reduced response to the pharmacological effects of alcohol. Recently, we investigated the possibility that this well-replicated effect may be mediated by reduced serotonin functioning. We found that monkeys with low CSF 5-HIAA concentrations require more pentobarbital anesthesia to maintain unconsciousness, and when they are given identical dosages of alcohol, they are rated as less intoxicated than subjects with high CSF 5-HIAA. When they are anesthetized with halothane and tested using PET scans, subjects with low CSF 5-HIAA concentrations show higher whole-brain arousal, particularly in the frontal cortex region. These findings suggest that one of the reasons that subjects with low CSF 5-HIAA are at risk to abuse alcohol and drugs may be because they drink at higher levels to induce the same effect as subjects with high CSF 5-HIAA.