News and Publications

Top of Page

Chairman's Overview

Ernest Beutler, M.D.

How do great scientific achievements come about? One view is that important scientific advances are the consequence of teamwork by a group of experts who, working in unison, create a great breakthrough. Such an approach led to the development of the nuclear bomb and to the success of space probes, and such successes have led the public at large and some politicians and some science administrators to embrace the idea that large teams are the key to success--the so-called critical mass. One of my teachers, Nobel laureate Charles Huggins, once said, "The critical mass in science is one scientist." In biomedical research, this statement, it seems to me, is closer to the truth. Collaborations and team efforts play a role in exploiting new ideas, but the origin of an important new concept can usually be attributed to one person. New ideas usually arise from existing observations, often observations that have lain fallow for a long time. Discussions with colleagues play an important role in developing new concepts, but it usually seems to be one person who formulates the critical idea and performs the critical experiment--one experiment that in retrospect is often extremely simple.

A basic philosophy that guides us in the Department of Molecular and Experimental Medicine is that it is the unfettered scientist who will make the important contribution. We now have 46 full-time faculty members in the department, and each member has his or her own research program. The department is composed of nine divisions, each headed by an accomplished scientist with his own program. However, these divisions are not teams working on the research projects of the division heads. Rather, each division is an administrative and intellectually cohesive unit composed of independent scientists working on related but distinct problems. All the heads of the divisions have achieved recognition as leaders in their respective fields of science. In every case, it was not because the head belonged to a successful "team"; it was because that person did something unique and unexpected that moved science forward. This year's overview focuses on the early work of these nine scientists and how that work has influenced research in the divisions they head.

In addition to serving as chairman of the department, I am also the head of the Division of Hematology. It was my own good fortune to be assigned directly from my clinical and laboratory studies at the University of Chicago to the Stateville Penitentiary at Joliet, Illinois, as a U.S. Army medical officer. There, working only with prisoners as laboratory assistants, and later at the University of Chicago, working with the part-time help of one technician, I was able to unravel much of the biochemistry of the erythrocyte defect that led to hemolytic anemia when 8-aminoquinoline antimalarial drugs were given to certain patients. The expression of this defect, X-linked glucose-6-phosphate dehydrogenase deficiency, in women heterozygous for the gene led to my formulation of the X-inactivation hypothesis in 1961, largely through my interactions with Susumu Ohno shortly after I arrived at the City of Hope. The genetics of common diseases and the physiology of erythrocytes have dominated my work ever since, and the study of iron storage disease (hemochromatosis), Gaucher disease, red cell enzyme defects, and chronic granulomatous disease have been major foci of research in the Division of Hematology in the past decade.

Bernard Babior, head of the Division of Biochemistry, studied enzymatic mechanisms as a Ph.D. candidate after receiving his M.D. degree at the University of California, San Francisco. Working in the laboratory of Konrad Bloch at Harvard, he unraveled the mechanism of enzymatic action of ethanolamine ammonia lyase and was the first to show the role of free radicals in an enzymatic mechanism. Subsequently, as a beginning assistant professor, he was the first to recognize the importance of free radicals of oxygen, specifically in antibacterial defense by phagocytes, but also in biology as a whole. Bernie's early insights in these areas have been the stimulus for work in literally hundreds of laboratories throughout the world. His own work in the biochemistry division has expanded on this theme, and he has brought the most modern technology to bear in gaining understanding of the delicate control mechanisms that determine whether a cell will fire off a burst of highly active oxygen radicals. The Division of Biochemistry was already well established when Bernie arrived at TSRI. The entirely independent world-class programs on mitochondrial function and on the cytochrome P-450 hydroxylating enzymes that were in place have thrived under his intellectual leadership.

Frank Chisari, head of the Division of Experimental Pathology, came to our institute as an M.D. trained in both pathology and internal medicine. As a postdoctoral fellow, he showed that it was not the hepatitis virus that destroyed liver cells, but the body's immune response. Recognizing that a thorough understanding of the techniques of molecular biology would be needed to unravel the pathogenesis of this viral disease, he spent a sabbatical leave at the Pasteur Institute and pioneered the production of the first transgenic mouse model of a human viral disease, a system that subsequently has provided unprecedented insight into the dynamics of the host-virus relationship during hepatitis B. This technique and the development of a method to selectively amplify the pool of extremely rare hepatitis B virus--specific cytotoxic T cells found in the blood of infected patients have kept the Chisari laboratory in the lead in unraveling the pathogenesis of hepatitis. One of Frank's long-term aims as a physician has been the control of this disease, and a novel synthetic lipopeptide vaccine that he developed is now being tested in chronically infected patients.

Zaverio Ruggeri, head of the Division of Experimental Hemostasis and Thrombosis, was a clinical hematologist in Milan when he became interested in patients who had congenital bleeding problems. He noticed that a group of patients with von Willebrand disease had a paradoxical result in a platelet aggregation test then considered essential for diagnosis of the disease. He showed that the defect responsible for this result was associated with a qualitative abnormality of von Willebrand factor, and in 1978, he came to Scripps to try to explain this paradox. This move was the beginning of almost 20 years of work that culminated in the crystallization of the A1 domain of von Willebrand factor, the location of the mutations for the variant type of von Willebrand disease that first piqued his interest. Understanding the mechanisms that regulate the formation of platelet thrombus may help explain how acute arterial thrombosis occurs and make it possible for Zaverio and his colleagues in the Division of Hemostasis and Thrombosis to devise better treatments to prevent this critical event in cardiovascular disease. Zaverio's work closely complements independent investigations under way in his division on thrombophilia, the mechanism of stroke and cellular signaling.

In 1970, Peter Vogt, head of the Division of Oncovirology, discovered src, the first oncogene. He showed that the Rous sarcoma virus that induced neoplastic transformation had a genome that was 2 kb larger than the genome of a nononcogenic strain of the same virus. A few years later, he and his colleagues showed that the cancer-producing genes of such retroviruses were modified cellular genes. His broad search for other cellular oncogenes led to the discovery of other oncogenes, including jun, which encodes Jun, the first DNA-binding protein with strong oncogenic potential. These early seminal findings have focused Peter's interest on control of gene expression and its role in the neoplastic cellular phenotype, a theme that dominates the work currently under way in the Division of Oncovirology.

Martin Lotz, head of the Division of Arthritis Research, was struck early in his career by the profound aging-associated changes in the function of human articular chondrocytes. Trained as a physician, he found these changes intriguing because of their possible role in the development of osteoarthritis, a disorder that affects a large percentage of older persons. This interest provided the basis for several current projects in his laboratory that address the role of changes in cell-cycle distribution on matrix gene expression, the contribution of nitric oxide and other radicals to aging of human cartilage, and the relationship between extracellular matrix and the regulation of chondrocyte differentiation and survival. Few programs anywhere deal with this important disorder, and the multipronged attack of the work in our new Division of Arthritis Research may yield important insights that will lead to prevention and treatment of a major crippling disease.

Jim Koziol, head of the Division of Biomathematics, has a long-standing interest in the analysis of cancer research data, both cell and tumor growth and clinical outcomes. In the late 1970s, he became interested in the analysis of growth curves of induced tumors in homogeneous populations of mice subjected to different immunotherapeutic regimens. The variability of the response of mice made a parametric method for analysis inappropriate, and in 1981 Jim published an appropriate distribution-free method of analysis based on fundamental rank permutation principles. This method and its extensions have become standards for the analysis of tumor growth curves and are also widely used with repeated measurements data as a powerful alternative to parametric modeling approaches. Currently, Jim and his group have been valuable collaborators not only with members of our department but also with others at TSRI and elsewhere in study design and analysis of a variety of clinical and nonclinical projects. These projects include studies on the use of 2-chlorodeoxyadenosine in the treatment of multiple sclerosis and lymphomas and investigations on interstitial cystitis, sleep, and tumor cell growth in animal models.

When Jim Hoch, head of the Division of Cellular Biology, first came to Scripps as a postdoctoral fellow, he was fascinated by the question of how a cell knows when to grow and to divide and when to shut down these activities and round up into a resistant spore. Bacteria are ideally suited for such studies, and he set out to characterize the genes that direct sporulation in Bacillus subtilis. His work attracted broad attention, because the genes that he found regulated developmental decisions defined new mechanisms of signal transduction that are now known to regulate many important bacterial processes, including virulence and antibiotic production. The discovery of signal transduction pathways that are unique to bacteria and not found in humans has made it possible to approach design of antibiotics in a unique manner. The development of such antibiotics would be another example of how basic scientific studies with no obvious practical applications can lead to results that are of great benefit to society.

Eng Tan, head of the Division of Research Rheumatology and director of the W.M. Keck Autoimmune Disease Center, joined the Rockefeller University as a postdoctoral fellow in 1962 after receiving his M.D. degree at Case Western Reserve, interning at Duke, and receiving further clinical training at Johns Hopkins. He worked in the laboratory of the late Henry G. Kunkel, one of the eminent pioneers in clinical immunology research. As a fellow, Dr. Tan characterized a nuclear autoantigen in systemic lupus erythematosus called Sm, showing that this antigen was localized in the nucleus and was not a histone protein. A second observation was the demonstration that DNA antigen and antibodies to the antigen appeared sequentially in the blood of a patient with lupus erythematosus, a finding that established the importance of the formation of antigen-antibody complexes in tissue injury and inflammation. Subsequent studies by several other investigators made use of autoantibodies to Sm to elucidate the role of small nuclear ribonucleoproteins in pre-mRNA splicing. Dr. Tan's laboratory has continued to determine and characterize other antigen-antibody systems in systemic autoimmune diseases. Investigators in this division are attempting to elucidate how drugs such as procainamide induce autoantibody responses to chromatin (DNA-histone complexes) and how cellular proteins involved in RNA processing, such as fibrillarin and SS-B/La, become transformed into immunogenic antigens capable of driving autoimmune responses. In this division, several independent investigators have made good use of spontaneously occurring autoantibodies as powerful reagents for elucidating autoimmune mechanisms and for determining proteins involved in important cellular functions.

I hope that these thumbnail sketches of the heads of the divisions put into some context the eclectic array of investigations described in the following pages.

Top of Page

Investigators' Reports

Division of Arthritis Research

Martin Lotz, M.D., Division Head

Research in this division is focused on pathogenetic mechanisms in arthritis. The major cause of disability in all forms of arthritis is loss of articular cartilage. In the inflammatory arthropathies, such as rheumatoid arthritis, the primary cause of cartilage degradation is a chronic immune and inflammatory process in the synovial membrane. By contrast, osteoarthritis, the most prevalent form of arthritis, is associated with intrinsic aging-associated changes in chondrocytes, leading to impaired function and degradation of the cartilage extracellular matrix.

Projects on inflammatory arthropathies address the role of cytokines and receptors for cytokines in the activation of leukocytes and mesenchymal cells. Projects on osteoarthritis analyze phenotypic changes in chondrocyte aging and molecular mechanisms responsible for these changes. Research on signal transduction in cartilage degradation is aimed at defining major intracellular messengers that induce catabolic responses in chondrocytes. The approaches used in this work are cell-based investigations with human tissues and studies in experimentally induced arthritis. Animal models are used to examine the consequence of gene deletions on joint inflammation and cartilage degradation and to develop methods for cartilage regeneration and chondrocyte transplantation.

Top of Page

Chondrocyte Aging and the Pathogenesis of Osteoarthritis

S. Hashimoto, R. Terkeltaub,^* M. Lotz

^* University of California, San Diego, CA

Cartilage is among the tissues with the highest prevalence of aging-related pathologic changes, and osteoarthritis is the most common musculoskeletal disease. Major changes associated with joint aging include abnormal deposition of crystals in articular cartilage and reduced cartilage thickness and cellularity. Sporadic chondrocalcinosis or pseudogout is an age-related disease. The deposition of calcium pyrophosphate dihydrate crystals in the articular cartilage of patients with pseudogout links abnormal or excessive production of pyrophosphate to the etiology of this disease. Thus, altered pyrophosphate metabolism appears to be a key event in the development of abnormal cartilage calcification in aging.

Transforming growth factor-ß1 (TGF-ß 1) is the only cytokine thus far known to elevate production of pyrophosphate. TGF-ß is also the major growth factor for human articular chondrocytes. To address the influence of age on pyrophosphate accumulation in human articular chondrocytes, we studied the effects of TGF-ß1 on pyrophosphate levels in the media and cell lysates of cultures of chondrocytes obtained from subjects in two different age groups: 15--55 and 56--91 years old. The same cell cultures were used to analyze the effects of TGF-ß on chondrocyte proliferation.

TGF-ß 1 increased pyrophosphate levels to a greater extent in chondrocytes from subjects in the older age group than in chondrocytes from younger subjects. Treatment of chondrocytes from subjects in both age groups led to a similar increase in total intracellular protein. Although TGF-ß increased nucleoside triphosphate pyrophosphohydrolase activity and decreased alkaline phosphatase activity, these effects did not differ between the two age groups. Analysis of the same cell preparations showed an age-related decrease in proliferation induced by TGF-ß and an increase in the production of pyrophosphate. These results indicate that aging differentially affects TGF-ß--induced accumulation of pyrophosphate vs TGF-ß--induced proliferation in human articular chondrocytes. These changes in the response to TGF-ß most likely contribute to the development of age-associated cartilage diseases.

Top of Page

---

Chondrocyte Apoptosis

S. Hashimoto, R.L. Ochs, M. Lotz

Degradation of the extracellular matrix of cartilage is a central event in the pathogenesis of arthritis and is thought to be related to the increased production of proteases. The extracellular matrix also regulates cell differentiation and survival. The aging-associated reduction in cartilage cellularity predisposes to degeneration of the matrix and eventually osteoarthritis. We have examined the relationship between apoptosis in chondrocytes and degradation of the extracellular matrix.

Chondrocyte apoptosis can be induced by proinflammatory cytokines, such as IL-1, IL-17, and TNF. These cytokines also induce the production of nitric oxide in chondrocytes, and nitric oxide is a mediator of cytokine-induced chondrocyte apoptosis. Activation of the Fas receptor also triggers apoptosis. Apoptosis induced by antibody to the Fas antigen does not depend on nitric oxide, and antibodies to Fas also did not induce production of nitric oxide. Articular cartilage is not vascularized and does not contain phagocytic cells, suggesting that apoptotic bodies that are generated as a result of chondrocyte death may remain in the tissue and potentially alter the extracellular matrix.

Electron microscopy showed that chondrocyte apoptosis was associated with degradation of the pericellular matrix that surrounds the chondrocytes. Isolated apoptotic bodies are functionally active and produce pyrophosphate, precipitate calcium, and contain matrix metalloproteinases. These results suggest that chondrocyte apoptosis can lead to the degradation and calcification of cartilage extracellular matrix.

Top of Page

---

Signal Transduction in Cartilage Degradation

Y. Geng, T. Shalom-Barak, J. Valbracht, M. Lotz

Interleukin-1 is prototypic for cytokines that induce synovitis and degradation of cartilage. The IL-1--activated signaling pathways show tissue-specific differences, and even in the same cell type, multiple IL-1--induced second messengers differentially regulate subsets of IL-1--responsive genes. Our previous studies suggested that activation of tyrosine kinase is an important signal transduction event in the response of chondrocytes to IL-1.

MAP kinases are major tyrosine phosphorylated proteins in IL-1--stimulated chondrocytes. Chondrocytes express the MAP kinases Erk-1, Erk-2, JNK-1, JNK-2, and p38. These kinases were time dependently activated by IL-1. Among other chondrocyte activators tested, only TNF activated all three of the MAP kinase subgroups. JNK and p38 were not activated by any of the other cytokines and growth factors tested. However, Erk was also activated by platelet-derived growth factor, insulin-like growth factor-1, and IL-6. Phorbol myristic acetate, calcium ionophore, and cAMP analogs increased Erk activity but had no significant effects on JNK or p38.

These results suggest differential activation of MAP kinase subgroups by extracellular stimuli. Erk is activated in response to qualitatively diverse extracellular stimuli and various second messenger agonists. In contrast, JNK and p38 are activated only by IL-1 or TNF, suggesting that these kinases participate in the induction of the catabolic program in cartilage.

Nitric oxide is produced during joint inflammation, and inhibition of the synthesis of nitric oxide reduces disease severity in different experimental models of arthritis. Chondrocytes are a major intraarticular cell source capable of expressing inducible nitric oxide synthase (iNOS) and of producing large quantities of nitric oxide. Guanylate cyclase is activated by IL-1 in nitric oxide--dependent and nitric oxide--independent pathways. In chondrocytes, IL-1 increases levels of cGMP, and this increase is, at least in part, mediated by nitric oxide.

We investigated the role of guanylate cyclase and phosphodiesterase in IL-1 activation of human articular chondrocytes. The guanylate cyclase inhibitors LY83583 and methylene blue inhibited IL-1--induced production of nitric oxide and expression of iNOS protein and mRNA. These effects were consistent with the rapid induction of cGMP by IL-1, which reached maximal levels after 5 minutes. The effects of the inhibitors were selective; they did not reduce IL-1--induced cyclooxygenase II protein and mRNA. Addition of the cGMP analog dibutyryl cGMP neither induced production of nitric oxide nor reversed the effects of LY83583, suggesting that cGMP-dependent protein kinase was not mediating the cGMP effects. Also, cGMP-dependent protein kinase activity was not detectable in resting or IL-1--stimulated cells. However, 8-bromo-cGMP, a slowly hydrolyzable analog of cGMP, induced production of nitric oxide and expression of iNOS protein and mRNA.

Furthermore, the phosphodiesterase inhibitor IBMX selectively blocked IL-1--induced expression of iNOS mRNA and release of nitric oxide but did not inhibit IL-1--induced expression of cyclooxygenase II mRNA. Chondrocytes contained extensive cGMP phosphodiesterase activity, which was approximately 100-fold higher than that of cAMP phosphodiesterase. The cGMP phosphodiesterase did not hydrolyze cAMP and had an inhibitor profile similar to those of the phosphodiesterase 5 and phosphodiesterase 6 families. Inhibitors specific for phosphodiesterase 5 suppressed IL-1--induced release of nitric oxide and expression of iNOS mRNA. These results suggest that increased cGMP metabolic flux is sufficient to induce iNOS and that guanylate cyclase and cGMP phosphodiesterase are required for IL-1 induction of iNOS expression via increases in coupled cGMP synthesis and hydrolysis.

The DNA-binding proteins involved in the regulation of iNOS expression include NF-B and the interferon regulatory factors (IRFs). In addition to iNOS, IRFs transcriptionally regulate IFN and many IFN-inducible genes. Several cytokines and other mediators regulated by IRF-1 are involved in the induction of inflammation. This information provided the basis for the hypothesis that deletion of IRF-1 inhibits or reduces inflammatory responses.

Mice with a targeted disruption of the gene for IRF-1 (IRF-1^-/-) were used to examine the role of this transcription factor in synovial inflammation and the production of nitric oxide. Intraarticular injection of IL-1 or lipopolysaccharide reduced the intensity of synovial lining hyperplasia and leukocyte infiltration significantly more in the IRF-1^-/- mice than in wild-type (control) mice of the same parental lineage. Nitric oxide is involved in the pathogenesis of arthritis, and IRF-1 regulates expression of iNOS in mononuclear phagocytes. Articular chondrocytes from IRF-1^-/- mice produced similar levels of nitric oxide in response to IL-1 or lipopolysaccharide. Furthermore, the synergistic induction of nitric oxide by IFN-gamma and IL-1 or lipopolysaccharide was almost identical in chondrocytes from wild-type and IRF-1^-/- mice. This finding was in contrast to the expected decrease in nitric oxide production by peritoneal macrophages from IRF-1^-/- mice, suggesting that IRF-1 is not required for iNOS expression in chondrocytes. IRF-1 has a tissue-specific role in the induction of iNOS. Inhibition of this transcription factor may represent a novel approach to the control of inflammatory diseases such as arthritis.

Top of Page

---

ILA, a Member of the Nerve Growth Factor/Tumor Necrosis Factor Receptor Family

M. Setareh, H. Schwarz,^* J. von Kempis,^** M. Lotz

^* University of Regensburg, Regensburg, Germany
^** University of Freiburg, Freiburg, Germany

The ILA gene is induced by lymphocyte activation. It is a member of the nerve growth factor/tumor necrosis factor receptor family and the human homolog of the murine 4-1BB gene. ILA has been mapped to chromosome 1p36, a region where the genes for several other members of this receptor family are located and one that contains deletions and rearrangements associated with hematopoietic neoplasms. ILA is expressed by activated but not by resting T lymphocytes, B lymphocytes, and blood monocytes. It can also be induced in nonlymphoid cells, including epithelial cells, hepatocytes, and chondrocytes. Studies of ILA expression in chondrocytes showed that it is inducible by proinflammatory cytokines, such as IL-1. The gene is not inducible in fibroblasts and thus can be used to detect chondrocyte-specific signaling events that are activated by IL-1.

Analysis of ILA function showed that this receptor is involved in the regulation of proliferation and survival of T lymphocytes. Antibodies to ILA increased T-lymphocyte proliferation induced by antibodies to CD3. Incubation of anti-CD3--stimulated cells on Chinese hamster ovary cells overexpressing ILA or on plates coated with a fusion protein that contains the extracellular part of ILA inhibited proliferation and induced lymphocyte apoptosis. Because the ILA ligand is also a membrane-associated protein with a large cytoplasmic domain, these findings indicate that cross-linking of the ILA ligand also regulates cell functions.

PUBLICATIONS

Geng, Y., Valbracht, J., Lotz, M. Activation of MAP kinase subgroups p38 and JNK is selectively associated with IL-1 or TNF stimulation of human articular chondrocytes. J. Clin. Invest. 98:2425, 1996.

Geng, Y., Zhou, L., Thompson, R., Lotz, M. A novel cyclic GMP phosphodiesterase required for interleukin-1 induced nitric oxide synthesis in human articular chondrocytes. J. Biol. Chem., in press.

Hashimoto, S., Setareh, M., Lotz, M. Fas/Fas ligand expression and induction of apoptosis in chondrocytes. Arthritis Rheum. 40:1749, 1997.

Lotz, M. Cytokines and their receptors. In: Arthritis and Allied Conditions, 13th ed. Koopman, W. (Ed.). Williams & Wilkins, Baltimore, 1997, p. 439.

Lotz, M. Mechanisms of pain induction in arthritis. In: Anesthesia: Biologic Foundations. Biebuyck, J.F., et al. (Eds.). Lippincott-Raven, Philadelphia, in press.

Lotz, M. Mechanisms of tissue destruction in rheumatoid arthritis. In: Rheumatology, 2nd ed. Klippel, J.H., Dieppe, P.A. (Eds.). Mosby, St. Louis, in press.

Lotz, M., Setareh, M., von Kempis, J., Schwarz, H. The nerve growth factor/tumor necrosis factor receptor family. J. Leukoc. Biol. 60:1, 1996.

Lotz, M., Villiger, P. Cellular responses to cytokines. In: Cytokines and the CNS. Ransohoff, R.M., Benveniste, E.N. (Eds.). CRC Press, Boca Raton, FL, 1996, p. 47.

Maier, R., Wisniewski, H.G., Vilcek, J., Lotz, M. Regulation of TSG-6 expression in human articular chondrocytes: Possible implications in inflammatory joint diseases. Arthritis Rheum. 39:552, 1996.

Rosen, F., McCabe, G., Quach, J., Solan, J., Terkeltaub, R., Seegmiller, J., Lotz, M. Aging differentially affects human chondrocyte responses to TGFß: Increased pyrophosphate production and decreased cell proliferation. Arthritis Rheum. 40:1275, 1997.

Schwarz, H., Blanco, F., Valbracht, J., von Kempis, J., Lotz, M. ILA, a member of the human NGF/TNF receptor regulates lymphocyte proliferation and survival. Blood 87:2839, 1996.

Schwarz, H., Lotz, M. ILA, a gene encoding a member of the nerve growth factor/ tumor necrosis factor receptor family is located on human chromosome 1p36. Biochem. Biophys. Res. Commun. 235:699, 1997.

Shiraishi, A., Dudler, J., Lotz, M. Reduced severity of synovial inflammation in IRF-1-deficient mice. J. Immunol. 159:3549, 1997.

von Kempis, J., Schwarz, H., Lotz, M. Stimulus-specific and differentiation-dependent expression of ILA, a member of the human NGF/TNF receptor family in cells of mesenchymal lineage: Evidence for a dominant transcriptional inhibitor in mesenchymal cell differentiation. Osteoarthritis Cartilage, in press.

Zvaifler, N.J., Tsai V., Alslameh, S., von Kempis, J., Firestein, G.S., Lotz, M. Pannocytes: Distinctive cells found in rheumatoid arthritis articular cartilage erosions. Am. J. Pathol. 150:1125, 1997.

Top of Page

---

Division of Biochemistry

Bernard M. Babior, M.D., Ph.D., Division Head

---

Oxygen-Dependent Bacterial Killing by Neutrophils

B.M. Babior

Neutrophils are small, highly motile cells that seek out and destroy invading pathogens. Among their microbicidal weapons is a remarkable variety of reactive oxidants that kill the invaders by chemical combustion. Among these oxidants are the hydroxyl radical and other highly reactive free radicals, hypochlorite and an almost infinite variety of chloramines, singlet oxygen, and a number of reactive oxides of nitrogen. These oxidants are all formed from superoxide (O₂^-), a compound generated by the one-electron reduction of oxygen.

Superoxide is produced by the leukocyte NADPH oxidase, an enzyme that is dormant in resting neutrophils but is activated when the cells encounter invading microorganisms. The activated oxidase catalyzes the production of O₂^- from oxygen and NADPH:

2 O₂ + NADPH 2 O₂^- + NADP⁺ + H⁺

The leukocyte NADPH oxidase is a complex enzyme containing at least five subunits, two in the plasma membrane and three in the cytosol. Those in the plasma membrane form a flavohemoprotein known as cytochrome b₅₅₈. Those in the cytosol include p47^PHOX (PHOX = phagocyte oxidase), which is important for activation of the enzyme; p67^PHOX, which probably contains the substrate-binding site; and p40^PHOX, an inhibitory subunit. When the cell is activated, the cytosolic subunits migrate to the membrane to assemble the working oxidase (Fig. 1).

This laboratory is studying the molecular mechanism of oxidase activation, in which phosphorylation plays an important part. Nine to 10 serines in the C-terminal quarter of one of the cytosolic oxidase subunits (p47^PHOX, see Fig. 1) become phosphorylated when neutrophils are stimulated, providing correlative evidence that phosphorylation is involved in oxidase activation. Our studies with serine-to-alanine mutants of p47^PHOX have furnished stronger evidence of the role of p47^PHOX phosphorylation in oxidase activation. The functionally important serines are Ser359 and Ser370, which must be phosphorylated before the rest of the molecule can take up phosphate or translocate to the plasma membrane, and Ser303 and Ser304, which must carry a negative charge for oxidase activity to be expressed, although their phosphorylation is not required for translocation. Phosphorylation of at least one of the two serines in each of these pairs is necessary for oxidase activation. All four of these serines are targets for protein kinase C, and this kinase is probably the enzyme responsible for their phosphorylation in intact cells.

The phosphorylation of p47^PHOX causes a conformational change that buries the C-terminal region of the protein, including the cysteine at position 378. The replacement of Cys378 with an alanine causes a major delay in oxidase activation in cells expressing this mutant p47^PHOX. A similar delay in oxidase activation accompanied by a major reduction in oxidase activity is seen in cells expressing the mutant p47^PHOX C111A. The similarity in the delay in activation seen with these two cysteine-to-alanine mutants hints at the possibility that oxidase activity may be controlled not only by phosphorylation but also by sulfhydryl oxidation.

We have developed a cell-free system in which the oxidase is activated by kinases. With this system, which contains neutrophil cytosol and plasma membranes, ATP, GTP, and magnesium, we can study, for the first time, oxidase activation by a physiologic route. We found that in addition to phosphorylation of the mutant p47^PHOX, activation requires the phosphorylation of a protein in the plasma membrane; the latter phosphorylation is accomplished by a membrane-associated kinase. The most extensively phosphorylated membrane protein is about 25 kD. Both activation of the oxidase and phosphorylation of this protein are partly inhibited by wortmannin, which inhibits phosphoinositide 3-kinase, and by GF109203X, which inhibits protein serine-threonine kinases, and are nearly completely inhibited by the tyrosine kinase inhibitor genistein. The correlation between oxidase inhibition and the inhibition of protein phosphorylation suggests that the 25-kD protein may be the one whose phosphorylation is required for oxidase activation.

The function of the cytosolic subunit p67^PHOX had been a mystery since the time of its discovery. We recently obtained evidence suggesting that this subunit carries the NADPH-binding site of the oxidase. Earlier studies from our laboratory showed that neutrophil cytosol contained an oxidase subunit that could be inactivated by NADPH dialdehyde, an affinity-label analog of NADPH. Measurements of K_m and K_i for this inactivation reaction suggested that the susceptible subunit contained the NADPH-binding site of the oxidase. We recently purified the labeled subunit and showed that it was p67^PHOX. In addition, we found that recombinant p67^PHOX could restore oxidase activity to cytosol that had been inactivated with NADPH dialdehyde and that the ability of p67^PHOX to correct the defect was abolished by NADPH dialdehyde. These results strongly suggest that the dialdehyde-susceptible oxidase component and p67^PHOX are one and the same.

PUBLICATIONS

Babior, B.M., Chanock, S.J., El Benna, J., Smith, R.M. The NADPH oxidase of leukocytes: The respiratory burst oxidase. In: Oxidative Stress and the Molecular Biology of Antioxidant Defenses. Scandalios, J.G. (Ed.). Cold Spring Harbor Press, Cold Spring Harbor, NY, 1997, p. 737.

Park, J.-W., Babior, B.M. Activation of the leukocyte NADPH oxidase subunit p47^phox by protein kinase C: A phosphorylation-dependent change in the conformation of the C-terminal end of p47phox. Biochemistry 36:7474, 1997.

Park, J.-W., Hoyal, C.R., El Benna, J., Babior, B.M. Kinase-dependent activation of the leukocyte NADPH oxidase in a cell-free system: Phosphorylation of membranes and p47^phox during oxidase activation. J. Biol. Chem. 272:11035, 1997.

Top of Page

---

Programmed Cell Death

R.A. Gottlieb, B.M. Babior

Programmed cell death, or apoptosis, refers to the process in which a cell responds to an external stimulus by committing suicide in a highly stereotyped manner. Events in apoptosis include cell shrinkage, degradation of the cell's chromatin and destruction of its genome, appearance of new antigens on the cell surface that mark the cell as a target for phagocytosis, cross-linking of cell proteins to form a leathery envelope at the cell surface, and eventually disintegration of the cell by blebbing and elimination of the resulting fragments by phagocytes. Apoptosis is one of the three general types of behavior of cells (the other two are replication and differentiation), and it plays an indispensable role in development, tissue homeostasis, and response to injury.

This rapidly expanding field can be broken down into several areas: the conditions under which apoptosis is induced (many different diseases manifest disorders of apoptosis, including myocardial infarction, neurodegenerative diseases, and cancer), the genetic control of cell death (many transcripts are induced in cells subjected to an apoptotic stimulus), and the biochemical mechanisms of cell destruction (death proteases, mitochondrial alterations, and the role of Bcl-2 and its relatives are currently the focus of intense scrutiny).

Our earlier work showed that neutrophils acidify their cytoplasm during apoptosis and that the cytokine granulocyte colony-stimulating factor protected against both acidification and apoptosis by upregulating a membrane-associated proton pump. We have shown that components of the proton pump are stabilized in the presence of granulocyte colony-stimulating factor and that cytosolic components translocate to the membrane to form active pumps. Similar studies in cardiomyocytes from adult rabbits have shown that the stress response known as preconditioning appears to require activity of the proton pump for protection against injury mediated by ischemia and reperfusion. Previously, we showed that ischemia-reperfusion injury resulted in apoptosis of cardiomyocytes. Caspase activation is a general and essential feature of apoptosis; we have been able to show that caspase inhibitors are highly protective against ischemia-reperfusion injury.

Acidification may be necessary for activation of the endonuclease responsible for degradation of DNA during apoptosis. In a cell culture model of cystic fibrosis, the elevated resting pH of the cells interferes with cytoplasmic acidification and apoptosis. Studies of apoptosis in the mouse model of cystic fibrosis (deletion of the gene that encodes the cystic fibrosis transmembrane conductance regulator) suggest that apoptosis is defective in exocrine tissues (Fig. 1). Defective apoptosis could be responsible for the high molecular weight DNA that contributes to the viscous mucus in the airways of patients who have cystic fibrosis.

Mitochondrial function is greatly impaired during apoptosis. Fas-mediated apoptosis results in defunctionalization of cytochrome c that involves generation of a cytosolic factor known as cytochrome c inactivating factor of apoptosis (CIFA). Bcl-2 can reverse the inactivation of cytochrome c, and purification of CIFA is under way. Entry of CIFA into the mitochondria is regulated by the mitochondrial outer membrane (Fig. 2). The basis for this effect is under investigation.

Mitochondrial abnormalities are a general feature of neurodegenerative diseases, and defects in cytochrome oxidase are implicated in Alzheimer's disease. Amyloid peptide appears to result in mitochondrial inactivation.

In some settings, the endonuclease responsible for DNA degradation during apoptosis is an acidic endonuclease similar or identical to DNase II. This enzyme is only active below pH 6.8 and could become active when cells acidify during apoptosis. We are purifying this enzyme for cloning and x-ray crystallographic studies.

PUBLICATIONS

Adachi, S., Cross, A.R., Babior, B.M., Gottlieb, R.A. Bcl-2 and the outer mitochondrial membrane in the inactivation of cytochrome c during Fas-mediated apoptosis. J. Biol. Chem. 272:21878, 1997.

Froelich, C.J., Orth, K., Turbov, J., Seth, P., Gottlieb, R.A., Babior, B.M., Shah, G.M., Dixit, V.M., Hanna, W.L. New paradigm for lymphocyte granule-mediated cytotoxicity. J. Biol. Chem. 271:29073, 1996.

Gottlieb, R.A. Cell acidification in apoptosis. Apoptosis 1:40, 1996.

Gottlieb, R.A., Babior, B.M. Regulation of Fas-mediated apoptosis. Curr. Top. Cell. Regul. 35:69, 1997.

Niessen, H., Meisenholder, G.W., Li, H.-L., Gluck, S.L., Lee, B.S., Bowman, B., Engler, R.L., Babior, B.M., Gottlieb, R.A. G-CSF upregulates the vacuolar proton ATPase in human neutrophils. Blood, in press.

Top of Page

---

Bioenergetics of Mitochondria and Bacteria

Y. Hatefi, A. Matsuno-Yagi, M. Yamaguchi, G.I. Belogrudov

The research of this laboratory is concerned with study of (1) the structure, composition, function, regulation, and interaction of the enzymes and enzyme complexes of the mitochondrial electron-transport and oxidative phosphorylation system and (2) the structure and mechanism of action of the mitochondrial and bacterial nicotinamide nucleotide transhydrogenase, which is a model for interconversion of substrate binding energy and transmembrane proton electrochemical potential via protein conformation changes.

MITOCHONDRIAL ATP SYNTHASE SUBUNITS

Recent studies have suggested that bovine ATP synthase contains 16 unlike polypeptides. Those previously characterized are the , ß, , , and subunits of the catalytic sector F₁; the ATPase inhibitor protein that binds to a ß subunit; and subunits a, b, c, d, OSCP, F₆, and A6L, which are contained in the membrane sector F_o and in the 45-Å-long stalk that connects F₁ to F_o. Subunits a and c are largely membrane intercalated. Each b subunit is considered to traverse the membrane twice, with its NH₂-terminal end (~30 residues) emerging from the membrane on the F₁ side. The remainder of each of the two b subunits (~130 residues) extends from F_o to F₁, to which are attached OSCP, d, and the two copies of F₆. A6L is anchored to the membrane via its NH₂-terminal 25--30 hydrophobic residues. The remaining half of the molecule is extramembranous on the F₁ side and resides near subunit d.

The newly characterized polypeptides found in preparations of bovine ATP synthase are three small molecules designated e, f, and g; the respective molecular masses are 8.2, 10.2, and 11.3 kD. To ascertain their involvement as bona fide subunits of the ATP synthase and to investigate their membrane topography and proximity to the previously described ATP synthase subunits, we prepared antibodies to the C-terminal amino acid residues 57--70 of e, 75--86 of f, and 91--102 of g. We found the following: (1) The subunits e, f, and g could be immunoprecipitated with anti-OSCP IgG from a fraction of submitochondrial particles enriched in oligomycin-sensitive ATPase. (2) The NH₂-termini of f and g are exposed on the matrix side of the mitochondrial inner membrane and can be curtailed by proteolysis. (3) The C-termini of all three polypeptides are exposed on the cytosolic side of the inner membrane. (4) The subunit f cross-links to A6L and to g, and e cross-links to g and appears to form an e-e dimer. Thus, the bovine ATP synthase complex appears to have 16 unlike subunits, twice as many as its counterpart in Escherichia coli.

INTERSUBUNIT INTERACTIONS IN BOVINE MITOCHONDRIAL COMPLEX I

The bovine mitochondrial complex I (NADH--ubiquinone oxidoreductase) is composed of more than 40 unlike subunits. We showed in 1967 that bovine complex I can be divided into three distinct fractions: flavoprotein, iron-sulfur protein, and hydrophobic protein. Bovine flavoprotein is composed of 3 polypeptides with molecular masses of 51, 24, and 9 kD. The 51-kD subunit binds NAD(H) and contains FMN and a tetranuclear iron-sulfur cluster. The 24-kD subunit contains a binuclear iron-sulfur cluster, and the 9-kD subunit carries no redox centers. The iron-sulfur protein contains 7 major polypeptides with molecular masses of 75, 49, 30, 18, 15, 13, and 11 kD; the 75-kD subunit contains a tetranuclear and a binuclear iron-sulfur cluster. Flavoprotein and iron-sulfur protein are water soluble. Hydrophobic protein is water insoluble and contains the 7 subunits of complex I that are encoded by the mitochondrial DNA. This tripartite arrangement appears also to be the architectural blueprint of complex I from Neurospora crassa (>= 32 subunits) and E. coli (13 subunits) and, possibly, of the enzyme from Paracoccus denitrificans (14 subunits).

Our cross-linking studies showed that each of the three flavoprotein subunits can be cross-linked to the other two; that the 75-kD subunit of iron-sulfur protein cross-linked to the 30-, the 18-, and the 13-kD subunits; and that the 30-, the 18-, and the 13-kD subunits also cross-linked to the 49-kD subunit. The 18- and the 13-kD subunits did not cross-link; nor did the 75- and the 49-kD subunits. Furthermore, the only link between flavoprotein and iron-sulfur protein involved the 51-kD subunit of the former and the 75-kD subunit of the latter, and iron-sulfur protein appeared to intervene between flavoprotein and hydrophobic protein.

We have now shown by ligand blotting that isolated iron-sulfur protein binds (1) only to the 51-kD subunit of flavoprotein and (2) to the 42-, 39-, 23-, 20-, and 16-kD subunits of hydrophobic protein. Because a 23-kD subunit and a 20-kD subunit of complex I are potential iron-sulfur proteins, these and our previous results are consistent with the following possible path of electrons in complex I: NADH 51- and 24-kD subunits of flavoprotein 75-kD subunit of iron-sulfur protein 23- and 20-kD subunits of hydrophobic protein ubiquinone.

FERROUS ION--CATALYZED FRAGMENTATION OF F₁-ATPase

Mitochondria produce large quantities of superoxide as a by-product of the terminal oxidation of foodstuffs by molecular oxygen. Dismutation of superoxide produces hydrogen peroxide, and Fenton chemistry produces hydroxyl radicals, which are potent oxidants capable of damaging proteins and DNA and of initiating lipid peroxidation. The rate of superoxide production is increased in diseases associated with mitochondrial dysfunction, of which a notable example is Alzheimer's disease. In laboratory experiments, ascorbate plus ferrous ions (Fe²⁺) can be used to produce the deleterious reactive oxygen species mentioned. We examined the effect of ascorbate plus Fe²⁺ on the catalytic sector F₁ of the ATP synthase complex.

We found that treatment of F₁ with Fe²⁺ plus ascorbate caused inactivation and fragmentation of the enzyme. Addition of magnesium ions or EDTA prevented inactivation and fragmentation. Both the and the ß subunits of F₁ were cleaved; most of the cleavage sites were on the subunit. Oxidative fragmentation of F₁ showed nucleotide dependence. Removal of nucleotides from F₁-ATPase and an excess of nucleotides in the medium dramatically altered the fragmentation pattern. The sizes of the fragments, their immunorecognition with antibodies against F₁ subunits, and the results of mild proteolysis of F₁ with trypsin suggested that the cleavage sites are located in the nucleotide-binding domains of both the and the ß subunits.

NUCLEOTIDE-BINDING DOMAINS OF RHODOSPIRILLUM RUBRUM TRANSHYDROGENASE

Nicotinamide nucleotide transhydrogenase is a proton pump of simple structure found in bacterial plasma and chromatophore membranes and in mammalian mitochondrial inner membranes. The enzyme catalyzes the direct and stereospecific transfer of a hydride ion between the 4A position of NAD(H) and the 4B position of NADP(H) in a reaction that is coupled to proton translocation with an H⁺/H^- stoichiometry of n = 1:

The bovine transhydrogenase is a half site--reactive homodimer of monomer molecular mass of 109 kD. The monomer is composed of three domains: an NH₂-terminal 430-residue-long hydrophilic domain I that binds NAD(H), a central 400-residue-long hydrophobic domain II that is largely membrane intercalated and harbors the enzyme's proton channel, and a 200-residue-long C-terminal hydrophilic domain III that binds NADP(H). Domains I and III protrude into the mitochondrial matrix, where together they form the enzyme's catalytic site. Other transhydrogenases have the same tridomain profile in two subunits (E. coli, Rhodobacter capsulatus), three subunits (R. rubrum), or a single subunit with altered arrangement of the domains (Eimeria tenella, Entamoeba histolytica).

We showed previously that the soluble substrate-binding domains I and III of bovine transhydrogenase (obtained by proteolysis from the purified enzyme or expressed in E. coli) interact, in the absence of domain II, to reconstitute transhydrogenation from NADPH to the NAD analog 3-acetylpyridine adenine dinucleotide (AcPyAD). Similarly, domain I (1 subunit) of R. rubrum transhydrogenase could be cross-reconstituted with domain III of the bovine enzyme. We have now shown that the expressed domains I and III of the R. rubrum transhydrogenase catalyze a rapid NADP(H)-dependent cyclic transhydrogenation from NADH to AcPyAD with a V_max of 214 mumol AcPyAD reduced (min x mg of domain I)^-1. As seen in Figure 1, cyclic NADH AcPyAD transhydrogenation differs from the normal, physiologic reaction (see the earlier equation) in two important respects: (1) During cyclic transhydrogenation, NADP(H) stays bound to the enzyme, and (2) as a consequence, the enzyme does not experience the difference in the binding energies of NADPH and NADP and does not undergo the conformational changes that result in proton translocation.

Thus, in agreement with the cyclic mechanism shown in Figure 1, the kinetics of the reaction were "ping-pong" with respect to NADH and AcPyAD, because these nucleotides bind interchangeably to domain I, and the stereospecificity of hydride ion transfer was from the 4A position of NADH to the 4A position of AcPyAD. The expressed domain I was dimeric, like the native 1 subunit of the enzyme, but the expressed domain III was monomeric and contained 0.94 mol of NADP(H) per mole.

PUBLICATIONS

Belogrudov, G.I. Mitochondrial ATP synthase: Fe²⁺-catalyzed fragmentation of the soluble F₁-ATPase. Arch. Biochem. Biophys. 335:131, 1996.

Belogrudov, G.I., Hatefi, Y. Intersubunit interactions in the bovine mitochondrial complex I as revealed by ligand blotting. Biochem. Biophys. Res. Commun. 227:135, 1996.

Belogrudov, G.I., Tomich, J.M., Hatefi, Y. Membrane topography and near-neighbor relationships of the mitochondrial ATP synthase subunits e, f, and g. J. Biol. Chem. 271:20340, 1996.

Matsuno-Yagi, A., Hatefi, Y. Ubiquinol:cytochrome c oxidoreductase: The redox reactions of the bis-heme cytochrome b in unenergized and energized submitochondrial particles. J. Biol. Chem., in press.

Yamaguchi, M., Hatefi, Y. High cyclic transhydrogenase activity catalyzed by expressed and reconstituted nicotinamide-binding domains of Rhodospirillum rubrum transhydrogenase. Biochim. Biophys. Acta 1318:225 1997.

Top of Page

---

Cytochrome P-450: Regulation, Structure, and Function

E.F. Johnson, J. Cosme, K.J. Griffin, M.-H. Hsu, F. Jung, W. Song

Our laboratory is working toward an understanding of the diversity of the cytochrome P-450 monooxygenases and of how these enzymes contribute to the ability to avoid the adverse effects of environmental chemicals. A striking feature of the P-450 enzymes is their capacity to catalyze the hydroxylation of a diverse array of lipophilic compounds, including hormones, carcinogens, and drugs. This metabolic transformation often provides a defensive function whereby toxic compounds can be converted to less toxic forms that are more easily excreted from the body.

PREDICTING P-450 SUBSTRATE SPECIFICITIES

A relatively large group of P-450 enzymes participate in the metabolism of drugs. In most cases, considerable overlap of substrate specificities exists, resulting in the metabolism of a drug by several P-450 enzymes. In some cases, a single enzyme may be involved, and genetic differences between individuals in the expression or efficiency of the enzyme can allow toxic concentrations of a drug to accumulate when normally safe doses are given, because the metabolism of the drug is diminished. The therapeutic ramifications of such interindividual variation in drug metabolism underscore the need not only to determine molecular characteristics of P-450 substrates that would be predictive of the enzymes most likely to metabolize a drug but also to understand the impact of genetic variation on catalytic activity. In addition, the rational design of new drugs would benefit from an understanding of determinants of efficient metabolism that may limit bioavailability of a candidate drug.

Earlier work in our laboratory showed that genetic variation leading to alterations in the substrate specificity of P-450s fits a framework model in which key residues function as the principal determinants of the active-site geometries. Empirically defined key residues correspond by sequence alignments with substrate-contacting residues in experimentally determined structures of bacterial P-450s. More explicit models incorporate conserved features of the protein folding shared by known structures for P-450s to yield detailed models of the active-site geometries of drug-metabolizing P-450s. We are currently validating and refining these models by using site-directed mutagenesis to predictably alter substrate specificity.

The basis for these models will also be greatly improved by experimental determinations of the structures of the microsomal drug-metabolizing P-450s. However, these enzymes are membrane bound and tend to aggregate when purified. These characteristics have hampered successful crystallization that would yield structural information on this enzyme group. We have examined several protein modifications designed to produce soluble versions of these enzymes that retain the enzymes' characteristic catalytic properties but do not aggregate extensively in solution. We have succeeded in generating modified microsomal P-450s that are no longer integral membrane proteins when heterologously expressed at high levels in Escherichia coli and that can be purified without the use of detergents. The modified enzymes have catalytic properties similar to those of the native enzymes and are dimers in solution. Crystallization trials with these modified P-450s are in progress.

ENZYME INDUCTION

A second, important aspect of the cytochrome P-450 system is that the expression of individual enzymes can increase dramatically in response to exposure to either drugs or toxic chemicals. This event leads to an increase in the concentration of the involved enzyme and often to an increased capacity to metabolize and detoxify the inducing agent. The main mechanism leading to increased enzyme expression is an increase in the transcription of the corresponding gene.

Peroxisome proliferators are a large class of foreign compounds that can induce a variety of peroxisomal enzymes and some microsomal P-450s. These compounds also increase the number of peroxisomes in sensitive species and can be carcinogenic. The P-450 enzymes induced by these compounds include those of subfamily 4A that catalyze hydroxylation of the terminal carbon of fatty acids. The product of this reaction can be oxidized further to a dicarboxylic acid that is metabolized in peroxisomes to shorter fatty acids that are more readily excreted. Dicarboxylic acids may in turn signal the cell to increase the number of peroxisomes. This mechanism burns off the caloric content of the fats less efficiently than other metabolic pathways do.

We have shown that induction of the P-450 gene 4A6 is mediated by a transcription factor, the peroxisome proliferator activated receptor (PPAR), that is related to the steroid hormone receptors. Multiple responsive elements in the promoter and 5´ flanking region of the 4A6 are involved. PPAR does not appear to interact with the gene alone. Rather it binds to the gene as a heterodimer with the retinoid X receptor (RXR).

Characterization of the peroxisome proliferator response elements (PPREs) in 4A6 indicated that they include imperfect direct repeats of the sequence AGGTCA that are also recognized by homodimers of other nuclear receptors such as RXR, apolipoprotein regulatory protein-1, and hepatocyte nuclear factor 4. Our studies indicate that PPREs contain an additional motif that may form an extended binding site that is not required by these other nuclear receptors.

The extended binding site for PPAR, AACT AGGTCA, is similar to the binding site for a different class of nuclear receptors that can function as monomeric transcription factors. Examination of known PPREs for other responsive genes suggested that the sequence of the direct repeat in PPREs generally deviates from that of strong binding sites for homodimeric nuclear receptors. The extended binding site for PPAR, which is conserved among PPREs, may compensate for the diminished binding strength of PPAR/RXR to the imperfect direct repeat. This compensatory binding interaction leads in turn to a greater specificity for the binding of PPAR/ RXR heterodimers over homodimers of RXR, apolipoprotein regulatory protein-1, and hepatocyte nuclear factor 4.

The molecular factors that activate PPARs are important in governing the balance between fat storage, glucose production from fats, and destruction of excess fats. Work is in progress to delineate these signal transduction pathways. A number of endogenous fatty acids appear to be ligands for PPARs but generally are weaker agonists than are oxidized metabolites of unsaturated fatty acids. PPAR is often expressed in tissues that express P-450s that catalyze the formation of epoxides of unsaturated fatty acids. These epoxides are rapidly metabolized to dihydroxylated compounds. Our studies indicate that dihydroxylated compounds formed from arachidonic acid are selective agonists of the different forms of PPAR, depending on the site of epoxidation. These agonists have relatively high affinities for the receptor, compared with the affinities of other peroxisome proliferators and other fatty acids, and thus may play a role in the signal transduction that regulates PPAR activity.

PUBLICATIONS

Brierley, C.H., Senafi, S.B., Clarke, D., Hsu, M.-H., Johnson, E.F., Burchell, B. Regulation of the human bilirubin UDP-glucuronosyltransferase gene. Adv. Enzyme Regul. 36:85, 1996.

Guengerich, F.P., Parikh, A., Johnson, E.F., Richardson, T.H., von Wachenfeldt, C., Cosme, J., Jung, F., Strassburg, C.P., Manns, M.P., Tukey, R.H., Pritchard, M., Fournel-Gigeux, S., Burchell, B. Heterologous expression of human drug-metabolizing enzymes. Drug Metab. Dispos., in press.

Johnson, E.F., Hsu, M.-H., Palmer, C.N.A. Peroxisome proliferator activated receptor: Transcriptional activation of the CYP4A6 gene. Ann. N.Y. Acad. Sci. 804:373, 1996.

Jung, F., Richardson, T.H., Raucy, J.L., Johnson, E.F. Diazepam metabolism by cDNA-expressed human 2C P-450s: Identification of P-4502C18 and P-4502C19 as low Km diazepam N-demethylases. Drug Metab. Dispos. 25:133, 1997.

Richardson, T.H., Griffin, K.J., Jung, F., Raucy, J.L., Johnson, E.F. Targeted antipeptide antibodies to cytochrome P-450 2C18 based on epitope mapping of an inhibitory monoclonal antibody to P-450 2C5. Arch. Biochem. Biophys. 338:157, 1997.

von Wachenfeldt, C., Richardson, T.H., Cosme, J., Johnson, E.F. Microsomal P-450 2C3 is expressed as a soluble dimer in Escherichia coli following modifications of its N-terminus. Arch. Biochem. Biophys. 339:107, 1997.

Top of Page

---

In Vivo Regulation of Hepatitis B Virus Biosynthesis by Peroxisome Proliferators

L.G. Guidotti, F.V. Chisari, A.K. Raney, A. McLachlan

The finding that the level of transcription from the nucleocapsid promoter is controlled by the nuclear hormone receptors retinoid X receptor and peroxisome proliferator activated receptor (PPAR) suggests that transcription and replication of hepatitis B virus (HBV) might be sensitive to disruption by antiviral agents that act as agonists or antagonists to the ligands for these receptors. In addition, the PPARs modulate transcriptional activity from promoters containing peroxisome proliferator response elements in response to intracellular metabolites such as polyunsaturated fatty acids, including arachidonic acid and its metabolic derivatives, which are synthesized at higher levels in response to inflammatory signals; hypolipidemic drugs such as fibric acid derivatives; xenobiotic compounds such as phthalate and adipate ester plasticizers; and herbicides. These observations suggest that HBV pregenomic RNA transcription and consequently HBV replication might be influenced by immunologic, nutritional, pharmacologic, and environmental compounds.

Because hepatitis B is an immune-mediated disease, the connection between the immune response of the host and viral transcription might be an important factor in determining the outcome of the infection. In the case of HBV, induction of an antiviral immune response in the liver would be predicted to activate the resident liver macrophages, the Kupffer cells, resulting in the increased synthesis of PPAR activators and of inflammatory mediators, the prostaglandins and leukotrienes. The uptake of prostaglandins and leukotrienes by hepatocytes suggests that the increase in the synthesis of these eicosanoids by activated Kupffer cells might result in the activation of PPAR in hepatocytes. In this way, viral transcription and replication might be modulated in a manner that reflects the extent of the ongoing antiviral immune response. This modulation may represent an evolutionary strategy adapted by the virus to evade elimination by the immune response of the host.

In order to examine this possibility, transgenic mice that replicate HBV in their livers at levels similar to those found in humans with chronic HBV infection were fed a diet containing the peroxisome proliferator clofibric acid, and the levels of viral RNAs and replication intermediates in their livers were determined. Analysis of liver RNA indicated that the level of the HBV 3.5-kb transcripts was higher in the mice treated with clofibric acid than in untreated control mice. In contrast, the level of the HBV 2.1-kb transcripts remained unaffected by the clofibric acid treatment. These observations reflect the result that might have been predicted on the basis of the transient transfection analysis of the effect of the retinoid X receptor and PPAR nuclear hormone receptors on HBV promoter activities. In these cell culture studies, selective activation of transcription from the nucleocapsid promoter occurred, but no significant alteration in the level of transcription from the major surface antigen promoter was detected.

Analysis of total liver DNA indicated that as a consequence of the increase in the HBV 3.5-kb transcripts, the level of HBV replication intermediates increased more than 20-fold in the animals treated with clofibric acid compared with the control mice. This finding suggests that clofibric acid--induced alterations in the rate of transcription from the nucleocapsid promoter result in greater increases in the level of intrahepatic viral replication intermediates than in the level of 3.5-kb HBV transcripts. However, it is clear that activation of PPAR by clofibric acid results in increased levels of HBV transcription and replication. This result shows that HBV biosynthesis is responsive to external stimuli in vivo in the HBV transgenic mouse model and suggests that HBV infection in humans may be subject to similar regulation.

PUBLICATIONS

Raney, A.K., Johnson, J.L., Palmer, C.N.A., McLachlan, A. Members of the nuclear receptor superfamily regulate transcription from the hepatitis B virus nucleocapsid promoter. J. Virol. 71:1058, 1997.

Raney, A.K., McLachlan, A. Characterization of the hepatitis B virus major surface antigen promoter hepatocyte nuclear factor 3 binding site. J. Gen. Virol., 78:3029, 1997.

Top of Page

---

The Biochemistry of Pteridines and the Short-Chain Dinucleotide-Requiring Dehydrogenases/Reductases

J.M. Whiteley, C.-F. Chang, T. Bray, K.I. Varughese, C.E. Grimshaw^*

* IGEN, Int., Gaithersburg, MD

Pteridines are bicyclic heterocyclic molecules that occur widely in nature. Among the most important ones are folic acid and biopterin and their respective derivatives. In humans, folic acid is ingested as one of the B vitamins, and biopterin is formed biosynthetically from a purine precursor (GTP). These two pteridines differ primarily in a side-chain substituent: Folates contain a p-aminobenzoylglutamate (or polyglutamate) group, whereas biopterin contains a dihydroxypropyl group. Both folic acid and biopterin perform their major biological actions when reduced.

The enzymes that participate in the metabolism of the folates have key functions in maintaining cellular viability, particularly in growing or proliferating cells, and are therefore major targets for the design of inhibitors with physiologic actions. Thus, compounds such as methotrexate, fluorouracil, dideazatetrahydrofolate, and mercaptopurine have been explored for their chemotherapeutic effects in many forms of cancer.

Biopterin, in its tetrahydro form, is essential for the enzymatic conversion of phenylalanine, tyrosine, and tryptophan, via hydroxylation reactions, to different members of the catecholamine family. Defects in any of these pathways can lead to severe mental disorders. Tetrahydrobiopterin is also an essential factor in the three forms of nitric oxide synthase, the product of which is an important physiologic messenger. However, whether this pteridine has a reactive or a dormant role in this important enzymatic function is unknown.

We have investigated the structure and mechanism of several of the enzymes in both the folate and the biopterin pathways. We have established both the structure and the possible mechanism of action of dihydropteridine reductase (DHPR), an enzyme that recycles quinonoid dihydrobiopterin, generated in the previously described hydroxylation reactions, back to the tetrahydro forms.

The structure and mechanism of DHPR are of particular interest because genetically induced errors of expression can lead to altered forms of phenylketonuria that do not respond to the usual therapy of a phenylalanine-free diet. Knowing the precise crystal structure of the enzyme and having the techniques to reproduce each lesion via bacterial expression have enabled us to analyze how such structural errors can influence the observed enzymatic activity.

In contrast to previous assumptions, we found that DHPR is not analogous to dihydrofolate reductase but is clearly a member of a large family of proteins known as short-chain dehydrogenases/reductases. This family now contains up to 100 members. In addition, the structure and mechanism of action of DHPR have close similarities to those of the carbohydrate epimerases and to those of the family of aldo-keto reductases. Thus, it appears that serendipitous comparative observations and the development of parallel lines of research from different laboratories have opened a new area for investigation.

Examination of this large family of proteins indicates that several important human enzymes are members. Because of both their biological importance and their intriguing mechanistic features, three have been chosen for closer study: 3ß-hydroxy steroid dehydrogenase ^5-4 isomerase, 15-hydroxyprostaglandin dehydrogenase, and sepiapterin reductase. An additional enzyme, PTR1, known to be elevated when the parasite Leishmania is treated with methotrexate, has an unusual composite of folate reductase activity and quinonoid dihydropteridine reductase structural features. Therefore, it has also been tentatively included in this family of proteins.

Structural and mechanistic investigations of homogeneous forms of the enzymes derived from gene-overexpression systems in Escherichia coli are ongoing. Although only 6 of the more than 200 amino acids that make up each enzyme sequence are strictly conserved, despite different substrates, the overall reaction pathway is similar. A major aim is to develop a cohesive molecular picture of how such a diverse group of apparently minimally sequence-related proteins can have similar folding patterns that ensure a few key amino acids are oriented to support the related reaction mechanisms. The human enzymes selected for deriving this information have important biological functions, and the resultant analytical picture could offer insight into metabolic control of the enzymes by improved drug design.

PUBLICATIONS

Feher, V.A., Zapf, J.W., Hoch, J.A., Whiteley, J.M., McIntosh, L.P., Rance, M., Skelton, N.J., Dahlquist, F.W., Cavanagh, J. High-resolution NMR structure and backbone dynamics of the Bacillus subtilis response regulator, Spo0F: Implications for phosphorylation and molecular recognition. Biochemistry 36:10015, 1997.

Kiefer, P.M., Grimshaw, C.E., Whiteley, J.M. The comparative interaction of (6R)-quinonoid dihydrobiopterin and an alternative dihydropterin substrate with wild-type and mutant rat dihydropteridine reductases. Biochemistry 36:9438, 1997.

Wang, J., Leblanc, E., Chang, C.F., Papadopoulou, B., Bray, T., Whiteley, J.M., Lin, S.X., Ouellette, M. Pterin and folate reduction by the Leishmania tarentolae H locus short-chain dehydrogenase/reductase PTR1. Arch. Biochem. Biophys. 342:198, 1997.

Zhou, X.Z., Madhusudan, Whiteley, J.M., Hoch, J.A., Varughese, K.I. Purification and preliminary crystallographic studies on the sporulation response regulatory phosphotransferase protein, Spo0B, from Bacillus subtilis. Proteins 27:597, 1997.

Top of Page

---

NADH Dehydrogenases

T. Kitajima-Ihara, A. Matsuno-Yagi, B. Seo, T. Yagi, T. Yano

STRUCTURE AND FUNCTION OF PROTON-TRANSLOCATING NADH-QUINONE OXIDOREDUCTASE

The proton-translocating NADH-quinone oxidoreductase (complex I) of mammalian mitochondria is composed of at least 43 different polypeptides and probably has the most intricate structure of any known membrane-bound enzyme complex. Of these subunits, 7 are encoded by mitochondrial DNA and are synthesized within the mitochondrion; the others are cytoribosomal products. Complex I contains one noncovalently bound FMN and at least five iron-sulfur clusters, detectable by electron paramagnetic resonance, as prosthetic groups. These five clusters are designated N1a, N1b, N2, N3, and N4. Clusters N1a and N1b are binuclear; N2, N3, and N4 are tetranuclear.

Studies of complex I are important because (1) the complex is the point of entry for a major fraction of electrons that traverse the respiratory chain, (2) this enzyme translocates H⁺ across the inner mitochondrial membrane from the matrix side to the cytoplasmic side for ATP synthesis, (3) the number of reports of human mitochondrial diseases and aging involving structural and functional defects at the level of this enzyme complex has increased, and (4) the complex is thought to be the most elaborate iron-sulfur protein.

Aerobically grown Paracoccus denitrificans, which is called "a free-living mitochondrion," expresses a mammalian mitochondrial--type respiratory chain. The proton-translocating NADH-quinone oxidoreductase (NDH-1) of Paracoccus membranes is quite similar to mammalian complex I in terms of electron carriers and amino acid sequences. However, in contrast to the mitochondrial enzyme, NDH-1 is composed of 14 dissimilar subunits, suggesting that the structure of Paracoccus NDH-1 is relatively simpler than that of its mitochondrial counterpart. Therefore, Paracoccus NDH-1 is a useful model system for studying the structure and function of complex I.

In previous studies, a gene cluster encoding the Paracoccus NDH-1 was cloned and sequenced. This cluster is composed of 14 structural genes (designated NQO1 to NQO14) and six unidentified reading frames. On the basis of the deduced primary structures of the Paracoccus NDH-1 subunits, subunits NQO1, 2, 3, 9, and possibly 6 are expected to bear iron-sulfur clusters. To verify these assumptions, we expressed single Paracoccus NDH-1 subunits in Escherichia coli and characterized the iron-sulfur clusters. The expressed NQO2 subunit contains a [2Fe-2S] cluster that most likely is N1a. Using site-directed mutagenesis, we showed that the cluster is coordinated by four conserved cysteine residues.

Electron paramagnetic resonance analysis of the expressed and reconstituted flavoprotein subcomplex, composed of the NQO1 and NQO2 subunits, suggests that the NQO1 subunit has a [4Fe-4S] cluster (N3). Furthermore, the expressed NQO3 subunit bears a [2Fe-2S] cluster (N1b), a [4Fe-4S] cluster (N4), and possibly an additional [4Fe-4S] cluster. Subunits NQO1--NQO6 in the membrane-bound Paracoccus NDH-1 were extracted by treatment at alkaline pH or with chaotropes, suggesting that these subunits are localized in the peripheral part of the enzyme complex. In addition, the subunit stoichiometry of NQO1--NQO6 of the Paracoccus NDH-1 is 1 mol each of the six subunits per mole of the Paracoccus NDH-1.

Thermus thermophilus HB-8, which was isolated from a hot spring in Japan, is an extremely thermophilic, obligatory aerobic, gram-negative chemoheterotrophic bacterium. The bacterium contains two NADH dehydrogenases: NDH-1 and a small non--energy-coupled enzyme referred to as NDH-2. We expect that thermostability will provide an advantage for structural studies of the NDH-1. Furthermore, because T. thermophilus bears only menaquinone, it will provide an exciting system for understanding the site 1 energy-coupling mechanism. In the past year, molecular cloning and DNA sequencing of the gene cluster encoding Thermus NDH-1 were done. This gene cluster is composed of 14 structural genes (designated NQO1 to NQO14) and no unidentified reading frames. All the structural genes encode subunits homologous to those of the Paracoccus NDH-1 enzyme complex.

Comparison of the deduced amino acid sequences with those of other organisms revealed that Thermus NDH-1 has principally the same molecular structure as other bacterial NDH-1 in terms of not only subunit composition but also cofactor-binding sites. In addition, the expressed Thermus NQO2 subunit bears a single [2Fe-2S] cluster with optical and electron paramagnetic resonance properties similar to those of the N1a cluster in the Paracoccus NQO2 subunit. The Thermus NQO2 subunit had a much higher stability than its mesophilic equivalents, and its iron-sulfur cluster remained intact even after incubation for 3 hours at 65°C under anaerobic conditions. As expected, Thermus NDH-1 provides a useful model system to investigate the structure-function relationships of the NDH-1 enzyme complexes.

MOLECULAR REMEDY OF COMPLEX I DEFECTS

Research in recent years has shown that structural and functional defects of complex I are involved in many human diseases. These diseases include Leber's hereditary optic neuropathy, Parkinson's disease, dystonia, severe lactic acidosis, various encephalomyopathies, and possibly Huntington's disease. Dysfunction of complex I causes three problems: (1) impairment of the ability of the respiratory chain to oxidize NADH back to NAD, which is required for operation of enzymes in the citric acid cycle and in fatty acid oxidation; (2) impairment of the ability of complex I to pump protons, resulting in a decrease in the rate of ATP synthesis; and (3) production of superoxide radicals, which cause mutations in mitochondrial DNA, peroxidation of lipids, and denaturation of proteins. Our overall goal is to find a remedy for the diseases caused by dysfunction of complex I.

As described earlier, complex I is composed of at least 43 unlike subunits and has the most intricate structure among membrane-bound enzyme complexes. Currently, mutations and deletions of the subunits encoded by mitochondrial DNA are difficult to correct, and mutations of plural subunits encoded by nuclear DNA are also difficult to repair.

Of the three problems mentioned here, impairment of proton pumping by one of the three proton translocation sites does not appear to be as severe a health hazard as the inability of mitochondria to oxidize NADH and damage by superoxide production. Yeast (Saccharomyces cerevisiae) mitochondria lack complex I and contain instead an NADH dehydrogenase composed of a single subunit (NDH-2). We intend to use yeast NDH-2 to transmit electrons from NADH to ubiquinone in mammalian mitochondria that lack a functional complex I.

First, to produce a strong antibody response to the yeast NDH-2, we tried to express in E. coli the NDI1 gene that encodes the yeast NDH-2. The expressed yeast NDH-2 was located in the membrane fraction of the bacteria but not in the soluble fraction or in the inclusion body fraction.

The activities involving NADH dehydrogenase of yeast NDH-2 expressed in E. coli membranes were significantly increased and inhibited by flavone (a specific inhibitor of yeast NDH-2). In addition, the NADH oxidase of the membranes was completely inhibited by potassium cyanide. These results indicate that the yeast NDH-2 is natively expressed in E. coli membranes and that it functions as a member of the E. coli respiratory chain. The expression of yeast NDH-2 in inner mitochondrial membranes of human cells is in progress.

PUBLICATIONS

Takano, S., Yano, T., Yagi, T. Structural studies of the proton-translocating NADH-quinone oxidoreductase (NDH-1) of Paracoccus denitrificans: Identity, property, and stoichiometry of the peripheral subunits. Biochemistry 35:9121, 1996.

Yano, T. NADH-quinone oxidoreductase: The hugest and most complicated membrane-bound enzyme complex. Tanpakushitsu Kakusan Koso 42:154, 1997.

Yano, T., Chu, S.S., Sled', V.D., Ohnishi, T., Yagi, T. The proton-translocating NADH-quinone oxidoreductase (NDH-1) of Thermus thermophilus HB-8: Complete DNA sequence of the gene cluster and thermostable properties of the expressed NQO2 subunit. J. Biol. Chem. 272:4201, 1997.

Top of Page

---

Normal and Leukemic Hematopoiesis

J. Yu, D.P. Dialynas, L.E. Shao, K.E. Dolter, M.J. Lee, J. Palyash, A. Galaviz

ACTIVIN A IN HEMATOPOIESIS AND INFLAMMATION

Activin A, a member of the transforming growth factor ß superfamily, is a protein consisting of two homodimeric ß_A subunits. Many studies have shown that activin A can potentiate the proliferation and differentiation of human erythroid progenitors in vitro and induce erythroleukemia cells to differentiate. Recent research indicated that proinflammatory cytokines and other regulators modulate the production of activin A not only in bone marrow stromal cells and monocytes but also in synoviocytes and chondrocytes. The upregulation by cytokines and suppression by glucocorticoids and retinoic acid imply that activin A may be involved in additional inflammation-related activities.

In one study, the levels of activin A in the synovial fluid were higher in patients with inflammatory arthropathies such as rheumatoid arthritis and gout than in patients with osteoarthritis. Further studies revealed that major biological activities induced by IL-6 that lead to systemic inflammatory responses were suppressed by activin A. These data suggest a role for activin A in the inflammation of joint tissue that parallels its known activity in the bone marrow microenvironment.

Currently, activin A is not known to play a role in inflammatory reactions. The previous findings, however, raise the possibility that additional inflammation-related activities of activin A should also be considered. For example, enhanced production of this molecule may increase host defenses by suppressing inflammatory manifestations in some pathologic disorders. All these studies, therefore, indicate that activin A may be involved in inflammation-related activities.

However, little is known about the regulation of activin A gene expression. Our studies of the 5´ ends of activin A RNAs in bone marrow stromal cells indicated that transcription start sites are located between 212 and 267 nucleotides upstream of the start codon. In addition, luciferase expression assays of a series of constructs containing overlapping 5´ sequences were used to determine regulatory sequences upstream of the activin A gene. An upstream sequence of 136 bp (-337 to -202 from the start codon) contains elements for basal expression, but several-fold higher expression requires an additional 71 bp (-670 to -600).

A promoter-regulatory region in HT1080 fibrosarcoma cells for the activin A gene may be located approximately 3 kb upstream from the start codon, suggesting the presence of one additional intron. We also found promoter activity between 3.6 kb and 2.5 kb upstream of the start codon in bone marrow stromal cells. However, using a primer located within the putative upstream intron, we observed a primer extension product with total RNA from HT1080 fibrosarcoma cells and bone marrow stromal cells. Therefore, some activin mRNA synthesized is presumably not spliced in the upstream region. These findings suggest that expression of activin A follows complex patterns of regulation.

CYTOKINES IN THE REGULATION OF HEMATOPOIESIS

It is thought that hematopoiesis is regulated by both stimulatory and inhibitory activities derived from the bone marrow microenvironment, which includes macrophages. Recently, we established a human macrophage line, 2MAC, that appears to be a better model of mature macrophage function than other established monocytic/macrophage lines. 2MAC cells are useful for studying the macrophage effector response and calcium mobilization that occur on ligation of HLA class II molecules.

We found that 2MAC cells constitutively secrete a reversible suppressor of early hematopoietic progenitors, designated NRH. The mechanism of NRH suppression appears to involve a marked decrease in the cycling of early progenitor cells. The data suggest that NRH activity corresponds to an acidic, heparin-binding glycoprotein with a molecular weight of about 20 kD and that NRH is a novel macrophage-derived negative regulator of hematopoiesis. This molecule may have future application in certain clinical settings as a chemoprotectant of primitive hematopoietic cells.

MOUSE MODELS FOR LEUKEMIC HEMATOPOIESIS

We have developed an experimental strategy for generating a murine model of human leukemias and other malignant neoplasms. Using a special preconditioning procedure, we reconstituted nonobese mice with severe combined immunodeficiency with cells from patients with T-cell acute lymphoblastic leukemia (T-ALL). Histopathologic and flow cytometric analyses of the mouse tissues showed that human T-ALL had spread from bone marrow to other major organs, including the meninges, resulting in a universally fatal leukemia. The progression and dissemination of the human leukemia in these mice mimicked the clinical features of patients, but the rate and the overall success of engraftment were substantially better than previously reported. In addition, these human leukemias were transferable to secondary murine recipients.

Using this murine model, we are investigating the efficacy of the newly developed anti-CD5, anti-CD7, and anti-CD38 immunotoxins conjugated with saporin for the treatment of human T-ALL. In addition, we will analyze the efficacy of l-alanosine, especially for cells from T-ALL patients with methylthioadenosine phosphorylase--deficient tumors. The current studies are part of our collaborative investigations with C. Carrera and A. Yu at the University of California, San Diego, on the use of 9p21 chromosome alterations in T-ALL therapy and with D. Flavell at the University of Southampton, England, on his new anti--T-ALL immunotoxins.

The ability to engraft leukemic cells or malignant tumors directly from patients into mice could be valuable for predicting the clinical course of the neoplastic disease, detecting residual tumor cells, and developing individualized therapeutic strategies. This model will also be used to study tumor biology, including the in vivo homing, engraftment, progression, and metastasis of the tumor and the disease heterogeneity of leukemia and other malignant neoplasms. In addition, the in vivo requirement of cytokines that support or suppress proliferation of leukemic and other malignant cells can be analyzed.

The preconditioning protocol used to prepare mice for injection with human leukemia cells included the formation of mixed chimerism. In addition, we recently found that one or more factors from fetal cord blood specifically promote the formation of colonies of T-ALL cells in a dose-dependent manner.

PUBLICATIONS

Batova, A., Diccianni, M.B., Nobori, T., Vu, T., Yu, J., Bridgeman, L., Yu, A.L. Frequent deletion in the methylthioadenosine phosphorylase gene in T-cell acute lymphoblastic leukemia: Strategies for enzyme-targeted therapy. Blood 88:3083, 1996.

Batova, A., Diccianni, M.B., Yu, J.C., Nobori, T., Link, M.P., Pullen, J., Yu, A.L. Frequent and selective methylation of p15 and deletion of both p15 and p16 in T-cell acute lymphoblastic leukemia (T-ALL). Cancer Res. 57:832, 1997.

Dialynas, D.P., Lee, M.J., Shao, L.-E., Tan, P.C., Yu, J. Phenotypic and functional characterization of a new human macrophage cell line, Klm, demonstrating immunophagocytic activity and signaling through HLA class II. Immunology 90:470, 1997.

Dialynas, D.P., Tan, P.C., Huhn, G.D., Yu, J. Characterization of a new human macrophage cell line 2MAC: 1. Expression of functional macrophage CD16/FcRIII; and tissue factor induction on ligation of HLA-DA. Cell. Immunol. 177:182, 1997.

Dialynas, D.P., Tan, P.C., Yu, J. Cytokine modulatable signalling through macrophage HLA class II: 1. IFN up-regulates the efficiency of Ca²⁺ mobilization in response to ligation of macrophage HLA-DP. J. Interferon Cytokine Res., in press.

Diccianni, M.B., Batova, A., Yu, J., Vu, T., Pullen, J., Amylon, M., Pollock, B., Yu, A.L. Shortened survival after relapse in T-cell acute lymphoblastic leukemia patients with p16/p15 deletions. Leuk. Res., in press.

Yu, J., Batova, A., Shao, L.E., Carrera, C.J., Yu, A.L. Presence of methylthioadenosine phosphorylase (MTAP) in hematopoietic stem/progenitor cells: Its therapeutic implication for MTAP (-) malignancies. Clin. Cancer Res. 3:433, 1997.

Yu, A.L., Chen, J., Diccianni, M.B., Batova, A., Yu, J. Exploitation of frequent p16 deletion in the treatment of T cell acute lymphoblastic leukemia. In: Molecular Biology of Hematopoiesis 5. Abraham, N.G., et al. (Eds.). Plenum, New York, 1996, p. 247.

Top of Page

---

Division of Biomathematics

James A. Koziol, Ph.D., Division Head

---

Components of the Scripps Neurologic Rating Scale: Responsiveness Over the Course of a Multiple Sclerosis Clinical Trial

J.A. Koziol, A. Lucero, J.C. Sipe, J.S. Romine, E. Beutler

A prerequisite for adoption of neurologic impairment scales as outcome measures in clinical studies is evaluation of the reliability, validity, and responsiveness of the scales. We recently completed a 2-year, randomized, placebo-controlled, double-blind crossover study of the efficacy of cladribine in the treatment of chronic progressive multiple sclerosis. In this trial, we used both the Kurtzke Extended Disability Status Scale and the Scripps Neurologic Rating Scale (SNRS) as outcome measures. On the basis of scores on these scales, we found that relative to placebo, cladribine appeared to influence favorably the course of chronic progressive multiple sclerosis. We have previously reported on certain psychometric properties of the SNRS instrument, in particular construct validity and reliability. We here address the responsiveness of the SNRS, that is, the ability of the instrument to detect significant changes in health status over time, within the context of our clinical trial.

At the start of the trial, patients were randomly assigned to two groups. The first group received cladribine for 12 months and then received placebo for the next 12 months. The second group received placebo for the first 12 months and then received cladribine for 12 months. The overall SNRS scores for the individual patients and the group estimates are shown in Figure 1. Most patients originally assigned to the group that received cladribine showed improvement in SNRS scores during the first 12 months of the trial. Analogously, most patients originally assigned to the group that received placebo showed improvement in SNRS scores during the second 12 months of the trial, when they received cladribine.

Tables 1 and 2 give the average effect sizes of the various components of the SNRS over the course of the clinical trial. A positive effect size indicates an improvement in that component; a negative effect size indicates a deterioration in that component.

Because of the crossover design of the clinical trial, one might anticipate that effect sizes in the group that received cladribine the first year would tend to be positive in months 1--12 (period 1) and show some deterioration in months 13--24 (period 2). This pattern seems to hold for most of the components. Exceptions are mentation and mood, which seems to decline rather severely in period 2 compared with the improvement in period 1, and DTRS (deep tendon reflexes, from various sites), which tends to show continued improvement in period 2. Conversely, one might expect some deterioration in effect sizes during period 1 and improvement during period 2 for the group that received placebo the first year and cladribine the second year. Again, this pattern seems borne out, with the notable exception of the components of mentation and mood and lower cranial nerves.

The effect sizes during period 2 of the group that initially received placebo might be compared with the effect sizes during period 1 of the group that initially received cladribine. In this comparison, the effect sizes for the group that initially received placebo are somewhat higher than those of the group that initially received cladribine. This finding reflects the fact that scores for the initial placebo group declined rather dramatically during period 1, so that baseline values for this group, which received cladribine during period 2, are smaller than the corresponding baseline values for the group that received cladribine during period 1. Finally, the direction of change is generally consistent across the individual components within each period for both treatment groups, and the effect sizes for the overall SNRS reflect this finding, tending to exceed those of the individual components in absolute magnitude.

The SNRS represents operationally a reduction of multivariate data--that is, assessment of function in different neurologic systems and assessment of patients' physical abilities--to a univariate quantity. In general, using a summary score has advantages, including the avoidance of multiplicity, clinical validity (so long as the individual components that make up the pooled index are each clinically important disease manifestations that have face validity), and improved sensitivity because of the expected reduction in measurement noise. Nevertheless, the fundamental question is whether the reduction of multivariate information to an overall univariate score is an adequate representation of the information available with the original assessments. In particular, it is not at all obvious how to determine the optimal weights or more generally the best method of combining the individual measures into a single index. We previously advocated using a multivariate statistical technique, principal components analysis, for this problem: this method identifies a linear combination of the individual measures that provides maximal spread (variation) among patients so as to maximize the information content of that combination.

The SNRS converts the neurologic examination into an ordinal impairment scale with an arbitrary weighting system. In particular, several potentially prognostic variables are combined into an overall risk score. It is therefore not surprising that various components of the SNRS appear to be more responsive than the overall summary measure and that other components seem to be less responsive. In particular, the various items that make up the SNRS are not expected to be homogeneous, because they measure different attributes of a complex clinical syndrome rather than different facets of the same attribute.

If one wished to develop a new assessment system, then the results presented here suggest items that are clearly unresponsive, marking the items as candidates for deletion: an excess of irrelevant items might well obscure the ability of a multiple-item index such as the SNRS to detect a change. In particular, the mentation and mood component evinces a pattern somewhat inconsistent with those of the other components, and the special category of bladder, bowel, or sexual dysfunction seems rather insensitive to change. Hence, these components are not likely to be useful as part of a multivariate or composite outcome measure; more detailed neuropsychologic assessment might be appropriate.

Top of Page

---

Division of Cellular Biology

James A. Hoch, Ph.D., Division Head

---

Signal Transduction in Cellular Differentiation

J.A. Hoch, M. Perego, C. Fabret, K. Kanamaru, M. Jiang

Formation of endospores in Bacillus subtilis is a model for understanding the mechanism of developmentally programmed gene expression. Several dozen genetically dispersed sporulation operons are regulated coordinately as temporal classes over the time required to complete the formation of spores. This complex developmental program is under the control of the spo0 genes, which control entry of the cell into sporulation. The protein Spo0A is the key master regulator of the initiation of developmental transcription. The activity of the protein is controlled by a reversible phosphorylation-dephosphorylation mechanism. In its unphosphorylated form, Spo0A is inactive. In its phosphorylated form, it is both an activator of the transcription of sporulation genes and a negative regulator of genes that prevent sporulation.

The key to understanding the initiation of sporulation is understanding the mechanism of Spo0A phosphorylation. We have shown that the pathway to Spo0A activation is a sequential series of phosphorylation reactions termed a multicomponent phosphorelay. The initial event in the phosphorelay is the activation of one of two kinases that phosphorylate the sporulation-specific response regulator Spo0F in response to environmental and metabolic signals. Spo0F acts as a secondary messenger, accumulating phosphate groups from developmentally activated kinases. Phosphorylated Spo0F (Spo0F~P) is the substrate for the Spo0B protein phosphotransferase that phosphorylates Spo0A. In this pathway, the signal transduction event is the activation (or perhaps the deinhibition) of the kinases to autophosphorylate. This step is followed by three sequential phosphotransferase reactions that produce Spo0A~P, the crucial transcription regulator for sporulation.

The flow of phosphate through the phosphorelay to Spo0A is highly controlled at several levels. Although the primary signals transduced by the two kinases, KinA and KinB, responsible for Spo0F phosphorylation remain obscure, genetic studies have revealed a series of genes unique for the activation of each kinase. The activity of each kinase is regulated by complex signal transduction pathways that respond to environmental, metabolic, and cell-cycle signals. KinB is mainly active during the exponential phase of growth. The kinase catalytic domains of this enzyme are fused to an amino domain with all the characteristics of a transport protein. We have discovered a new type of two-component kinase inhibitor protein for KinA that binds to the ATP-binding site of the KinA catalytic domain. This protein, KipI, may also bind to another protein, KipA, that regulates the activity of KipI on KinA. This inhibition system is regulated by the level of available nitrogen.

It is now clear that all the signals that affect sporulation cannot be processed by the kinases. Access to the phosphorelay for additional signals is provided by two families of phosphatases, which dephosphorylate either Spo0F~P or Spo0A~P and act to prevent sporulation. One of these phosphatases is controlled by the competence pathway, a finding that suggests that alternative physiologic processes induced at the end of exponential growth compete with sporulation by preventing activation of the major sporulation transcription factor.

Thus, the probability of initiating sporulation depends on the competition between kinases and phosphatases. The activity of the kinases and phosphatases on the phosphorelay acts as a signal integration circuit, allowing input from a variety of environmental, metabolic, and cell-cycle sources with a single output, the cellular level of Spo0A~P. By placing the developmental fate of the cell in the cellular level of a single phosphorylated transcription factor and coupling all signal inputs to this level, a large number of signal inputs can be accommodated with incremental effects that lead to a level of Spo0A~P that reflects the sum of all positive and negative factors.

PUBLICATIONS

Dartois, V., Djavakhishvili, T., Hoch, J.A. KapB is a lipoprotein required for KinB signal transduction and activation of the phosphorelay to sporulation in Bacillus subtilis. Mol. Microbiol., in press.

Dartois, V., Liu, J., Hoch, J.A. Alterations in the flow of one-carbon units affect KinB-dependent sporulation in Bacillus subtilis. Mol. Microbiol. 25:39, 1997.

Manson, M.D., Armitage, J.P., Hoch, J.A,. Macnab, R.M. Bacterial locomotion and signal transduction. J. Bacteriol., in press.

Wang, L., Grau, R., Perego, M., Hoch, J.A. A novel histidine kinase inhibitor regulating development in Bacillus subtilis. Genes Dev. 11:2569, 1997.

Top of Page

---

Molecular Recognition in Signal Transduction

J.A. Hoch, Y.-L. Tzeng, X.Z. Zhou, Madhusudan, K.I. Varughese

Two-component signal transduction systems are the major mechanism for environmental signal recognition in bacteria and the induction of virulence in pathogens. Microorganisms such as Escherichia coli and Bacillus subtilis contain about 50 different pairs of kinase-response regulator two-component systems that are highly related in sequence and structure and that process a variety of signals. Each signal activates a specific kinase that phosphorylates only its own cognate response regulator despite the presence of many response regulators of identical structure. This circuitry is kept intact by the exquisite recognition properties of the kinase for its response regulator. We are using a multifaceted approach of mutational and structural analysis to study the molecular basis of recognition between the kinases and response regulators.

The Spo0F response regulator is the key intermediate in the signal transduction phosphorelay. It interacts with two different kinases, at least two phosphatases, and the Spo0B response regulator phosphotransferase. The complete structure of Spo0F has been determined with both crystallography and nuclear magnetic resonance. This structure is the basis for studies that use the alanine-scanning technique to determine the roles of amino acid side chains in molecular recognition. All the surface residues of the Spo0F protein were changed individually to alanine, and we determined how deleting the side chain past the ß carbon affected the interaction of Spo0F with kinases, phosphatases, and Spo0B. These studies revealed that residues important for recognition by all these proteins are clustered around the aspartate triad active site of phosphorylation. Furthermore, individual residues could be identified that are specific for interaction with one or more of the kinases, phosphatases, or Spo0B.

Spo0B is a unique phosphotransferase capable of transferring phosphate from one response regulator to another with specificity for both donor and recipient. The alanine-scanning studies of the donor Spo0F~P protein are being complemented by structural analyses of its primary phosphate recipient, Spo0B. Mechanistic studies of the Spo0B phosphotransferase reaction showed that a histidine-phosphate is an obligatory intermediate in the reaction, and the identity of this histidine has been established. Spo0B has been crystallized with two molecules in the asymmetric unit. Diffraction data to a resolution of 2.6 Å have been acquired for a native crystal and for five heavy-atom derivatives. The interface of the dimer is formed by a four-helix bundle, two helices from each monomer.

PUBLICATIONS

Madhusudan, Zapf, J.W., Hoch, J.A., Whiteley, J.M. A response regulatory protein with the site of phosphorylation blocked by an arginine interaction: crystal strucure of Spo0F from Bacillus subtilis. Biochemistry 36:12739, 1997.

Madhusudan, Zapf, J.W., Whiteley, J.M., Hoch, J.A., Xuong, N.H., Varughese, K.I. Crystal structure of a phosphatase-resistant mutant of sporulation response regulator Spo0F from Bacillus subtilis. Structure 4:679, 1996.

Tzeng, Y.-L., Hoch, J.A. Molecular recognition in signal transduction: The interaction surfaces of the Spo0F response regulator with its cognate phosphorelay proteins revealed by alanine scanning mutagenesis. J. Mol. Biol. 272:200, 1997.

Zhou, X.Z., Madhusudan, Whiteley, J.M., Hoch, J.A., Varughese, K.I. Purification and preliminary crystallographic studies on the sporulation response regulatory phosphotransferase protein, Spo0B, from Bacillus subtilis. Proteins 27:5597, 1997.

Top of Page

---

Molecular Dynamics of Response Regulators

J.A. Hoch, Y.-L. Tzeng, V. Feher,^* J. Cavanagh^*

^* New York State Department of Health, Albany, NY

Spo0F belongs to a large class of proteins, the response regulators, that participate in many different bacterial signal transduction pathways. Response regulators have diverse functions, although common to all is a regulatory domain of approximately 120 residues that becomes phosphorylated at a conserved aspartate residue in a magnesium-dependent reaction with a histidine autokinase. We have solved the high-resolution three-dimensional solution structure of Spo0F. The overall fold of Spo0F consists of five -helices surrounding five parallel ß-strands, forming a hydrophobic sheet, as for other response regulators. The fold brings three aspartic acid residues into proximity to form the binding pocket to accept the phosphoryl group. From the structural studies, we have been able to determine differences in the orientation of secondary structure elements in the putative recognition surfaces and the relative charge distribution of residues surrounding the site of phosphorylation.

Additionally, we have been studying the backbone dynamics of the Spo0F protein. In conjunction with alanine-scanning mutagenesis studies, our dynamics studies have enabled us to propose a model in which communication of information through the core of the protein, between buried and surface-bound residues, is responsible for the dissociation of the cognate kinase of the protein after phosphorylation. The helix-4--strand-5 loop contains a primary recognition site for the kinase involving residues Tyr84, Glu86, and Leu87. The structural and dynamics studies show that this region has a propensity for multiple conformers. We defined a region on the protein, including helix-4, part of helix-3, ß-strand-5, and the helix-4--strand-5 loop, that moves in a dynamically concerted fashion, driven by the motion of the imidazole ring of His101.

We propose that the imidazole ring moves from a buried position under the helix-4--strand-5 loop to a more solvent, exposed position in response to a conformational change in the aspartic acid binding pocket upon phosphorylation. Movement of the ring disrupts packing interactions, a condition that alters the topology of the kinase recognition site, thereby causing the kinase to dissociate. This model is one of the first connections between protein dynamics and their specific biological function.

PUBLICATIONS

Feher, V.A., Zapf, J.W., Hoch, J.A., Whiteley, J.M., McIntosh, L.P., Rance, M., Skelton, N.J., Dahlquist, F.W., Cavanagh, J. High-resolution solution structure and dynamics of Bacillus subtilis response regulator, Spo0F: Implications for phosphorylation and molecular recognition. Biochemistry 36:10015, 1997.

Top of Page

---

Protein Aspartate Phosphatases in Bacterial Development

M. Perego

Recognition of the diverse signals involved in the initiation of sporulation is carried out by proteins that alter the level of phosphorylation of the Spo0A transcription factor. The pathway to Spo0A activation is a sequential series of phosphorylation and phosphotransfer reactions that occur in the multicomponent phosphorelay. Initially, it was thought that positive sporulation signals activated the kinases to phosphorylate Spo0F and that negative signals regulated the phosphatase activities of the input kinases and the Spo0A~P phosphatase Spo0E. However, it is now clear that signal interpretation for such a complex cellular process involves many more proteins.

Access to the phosphorelay for additional signals is provided by a recently discovered family of response regulator aspartyl phosphate phosphatases. The best characterized proteins of this family, RapA and RapB, specifically dephosphorylate the Spo0F~P intermediate component of the phosphorelay. Because the Spo0B phosphotransferase activity is reversible, dephosphorylation of Spo0F~P results in dephosphorylation of Spo0A~P, thus inhibiting sporulation.

The finding that rapA and rapB loci are regulated differently gives a key to their function. The rapA locus is induced at the end of exponential growth by the competence pathway and is repressed by Spo0A. The rapB locus is under control of vegetative growth signals. Different environmental and physiologic conditions are required for the induction of each locus and, therefore, for each regulator to prevent sporulation. The phosphatase activity of these proteins is also differentially regulated by effector molecules.

Phosphatase activity of RapA is regulated by a cotranscribed gene, phrA, that encodes a 44 amino acid protein. This protein is proteolytically processed from an inactive preinhibitor peptide to an active peptide. In fact, the 44 amino acid preinhibitor is first cleaved by the protein export apparatus to a putative proinhibitor of 19--20 amino acids. This putative proinhibitor is further processed to the active inhibitor peptide, which is internalized by the oligopeptide transport system. This export-import circuit for production of the PhrA pentapeptide inhibitor is a mechanism for timing the RapA phosphatase activity. The PhrA pentapeptide, once internalized, directly and specifically inhibits the RapA phosphatase activity (Fig. 1). The specificity of PhrA for its target protein RapA is determined by the amino acid sequence of the peptide.

The Rap family of phosphatases consists of 11 chromosomal genes. The Rap proteins show 40--50% identities in amino acid residues, and five of the proteins are associated with a PhrA-like protein. Preliminary studies indicate that not all the Rap phosphatases affect sporulation, but they most likely have roles in processes, other than sporulation, that are controlled by two-component signal transduction systems.

PUBLICATIONS

Glaser, P., Sharpe, M.E., Raether, B., Perego, M., Ohlsen, K., Errington, J. Dynamic, mitotic-like behavior of a bacterial protein required for accurate chromosome partitioning. Genes Dev. 11:1160, 1997.

Perego, M. A peptide export-import control circuit modulating bacterial development regulates protein phosphatases of the phosphorelay. Proc. Natl. Acad. Sci. U.S.A. 94:8612, 1997.

Reizer, J., Reizer, A., Perego, M., Saier, M.H., Jr. Characterization of a family of bacterial response regulator aspartyl-phosphate (Rap) phosphatases. Microb. Comp. Genomics 2:103, 1997.

Top of Page

---

Structure-Function Studies of a Global Regulator of Gene Expression

K. Xu, M.A. Strauch

AbrB is a global regulator of the Bacillus subtilis genes that are expressed in response to nutrient limitation and other environmental stresses. AbrB exerts its effects at the transcriptional level by binding to DNA regions present on numerous genes responsible for functions necessary for the assembly of alternative metabolic and developmental pathways. We have detected and characterized, both genetically and biochemically, more than 40 AbrB-binding regions in the genome, but a variety of additional experimental observations indicate that many more sites of AbrB action must also exist. Obviously, AbrB occupies a central position in coordinating various aspects of cellular physiology.

Our previous work led to the hypothesis that AbrB achieves binding specificity through recognition of a subtle three-dimensional DNA structure produced by members of a finite subset of different base sequences. The results of chemical footprinting to define contact points between DNA and AbrB suggest that the protein has a structural flexibility that allows it to accommodate the heterogeneity of base sequences present in the recognition subset. The ability to productively interact with a variety of base sequences allows AbrB to achieve regulatory effects in a variety of contexts (e.g., promoter structure, superposition of other regulatory sites) and ensures that different relative degrees of desired control can be economically brought about by a single protein.

Although we are continuing molecular genetic investigations on the physiologic function of AbrB, much of our recent effort has been directed at elucidation of the protein's structure. Originally, we thought that AbrB was composed of homohexameric subunits. We now have evidence that it may actually be an unusually shaped tetramer of identical 10.5-kD polypeptides. Analysis of mutants and truncated parts of AbrB has indicated that the N-terminal half of the monomer contains the DNA-binding determinants and that the C-terminal half is responsible for multimerization. The N-terminal domain defines a new class of DNA-binding motif that is also present in two recently discovered B. subtilis regulatory proteins. Cross-linked dimers of the domain have full DNA-binding activity and specificity.

To investigate the biophysical basis of AbrB's unique DNA-binding properties, we are using nuclear magnetic resonance to determine the structure of the N-terminal domain. This work is being done in collaboration with J. Cavanagh, New York State Department of Health. Initial secondary structure assignments of monomeric domains show an arrangement of four ß-sheets and a single -helix that is distinct from previously studied motifs present in DNA-binding proteins. The full three-dimensional structure should soon be solved, and we have begun preparation and study of the high-affinity dimeric form and its DNA conjugate.

PUBLICATIONS

Strauch, M.A. Hoch, J.A. Spore, sporulation. In: The Encyclopedia of Molecular Biology. Creighton, T.E. (Ed.). Wiley, New York, in press.

Xu, K., Strauch, M.A. Identification, sequence, and expression of the gene encoding -glutamyltranspeptidase in Bacillus subtilis. J. Bacteriol. 178:4319, 1996.

Top of Page

---

Bacterial Response to Environmental Stimuli

K.I. Varughese, J.M. Whiteley, Madhusudan, J. Zapf, J.A. Hoch

Gene activation in response to environmental stimuli is characteristic of all life forms, from bacteria to humans. Bacteria respond to environmental stress by activating transcription of genes that code for products that enable the bacteria to adapt to the new environment. Common stresses include osmotic shock; starvation of nitrogen, phosphate, and various carbon sources; and responses to oxygen concentration. These and common bacterial processes such as motility and chemotaxis; secretion of enzymes; transport of hexoses, dicarboxylates, and tricarboxylates; and the capacity for virulence of some human and plant pathogens involve the induction of specific metabolic pathways and biochemical systems so that the organisms can cope with the new environmental status.

In all these systems, a signal induces a transcriptional change via a so-called two-component regulatory switch. These switches consist of a histidine protein kinase and a response regulator. The signal activates the kinase to autophosphorylate, and then transfer of this phosphoryl group to the amino-terminal domain of the regulatory protein activates the transcription of genes. The amino-terminal domains of the phosphoreceptor proteins are highly conserved. The two-component system was initially thought to occur only in bacteria; however, similar systems have been found in species of yeast and certain other eukaryotes.

Bacteria such as Bacillus subtilis respond to environmental stress by forming spores, an action that is regulated by a somewhat more complex phosphorelay system. This system contains four major components-- a kinase (KinA), a receptor (Spo0F), a transfer protein (Spo0B), and a transcription activator (Spo0A)--and some additional phosphatase-regulating features. Nevertheless, KinA, Spo0F, and Spo0A still retain the sequence homology of two-component systems. The genes for each of these proteins have been expressed in Escherichia coli, and the proteins have been isolated and characterized by chromatographic procedures. Both a Y13S mutant and the wild-type Spo0F have been crystallized, and their structures have been analyzed at high resolution (Fig. 1).

The three-dimensional structure of Spo0F is similar to that of CheY, a response regulatory protein in the chemotaxis pathway. Despite structural and sequence similarities between these two response regulators, Spo0F~P was nearly three orders of magnitude more stable than CheY~P to phosphate hydrolysis. Response regulator proteins are activated by phosphorylation, and phosphate hydrolysis or protein dephosphorylation is thought to be due to an autophosphatase activity that ensures that these proteins do not remain permanently activated by phosphorylation. The "stability" or the magnitude of the autophosphatase activity of Spo0F~P and CheY~P is appropriately matched to suit their respective biological roles. Sporulation is a slow process that occurs over an hour or hours, whereas chemotaxis is a rapid response that occurs in a matter of seconds.

Subtle differences between the structures of Spo0F and CheY in the regions adjacent to the phosphorylation active site may account for the divergent phosphorylation properties of these two response regulators. A possible model for Spo0F~P designed by using computer graphics highlighted the importance of nonconserved residues located adjacent to the phosphorylation active site. This model showed that the side chains of Lys56 and Gln12 could interact with the phosphoryl oxygens and stabilize the phosphoryl group. Indeed, mutation of Lys56 to Asn56, the residue occurring at the equivalent position in CheY, increased the autophosphatase activity 23-fold. Response regulators naturally containing an asparagine residue at the position equivalent to 56 in Spo0F also hydrolyze phosphate in a matter of minutes. These results suggest that the type of residue at the position equivalent to 56 in Spo0F "tunes" the response regulator autophosphatase activity to suit a particular biological role.

In addition, we have analyzed the structure of Spo0B, a unique phosphotransferase protein (Fig. 2). The protein crystallized as a dimer in the asymmetric unit, and the crystals diffracted to 2.6-Å resolution. The structure was solved by multiple isomorphous method with five heavy atom derivatives and refined to an R-factor of 21.7%. Little is known about the structure of histidine kinases, and the structure of Spo0B will serve as a possible model for these enzymes.

We have crystallized the DNA-binding domain of Spo0A complexed with an oligonucleotide duplex containing the recognition site. Earlier information obtained by chromatographic techniques proved that the Spo0A protein could be cleaved at a midpoint site by trypsin to yield two peptides, one of which could still be phosphorylated (N-terminal) and a second (C-terminal) that retained DNA-binding affinity for the abrB promoter region. Because of solubility and stability factors, the B. subtilis system is an excellent model for examining all bacterial regulatory controls. This system could have particular value in understanding other regulatory pathways, for example, those of the bacterial virulence factors, in which external control might be beneficial in preventing infection.

PUBLICATIONS

Zhou, Z.X., Madhusudan, Whiteley, J.M., Hoch, J.A., Varughese, K.I. Purification and preliminary crystallographic studies on sporulation response regulatory phosphotransferase protein, Spo0B, from Bacillus subtilis. Proteins 27:597, 1997.

Top of Page

---

Structural Studies of the von Willebrand Factor A1 Domain in Complex With the Function-Blocking NMC-4 Fab and Modulators of von Willebrand Factor Function

K.I. Varughese, R. Celikel, S. Vasudevan, Z.M. Ruggeri

Normal platelet function needed to arrest hemorrhage requires binding of the membrane glycoprotein (GP) Ib, a component of the GP Ib--IX--V receptor complex, to the A1 domain of von Willebrand factor. This binding is the sole adhesive interaction capable of mediating the initial tethering of circulating platelets to thrombogenic surfaces exposed to rapid blood flow. The phenotypes of von Willebrand disease, the most prevalent congenital bleeding disorder in humans, and of the less common Bernard-Soulier syndrome clearly show the physiologic hemostatic function mediated by GP Ib and the A1 domain. Their interaction, however, may contribute to acute thrombotic occlusion of atherosclerotic stenosed arteries, causing catastrophic organ damage, as in myocardial infarction in patients with coronary artery disease.

We obtained crystals of the active A1 domain in complex with the Fab fragment of NMC-4, a monoclonal antibody that binds to the domain with high affinity, and used x-ray diffraction data to generate a structural model to 2.2-Å resolution (Fig. 1). The complex is formed through the interactions of the helix 4 of the A1 domain with loops L3, H2, and H3 of the Fab complementarity-determining regions.

One of the most intriguing aspects of the interaction between the A1 domain and GP Ib is the rapid rate of bond formation with temporary high resistance to tensile stress and eventual dissociation. The current crystal structure suggests a possible mechanism for this biological process, based on the observation that the A1 domain binds to NMC-4 by extending the region N-terminal to helix 4, formed by residues 634 to 643, into the antigen-binding site (Fig. 1). The most interesting feature of the peptide conformation at this region is the formation of a type I ß-turn formed by residues 631--634 with Arg⁶³² and Asn⁶³³ at the corners of the bend. There is a strong (4-1) internal hydrogen bond between the carbonyl of Ser⁶³¹ and the amide of Phe⁶³⁴, with an N-O distance of 2.74 Å. Formation of this ß-bend places the side chain of Arg⁶³² in a suitable orientation for interacting with NMC-4. Interestingly, the residues 628--631 located toward the N-terminal of the ß-bend assume a tighter 3₁₀ helical conformation that, in turn, places the side chains of Arg⁶²⁹ and Gln⁶²⁸ in a proper location for interaction with NMC-4.

Thus, it appears that this region of the peptide chain can adopt two conformations, possibly one for binding and the other resulting in dissociation. The side chain of Phe⁶³⁴ is located in the interior of the protein and may serve as an anchoring point. In the proposed second conformation, the helix would extend toward the N-terminal side of Phe⁶³⁴, eliminating the ß-turn and rotating the side chains of residues 628, 629, and 632 out of place for binding. On the basis of the structure and the current knowledge of mutations that lead to loss of bonding, we predict that the binding surface of GP Ib extends from 4 to ß3 in the A1 domain.

Botrocetin, one of the modulators of the function of von Willebrand factor in vitro, has been purified and crystallized, and diffraction data to a resolution of 2.7 Å have been collected. The crystal of botrocetin belongs to orthorhombic space group P2₁2₁2₁, and each asymmetric unit has two molecules. Searches for heavy-atom derivatives to enable crystal structure solution are in progress.

PUBLICATIONS

Celikel, R., Madhusudan, Varughese, K.I., Shima, M., Yoshioka, A., Ware, J., Ruggeri, Z.M. Crystal structure of NMC-4 Fab anti-von Willebrand factor A1 domain. Blood Cells Mol. Dis. 23:123, 1997.

Top of Page

---

Division of Experimental Hemostasis and Thrombosis

Zaverio M. Ruggeri, M.D., Division Head

Cardiovascular disease and stroke remain the major health concerns in adults in developed countries, including the United States. Investigators in the Division of Experimental Hemostasis and Thrombosis are conducting basic research to address the main pathogenetic mechanisms responsible for arterial and venous thrombosis, laying the foundation for novel and more efficient therapeutic approaches. Efforts are also devoted to further the understanding of endothelial cell function in normal and pathologic conditions and to clarify the role of the immune system in determining the success or failure of organ transplantation. Thus, several themes relevant to vascular biology are being developed, by using all the advanced tools of cell, molecular, and structural biology.

---

Structure of Adhesive Proteins Mediating Platelet Function

K.I. Varughese, R. Celikel, S. Vasudevan, R.A. McClintock, J.R. Roberts, Z.M. Ruggeri

Knowledge of the detailed three-dimensional structure of adhesive proteins that mediate formation of platelet thrombi is indispensable to explain in detail the mechanisms responsible for normal hemostasis and pathologic thrombosis. The crystallographic work performed in the Division of Experimental Hemostasis and Thrombosis has been the result of a close collaboration with K.I. Varughese that started when he was still at the University of California, San Diego, before he moved to TSRI. Now, a team of investigators is working on fundamental structural and biomechanical issues of platelet adhesion. Two projects have been completed, setting the bases for elucidating the molecular mechanisms that support the initiation of formation of platelet thrombi through the interaction between the adhesive protein von Willebrand factor and the platelet receptor glycoprotein Ib.

Top of Page

---

Crystal Structure of an Fab Fragment of an Antibody to von Willebrand Factor A1 Domain

R. Celikel, Madhusudan, K.I. Varughese, M. Shima,^* A. Yoshioka,^* J. Ware, Z.M. Ruggeri

^* Nara Medical College, Nara, Japan

Glycoprotein (GP) Ib, a component of the GP Ib-IX-V receptor complex, and von Willebrand factor (vWF) have a unique role in platelet function. Their interaction is necessary to initiate platelet deposition at sites of vascular injury when blood flow is elevated, as in arterioles during hemostasis, or in stenosed atherosclerotic coronary arteries. A fast rate of interaction and high resistance to tensile stress are the key properties that allow the bond between GP Ib and the A1 domain of vWF that supports the initial tethering of rapidly flowing platelets to thrombogenic surfaces, thus reducing drastically the velocity of the platelets relative to the vessel wall. Ensuing activation then results in interaction of the integrin _IIbß₃ (GP IIb-IIIa complex) with the Arg-Gly-Asp sequence in the carboxyl-terminal C1 domain of vWF. This second bond makes adhesion irreversible and favors the accrual of additional platelets.

As a step toward elucidating the molecular bases of the function of vWF and GP Ib, we have solved the crystal structure of the Fab fragment of NMC-4, a monoclonal antibody that binds to the A1 domain of vWF with high affinity, blocking interaction with GP Ib. Two aspartic acid and three tyrosine residues in complementarity-determining regions 1 and 3 of the heavy chain exhibited a spatial orientation suggestive of a dominant role in establishing contact with the antigen. A cluster of aspartic acid and tyrosine residues occurs also in a region of the GP Ib amino-terminal domain known to be critically involved in binding vWF. These results define an antibody structure that may reflect the spatial orientation of corresponding GP Ib residues involved in supporting contact with the A1 domain of vWF.

Top of Page

---

Crystal Structure of the von Willebrand Factor A1 Domain in Complex With the Function-Blocking NMC-4 Fab

R. Celikel, K.I. Varughese, Madhusudan, A. Yoshioka,^* J. Ware, Z.M. Ruggeri

^* Nara Medical College, Nara, Japan

Determining the contact interface between the A1 domain of von Willebrand factor (vWF) and the amino-terminal region of glycoprotein Ib is essential to elucidate the structural basis of an interaction that plays a key role in normal platelet function and may be a major pathogenetic factor in arterial thrombosis. We have crystallized the complex consisting of the recombinant vWF A1 domain and an Fab fragment of NMC-4, a monoclonal antibody that binds to the A1 domain, and have solved the structure of the complex at 2.2-Å resolution by using a combination of molecular replacement and single isomorphous replacement techniques.

The vWF A1 domain has an /ß Rossmann fold with a six-strand central ß-sheet sandwiched between two sets of three -helices, one helix on each side. NMC-4 interacts with the A1 domain, bridging to one of these helices, 4, which is separated by a break that occurs after the first turn into two segments, a and b, that have different longitudinal axes. The short segment 4a is a 3₁₀ helix and comprises residues 627--631 of mature vWF. Residues 631--634 form a type I ß-turn, and residues 634--643 form 4b. The entire segment of 4a, the ß-turn, and the N-terminal part of 4b wedge into the antigen-binding site of NMC-4 created primarily by loops L3, H2, and H3 of the complementarity-determining regions assembled as a tripod.

Work is still in progress on this structure to elucidate aspects that may be relevant to the interaction of the vWF A1 domain with platelet glycoprotein Ib. In particular, targeted mutagenesis studies are being used to confirm the functional and structural roles of specific residues with key roles in the formation of the complex of the A1 domain and NMC-4.

Top of Page

---

Synergism of Adhesive Mechanisms for Development of Platelet Thrombi in Flowing Blood

E. Saldívar, B. Savage, J.A. Dent, Z.M. Ruggeri

Thrombi consisting of aggregated platelets support hemostasis but may occlude atherosclerotic arteries, causing ischemic damage to vital organs. We have developed a new approach, based on laser confocal microscopy, to measure formation of thrombus in real time during blood flow in experimental chambers. We are obtaining three-dimensional volumetric information and are elucidating the effects of blood flow on the mechanisms of platelet aggregation.

When blood flows slowly, at the venous shear rate of 100 s^-1, the integrin _IIbß₃, but not glycoprotein Ib, is essential for attachment of platelets to one another, and fibrinogen is the most efficient bridging ligand. In contrast, when blood flows rapidly, at the arteriolar shear rate of 1500 s^-1, formation of thrombus is totally dependent on von Willebrand factor and its two receptors, glycoprotein Ib and _IIbß₃. At intermediate velocity, with a wall shear rate of 300 s^-1, the size of a thrombus is reduced by more than 50% if the function of glycoprotein Ib is blocked. These results indicate that synergistic adhesive mechanisms support platelet aggregation and determine the rate of thrombus growth, acting as continuous variables dependent on blood flow conditions. Ongoing work in this area may substantiate the concept that optimal antithrombotic therapy depends on the combination of different inhibitors, specifically targeting the distinct adhesive interactions that are responsible for the biomechanical attributes of a developing thrombus.

Top of Page

---

Genetic Analysis of Platelet Receptor Expression

B. Zieger,^* H. Fujita, J. Ware

^* Universitäts Kinderklinik, Freiburg, Germany

Our long-term objective is to gain an understanding of the molecular events that control formation of megakaryocytes and production of platelets. Acquiring basic information on these two processes has been slow because of a lack of cell lines that mimic the normal progenitor cell of megakaryocytes and an inability to experimentally manipulate the formation of platelets. Clearly, thrombopoietin is essential for the differentiation of the pluripotent stem cell to a polyploid megakaryocyte, but the late events of the process, specifically those controlling the release of platelets, are still poorly understood. However, the expression of megakaryocyte- or platelet-specific antigens is one strictly unique aspect of the formation of megakaryocytes that can be examined. Such experiments define the molecular events and factors that regulate the commitment of cells to the megakaryocytic lineage and, ultimately, the release of platelets into the bloodstream.

With regard to this last point, the glycoprotein (GP) Ib-IX-V complex is an essential multisubunit platelet membrane receptor critical for hemostasis. This complex has been implicated in normal platelet release and structure as shown by the congenital absence of the complex and the release of abnormal or "giant" platelets, a condition referred to as the Bernard-Soulier syndrome.

The genes that encode GP Ib and the smaller ß-subunit of GP Ib, GP Ibß, both contain promoter elements necessary for megakaryocytic gene expression. GP Ib is composed of two subunits, Ib and Ibß, each synthesized from separate genes. The 206 amino acid precursor of GP Ibß is synthesized from a 1.0-kb mRNA expressed by megakaryocytes and was originally characterized on the basis of cDNA clones of mRNA from HEL cells, a human erythroleukemic cell line that has megakaryocyte-like properties. The cell line CHRF-288-11 also exhibits megakaryocyte-like properties, but these cells synthesize two related GP Ibß mRNA species of 3.5 and 1.0 kb.

We used cDNA cloning to determine the origin of the 3.5-kb transcript and its relationship to the 1.0-kb GP Ibß mRNA found in megakaryocytes, platelets, and HEL cells. The results showed that the longer transcript results from a nonconsensus polyadenylation recognition sequence, ^5´AACAAT^3´, within a separate gene located 250 nucleotides 5´ to the transcription start site of the platelet gene for GP Ibß. In the absence of normal polyadenylation, the more 5´ gene uses the polyadenylation site within its 3´ neighbor, the platelet GP Ibß gene. This newly determined 5´ gene contains an open reading frame that encodes 369 amino acids with a high degree of sequence similarity to an expanding family of GTP-binding proteins referred to as septins.

Septins were originally described as essential for yeast budding, and their functional relevance in higher eukaryotes has been examined in only one case: a Drosophila protein designated PNUT that is embryonically lethal in the homozygous state and produces an overaccumulation of embryonic polyploid cells. However, a unifying functional property of septin proteins is their involvement in cytokinesis. The spatial proximity of the newly detected human septin gene and the human platelet GP Ibß gene is also present within the homologous mouse locus, with a remarkable conservation of gene sequence and structure. Future studies will examine the structural and functional roles of human septins and the ability of these proteins to influence the expression of the GP Ib-IX-V platelet receptor.

Top of Page

---

Mechanisms of Tumor Cell Interaction With Vascular Cells

B. Felding-Habermann, T.J. Kunicki, R. Billetta,^* S. Ferrone,^** G.H. Ryu,^*** Z.M. Ruggeri

^* University of California, San Diego, CA
^** New York Medical College, Valhalla, NY
^*** Bureau of Medical and Radiation Health, Seoul, Korea

Hematogenous metastasis requires attachment of tumor cells to the vessel wall during blood flow. Using an in vitro perfusion system, we found that tumor cell arrest involved interaction between tumor cells and activated platelets via ß₃ integrins and cross-linking plasma proteins. Recognition between ß₃ integrins and their ligands involves the Arg-Gly-Asp (RGD) ligand motif. Therefore, we analyzed critical flanking sequences surrounding the RGD site and the conformational constraints of the RGD domain that are crucial in determining the ligand specificity for tumor cell integrin _vß₃ and platelet integrin _IIbß₃. This information will be important for the development of agents that inhibit interactions between tumor cells and platelets.

Another receptor of potential importance in adhesion of tumor cells to vascular cells is intracellular adhesion molecule-1. We found a clear correlation between expression of this molecule and tumor progression in melanoma patients, indicating a role for the molecule in adhesion of melanoma cells during the metastatic cascade. Because of the clinical significance of the expression of integrin _vß₃ and intracellular adhesion molecule-1 in melanoma lesions, these receptors are targets for future studies on melanoma cell arrest and extravasation. For this purpose, we have developed an experimental system that enables us to analyze both tumor cell attachment to the endothelium and transendothelial migration of tumor cells during blood flow.

Critical parameters in this in vitro system are stimulation of the endothelium and its presentation of a shear-resistant three-dimensional cushion that enable us to analyze transmigrated cells. We will use three-dimensional reconstruction of confocal images acquired during the perfusion to detect and quantify attached and penetrated cells. This technique will enable us to study vascular cells and their receptors and ligands that are involved in adhesive and invasive tumor cell interaction with the vessel wall and that affect the rate of hematogenous tumor metastasis.

PUBLICATIONS

Celikel, R., Madhusudan, Varughese, K.I., Shima, M., Yoshioka, A., Ware, J., Ruggeri, Z.M. Crystal structure of NMC-4 Fab anti-von Willebrand factor A1 domain. Blood Cells Mol. Dis. 23:123, 1997.

Federici, A.B., Mannucci, P.M., Stabile, F., Canciani, M.T., Di Rocco, N., Miyata, S., Ware, J., Ruggeri, Z.M. A type 2B von Willebrand disease mutation (Ile⁵⁴⁶*Val) associated with an unusual phenotype. Thromb. Haemost. 78:1132, 1997.

Hayashi, T., Ware, J., Niiya, K., Sakuragawa, N. Isolated recombinant domain of von Willebrand factor displaying increased sensitivity to ristocetin Am. J. Hematol. 52:248, 1996.

Kunicki, T.J, Annis, D.S, Felding-Habermann, B. Molecular determinants of Arg-Gly-Asp ligand specificity for ß₃ integrins. J. Biol. Chem. 272:4103, 1997.

Lanza, P., Felding-Habermann, B., Ruggeri, Z.M, Zanetti, M., Billetta, R. RGD conformationally-constrained in an antibody loop interacts selectively with the alpha_vß₃ integrin on tumor cells. Blood Cells Mol. Dis. 23:230, 1997.

Natali, P.G., Hamby, C.V., Felding-Habermann, B., Liang, B., Nicotra, M.R., Di Filippo, F., Giannarelli, D., Temponi, M., Ferrone, S. Clinical significance of alpha_vß₃ integrin and intercellular adhesion molecule-1 expression in cutaneous malignant melanoma lesions. Cancer Res. 57:1554, 1997.

Pareti, F.I., Cattaneo, M., Carpinelli, L., Zighetti, M.L., Bressi, C., Mannucci, P.M., Ruggeri, Z.M. Evaluation of the abnormal platelet function in von Willebrand disease by the blood filtration test. Thromb. Haemost. 75:460, 1996.

Ware, J., Hashimoto, Y., Zieger, B., Russell, A. Controlling elements of platelet glycoprotein Ibalpha expression. C. R. Acad. Sci. 319:811, 1996.

Zieger, B., Hashimoto, Y., Ware., J. Alternative expression of platelet glycoprotein Ibß mRNA from an adjacent 5´ gene with an imperfect polyadenylation signal sequence. J. Clin. Invest. 99:520, 1997.

Zieger, B., Ware, J. Cloning and deduced amino acid sequence of human nicotinamide nucleotide transhydrogenase. DNA Seq. 7:369, 1997.

Top of Page

---

Microvascular Responses to Experimental Focal Cerebral Ischemia

G.J. del Zoppo, B.R. Copeland,^* T. Abumiya, M. Tagaya, S. Wagner, V. Quaranta, T.E. Hugli, J.H. Garcia^**

^* Scripps Clinic, La Jolla, CA
^** Henry Ford Hospital, Detroit, MI

The rapid appearance of signs and symptoms of focal cerebral ischemia in patients and the early appearance of neuronal injury in primate and rodent models of experimental stroke suggest that neurons are selectively vulnerable to cessation of blood flow. Work in this laboratory indicates that alterations in structural and signaling relationships within the microvascular complex, which includes the endothelium, the basal lamina or extracellular matrix, and astrocyte end-feet, are relevant to the development of neuronal dysfunction. Recent work showed that microvascular responses to cessation of blood flow (and reperfusion) are equally rapid in neuronal injury in the corpus striatum of larger primate species, including humans.

Differences between species and models in the neuronal response to focal cerebral ischemia may be due to differences in the expression of microvascular adhesion receptors. Using in situ incorporation of dUTP to detect cells with DNA damage, we found significant topographic differences between rodent and nonhuman primate models of occlusion and reperfusion of the middle cerebral artery. In both models, cells that incorporate dUTP could be detected in areas of focal ischemia. In the rodent model, the affected areas expanded; in the nonhuman primate model, the areas remained static from the onset of ischemia. In the primate model, 80% of dUTP-labeled cells expressed MAP2 antigen 2 hours after occlusion of the middle cerebral artery, and 1.8% were associated with microvessels 24 hours after reperfusion of the artery. The number and density of dUTP-labeled cells increased with time from occlusion of the artery and were dramatically different between the species but became nearly identical by 24 hours after reperfusion.

Distinct temporal, topographic, and density patterns in the corpus striatum of both species suggest differences in cellular (neuronal) injury and repair mechanisms and microvascular flow dynamics. The early development of neuronal injury parallels the loss of microvascular expression of integrins, which may be relevant to intercellular signaling between cells within the microvasculature and neurons in response to ischemia.

Intercellular interactions for nutrient exchange and signaling are suggested by the proximity of the endothelial, basal lamina, and astrocyte components of the microvasculature. However, the exact structural and signaling relationships among these components are unknown. Previous work showed that integrin ₁ß₁ may tie the microvascular endothelium to the subjacent basal lamina. Investigation of the potential role of ₆ß₄ in the brain showed that the integrin is localized at the interface between astrocytes and cells in the basal lamina, where it may tie astrocyte end-feet to the extracellular matrix. This integrin accounted for 59% of ₆ antigen in cerebral microvessels less than 100 mum in diameter. By 2 hours after occlusion of the middle cerebral artery, a significant reduction in the expression of ₆ was accompanied by less pronounced parallel changes in expression of laminin-5 and laminin-1 by the extracellular matrix. Decreases in ß₄:laminin-5 and ₆ß₄:laminin-5 expression ratios in regions with dUTP-labeled cells coincided by 24 hours. Furthermore, a topographic gradation of loss of ₆ß₄ expression occurred in the ischemic corpus striatum by 2 hours after occlusion of the middle cerebral artery.

The temporal decreases in ₆ß₄ expression followed in the topographic order of normal area, ischemic zone with no dUTP-labeled cells, and then ischemic zone with dUTP-labeled cells. Histologic examination of zones peripheral to the area of neuronal incorporation of dUTP showed little change in microvascular appearance. Nonetheless, the disappearance of ₆ß₄ relative to its ligand (laminin-5) reflects early loss of integrity between astrocytes and the basal lamina and coincides with the swelling and detachment of astrocyte end-feet from the endothelium during focal ischemia. Therefore, rapid, subtle significant alterations between astrocytes and the extracellular matrix occur in parallel with neuronal injury after cessation of blood flow. This finding suggests that approaches that preserve microvascular integrity might reduce neuronal injury during focal cerebral ischemia. A second suggestion is that windows for intervention in patients may be defined by microvascular integrin responses to cessation in blood flow.

PUBLICATIONS

Abumiya, T., Masuda, J., Kawai, J., Suzuki, T., Ogata, J. Heterogeneity in the appearance and distribution of macrophage subsets and their possible involvement in hypertensive vascular lesions in rats. Lab. Invest. 75:125, 1996.

Balousek, P.A., Knowles, H.J., Higashida, R.T., del Zoppo, G.J. New interventions in cerebrovascular disease: The role of thrombolytic therapy and balloon angioplasty. Curr. Opin. Cardiol. 11:550, 1996.

del Zoppo, G.J. Reperfusion damage: The role of PMN leukocytes. In: Primer on Cerebrovascular Diseases. Welch, K.M.A., et al. (Eds.). Academic Press, New York, 1997, p. 217.

del Zoppo, G.J., Hamann, G., Fitridge, R., Okada, Y. Thrombolytic therapy. In: Stroke Therapy. Fisher, M. (Ed.). Butterworth-Heinemann, Boston, 1996.

del Zoppo, G.J., Okada, Y., Haring, H.-P., Tagaya, M., Wagner, S., Schmid-Schönbein, G.W. The role of adhesion molecules in acute cerebral ischemia. In: Pharmacology of Cerebral Ischemia. Krieglstein, J. (Ed.). MedPharm Scientific, Stuttgart, 1996, p. 393.

Hamann, G.F., Okada, Y., del Zoppo, G.J. Hemorrhagic transformation and microvascular integrity during focal cerebral ischemia/reperfusion. J. Cereb. Blood Flow Metab. 16:1373, 1996.

Haring, H.-P., Berg, E.L., Tsurushita, N., Tagaya, M., del Zoppo, G.J. E-selectin appears in non-ischemic tissue during experimental focal cerebral ischemia. Stroke 27:1386, 1996.

Haring, H.-P., del Zoppo, G.J. Thrombosis. In: Primer on Cerebrovascular Diseases. Welch, K.M.A., et al. (Eds.). Academic Press, New York, 1997, p. 148.

Härtl, R., Schörer, L., Schmid-Schönbein, G.W., del Zoppo, G.J. Experimental antileukocyte interventions in cerebral ischemia. J. Cereb. Blood Flow Metab. 16:1108, 1996.

Larrue, V., von Kummer, R., del Zoppo, G.J., Bluhmki, E. Hemorrhagic transformation in acute ischemic stroke: Potential contributing factors in the European Cooperative Acute Stroke Study. Stroke 28:957, 1997.

Okada, Y., Copeland, B.R., Hamann, G.F., Koziol, J.A., Cheresh, D.R., del Zoppo, G.J. Integrin alpha_Vß₃ is expressed in selected microvessels after focal cerebral ischemia. Am. J. Pathol. 149:37, 1996.

Schmid-Schönbein, G.W., Suematsu, M., Suzuki, H., del Zoppo, G.J. Microvascular cell activation in pathogenesis of ischemic disease. In: Pharmacology of Cerebral Ischemia. Krieglstein, J. (Ed.). MedPharm Scientific, Stuttgart, 1996, p. 437.

Tagaya, M., del Zoppo, G.J. Embryogenesis and angiogenesis. In: Primer on Cerebrovascular Diseases. Welch. K.M.A., et al. (Eds.). Academic Press, New York, 1997, p. 5.

Wagner, S., Tagaya, M., Koziol, J.A., Quaranta, V., del Zoppo, G.J. Rapid disruption of an astrocyte interaction with the extracellular matrix mediated by alpha₆ß₄ during focal cerebral ischemia/reperfusion. Stroke 28:858, 1997.

Top of Page

---

The Antithrombotic Protein C Pathway

J.H. Griffin, M.J. Heeb, J.A. Fernández, A.J. Gale, J.S. Greengard, A. Gruber, T.M. Hackeng, K. Kojima, Y. Kojima, J. Petäjä, H. Deguchi, S. Yegneswaran, X. Xu, Y. Montejano, B.N. Bouma

Research in our laboratory focuses on the molecular regulation of blood coagulation pathways and of one of the body's major antithrombotic mechanisms, the protein C pathway. Both basic and clinical studies help define how the blood coagulation pathways operate to achieve hemostasis under normal conditions or to promote thrombosis under pathologic conditions. During normal functioning of the protein C pathway, protein S and activated protein C (APC) are used to help the body defend itself against thrombosis.

We studied interactions between standard heparin and APC. The ability of heparin to prolong the activated partial thromboplastin time and the factor Xa--one-stage clotting time of normal plasma was markedly enhanced by adding purified APC. Heparin enhanced by fourfold the phospholipid-dependent inactivation of factor V by APC but had no effect on APC inactivation of thrombin-activated factor Va. In addition, heparin enhanced the rate of APC proteolysis of factor V but not factor Va. Low molecular weight heparin preparations had similar properties. In short, heparin and APC showed significant anticoagulant synergy in plasma because of three mechanisms that simultaneously decreased generation of thrombin: (1) heparin enhancement of antithrombin III--dependent inhibition of factor V activation by thrombin; (2) the inactivation of membrane-bound factor Va by APC; and (3) the proteolytic inactivation of membrane-bound factor V by APC, an event that is enhanced by heparin. This synergy between APC and heparin may contribute to the physiologic action of heparin and may lead to new therapies that combine these two agents.

Previously, we found that in patients who had a stroke within 1 week after infection or inflammation, a significant percentage (10%) had APC resistance without an R506Q factor V polymorphism and that circulating APC levels were decreased, in correlation with antibodies to phospholipids. Moreover, high-density but not low-density lipoprotein enhances the anticoagulant activity of APC and protein S. The emerging epidemiologic picture indicates that a poor anticoagulant response to APC is correlated with increased cholesterol levels and increased risk of ischemic stroke, whereas patients whose plasma shows a hyperresponse to APC have low cholesterol levels.

To study the potential influence of the carbohydrate moieties of factors V and Va on activation of factor V by thrombin and inactivation of factor Va by APC, we subjected factor V to mild deglycosylation under nondenaturing conditions. Deglycosylation had no effect on the procoagulant activity of factor V and activation of the factor by thrombin, whereas carbohydrate removal increased the susceptibility of factor Va to inactivation by APC. Thus, variability in carbohydrate could account for variability in APC resistance ratios, including the presence of borderline or low APC resistance ratios among patients who lack the R506Q mutation.

Human group II secretory phospholipase A₂ (sPLA₂) is an enzyme found in the -granules of platelets and at inflammatory sites. Although its physiologic function is unclear, sPLA₂ can inhibit blood coagulation reactions independent of its lipolytic action. To study the molecular basis of PLA₂ activities, in collaboration with S. Kent, Department of Cell Biology, and C. Bon and C. Mounier, Institut Pasteur, we developed a method for total chemical synthesis of sPLA₂ by chemical ligation of large unprotected peptides. Synthetic sPLA₂, like recombinant sPLA₂, inhibited generation of thrombin from prothrombinase complex (factors Xa, V, and II; calcium; and phospholipids). In the absence of phospholipids, both synthetic and recombinant sPLA₂ inhibited by 70% prothrombin activation by factors Xa and Va and calcium. Thus, like recombinant or natural sPLA₂, synthetic sPLA₂ is a phospholipid-independent anticoagulant. This study showed that chemical synthesis of sPLA₂ yields fully active nativelike enzyme and is a straightforward tool to provide sPLA₂ analogs for structure-function studies of the anticoagulant, lipolytic, or inflammatory activities of this enzyme.

To study whether circulating APC could regulate thrombin activity in basal physiologic conditions, we determined levels of fibrinopeptide A and APC in plasma samples obtained from 40 healthy adults. The results showed that the levels of these two components were inversely correlated (R = -0.487, P = .0023). These findings support the hypothesis that APC downregulates thrombin activity in vivo in subjects with normal levels of protein C.

Top of Page

---

Mechanisms of Regulation of Blood Coagulation Factors V and Va

M.J. Heeb, Y. Montejano, Y. Kojima, A. Rehemtulla,^* R.J. Kaufman^*

^* University of Michigan, Ann Arbor, MI

The mutation R506Q in blood coagulation factor V is associated with activated protein C (APC) resistance and thrombotic risk, but the relative importance of APC cleavages at R306, R506, and R679 for inactivation of factor Va vs inactivation of factor V is unclear, and the reasons that any of these cleavages cause inactivation are unclear. To address these questions, we prepared and purified the following recombinant forms of factor V: wild-type rFV, Q306-rFV, Q506-rFV, and Q306/Q506-rFV. The activated forms of these recombinants are designated as follows: rFVa, Q306-rFVa, Q506-rFVa, and Q306/Q506-rFVa.

Q506-rFV, Q506-rFVa, and Q306-rFVa were APC resistant, with low APC resistance ratios (clotting times with and without APC in factor V--deficient plasma). In assays with purified components of the prothrombinase complex, Q306-rFVa was inactivated less than 40% under conditions in which Q506-rFVa was inactivated more than 90%, showing that cleavage at R306 is most important for normal efficient inactivation of factor Va. Q306/Q506-rFV and Q306/Q506-rFVa were almost completely APC resistant in clotting assays and during incubation with excess APC, indicating that cleavage at either R306 or R506 is essential for any significant inactivation of factor V or factor Va. Rate constants for cleavages at R506, R306, and R679 in factor Va were 55, 9, and 0.6 (x 10⁶ M^-1 s^-1), respectively.

Surprisingly, Q306-rFV was inactivated by APC at near-normal rates and had a near-normal APC resistance ratio. Furthermore, rates of rFV, Q306-rFV, and Q506-rFV inactivation were not much slower (less than two times) than the rate of inactivation of normal factor Va. Immunoblotting showed that the relative rates of APC cleavage at the three sites in factor V were more similar than in factor Va, and R679 was cleaved approximately 10 times faster in factor V than in factor Va. No additional cleavage fragments were detected to explain the increased susceptibility to APC of Q306-rFV vs Q306-rFVa. Q306/Q506-rFV was activated by factor Xa at about 5% of the rate of rFV, and thrombin-activated Q306/Q506-rFVa had 50--77% of normal activity.

We conclude that mechanisms of APC inactivation of factor Va and factor V differ, but the rates of inactivation are not greatly different, partly because of the rate of cleavage at R679 is faster in factor V than in factor Va. This finding suggests that regulation at the level of factor V may be more important than previously thought. APC cleavage at R306 is most important for normal inactivation of factor Va but not of factor V. These results and our previous studies showing the presence of binding sites for factor Xa near 306 and 506 in factor V suggest that binding of factor Xa is impaired by mutations at both of these sites and may also be impaired by APC cleavage at these sites.

PUBLICATIONS

Fernández, J.A., Hackeng, T.M., Kojima, K., Griffin, J.H. The carbohydrate moiety of factor V modulates inactivation by activated protein C. Blood 89:4348, 1997.

Fernández, J.A., Heeb, M.J., Radtke, K.-P., Griffin, J.H. Potent blood coagulant activity of human semen due to prostasome-bound tissue factor. Biol. Reprod. 56:757, 1997.

Fernández, J.A., Petäjä, J., Gruber, A., Griffin, J.H. Activated protein C correlates inversely with thrombin levels in resting healthy individuals. Am. J. Hematol. 56:29, 1997.

Gale, A.J., Sun, X., Heeb, M.J., Griffin, J.H. Nonenzymatic anticoagulant activity of the mutant serine protease Ser360Ala-activated protein C mediated by factor Va. Protein Sci. 6:132, 1997.

Gillespie, D.L., Carrington, L.R., Griffin, J.H., Alving, B.M. Resistance to activated protein C: A common, inherited cause of venous thrombosis. Ann. Vasc. Surg. 10:174, 1996.

Griffin, J.H., Petäjä, J. Activated protein C resistance. In: Molecular Mechanisms of Hypercoagulable States. Schafer, A.I. (Ed.). Landes Bioscience and Chapman & Hall, New York, 1997, p. 101.

Hackeng, T.M., Mounier, C.M., Bon, C., Dawson, P.E., Griffin, J.H., Kent, S.B.H. Total chemical synthesis of enzymatically active human type II secretory phospholipase A₂. Proc. Natl. Acad. Sci. U.S.A. 94:7845, 1997.

Macko, R.F., Ameriso, S.F., Gruber, A., Griffin, J.H., Fernández, J.A., Barndt, R., Quismorio, F.P., Weiner, J.M., Fisher, M. Impairments of the protein C system and fibrinolysis in infection-associated stroke. Stroke 27:2005, 1996.

Petäjä, J., Fernández, J., Gruber, A., Griffin, J.H. Anticoagulant synergism of heparin and activated protein C in vitro: Role of a novel anticoagulant mechanism of heparin, enhancement of inactivation of factor V by activated protein C. J. Clin. Invest. 99:2655, 1997.

Petäjä, J., Griffin, J.H. Activated protein C resistance: What have we learned after the dust has settled? Ann. Med., in press.

Smiley, S.T., Boyer, S.N., Heeb, M.J., Griffin, J.H., Grusby, M.J. Protein S is interleukin-4-inducible in T cells and inhibits lymphoid procoagulant activity. Proc. Natl. Acad. Sci. U.S.A., in press.

Wang, L., Zoppe, M., Hackeng, T., Griffin, J.H., Lee, K.-F., Verma, I.M. A factor IX-deficient mouse model for hemophilia B gene therapy. Proc. Natl. Acad. Sci. U.S.A., in press.

Zivelin, A., Griffin, J.H., Xu, X., Pabinger, I., Samama, M., Conard, J., Brenner, B., Eldor, A., Seligohn, U. A single genetic origin for a common caucasian risk factor for venous thrombosis. Blood 89:397, 1997.

Top of Page

---

Molecular Determinants of Arginine--Glycine--Aspartic Acid Ligand Specificity for ß₃ Integrins

T.J. Kunicki, D.S. Annis, B. Felding-Habermann

The arginine--tyrosine--aspartic acid (RYD) and arginine--glycine--aspartic acid (RGD) sequences within the third complementarity-determining region of the heavy chain (H3) of murine recombinant Fab molecules OPG2 and AP7, respectively, are responsible for the specific binding of the Fabs to the platelet integrin _IIbß₃. We evaluated the influence of divalent cations and single amino acid substitutions at key positions within H3 on the selectivity of these Fab molecules for _IIbß₃ and for the vitronectin receptor _vß₃.

The parent Fab molecule OPG2 (H3 sequence, HPFYRYDGGN) binds selectively to _IIbß₃ and not at all to any other RGD-cognitive integrin, particularly _vß₃, under any divalent cation conditions. The binding of AP7 (HPFYRGDGGN) to _IIbß₃ is not affected by the relative composition of calcium, magnesium, or manganese. However, AP7 binding to _vß₃, either expressed by M21 cells or as the purified integrin, is supported by manganese and inhibited by calcium.

If the flanking asparagine¹⁰⁸ residue within the AP7 H3 loop is replaced by alanine (HPFYRGDGGA), the resulting Fab molecule AP7.4 binds selectively to _vß₃ in a cation-dependent manner but does not bind at all to _IIbß₃ under any conditions. AP7.4 binding to _vß₃ is supported by manganese, completely inhibited by calcium, and largely unaffected by magnesium. This behavior mimics that of the adhesive protein, osteopontin, another ligand that binds preferentially to _vß₃.

Despite these differences in specificity for _IIbß₃ and _vß₃, AP7 and AP7.4 remain selective for the ß₃ integrins and do not bind to cell lines that express the RGD-cognitive integrins _vß₅ or ₅ß₁. These results confirm that subtle changes in the amino acid composition immediately flanking the RGD or RYD motifs can have a profound effect on ß₃ integrin specificity, most likely because the changes influence the juxtaposition of the arginine and aspartate side chains within the extended RGD loop sequence.

Top of Page

---

Ultrastructural Studies of Platelet Aggregates From Human Subjects

M. Humbert,^* P. Nurden,^* C. Bihour,^* J.-M. Pasquet,^* J. Winckler,^* E. Heilmann,^* P. Savi,^** J.-M. Herbert,^** T.J. Kunicki, A.T. Nurden^*

^* Hôpital Cardiologique, Pessac, France
^** Sanofi Recherche, Toulouse, France

The ultrastructure of platelet aggregates induced by high-dose ADP in platelet-rich plasma of healthy subjects receiving the antithrombotic drug clopidogrel (75 mg/day for 7 days) was compared with that of platelet aggregates obtained from a patient who has a congenitally reduced and reversible platelet response to ADP. Electron microscopy showed that when untreated normal platelets were stimulated with 10 muM ADP, large and stable aggregates were often surrounded by a layer of degranulated platelets. A similar layer was not seen on aggregates of platelets from subjects receiving clopidogrel or from the patient with the congenital defect; instead, ADP-stimulated platelets formed loose aggregates with few contact points. Immunogold staining of ultrathin sections of aggregates showed a much decreased surface staining on platelets within the aggregates for both a polyclonal antibody to fibrinogen and for AP-6, a monoclonal antibody to the 211--222 amino acid sequence of glycoprotein IIIa. Disaggregation occurred in both situations, and many of the platelets regained a normal morphology.

Flow cytometric analysis of the activation-dependent binding of fibrinogen labeled with fluorescein isothiocyanate or of AP-6 to unstirred suspensions of clopidogrel-treated platelets confirmed that binding was much reduced and that full activation of glycoprotein IIb-IIIa did not occur. In contrast, other results showed that ADP-induced changes in platelet shape and mobilization and influx of calcium were unaffected by clopidogrel. On the basis of these results, we speculate that platelet activation by ADP is biphasic and that the receptor-induced activation pathway that is affected by clopidogrel or is congenitally impaired in the patient is necessary for the full activation of glycoprotein IIb-IIIa and the formation of stable macroaggregates.

Top of Page

---

Visualization of Activation-Dependent Epitopes on Glycoprotein IIb-IIIa Complexes of Platelets Stimulated by Thrombin

M. Humber,^* T.J. Kunicki, C. Bihour,^* J. Winckler,^* A.T. Nurden,^* P. Nurden^*

^* Hôpital Cardiologique, Pessac, France

We used AP-6, an IgM monoclonal antibody that recognizes the sequence 211--221 of glycoprotein IIIa, to study the distribution of this epitope on unstimulated or thrombin-activated platelets. We used flow cytometry to evaluate the expression of the epitope on platelets and immunogold staining of ultrathin sections to analyze the distribution within the cell.

We found little or no binding of AP-6 to unstimulated platelets, but immunogold staining showed labeling associated with the membranes of -granules. The binding of AP-6 to thrombin-stimulated platelets was compared with the binding of monoclonal antibodies to receptor-induced binding sites and of polyclonal antibodies to fibrinogen. Increased expression of the AP-6 epitope was observed on membranes of the surface-connected canalicular system and at the periphery of the cell as early as 10--15 seconds after platelet activation by thrombin. At the same time, binding of antibodies to receptor-induced binding sites confirmed that at least part of the endogenous fibrinogen had left the -granules and was bound to platelet membranes. Staining with polyclonal antibodies to fibrinogen also revealed fibrinogen in vesicles resulting from granule fusion. The pool of internal membranes with bound fibrinogen rapidly became exposed to the outside of the platelet, a process involving the unfolding of membranes and formation of pseudopods.

Our results provide evidence that activation of glycoprotein IIb-IIIa and binding of fibrinogen to platelet membranes can occur before the expression of these proteins at the platelet surface. They also suggest the presence of an -granule pool of glycoprotein IIb-IIIa linked to fibrinogen in unstimulated platelets.

PUBLICATIONS

Humbert, M., Kunicki, T.J., Bihour, C., Winckler, J., Nurden, A.T., Nurden, P. Visualization of activation-dependent epitopes on glycoprotein IIb-IIIa complexes of platelets stimulated by thrombin: Immunogold staining of ultrathin sections. Semin. Thromb. Hemost. 22:279, 1996.

Humbert, M., Nurden, P., Bihour, C., Pasquet, J.-M., Winckler, J., Heilmann, E., Savi, P., Herbert, J.-M., Kunicki, T.J., Nurden, A.T. Ultrastructural studies of platelet aggregates from human subjects receiving clopidogrel and from a patient with an inherited defect of an ADP-dependent pathway of platelet activation. Arteriosclerosis 16:1532, 1996.

Kunicki, T.J., Annis, D.S., Deng, Y.J., Loftus, J.C., Shattil, S.J. A molecular basis for affinity modulation of Fab ligand binding to the integrin alpha_IIbß₃. J. Biol. Chem. 271:20315, 1996.

Kunicki, T.J., Felding-Habermann, B., Annis, D.S. Molecular determinants of RGD ligand specificity for ß₃ integrins. J. Biol. Chem. 272:4103, 1997.

Kunicki, T.J., Kritzik, M., Annis, D.S., Nugent, D.J. Hereditary variation in platelet integrin alpha₂ß₁ density is associated with two silent polymorphisms in the alpha₂ gene coding sequence. Blood 89:1939, 1997.

Marcinkiewicz, C., Rosenthal, L.A., Mosser, D.M., Kunicki, T.J., Niewarowski S. Immunological characterization of eristostatin and echistatin binding sites on alpha_IIbß₃ and alpha_vß₃ integrins. Biochem. J. 317:817, 1996.

McKay, B.S., Annis, D.S., Honda, S., Christie, D., Kunicki, T.J. Molecular requirements for assembly and function of a minimized human integrin alpha_IIbß₃. J. Biol. Chem. 271:30544, 1996.

Nurden, P., Humbert, M., Piotrowicz, R.S., Bihour, C., Nurden, A.T., Kunicki, T.J. The distribution of ligand-occupied alpha_IIbß₃ in resting and activated human platelets determined by expression of a novel class of ligand-induced binding site (LIBS) recognized by monoclonal antibody AP6. Blood 88:887, 1996.

Top of Page

---

Release of Basic Fibroblast Growth Factor From Endothelial Cells

R.S. Piotrowicz, J.L. Martin,^* W.H. Dillman,^* E.G. Levin

^* University of California, San Diego, CA

Basic fibroblast growth factor is an important mitogenic and angiogenic factor that stimulates growth and migration of endothelial cells. It is not secreted via the classical vesicular pathway, and the intracellular proteins that facilitate its release have not been determined. Transfection and expression of the 27-kD human heat-shock protein in bovine arterial endothelial cells doubles the rate of estrogen-induced secretion of basic fibroblast growth factor, preferentially inducing the release of high molecular weight forms. The secreted factor is mitogenic to breast adenocarcinoma cells cultured in conditioned medium obtained from transfected endothelial cells. In contrast, decreasing the level of the endogenous heat-shock protein homolog with an antisense vector markedly decreases release of the factor. Antibodies to heat-shock protein or to basic fibroblast growth factor coprecipitate both proteins from extracts of endothelial cells, showing a direct association between the two proteins. This interaction is most likely an important step in the mechanism of secretion of basic fibroblast growth factor.

Top of Page

---

Effect of ß-Estradiol on Platelet Adhesion to Collagen

R.S. Piotrowicz, T.J. Kunicki, E.G. Levin

The 27-kD heat-shock protein (HSP27) is an estrogen-responsive actin-binding protein. It acts as a microfilament capping protein in vitro and inhibits membrane-associated polymerization of F-actin in vivo. Cells of the vasculature, including platelets, constituitively express HSP27. To determine if expression of HSP27 by platelets is influenced by serum levels of estrogen, we used immunoblotting to analyze platelet lysates from a panel of five female donors; platelets were obtained at 3- to 4-day intervals for 1 month. Expression of HSP27 cycled with serum levels of ß-estradiol; peak HSP27 expression occurred 4--9 days after the beginning of the menstrual cycle. The magnitude in the range of HSP27 expression varied between donors. Levels of peak expression were 1.5 to 6 times that of the lowest level of HSP27 expression during the cycle.

Adhesion of these platelets to collagen I or III but not to fibronectin also cycled with serum levels of ß-estradiol, with peak adherence correlating with the ebb in HSP27 levels. Conversely, at times when peak levels of HSP27 were expressed, platelet adhesion to collagen I and III was at its minimum. Expression of the ₂ß₁ integrin, the receptor that mediates platelet adhesion to collagen, did not vary from times of peak adhesion to times of minimal adhesion. Transfection of HSP27 into bovine arterial endothelial cells generated decreased adhesion to collagen but not to fibronectin, indicating that enhanced expression of HSP27 is sufficient for the alteration in substrate-specific adhesiveness. Thus, induction of platelet HSP27 by ß-estradiol, by reducing platelet adhesiveness to collagen, may represent a mechanism by which ß-estradiol exerts a protective effect on the cardiovascular system.

PUBLICATIONS

Levin, E.G., Osborn, K.G. The expression of endothelial tissue plasminogen activator in vivo: A function defined by vessel size and anatomic location. J. Cell Sci. 11:139, 1997.

Piotrowicz, R.S., Levin, E.G. Baso-lateral membrane-associated heat shock protein 27 kDa and microfilament polymerization. J. Biol. Chem., 272:25920, 1997.

Piotrowicz, R.S., Martin, J.L., Dillman, W.H., Levin, E.G. The 27-kDa heat shock protein facilitates basic fibroblast growth factor release from endothelial cells. J. Biol. Chem. 272:7042, 1997.

Top of Page

---

Integrins in Cell Activation, Differentiation, and Transplantation

D.R. Salomon, L. Crisa, J.K. Ishii, B.E. Torbett,^* K.A. Smith,^* K. Conners,^* A. Hayek,^** V. Cirruli^**

^* Department of Immunology, TSRI
^** University of California, San Diego, CA

Integrins are a major family of cell adhesion molecules that include receptors for an array of proteins on the extracellular matrix (e.g., fibronectin) and for cell-surface proteins (e.g., vascular cell adhesion molecule-1).

STEM CELL DEVELOPMENT

Both thymus and bone marrow direct development of stem cells by a series of signals delivered via stromal cells located in unique microenvironments. Therefore, the function of these developmental compartments must be understood in the context of their structure. Unfortunately, little is known about the molecular mechanisms that mediate movement of progenitor cells in and out of these unique compartments or that regulate interactions between progenitor cells and stromal cells during the developmental program.

We have developed several lines of evidence that implicate interactions between integrins and their ligands. For example, a unique subset of immature thymocytes expresses an activated form of very late activation antigen 4 (also known as integrin ₄ß₁) that mediates adhesion to fibronectin and vascular cell adhesion molecule-1. In contrast, more mature double-positive cells use very late activation antigen 4 and upregulate the activation state of another integrin, very late activation antigen 5 (₅ß₁), to mediate migration. Thus, a balance of adhesion and migration is regulated by integrin activation during maturation.

We now have evidence for differential expression of several chemokine receptors on subsets of thymic cells and for chemokine production by thymic epithelial cells. These chemokines selectively activate integrins and may regulate traffic of progenitor cells during development. We also found that thymocytes produce several metalloproteinases, which may also regulate the efficiency of cell migration. We have successfully transplanted human thymus into immunodeficient mice reconstituted with human hematopoietic stem cells. We are using this model to study stem cell traffic, the role of chemokines, allogeneic bone marrow transplantation, gene therapy, and implications for HIV infection and therapy (Fig. 1).

PANCREATIC ISLET CELL TRANSPLANTATION

Human pancreatic islet cells express an array of integrins that are regulated in development and mediate cell adhesion and migration. Although it is axiomatic that islet structure determines islet function, little is known about the determinants of that structure. We propose that integrins shape the islet around a matrix framework secreted during fetal development of the islet's arteriolar network. In fact, confocal microscopy shows the presence of such frameworks in human islets that colocalize with areas of integrin expression.

We have developed models for islet transplantation to study the role of integrins in islet development, structure, and function. Most recently, we used a novel three-dimensional mesh technology integrated with matrix adhesive-site analogs to enhance the function of islet cells in long-term culture and after transplantation. Experiments are under way to determine the molecular regulation of islet integrin function during development from endocrine progenitor cells and after transplantation of mature islets. These studies are being integrated with new strategies for islet xenotransplantation.

PUBLICATIONS

Cooperative Clinical Trials in Transplantation Research Group. Murine OKT4a immunosuppression in cadaver donor renal allograft recipients. Transplantation 63:1243, 1997.

Cramer, D.V., Wu, G.D., Chapman, F.A., Marboe, C., Salomon, D.R. 2-Chlorodeoxyadenosine in combination with cyclosporine inhibits the development of transplant arteriosclerosis in rat cardiac allografts. Transplant. Proc. 29:616, 1997.

Crisa, L., McMaster, M.T., Ishii, J.K., Fisher, S.J., Salomon, D.R. Identification of a thymic epithelial cell subset sharing expression of the class Ib HLA-G molecule with fetal trophoblasts. J. Exp. Med. 186:289, 1997.

Salomon, D.R. Adhesion molecules in immunoregulation In: Adhesion Molecules in Health and Disease. Paul, L.C., Issekutz, T.B. (Eds.). Marcel Dekker, New York, 1997, p. 223.

Salomon, D.R. The molecular basis of transplantation immunity. In: Kidney Transplant Rejection, 3rd ed. Racusen, L.C., et al. (Eds.). Marcel Dekker, New York, in press.

Salomon, D.R. Rationale and mechanisms of immunosuppression. Kidney Int. 51(Suppl. 58):S48, 1997.

Salomon, D.R., Crisa, L., Mojcik, C.F., Ishii, J.K., Klier, G., Shevach, E.M. Vascular cell adhesion molecule-1 is expressed by cortical thymic epithelial cells and mediates thymocyte adhesion: Implications for the function of alpha₄ß₁ (VLA4) integrin in T-cell development. Blood 89:2461, 1997.

Todd, S.C., Lipps, S.G., Crisa, L., Salomon, D.R., Tsoukas, C.D. CD81 expressed on human thymocytes mediates integrin activation and IL2-dependent proliferation. J. Exp. Med. 184:2055, 1997.

Top of Page

---

Division of Experimental Pathology

Francis V. Chisari, M.D., Division Head

---

Molecular Biology of Hepatitis B and C Viruses and the Immune Response to Their Antigens

F.V. Chisari, L.G. Guidotti, C. Pasquinelli

Our laboratory focuses on the immunobiology and pathogenesis of viral clearance and liver cell injury during infections with hepatitis B and C viruses and the mechanisms whereby these viruses cause hepatocellular carcinoma. Our studies are multidisciplinary, involving molecular, virologic, and immunologic approaches, and include research with infected patients and transgenic mice.

---

Cytotoxic T-Lymphocyte Response in Humans and Chimpanzees Infected With Hepatitis B Virus

R. Bertoni, B. Rehermann, A. Sette,^* R. Purcell,^** F.V. Chisari
^* Cytel Corporation, San Diego, CA

^** National Institute of Allergy and Infectious Diseases, Bethesda, MD

Most patients who recover from acute infection with hepatitis B virus (HBV) have a polyclonal HLA class I--restricted cytotoxic T-lymphocyte (CTL) response to multiple HBV-encoded epitopes that map to the polymerase, envelope, and nucleocapsid proteins. Recently, we showed that traces of HBV persist in the blood for decades after recovery and that patients maintain a strong CTL response that lasts indefinitely after clinical recovery, illustrating that sterilizing immunity need not occur after successful resolution of HBV infection. Low-level persistence of virus is not due to mutational escape from the CTL response, suggesting that the virus persists in an immunologically privileged reservoir that seeds the circulation and maintains the immune response in an active state, perhaps providing protection from reinfection.

In contrast to these results, in patients who do not clear HBV, the HBV-specific CTL response is weak or undetectable. Nonetheless, CTL precursors are present, because a strong CTL response develops during spontaneous or interferon-induced remission. These observations suggest that therapeutic augmentation of the CTL response to HBV most likely will result in viral clearance in chronically infected patients.

To test this hypothesis, Cytel Corporation has undertaken a phase II clinical trial to determine the safety and immunogenicity of a lipopeptide formulation containing an immunodominant HBV-derived, HLA-A2-- restricted CTL epitope that is commonly targeted in patients who successfully clear HBV. Recent results indicate that this vaccine is immunogenic in these patients and that CTL responsiveness depends on the dose of the peptide and is associated with clinical, serologic, and virologic improvement.

Despite these promising results, several epitopes must be included in a synthetic peptide vaccine to extend its usefulness to patients with other HLA haplotypes and to minimize the chances of inducing a monospecific CTL response that could select for escape mutants. Recently, using HBV-derived synthetic peptides that can bind to multiple HLA class I alleles, we determined several new HBV-specific CTL epitopes that are restricted by the HLA-A2, HLA-A3, and HLA-B7 superfamilies, thereby extending the potential usefulness of this therapeutic approach in an ethnically unbiased manner to more than 90% of the world's population.

One of the most important unanswered questions in HBV immunobiology is the nature of the immune response to HBV during the incubation period, because these events probably determine the ultimate outcome of the infection. Analysis of this period in humans is difficult. Therefore, we are developing systems to study the T-cell response to HBV in animal models of HBV infection. In collaboration with R. Purcell and his colleagues at the National Institutes of Health, we infected two chimpanzees with serum from transgenic mice that replicate the HBV genome (see following reports). Typical cases of acute viral hepatitis developed in both chimpanzees, and both animals ultimately cleared the infection. Peripheral blood lymphocytes obtained from both animals at the peak of infection showed proliferative T-cell responses to the particulate HBV core antigen and CTL responses to viral peptides that we previously defined as CTL epitopes in humans infected with HBV.

In collaboration with A. Sette at Cytel Corporation, we showed that these immunogenic viral peptides bind with high affinity to class I molecules isolated from the corresponding chimps. Using this peptide-binding assay, we have identified additional chimps with class I molecules that bind these peptides. These animals should respond to the peptides and thus enable us to define the kinetics of the CTL response to HBV during the early incubation period of HBV infection and to test the protective and therapeutic effects of the corresponding lipopeptide vaccine cocktails.

Top of Page

---

The Cytotoxic T-Lymphocyte Response in Patients Infected With Hepatitis C Virus

K.M Chang, B. Rehermann, F.V. Chisari

Recently, we showed that patients with chronic hepatitis C mount a polyclonal cytotoxic T-lymphocyte (CTL) response to hepatitis C virus (HCV). Although the virus routinely persists despite this response, an inverse relationship exists between the strength of the CTL response and viral load in these patients, suggesting that the CTL response can control HCV, at least to some extent. If this hypothesis is correct, CTL selection of HCV escape variants might play a role in HCV persistence. Accordingly, we compared the peripheral blood CTL responsiveness to a panel of 10 prototype HCV peptides with the corresponding peptide sequences encoded by the infecting viruses in several patients with chronic HCV infection.

Variant viral peptide sequences were present three times more often in the presence of CTL pressure than in its absence, and prototype-specific CTLs were detected nearly twice as often when variant peptide sequences were present than when prototype peptide sequences were present. In most instances, the amino acid substitutions in the variant viral peptides caused substantial reduction in HLA binding affinity. Collectively, these observations suggest that CTL selection of epitope variants may have occurred in these patients before the patients became subjects in the study. Surprisingly, follow-up analysis over a period of up to 4 years revealed that the subsequent emergence rate of additional CTL escape variants was low, in contrast to the high frequency of escape variants that we initially observed. We conclude that CTL selection of escape variants does occur, probably as an early event in HCV infection, and that it may play a role in HCV persistence.

Top of Page

---

Host-Virus Interactions in Mice Transgenic for Hepatitis B Virus

L.G. Guidotti, V. Cavanaugh, Y. Shimizu, Y. Nakamoto, V. Pasquetto, A. Franco, F.V. Chisari

To directly examine the consequences of the cytotoxic T-lymphocyte (CTL) response to hepatitis B virus (HBV), we have monitored the cytopathic and antiviral consequences of passive transfer of HBV-specific CTL clones into transgenic mice that contain the complete viral genome, express all the viral gene products, and replicate the virus episomally in hepatocytes. We found that murine HBV-specific CTLs induce a necroinflammatory liver disease in HBV transgenic mice and that the CTLs initiate this process by triggering hepatocytes to undergo programmed cell death, or apoptosis.

To determine the signaling pathways involved in this process, we used CTL clones derived from genetically defined mice that were Fas-ligand deficient (gld), perforin deficient (pko), or wild type. Although all CTL clones had comparable cytolytic activity in vitro, their pathogenic and antiviral potential differed dramatically when they were injected into the transgenic mice. Specifically, the wild-type CTLs induced a moderately severe liver disease, whereas the gld and pko CTLs were virtually nonpathogenic. These results indicate that both Fas ligand-- and perforin-dependent death pathways must be activated by the same CTLs to kill the hepatocyte in vivo, whereas either pathway alone is sufficient in vitro.

Although the gld and pko CTLs do not cause any liver disease, they abolish HBV replication in the liver by secreting TNF- and IFN- when they recognize antigen presented by hepatocytes. Unexpectedly, lines and clones of type 1 T-helper cells that are MHC class II (IA^d) restricted, CD4⁺, and specific for the HBV surface antigen also inhibit HBV replication in hepatocytes. Similar to the CTLs, the type 1 T-helper cells inhibit HBV replication by secreting IFN- and TNF- in response to antigen, but in this case, the antigen is presented by hepatic macrophages, not hepatocytes. In related experiments, we showed that the inflammatory cytokines activate the hepatocyte to destroy the HBV nucleocapsid particles in the cytoplasm and to posttranscriptionally destabilize the viral pregenomic and messenger RNA in the nucleus. These results suggest that viral clearance during HBV infection may be mediated principally by noncytolytic signals delivered by both limbs of the T-cell response.

We explored the concept of bystander suppression of HBV replication by using several nonspecific stimuli in the transgenic mice. For example, we showed that the amount of HBV replicative forms and HBV core protein in the liver of HBV transgenic mice is transiently reduced during massive liver regeneration after partial hepatectomy, whereas the steady-state amount of viral RNA is unchanged. This antiviral effect is triggered by interferons and TNF- that are induced in the liver after hepatectomy and either prevent the formation or accelerate the degradation of viral nucleocapsids in the cytoplasm of hepatocytes. Despite massive hepatocellular turnover, this effect is independent of liver cell division, indicating that HBV replicates efficiently in both resting and dividing hepatocytes.

We also found that HBV replication is abolished within 12 hours of acute infection with lymphocytic choriomeningitis virus (LCMV), because of the induction of TNF- and IFN-/ß expression by the hepatic macrophages in response to LCMV. Furthermore, expression and replication of HBV genes are permanently abolished in mice persistently infected with LCMV that produce TNF- and IFN- in the liver at high levels for life without any attendant liver disease. Hepatic HBV replication is also inhibited in transgenic mice infected by hepatotropic members of the adenovirus and herpesvirus families, a finding that supports the bystander suppression hypothesis. The antiviral effect coincides with the induction of IFN-/ß in the liver by the adenovirus. As in LCMV infection, this noncytolytic antiviral process continues for several days until it is superseded by an immune-mediated adenovirus-specific liver disease characterized by the induction of IFN- and TNF-.

A similar process is also induced by infection with murine cytomegalovirus and by the administration of recombinant murine IL-12 in these transgenic mice. In both these cases, suppression of HBV replication extends to extrahepatic sites such as the kidney. These results suggest that IL-12 could play an important role in the immune response to HBV during natural infection. If these properties extend to chronically infected humans, IL-12 may have therapeutic value both as an antiviral agent and as an immunostimulant in patients with chronic hepatitis.

Finally, because they are immunologically tolerant to HBV proteins, these transgenic animals resemble chronically infected patients whose T-cell response to HBV is severely attenuated or undetectable. As such, the mice can be used to assess the relative efficacy of various immunization strategies designed to break this state of tolerance, with the hope that similar strategies may be useful for the treatment of patients with chronic hepatitis. To this end, we have compared the CTL responses of these animals and nontransgenic littermate controls after immunization with plasmid vectors that express HBV proteins, with lipopeptides containing previously defined CTL epitopes, and with activated dendritic cells that have been pulsed with the corresponding proteins and peptides.

The transgenic mice did not produce HBV-specific CTL responses after immunization with plasmid DNA even though plasmid DNA was quite immunogenic in nontransgenic littermates. In contrast, HBV-specific CTL responses were routinely induced after immunization with lipopeptide or dendritic cells. These findings suggest that high epitope density and professional antigen-presenting cells are required to break immunologic tolerance in these animals, perhaps presaging the existence of similar requirements for the development of immunotherapeutic strategies for termination of persistent HBV infection in humans.

Top of Page

---

Analysis of Hepatitis C Virus Pathogenesis in Transgenic Mice

C. Pasquinelli, K. Kakimi, K.M. Chang, L.G. Guidotti, F.V. Chisari

To characterize the pathogenetic mechanisms responsible for liver cell injury and the development of chronic hepatitis associated with infection with hepatitis C virus (HCV), we have produced transgenic mice that express the HCV core and E2 proteins in the liver under the control of liver-specific promoters. Both the HCV core and E2 transgenic mice express the HCV proteins in their hepatocytes at levels comparable to those detected in naturally infected humans. However, the mice have no histologic or biochemical evidence of liver disease or hepatocellular carcinoma, suggesting that these HCV proteins are not directly cytopathic or transforming for primary hepatocytes in vivo.

To test the hypothesis that HCV-induced liver disease is due to an immune response to viral antigens expressed at the surface of hepatocytes, we studied the cytopathic potential of HCV core--specific cytotoxic T lymphocytes (CTLs) at various times after intravenous injection of the CTLs into syngeneic HCV transgenic mice. Although in vitro these CTLs efficiently kill target cells expressing HCV core antigen, they cause only mild liver disease when adoptively transferred to HCV core transgenic mice. The disease is characterized histologically by an early appearance of apoptotic hepatocytes that is followed by a lymphomononuclear infiltrate that becomes detectable 24 hours after the injection of CTLs; the disease subsides 3 days later. The possibilities that the HCV core protein is inefficiently processed or that it globally inhibits antigen processing or expression of MHC class I molecules by primary hepatocytes are under investigation.

Yeast two-hybrid analysis indicated that HCV core protein binds to the cytoplasmic domain of the lymphotoxin ß receptor, and functional analysis of cell lines stably transfected with core expression vectors showed that the core protein enhances Fas-mediated apoptosis in vitro. To determine whether the core protein can sensitize primary hepatocytes to Fas-induced apoptosis in the liver, we injected antibody to Fas intravenously into HCV core transgenic mice and into nontransgenic littermates of these mice and monitored the induction of hepatocellular apoptosis 24 hours later. The HCV core transgenic mice had twice as many apoptotic hepatocytes as the nontransgenic controls did and had a correspondingly more severe liver disease, indicating that the presence of HCV core protein in the hepatocytes may sensitize the cells to Fas-mediated apoptosis. The survival advantage of this property for the virus is not immediately obvious; hence, the property may be linked to other events that favor viral persistence.

Top of Page

---

Mechanisms of Transcriptional Regulation in Xenopus laevis

P. Labhart

My laboratory is interested in basic mechanisms of transcriptional regulation of gene expression in Xenopus laevis, a model system that is amenable to experimental manipulation of oogenesis and embryogenesis. We focus on two aspects: the role of DNA-dependent protein kinase (DNA-PK) in repression of transcription by RNA polymerase I and the functional analysis of multiple isotypes of elongation factor TFIIS in class- and gene-specific transcription.

Phosphorylation of transcription factors is thought to be an important mechanism of transcriptional regulation. We recently showed that DNA-PK represses transcription by RNA polymerase I in vitro but has no effect on transcription by the other classes of RNA polymerase. DNA-PK, a nuclear kinase that is activated by double-stranded DNA, is known to be involved in signaling pathways related to the repair of double-stranded DNA breaks, but a physiologic role in transcriptional regulation has not been proved. Our findings indicate that the repression of RNA polymerase I transcription is mediated by the ribosomal transcription factor Rib1.

We found that TATA-box binding protein, a universal eukaryotic transcription factor that is a component of Rib1, is indeed a substrate for DNA-PK. We recently isolated a segment of the gene for Xenopus DNA-PK and are planning to inhibit DNA-PK expression in oocytes and embryos by injecting antisense oligonucleotides. These experiments will directly test the in vivo significance of the observed in vitro effects of DNA-PK on ribosomal gene transcription.

TFIIS is a transcription factor that stimulates elongation by RNA polymerase II. It does so by helping the elongation complex transcribe through various kinds of impediments on the template. In vitro, TFIIS induces cleavage of the nascent RNA in a stalled elongation complex, and the concomitant backward movement of RNA polymerase is thought to be important for the stimulation of transcriptional readthrough. We recently showed that an activity similar to but distinct from that of TFIIS is also associated with RNA polymerase I.

While searching for TFIIS-related genes, we isolated two cDNA clones that encode novel isotypes of Xenopus TFIIS. Xenopus thus has at least three isotypes of this elongation factor, and we found that these three genes display a tissue- and developmental stage--specific expression pattern. However, all three isoforms interact with the RNA polymerase II elongation complex, and the RNA polymerase I--specific factor remains to be identified.

PUBLICATIONS

Bertoni, R., Sidney, J., Fowler, P., Chisari, F.V., Sette, A. HLA-binding supermotifs predict broadly cross-reactive cytotoxic T lymphocyte responses in patients with acute hepatitis. J. Clin. Invest., in press.

Cavanaugh, V.J., Guidotti, L.G., Chisari, F.V. Interleukin-12 inhibits hepatitis B virus replication in HBV transgenic mice. J. Virol. 71:3236, 1997.

Chang, K.-M., Rehermann, B., Chisari, F.V. Immunopathology of hepatitis C. Semin. Immunopathol., in press.

Chemin, I., Takahashi, S., Belloc, C., Lang, M.A., Ando, K., Guidotti, L.G., Chisari, F.V., Wild, C.P. Differential induction of carcinogen metabolizing enzymes in a transgenic mouse model of fulminant hepatitis. Hepatology 24:649, 1996.

Chisari, F.V. Cytotoxic T Cells and viral hepatitis. J. Clin. Invest. 99:1472, 1997.

Chisari, F.V. Hepatitis B virus pathogenesis: What have we learned from the transgenic mouse? In: 2nd International Meeting on Therapy in Liver Diseases. Arroyo, V. (Ed.). Masson--Little, Brown, Barcelona, Spain, in press.

Chisari, F.V. Immunobiology and pathogenesis of viral hepatitis. In: Proceedings of the IX Triennial International Symposium on Viral Hepatitis and Liver Disease. Vyas, G.N., Buschefelde, M.Z. (Eds.). Rome, 1997, p. 405.

Chisari, F.V., Ferrari, C. Viral hepatitis. In: Viral Pathogenesis. Nathenson, N., et al. (Eds.). Lippincott-Raven, Philadelphia, 1996, p. 745.

Chisari, F.V., Lai, M.M.C. Hepatitis B and virus pathogenesis. In: Persistent Viral Infections. Ahmed, R., Chen, I. (Eds.). Wiley, New York, in press.

Fernandes, C.L., Dong, J.-H., Roebuck, B.D., Chisari, F.V., Montali, J.A., Schmidt, D.E., Jr., Prochaska, H.J. Elevations of hepatic quinone reductase, glutathione, and alpha- and mu-class glutathione S-transferase isoforms in mice with chronic active hepatitis: A compensatory response to injury. Arch. Biochem. Biophys. 331:104, 1996.

Franco, A., Guidotti, L.G., Hobbs, M.V., Pasquetto, V., Chisari, F.V. Pathogenetic effector function of CD4-positive T helper 1 cells in hepatitis B virus transgenic mice. J. Immunol., in press.

Guidotti, L.G., Chisari, F.V.To kill or to cure: Options in host defense against viral infection. Curr. Opin. Immunol. 8:478, 1996.

Guidotti, L.G., Chisari, F.V. The transgenic mouse model of hepatitis B virus infection. In: FORUM: Trends in Experimental and Clinical Medicine, Vol. 6. Frati, L., Santi, L. (Eds.). Scuola Superiore di Oncologia E Scienze Biomediche, Genova, Italy, 1996, p. 189.

Guidotti, L.G., Matzke, B., Chisari, F.V. HBV replication is cell cycle independent during liver regeneration in transgenic mice. J. Virol. 71:4804, 1997.

Guidotti, L.G., Matzke, B., Pasquinelli, C., Shoenberger, J.M., Rogler C.E. Chisari, F.V. The hepatitis B virus precore protein inhibits HBV replication in transgenic mice J. Virol. 70:7056, 1996.

Kaplanski, C., Chisari, F.V., Wild, C.P. Minisatellite rearrangements are increased in liver tumours induced by transplacentally aflatoxin B₁ treatment of hepatitis B virus transgenic mice but not in spontaneously arising tumours. Carcinogenesis 18:633, 1997.

Kuhröber, A., Wild, J., Pudollek, H.-P., Chisari, F.V., Reimann, J. DNA vaccination with plasmids encoding secreted precore (HBeAg) or intracellular core (HBcAg) antigen of hepatitis B virus primes T cell responses to two overlapping K^b- and K^d-restricted epitopes. Int. Immunol., in press.

Labhart, P. Transcript cleavage in an RNA polymerase I elongation complex: Evidence for a dissociable activity similar to but distinct from TFIIS. J. Biol. Chem. 272:9055, 1997.

Livingston, B.D., Crimi, C., Grey, H., Ishioka, G., Becorest, A., Francis, L., Chisari, F.V., Hermanson, G., Fikes, J., Grey, H., Chesnut, R.W., Sette, A. The HBV-specific CTL responses induced by lipopeptide vaccination are comparable to those elicited by acute viral infection. J. Immunol., in press.

Nakamoto, Y., Guidotti, L.G., Pasquetto, V., Schreiber, R.D., Chisari, F.V. Differential target cell sensitivity to cytotoxic T lymphocyte-activated death pathways in vivo. J. Immunol., in press.

Pasquinelli, C., Shoenberger, J., Chung, J., Chang, K.-M., Guidotti, L.G., Selby, M., Berger, K., Lesniewski, R., Houghton, M., Chisari, F.V. Hepatitis C virus core and E2 protein expression in transgenic mice. Hepatology 25:719, 1997.

Penna, A., Artini, M., Cavalli, A., Levrero, M., Bertoletti, A., Pilli, M., Chisari, F.V., Rehermann, B., Prete, G.D., Fiaccadori, F., Ferrari, C. Long-lasting memory T cell responses following self-limited acute hepatitis B. J. Clin. Invest. 98:1185, 1996.

Rehermann, B., Chang, K.-M., McHutchison, J.G., Kokka, R., Houghton, M., Chisari, F.V. Quantitative analysis of the peripheral blood cytotoxic T lymphocyte response, disease activity and viral load in patients with chronic hepatitis C virus infection. J. Clin. Invest. 98:1432, 1996.

Rehermann, B., Chang, K.-M., McHutchison, J., Kokka, R., Houghton, M., Rice C.M., Chisari, F.V. Differential cytotoxic T lymphocyte responsiveness to the hepatitis B and C viruses in chronically infected patients. J. Virol. 70:7092, 1996.

Rehermann, B., Chisari, F.V. Hepatitis B virus immunopathology: Experimental and clinical features. In: Viral Hepatitis. Willson, R.A. (Ed.). Marcel Dekker, New York, 1997, p. 85.

Rehermann, B., Chisari, F.V. The immunology of chronic hepatitis B virus. In: Hepatitis B Virus: Molecular Mechanisms in Disease and Novel Strategies for Therapy. Koshy, R., Caselmann, W. (Eds.). Imperial College Press, London, in press.

Rehermann, B., Ferrari, C., Pasquinelli, C., Chisari, F.V. The hepatitis B virus persists for decades after recovery from acute viral hepatitis despite active maintenance of a cytotoxic T lymphocyte response. Nature Med. 2:1104, 1996.

Top of Page

---

Division of Hematology

Ernest Beutler, M.D., Division Head

---

Phosphatidylserine Exposure in Red Cell Aging

F.E. Boas, L. Forman, E. Beutler

Human red blood cells are normally produced in the bone marrow and after being released into the bloodstream circulate for an average of 120 days. They are then removed from the bloodstream by tissue macrophages. The signal that leads to their recognition by macrophages as senescent cells has not been established. Many possible signals have been proposed, including the binding of antibodies by clustered band 3, loss of deformability, the exposure of a "senescent antigen" that binds circulating antibodies, and loss of ATP.

Most of the hypotheses have been supported chiefly by studies based on separation of cells by density, a method that we do not consider satisfactory for the isolation of old erythrocytes. For a long time, evaluation of these hypotheses was hampered by the lack of methods for in vitro separation of old cells from young cells. G. Dale, Department of Molecular and Experimental Medicine, developed technology for in vivo labeling of red cells in rabbits that allows separation of newly produced (unbiotinylated) cells from older (biotinylated) cells for the first time. This method allowed examination of the existing hypotheses. None of them could be confirmed.

Phosphatidylserine acts as a signal that allows macrophages to detect apoptotic cells. Although red cells, being nonnucleated, cannot undergo apoptosis, we thought that the same signal might be used for their removal. The availability of fluorescently labeled annexin V as a probe for phosphatidylserine and of fluorescently labeled streptavidin as a probe for biotin has made it possible to examine this hypothesis. Rabbits were used in these studies; the mean life span of rabbit red cells is normally about 62 days. We tagged the red cells in rabbits in vivo with NHS-biotin and then obtained samples of blood. The blood samples were treated in vitro with fluorescently labeled streptavidin to detect the biotinylated cells and with fluorescently labeled annexin V to detect the exposed phosphatidylserine. Flow cytometric analysis of red cells obtained at 10-day intervals up to 70 days after biotinylation (Fig. 1) indicated that older, biotinylated cells expose more phosphatidylserine than younger cells do.

We developed a simple model to test the correlation between phosphatidylserine exposure and the removal of biotinylated red cells from the circulation. This model of red cell aging assumes both a random destruction of red cells and an age-dependent destruction of senescent red cells. We assume that the removal of senescent red cells is preceded by a fixed period of phosphatidylserine exposure and that the age of senescent death follows a distribution, a widely used lifetime distribution that successfully models time between inventory restockings, time to recalibration of instruments, and the survival time of irradiated mice. Finally, we assume that red cells are produced and destroyed at a constant rate.

The four parameters of the model--rate of random destruction, mean and standard deviation of the age of senescent death, and duration of phosphatidylserine exposure before the cells are removed from the circulation--were adjusted to provide a least squares fit to the data on phosphatidylserine exposure (Fig. 2). The model provided an excellent fit to the data (r² > 0.98). Moreover, the best fit parameters agreed closely with previously published data (Brown and Eadie, Journal of Physiology, 1953;36:327) derived by using totally different methods (Table 1).

PUBLICATIONS

Baronciani, L., Zanella, A., Bianchi, P., Zappa, M., Alfinito, F., Iolascon, A., Tannoia, N., Beutler, E., Sirchia, G. Study of the molecular defects in glucose phosphate isomerase-deficient patients affected by chronic hemolytic anemia. Blood 88:2306, 1996.

Beutler, E. Disorders of galactose metabolism. In: The Molecular and Genetic Basis of Neurological Disease. Rosenberg, R.N., et al. (Eds.). Butterworth Heinemann, Boston, 1997, p. 1099.

Beutler, E. Disorders of hemoglobin. In: Harrison's Principles of Internal Medicine. Fauci, A.S., et al. (Eds.). McGraw-Hill, New York, 1997, p. 645.

Beutler, E. Gaucher disease. Curr. Opin. Hematol. 4:19, 1997.

Beutler, E. Genetic irony beyond haemochromatosis: Clinical effects of HLA-H mutations. Lancet 349:296, 1997.

Beutler, E. Hair iron content: Possible marker to complement monitoring therapy of iron deficiency in patients with chronic inflammatory bowel disease? Clin. Chem. 43:1099, 1997.

Beutler, E. How little we know about the absorption of iron. Am. J. Clin. Nutr. 66:419, 1997.

Beutler, E. The significance of the 187G (H63D) mutation in hemochromatosis. Am. J. Hum. Genet. 61:762, 1997.

Beutler, E., Carson, D., McMillan, R., Longmire, R. Development of 2-chlorodeoxyadenosine (2-CdA) as a chemotherapeutic and immunosuppressive drug. In: Nucleoside Analogs in Cancer Therapy. Cheson, B.D., Keating, M.J., Plunkett, W. (Eds.). Marcel Dekker, New York, 1997, p. 159.

Beutler, E., Carson, D., Sipe, J.C., Romine, J.S., McMillan, R., Younossi, Z.M., Rodgers, V.D., Koziol, J.A., Zyroff, J. 2-Chlorodeoxyadenosine in nonmalignant disorders. In: Nucleoside Analogs in Cancer Therapy. Cheson, B.D., Keating, M.J., Plunkett, W. (Eds.). Marcel Dekker, New York, 1997, p. 359.

Beutler, E., Gelbart, T. Hematologically important mutations: Gaucher disease. Blood Cells Mol. Dis. 23:2, 1997.

Beutler, E. Gelbart, T. HLA-H mutations in the Ashkenazi Jewish population. Blood Cells Mol. Dis. 23:95, 1997.

Beutler, E., Gelbart, T., West, C., Lee, P., Adams, M., Blackstone, R., Pockros, P., Kosty, M., Venditti, C.P., Phatak, P.D., Seese, N.K., Chorney, K.A., Ten Elshof, A.E., Gerhard, G.S., Chorney, M. Mutation analysis in hereditary hemochromatosis. Blood Cells Mol. Dis. 22:187, 1996.

Beutler, E., Vulliamy, T., Luzzatto, L. Hematologically important mutations: Glucose-6-phosphate dehydrogenase. Blood Cells Mol. Dis. 22:49, 1996.

Beutler, E., West, C. New diallelic markers in the HLA region of chromosome 6. Blood Cells Mol. Dis. 23:219, 1997.

Beutler, E., West, C., Gelbart, T. HLA-H and associated proteins in patients with hemochromatosis. Mol. Med. 3:397, 1997.

Beutler, E., West, C., Gelbart, T., Felitti, V.J. Genetic and clinical description of hemochromatosis probands and heterozygotes: Evidence that multiple genes linked to the major histocompatibility complex are responsible for hemochromatosis. Blood Cells Mol. Dis. 23:145a, 1997. Commentary.

Beutler, E., Westwood, B., van Zwieten, R., Roos, D. G*T transition at cDNA nt 110 (K37Q) in the PKLR (pyruvate kinase) gene is the molecular basis of a case of hereditary increase of red blood cell ATP. Hum. Mutat. 9:282, 1997.

Corey, S.J., Beutler, E. Clinical, genetic, and biochemical features of G-6PD West Virginia. Am. J. Hematol. 53:35, 1996.

Dawson, A., Elias, D.J., Rubenson, D., Bartz, S., Garver, P., Kay, A.C., Bloor, C.M., Beutler, E. Development of pulmonary hypertension after alglucerase therapy in two patients with hepatopulmonary syndrome complicating type 1 Gaucher disease. Ann. Intern. Med. 125:901, 1996.

Le Coutre, P., Demina, A., Beutler, E., Beck, M., Petrides, P.E. Molecular analysis of Gaucher disease: Distribution of eight mutations and the complete gene deletion in 27 patients from Germany. Hum. Genet. 9:816, 1997.

Lee, P.L., Gelbart, T., West, C., Adams, M., Blackstone, R., Beutler, E. Three genes encoding zinc finger proteins on human chromosome 6p21.3: Members of a new subclass of the Krüppel gene family containing the conserved SCAN box domain. Genomics 43:191, 1997.

McCabe, E.R.B., Fine, B.A., Golbus, M.S., Greenhouse, J.B., McGrath, G.L., New, M., O'Brien, W.E., Rowley, P.T., Sly, W.S., Spence, M.A., Stockman, J.A. III, Whyte, M., Wilson, W., Wolf, B., Aerts, J.M.F.G., Barranger, J.A., Barton, N.W., Beutler, E., Brady, R.O., Cox, T.M., Ekstein, J., Eng, C.M., Erikson, A., Findling, D.M. Gaucher disease: Current issues in diagnosis and treatment. JAMA 275:548, 1996.

Medina, M.D., Vaca, G., Lopez-Guido, B., Westwood, B., Beutler, E. Molecular genetics of glucose-6-phosphate dehydrogenase deficiency in Mexico. Blood Cells Mol. Dis. 23:88, 1997.

Niazi, G.A., Adeyokunnu, A., Westwood, B., Beutler, E. Neonatal jaundice in Saudi newborns associated with G6PD Aures. Ann. Trop. Pediatr. 16:33, 1996.

Romine, J.S., Sipe, J.C., Koziol, J.A., Zyroff, J., McMillan, R., Beutler, E. Cladribine: Use in therapy of multiple sclerosis. BioDrugs 7:386, 1997.

Vulliamy, T., Luzzatto, L., Beutler, E. Hematologically important mutations: Glucose-6-phosphate dehydrogenase. Blood Cells Mol. Dis. 23:302, 1997.

Xu, W., Beutler, E. A new exonic polymorphism in the human glucose phosphate isomerase (GPI) gene. Blood Cells Mol. Dis. 23:377, 1997.

Zimmerman, S.A., Ware, R.E., Forman, L., Westwood, B., Beutler, E. Glucose-6-phosphate dehydrogenase Durham: A de novo mutation associated with chronic hemolytic anemia. J. Pediatr. 131:284, 1997.

Zimran, A., Sorge, J., Beutler, E. Gaucher disease, molecular genetics. In: Encyclopedia of Human Biology. Delbecco, R. (Ed.). Academic Press, San Diego, 1997, p. 165.

Top of Page

---

Regulation of Superoxide Production by Human Neutrophils

A.R. Cross, P.G. Heyworth, D.E. Schiff, A. Görlach, B.A. Ellis, D. Noack, P. Lee, J.T. Curnutte

Neutrophils are the most numerous phagocytes in the circulating blood and are the first cells to arrive at sites of infection. At these sites, neutrophils ingest invading microorganisms and sequester them within phagocytic vacuoles. Concomitantly, neutrophils activate a highly specialized metabolic pathway, the respiratory burst NADPH oxidase, that results in the production of extremely reactive derivatives of oxygen, such as superoxide, hydrogen peroxide, and hypochlorous acid. These agents have powerful antimicrobial activity and are highly effective in destroying life-threatening pathogens. In contrast to this beneficial role, when released from the cell, the agents are equally capable of causing serious damage to surrounding normal tissues in inflammatory diseases such as ischemia-reperfusion injury and adult respiratory distress syndrome. It is not surprising, therefore, that neutrophils have an elaborate regulatory mechanism that guards against the inappropriate activation of this pathway but still allows its rapid activation when virulent microbes are present.

One of the two major goals of our laboratory is to elucidate the biochemical details of the mechanisms that regulate NADPH oxidase, the phagocytic enzyme that generates superoxide. As we understand this pathway better, it may be possible to develop drugs capable of enhancing its activity in such a way that the antimicrobial efficiency is improved. Conversely, agents that depress the production of oxygen radicals could be used clinically as a new class of antiinflammatory drugs. The other major goal is to understand the molecular basis of chronic granulomatous disease (CGD), in which superoxide generation by NADPH oxidase is absent or severely diminished, and of related phagocytic disorders in which the production of oxygen radicals is defective.

REGULATION OF NADPH OXIDASE ACTIVITY

NADPH oxidase becomes catalytic when the cytosolic proteins, p47^phox and p67^phox, become associated with the plasma membrane and form a multimeric complex with the two subunits of cytochrome b₅₅₈ (gp91^phox and p22^phox). In addition to heme, the cytochrome probably contains an NADPH-binding site and a flavin group and serves as the electron-carrying component of the system, transferring electrons from NADPH to oxygen. Together with a recently identified protein termed p40^phox, the components p47^phox and p67^phox exist in a 260-kD complex in the cytosol of resting neutrophils. The three proteins translocate to the membrane in a process mediated by interactions between the p47^phox subunit and cytochrome b₅₅₈.

We are actively trying to determine the components of the 260-kD cytosolic complex, additional proteins with which the components interact, and the domains in cytochrome b₅₅₈ responsible for the docking of this complex. In addition to the components already described, the activity of NADPH oxidase is also dependent on Rac, a GTP-binding protein of the Ras superfamily. Current studies are directed toward determining the role of Rac2 in regulating the assembly and activation of human neutrophil NADPH oxidase.

In collaboration with J.A. Badwey, Boston Biomedical Research Institute, we are studying the role of protein phosphorylation in NADPH oxidase activation. This study has broadened in scope to include cofilin, a ubiquitous actin-binding protein and a potent depolymerizing agent of filamentous actin. Cofilin is phosphorylated in resting neutrophils and rapidly dephosphorylated when the cells are stimulated to generate superoxide. Using immunofluorescence, we showed that in activated cells cofilin colocalizes with filamentous actin in membrane ruffles and lamellipodia, which are also major sites of NADPH oxidase activity. Our current hypothesis is that cofilin forms a link, long known to exist but never identified, between the cytoskeleton and the phagocyte NADPH oxidase.

Studies are also under way to determine the structure of cytochrome b and the way in which electron transfer between the flavin and heme redox centers is regulated by the cytosolic factors. We discovered that the cytochrome contains two separate heme centers with closely spaced redox potentials. We have also found that p47^phox and p67^phox have distinct roles in controlling electron flow from NADPH to oxygen, by regulating electron flow both from NADPH to flavin and from flavin to heme.

These studies suggest multiple ways in which the activity of NADPH oxidase could be artificially controlled. In addition to its capacity to generate superoxide, the oxidase has a mechanism for efficiently transferring protons generated internally from NADPH oxidation to the exterior of the cell to maintain pH homeostasis. We are investigating the mechanisms by which this process can occur and how it is controlled.

MOLECULAR GENETICS OF CHRONIC GRANULOMATOUS DISEASE

Substantial progress has been made both in defining the molecular basis of CGD and in developing improved ways of treating the disease. The CGD phenotype of absent superoxide production can result from genetic lesions affecting any one of four oxidase components (p47^phox, p67^phox, gp91^phox, or p22^phox). Patients with CGD are highly susceptible to life-threatening microbial infections. More than 120 CGD patients have been assessed in the Scripps General Clinical Research Center. Of these, more than 100 have been characterized at the molecular genetic level, and a wide variety of mutations have been detected. During the course of these studies we found that control subjects who do not have CGD carry not only a normal copy of the gene for p47^phox but also a nonfunctional pseudogene. It appears that recombination events between the gene and the pseudogene account for most, if not all cases of CGD in which p47^phox is affected.

Mutation analysis has enabled us to provide prenatal diagnosis to a limited number of pregnant female relatives of CGD patients. In conjunction with the Scripps Reference Laboratory, we are developing clinical assays for the diagnosis of CGD and for the detection of specific mutations in patients and families affected by the X-linked (gp91^phox-deficient) form of the disease.

In other clinical work, following our findings that recombinant IFN- decreases the occurrence of serious infections in CGD patients by 70%, we are investigating the mechanisms by which this cytokine mediates this beneficial effect. The improvement in phagocytic function is not due to a reversal in the underlying defect in superoxide production. This finding is not surprising, because most CGD mutations are in the coding regions of oxidase components. In a study with healthy human volunteers who do not have CGD, we showed that in vivo treatment with recombinant IFN- is associated with increased expression by neutrophils of FcRI, the high-affinity Fc receptor for IgG. This increased expression shows a strong correlation with the improved ability of neutrophils to phagocytose bacteria in an Fc receptor--mediated manner. These studies may provide insights into other clinical uses for recombinant IFN-.

PUBLICATIONS

Adachi, S., Cross, A.R., Babior, B.M., Gottleib, R.A. Bcl-2 and the outer mitochondrial membrane in the inactivation of cytochrome c during Fas-mediated apoptosis. J. Biol. Chem., in press.

Chanock, S.J., Roesler, J., Hopkins, P., Lee, P., Barrett, D.T.,Christensen, B.L., Curnutte, J.T., Görlach, A. Characterization of the genomic structure of the p47-phox gene. Blood, in press.

Cross, A.R., Curnutte, J.T., Heyworth, P.G. Hematologically important mutations: The autosomal-recessive forms of chronic granulomatous disease. Blood Cells Mol. Dis. 22:268, 1997.

Görlach, A., Lee, P., Roesler, J., Hopkins, P.J., Green, E.D., Chanock, S.J., Curnutte, J.T. The p47-phox gene has a pseudogene carrying the most common mutation for p47-phox deficient chronic granulomatous disease. J. Clin. Invest. 100:1907, 1997.

Heyworth, P.G., Robinson, J.M., Ding, J., Ellis, B.A., Badwey, J.A. Cofilin undergoes rapid dephosphorylation in stimulated neutrophils and translocates to ruffled membranes enriched in products of the NADPH oxidase complex: Evidence for a novel cycle of phosphorylation and dephosphorylation. Histochem. Cell Biol., in press.

Roos, D., Curnutte, J.T. Chronic granulomatous disease. In: Primary Immunodeficiency Diseases: A Molecular and Genetic Approach. Ochs, H., Puck, J., Smith, E. (Eds.). Oxford University Press, New York, in press.

Schiff, D.E., Rae, J., Martin, T.R., Davis, B.H., Curnutte, J.T. Increased phagocyte FcRI expression and improved Fc-receptor-mediated phagocytosis following in vivo recombinant human interferon- treatment of normal human subjects. Blood, in press.

Youngson, C., Nurse, C., Yeger, H., Curnette, J.T., Vollmer, C., Wong, V., Cutz, E. Immunocytochemical localization of O₂-sensing protein (NADPH oxidase) in chemoreceptor cells. Microsc. Res. Tech. 37:101, 1997.

Zeigelstein, R.C., Zheng, G., Wang, P., Dodd, J.M., Ocampo, C.J., Silverman, H.S., Cross, A.R., Zweier, J.L. Vasoconstriction mediated by vascular NADPH oxidase-derived superoxide. Circ. Res., in press.

Top of Page

---

Division of Oncovirology

Peter K. Vogt, Ph.D., Division Head

---

Molecular Genetics of Cancer

P.K. Vogt, J. Li, M. Aoki, H.W. Chang, B. Freyaldenhoven, M. Freyaldenhoven, S.-L. Fu, K. Hollen, B.-H. Jiang, U. Kruse, M. Nishizawa, B. Nowakowski, S. Scheidler

Most cancers are due to genetic changes in somatic cells. These changes affect both positive and negative regulators of growth: oncogenes and tumor suppressor genes, respectively. In cancer, oncogenes usually undergo a gain of function and act in a dominant fashion. The tumor suppressor genes suffer loss of function; they act as recessive control elements of growth. Work in the Division of Oncovirology concentrates on both types of cancer genes, with an emphasis on the positive regulators, the oncogenes.

Oncogenes code for components of cellular signal transduction. These signals originate at the periphery of the cell, traverse the cytoplasm, and end in the nucleus. Transit through the cytoplasm proceeds along sequential and programmed steps of protein-protein interactions that are mediated by modular protein recognition domains and are driven by protein phosphorylation. The final recipient of the signal in the nucleus is a transcription factor that converts the signal into a pattern of gene expression by upregulating or downregulating specific genes.

Transcription factors, whether carried by a retrovirus or expressed from the cellular genome, have oncogenic potential. Because of their role as signal converters, transcription factors occupy a privileged place in cellular signaling. This place defines the promise and the problems of oncogenic transcription factors. The promise is an understanding of oncogenesis at the level of gene regulation and gene expression. The problems arise from our ignorance of what lies downstream of transcription factors. We do not know the target genes whose upregulation or downregulation controls the oncogenic phenotype of the cell, and we do not know the protein-protein interactions that transmit the signal from the DNA-bound transcription factor to the basic transcriptional machinery. These questions define the principal challenges in transcription and cancer.

TARGETS OF JUN

Oncogenic versions of the transcriptional activator Jun induce cancer by upregulating and possibly downregulating various specific target genes. A subset of these targets controls the oncogenic phenotype of the cancer cell. We have isolated several genes that are differentially expressed in Jun-transformed cells. Some of these show homology to growth regulatory genes. To characterize these Jun targets further, we constructed a regulatable Jun that transactivates and induces oncogenic transformation only in the presence of a steroid hormone. Turning on this regulatable Jun induces the expression of the target genes.

Direct and indirect targets are distinguished by their sensitivity to inhibitors of protein synthesis during induction. Direct targets can be induced without de novo protein synthesis. In contrast, indirect targets depend on prior synthesis of a protein, because they are targets of targets. Currently, the set of known targets of oncogenic jun is small and incomplete. We are adapting more comprehensive techniques for the isolation and characterization of these targets and will test targets for their functional role in oncogenic transformation.

TRANSCRIPTIONAL REPRESSION BY QIN AND ONCOGENICITY

The oncogene qin was isolated from an avian retrovirus. Its product, the protein Qin, is the avian homolog of brain factor I, a developmental regulator belonging to the winged-helix family of transcription factors. Qin represses transcription. The oncogenic viral version of the protein is a more active repressor than its nononcogenic cellular counterpart. An important role of transcriptional repression in oncogenic transformation is strongly suggested by the properties of chimeric constructs that fuse the Qin DNA-binding domain either to the heterologous repressor domain derived from the Drosophila Engrailed protein or to the activator domain of the herpes simplex virus protein VP16. The Qin-Engrailed chimera represses transcription and is oncogenic. The Qin-VP16 chimera activates transcription and not only does not induce oncogenic transformation but also can actively interfere with transformation induced by Qin.

This correlation between transcriptional repression and oncogenicity makes it important to determine target genes of Qin that are downregulated and induce the oncogenic phenotype of the cell. These targets are tumor suppressors. As with other oncogenic transcription factors, the protein-protein interactions that mediate repression are also still unknown for Qin. The determination and isolation of Qin-binding proteins will be important for the understanding of Qin-induced oncogenicity.

EFFECT OF THREE NEW CHICKEN WINGED-HELIX PROTEINS ON CELL GROWTH

We have isolated three new winged-helix transcription factor genes from the chicken genome. They are related to qin and to other known members of the winged-helix family. Overexpression of these genes in chicken embryo fibroblasts leads to a high saturation density in cell culture, and two of the genes also stimulate formation of agar colonies by the overexpressing cells. This growth-promoting activity can be turned into frank oncogenicity by mutation. Studying the nature of these oncogenic mutations and their effects on DNA binding and transcriptional regulation will reveal important aspects of oncogenic transformation induced by these transcription factors.

Interestingly, the three new chicken winged-helix proteins all act as transcriptional repressors, and the repressor domain of one of these proteins has now been mapped. It coincides with a repressor domain that occupies the C-terminal part of Qin.

ONCOGENICITY OF PHOSPHATIDYLINOSITOL-3-KINASE

We have cloned the genome of ASV 16, an avian retrovirus that induces hemangiosarcomas in chickens, and from this replication-defective retroviral genome, we have isolated a new oncogene that shows close homology to the catalytic subunit of the cellular phosphatidylinositol-3-kinase (PI 3-kinase). The new oncogene, termed p3k, has been expressed from the retroviral vector RCAS. It induces oncogenic transformation in cell culture and hemangiosarcomas in chickens.

The viral version of p3k shows several mutations compared with its cellular counterpart. These probably account for the enhanced oncogenicity of the viral gene. The cellular p3k gene is also oncogenic when overexpressed from the retroviral vector, but much less so than viral p3k. The viral version of p3k is fused at its N-terminus to gag sequences of the transducing retrovirus. The fusion protein acts as a strong phosphoinositide 3-kinase. Cells transformed by p3k have elevated levels of phosphatidylinositol-3,4-bisphosphate and phosphatidylinositol-3,4,5-trisphosphate. A known downstream target of PI 3-kinase, the serine kinase Akt, is activated in p3k-transformed cells.

PI 3-kinase lies at a nodal point of cellular signaling. Signals that involve or pass through this kinase affect numerous cellular activities. However, its oncogenic effect is highly specific, limited to the induction of a single tumor type, the hemangiosarcoma. Work in progress focuses on the interactions of p3k with angiogenic mechanisms of the host and on the mutations that activate the oncogenic potential of PI 3-kinase.

PUBLICATIONS

Chang, H.W., Aoki, M., Fruman, D., Auger, K.R., Bellacosa, A., Tsichlis, P.N., Cantley, L.C., Roberts, T.M., Vogt, P.K. Transformation of chicken cells by the gene encoding the catalytic subunit of PI 3-kinase. Science, in press.

Chang, H.W., Li, J., Vogt, P.K. Domains of the Qin protein required for oncogenic transformation. Oncogene 13:441, 1996.

Freyaldenhoven, B.S., Freyaldenhoven, M.P., Iacovoni, J.S., Vogt, P.K. Aberrant cell growth induced by avian winged helix proteins. Cancer Res. 57:123, 1997.

Freyaldenhoven, B.S., Freyaldenhoven, M.P., Iacovoni, J.S., Vogt, P.K. Avian winged helix proteins CWH-1, CWH-2, and CWH-3 repress transcription from Qin binding sites. Oncogene, in press.

Gao, M., Morgan, I.M., Vogt, P.K. Differential and antagonistic effects of v-Jun and c-Jun. Cancer Res. 56:4229, 1996.

Qian, Z., Brunovskis, P., Lee, L., Vogt, P.K., Kung, H.-J. Novel DNA-binding specificities of a putative herpesvirus bZIP oncoprotein. J. Virol. 70:7161, 1996.

Scheidler, S., Fredericks, W.J., Rauscher, F.J. III, Barr, F.G., Vogt, P.K. The hybrid PAX3-FKHR fusion protein of alveolar rhabdomyosarcoma transforms fibroblasts in culture. Proc. Natl. Acad. Sci. U.S.A. 93:9805, 1996.

Vogt, P.K. Gene regulation and cancer. In: Challenges in Medicine. Herausforderungen der Medizin Symposium aus Anlaß des 100 Geburtstages von Ernst Jung. Siegenthaler, W., Haas, R. (Eds.). Thieme Verlag, Stuttgart, 1996, p. 144.

Vogt, P.K. Peyton Rous: Homage and appraisal. FASEB J. 10:1559, 1966.

Vogt, P.K., Chang, H.-W., Goller, M.E., Iacovoni, J.S., Kruse, U., Li, J., Scheidler, S. Onkogene: Fortschritte in Zytoplasma und Nukleus. Acta Leopoldina, in press.

Top of Page

---

Division of Research Rheumatology

W.M. Keck Autoimmune Disease Center
Eng M. Tan, M.D., Division Head

---

Autoantibodies in Chronic Fatigue Syndrome

K. Konstantinov, A. von Mikecz, J. Jones,^* D. Buchwald,^** L. Gerace, E.M. Tan

^* National Jewish Hospital and Research Center, Denver, CO
^** University of Washington, Seattle, WA

Chronic fatigue syndrome (CFS) is characterized by prolonged debilitating fatigue and numerous nonspecific signs and symptoms. Currently, no established cause or pathophysiologic abnormality is associated with the syndrome, and no definitive diagnostic test or proven treatment is available. The diagnosis is ascertained by exclusion of many other diseases known to be associated with chronic fatigue. However, criteria for the diagnosis of CFS were established at an international meeting held at the Centers for Disease Control and Prevention, Atlanta, Georgia, in 1994. Some evidence suggests that patients with CFS have a low-grade chronic activation of the immune system.

Our studies have shown that CFS patients have relatively low titers of autoantibodies primarily to a group of insoluble cytoskeletal and karyoskeletal antigens operationally defined as matrix-associated proteins. This finding is in distinct contrast to findings in classical autoimmune diseases such as systemic lupus erythematosus, scleroderma, and Sjögren's syndrome. In these three diseases, titers of autoantibodies are much higher, and the antibodies are directed at different intracellular antigens.

Three major antigen-antibody systems in CFS have been characterized. Most patients with CFS have autoantibodies to the nuclear envelope protein lamin B1, the cytoskeletal protein vimentin, and a nuclear matrix antigen that colocalizes with a group of proteins involved in pre-mRNA splicing. Our studies on these autoantibodies are being done in collaboration with investigators at the University of Washington in Seattle and at Harvard University. We are developing a quantitative assay for serologic markers that may help to distinguish patients with CFS from persons without CFS and from patients with other diseases with features of chronic fatigue.

Top of Page

---

Differential Cleavage of Nuclear Autoantigens During Apoptosis and Necrosis

C.A. Casiano, S.J. Martin,^* D.R. Green,^* R.L. Ochs, E.M. Tan
^* La Jolla Institute of Allergy and Immunology, La Jolla, CA

Apoptosis and necrosis are morphologically different pathways of cell death. A central mechanism in apoptosis is activation of the ICE/CED-3 protease family. Although protease activation has been implicated in necrosis, the proteases involved and the specific substrates cleaved during the early stages of necrosis are unknown.

Using autoantibodies from patients with systemic autoimmune diseases as probes, we observed that a defined and somewhat similar set of nuclear protein autoantigens, including poly(ADP-ribose)-polymerase and topoisomerase 1, are cleaved into different molecular size fragments during apoptosis and necrosis of cultured cells induced by various stimuli (Fig. 1). The resulting cleavage products can be clearly distinguished in the two different cell death pathways. Cleavage of these nuclear proteins is blocked by inhibitors of ICE/CED-3 proteases during apoptosis but not during necrosis. These findings indicate that spontaneously occurring autoantibodies of defined specificities can be used to distinguish between these two types of cell death. Studies are in progress to define proteolytic mechanisms associated with necrosis and to explore the possible role of autoantigen cleavage in the induction of autoimmune responses.

Top of Page

---

Autoimmune Response and Autoimmunity in Malignant Neoplasms

W.-G. Zhu, J.-Y. Zhang, G. Covini,^* E.K.L. Chan, M. Nishioka,^** S.I. Reed, E.M. Tan

^* Istituto Clinico Humanitas, Milan, Italy
^** Kagawa Medical University, Kagawa, Japan

The nuclear autoantigen SG₂NA was detected by using autoantibodies from a patient with bladder and lung cancer. Expression of this nuclear protein is associated with the cell cycle and occurs primarily in the S and G₂ phases. A full-length cDNA of 713 amino acids of the antigen was isolated. The carboxyl-terminal half of SG₂NA contains six copies of WD-40 motifs characteristic of a large family of proteins with diverse biological cell functions. We showed that the amino-terminal half of SG₂NA (amino acids 1--391) acts as a strong transcriptional activator when fused to the GAL-4 DNA-binding domain in yeast with the LACZ and HIS reporter system. No transcriptional activity was found in the carboxyl-terminal half (amino acids 392--713).

The ability of the N-terminal half of SG₂NA to activate transcription was confirmed in the COS-7 mammalian cell line by cotransfection experiments in which the expression of the GAL-4 fusion protein activated the luciferase reporter gene. Further studies are under way to determine how this ability to activate transcription might be related to the S and G₂ phases of the cell cycle and how it might be involved in malignant transformation.

In previous studies, a nuclear autoantigen called HCC1 was detected and cloned. This nuclear protein contains serine-arginine motifs found in a family of alternative splicing factors. Recently, using sera from Chinese patients with hepatocellular carcinoma, we detected a 90-kD cytoplasmic antigen. A cDNA clone encoding part of the 90-kD protein was isolated by using antibodies from these patients. The detection of two RNA-binding domains (consensus RBD and KH domains) at the N-terminal region suggests that this protein might be involved in RNA binding and processing. Further characterization of this novel cytoplasmic antigen related to hepatocellular carcinoma by way of an autoimmune reaction is in progress.

PUBLICATIONS

Andrade, L.E.C., Chan, E.K.L., Peebles, C.L., Tan, E.M. Two major autoantigen-antibody systems of the mitotic spindle apparatus. Arthritis Rheum. 39:1643, 1996.

Baez, A., Torres, K., Tan, E.M., Pommier, Y., Casiano, C.A. Expression of proliferation-associated nuclear autoantigens, p330^d/CENP-F and PCNA, in differentiation and in drug-induced growth inhibition using two-parameter flow cytometry. Cell Prolif. 29:183, 1996.

Casiano, C.A., Martin, S., Green, D.R., Tan, E.M. Selective cleavage of nuclear autoantigens during CD95 (Fas/APO-1)-mediated T cell apoptosis. J. Exp. Med. 184:765, 1996.

Covini, G., Chan, E.K.L., Nishioka, M., Morshed, S.A., Reed, S.I., Tan, E.M. Immune response to cyclin B1 in hepatocellular carcinoma. Hepatology 25:75, 1997.

Furuta, K., Chan, E.K.L., Kiyosawa, K., Reimer, G., Luderschmidt, C., Tan, E.M. Heterochromatin protein HP1^HSB (p25ß) and its localization with centromeres in mitosis. Chromosoma, in press.

Konstantinov, K. Nuclear envelope protein autoantibodies. In: Autoantibodies. Peters, J.B., Schoenfeld, Y. (Eds.), Elsevier, New York, 1996, p. 561.

Landberg, G., Erlanson, M., Roos, G., Tan, E.M., Casiano, C.A. The nuclear autoantigen p330d/CENP-F: A marker for cell proliferation in human malignancies. Cytometry 25:90, 1996.

Top of Page

---

Cytoskeletal Regulation of IgE-Mediated Degranulation in Mast Cells

J.R. Apgar, L.G. Frigeri, T. Kawakami^*

^* La Jolla Institute of Allergy and Immunology, La Jolla, CA

The major purpose of our work is to determine the role of actin microfilaments in the regulation of degranulation in mast cells, which are involved in a variety of allergic disorders. These cells express an Fc receptor on the cell surface that binds to IgE with high affinity. In this state, the cells are sensitized but inactive. Exposure to multivalent antigen causes the cross-linking of the sensitized receptors, the critical event in the activation of mast cells. Activation leads to degranulation, the release of preformed mediators such as histamine, and the de novo production of eicosanoids and cytokines. All these mediators are important in the inflammatory response.

In our studies, we use rat basophilic leukemia cells. These cells originated from mucosal mast cells and are one of the major models for studying IgE-mediated allergic reactions. Rat basophilic leukemia cells adhere to fibronectin. Adherence causes polymerization of actin, which in turn leads to spreading of the cells. Adherence does not cause any degranulation on its own, but it does greatly augment IgE-mediated responses. We have determined that it is the spreading of the cells and not simply adhesion that is important in the increased activation of the cells. Stimulation of spread cells also causes increases in the activity of phospholipases A₂, C, and D. We are investigating the activation status of three tyrosine kinases (Lyn, Syk, and Btk), which are among the earliest signaling molecules used.

Cross-linking of the receptors also causes further polymerization of actin. Unlike the microfilaments formed on adhesion, this actin is probably involved in downregulation or desensitization of the receptors. We are interested in the signaling pathways leading to IgE-mediated formation of microfilaments. Antigen-induced polymerization of actin appears to be biphasic. The early phase is short-lived and requires activation of protein kinase C. Currently, we are studying the upstream signals that may involve the sphingomyelinase pathway.

The late phase depends on tyrosine kinases, which then activate phospholipase D. This activation leads to the production of phosphatidic acid, which is converted into diacylglycerol. This conversion in turn activates protein kinase C. Protein kinase C is then responsible either directly or indirectly for the activation of phosphatidylinositol 4-kinase and phosphatidylinositol-4-phosphate 5-kinase, which produce phosphatidylinositol monophosphate and phosphatidylinositol bisphosphate, respectively. Increased production of these polyphosphoinositides, which can dissociate actin from many actin capping proteins, correlates well with actin polymerization.

PUBLICATIONS

Apgar, J.R. Increased degranulation and phospholipase A₂, C, and D activity in RBL cells stimulated through FcR1 is due to spreading and not simply adhesion. J. Cell Sci. 110:771, 1997.

Top of Page

---

Molecular and Cell Biology of Autoantigens

E.K.L. Chan, D. Wang, M. Lu, J.C. Hamel, M.J. Fritzler,^* J.P. Buyon^**

^* University of Calgary, Calgary, Alberta
^** Hospital for Joint Diseases, New York, NY

---

CARDIAC EXPRESSION OF 52ß

In the fetus, complete heart block associated with maternal autoantibodies to 60-kD SS-A/Ro, 52-kD SS-A/Ro, and 48-kD SS-B/La is most often detected between 18 and 24 weeks of gestation and is considered a consequence of the transplacental passage of these antibodies into the fetal circulation, which results in tissue injury. Despite their exposure to the identical antibodies, complete heart block does not occur in the mothers. We recently reported that an alternatively spliced 52-kD SS-A/Ro mRNA, derived from the skipping of exon 4 encoding amino acid residues 168--245 inclusive of the leucine zipper, that encodes a smaller protein, 52ß, with a predicted size of 45 kD, was recognized by autoantibodies in 26 of 30 mothers of fetuses with complete heart block. To identify any potential relationship between the expression of 52ß and full-length 52 in the development of complete heart block in human fetuses, we used reverse transcription and polymerase chain reaction with primers flanking exon 4 and mRNA from the hearts of 22 fetuses and three adults.

Expression of 52ß was greatest between 14 and 16 weeks of gestation. Specifically, between 11 and 13 weeks, expression of 52 mRNA was dominant. In contrast, expression of 52ß was dominant in one of two hearts at 14 weeks and in one of two hearts at 15 weeks. Equal expression of 52 and 52ß was observed in two of three hearts at 16 weeks. Although 52ß was still readily detectable at 17 and 18 weeks, expression of 52 was greater. By 22--25 weeks, the 52ß transcript was decreased, and 52 clearly dominated. This last pattern was also observed in the three hearts obtained from adults.

We found 52ß mRNA in the brain, kidney, lung, and spleen of a 15-week-old fetus, but its expression relative to 52 was greatest in the heart. In 18- and 24-week-old fetuses, expression of 52ß relative to 52 was decreased in all tissues except the kidney and brain. By comparison, we detected no differences in the expression of 60-kD SS-A/Ro or SS-B/La mRNA in any fetal tissue or adult heart.

In conclusion, the expression of the alternatively produced 52ß is greatest at the time of cardiac ontogeny, when maternal antibodies gain access to the fetal circulation just before the clinical detection of complete heart block. Although other fetal tissues express 52ß, differences may exist in the accessibility of antigen or in regenerative capacities. Our data implicate a possible role for 52ß in the development of complete heart block. Further work is needed to address the significance of these observations.

A NOVEL 97-kD GOLGI COMPLEX AUTOANTIGEN ASSOCIATED WITH SJÖGREN'S SYNDROME

The serum from a patient with Sjögren's syndrome with antibodies to the Golgi complex was used as a probe to isolate a 3.7-kb cDNA insert from a HeLa cDNA library. The encoded 56-kD recombinant protein was immunoprecipitated by the human anti-Golgi serum and by immune rabbit serum but not by normal human serum or preimmune rabbit serum. Western blot analysis showed that the immune rabbit sera recognized a cellular protein of 97 kD (golgin-97), suggesting that the isolated clone contained a partial cDNA. The 5´ upstream sequence was obtained by using rapid amplification of the cDNA end, and the combined cDNA contained 4860 bp encoding a protein of 88 kD. Like golgin-245, golgin-97 has a granin signature sequence that suggests a role associated with secretory granules. Antibodies to golgin-97 were found in 12 (20%) of 60 sera with autoantibodies to the Golgi complex, and the majority of the serum donors (8 of 12 or 75%) had Sjögren's syndrome. Of four cloned Golgi complex autoantigens analyzed to date, golgin-97 is the most common target associated with Sjögren's syndrome. Although autoantibodies to the Golgi complex are rare in patients with Sjögren's syndrome, detection of these antibodies may indicate a subgroup of patients with interesting defects associated with the function of the Golgi complex.

PUBLICATIONS

Andrade, L.E.C., Chan, E.K.L., Peebles, C.L., Tan, E.M. Two major autoantigen-autoantibody systems of the mitotic spindle apparatus. Arthritis Rheum. 39:1643, 1996.

Buyon, J.P., Tseng, C.-E., Di Donato, F., Rashbaum, W., Morris, A., Chan, E.K.L. Cardiac expression of 52ß, an alternative transcript of the congenital heart block-associated 52-kD SS-A/Ro autoantigen, is maximal during fetal development. Arthritis Rheum. 40:655, 1997.

Chan, E.K.L. RACE to power cloning. Trends Genet. 13:288, 1997.

Chan, E.K.L., Pollard, K.M. Detection of autoantibodies to ribonucleoprotein particles by immunoblotting. In: Manual of Clinical Laboratory Immunology, 5th ed. Rose, N.R., et al. (Eds.). American Society for Microbiology Press, Washington, DC, 1997, p.928.

Covini, G., Chan, E.K.L., Nishioka, M., Ahmed Morshed, S., Reed, S.I., Tan, E.M. Immune response to cyclin B1 in hepatocellular carcinoma. Hepatology 25:75, 1997.

Covini, G., von Mühlen, C.A., Pacchetti, S., Colombo, M., Chan, E.K.L., Tan, E.M. Diversity of antinuclear antibody responses in hepatocellular carcinoma. J. Hepatol. 26:1255, 1997.

Furuta, K., Chan, E.K.L., Kiyosawa, K., Reimer, G., Luderschmidt, C., Tan, E.M. Heterochromatin protein HP1^Hsß (p25ß) and its localization with centromere in mitosis. Chromosoma 106:11, 1997.

Furuta, K., Matsuoka, S., Hildebrandt, B., Kiyosawa, K., Reimer, G., Luderschmidt, C., Chan, E.K.L., Tan, E.M. Immunologic characterization of heterochromatin protein p25ß autoantibodies and relationship with centromere autoantibodies and pulmonary fibrosis in systemic scleroderma. J. Mol. Med., in press.

Griffith, K.J., Chan, E.K.L., Lung, C.C., Hamel, J.C., Guo, X.Y., Miyachi, K., Fritzler, M.J. Molecular cloning of a novel 97 kDa Golgi complex autoantigen associated with Sjögren's syndrome. Arthritis Rheum. 40:1693, 1997.

Lung, C.C., Chan, E.K.L. Simplifying and optimizing 5´ RACE in the hunt for full-length cDNAs. Trends Genet. 12:389, 1996.

Lung, C.C., Chan, E.K.L., Zuraw, B.L. Analysis of an exon 1 polymorphism of the B2 bradykinin receptor gene and its transcript in normal subjects and patients with C1 inhibitor deficiency. J. Allergy Clin. Immunol. 99:134, 1997.

Matera, A.G., Wu, W., Imai, H., O'Keefel, C.L., Chan, E.K.L. Molecular cloning of the RNA polymerase I transcription factor hUBF/NOR-90 (UBTF) gene and localization to 17q21.3 by fluorescence in situ hybridization and radiation hybrid mapping. Genomics 41:135, 1997.

Tseng, C., Chan, E.K.L., Miranda, E., Gross, M., Di Donato, F., Buyon, J.P. The 52kD protein is a target of intermolecular spreading of the immune response to components of the SS-A/Ro-SS-B/La complex. Arthritis Rheum. 40:936, 1997.

Wang, D., Buyon, J.P., Chan, E.K.L. Cloning and expression of mouse 60-kDa ribonucleoprotein SS-A/Ro. Mol. Biol. Rep. 23:205, 1996.

Top of Page

---

Human Autoantibodies: Biological Markers of Disease

R.L. Ochs, F. Alcantara, C.L. Parsons,^* Y. Muro,^** D. Iglehart,^*** E.K.L. Chan, E.M. Tan

^* University of California, San Diego, CA
^** Nagoya University School of Medicine, Nagoya, Japan
^*** Duke University Medical Center, Durham, NC

We use human autoantibodies as probes for investigating basic cell biology and as immunologic tools for studying the autoimmune response in various diseases.

AUTOANTIBODIES IN INTERSTITIAL CYSTITIS AND ATOPIC DERMATITIS

Interstitial cystitis literally means bladder wall (interstitial) inflammation (cystitis). Interstitial cystitis is a chronic inflammatory disease characterized by pain, frequent urination (small-capacity bladders), and focal ulcers (so-called Hunner's ulcers). The disease affects primarily females (90%), and the average age at onset is 30--45 years. The etiology of interstitial cystitis is unknown but may involve bacteria, viruses, toxic substances, physical trauma, or autoimmune mechanisms. Any one of these factors or agents can cause the chronic inflammation characteristic of this disease.

Previously, we reported that 36% of patients with interstitial cystitis were seropositive for antinuclear antibodies. A number of the positive sera had a pattern of antibody distribution (nucleoplasmic dense fine speckles) that represented reactivity with a 70-kD protein, DFS70, by Western blotting. These findings indicate that DFS70 may be a major interstitial cystitis autoantigen, with autoantibodies to it appearing in up to 25% or more of all patients with interstitial cystitis.

While investigating various chronic inflammatory conditions for the presence of autoantibodies, we found that 26 (41%) of 64 patients with the inflammatory skin disease atopic dermatitis had antinuclear antibodies and 18 (28%) had autoantibodies to DFS70. Atopic dermatitis is characterized as a chronic, relapsing, pruritic skin disorder that generally first appears in childhood and often progresses to asthma or allergic rhinitis in adulthood. The disease has no objective laboratory marker, but patients with this dermatitis often have elevated serum levels of IgE and allergic reactivity to foods and other common allergens such as pollens, molds, and insects. The prevalence of atopic dermatitis appears to be on the increase; 10--15% of the population are affected at some time during childhood. These studies show that DFS70 is a major autoantigen targeted by autoantibodies in both atopic dermatitis and interstitial cystitis and indicate that common disease mechanisms may exist for these two seemingly unrelated conditions.

AUTOANTIBODIES IN BREAST CANCER

The search for autoantibodies in sera from women with breast cancer revealed an immunologic response to specific autoantigens. This response may reflect the presence of molecules altered by mutation or overexpression whose function may be important in the development or maintenance of cancer. Sera from 158 patients with breast cancer were initially screened by using indirect immunofluorescence and immunoblotting against T47D whole-cell extracts. Autoantibodies were detected in 78 (49%) of the 158 sera studied to date. According to the immunofluorescence patterns, the distribution of the autoantibodies was classified as small nuclear bodies (6 patients), cytoplasmic dots (13), nuclear matrix (11), nucleolus (13), mitochondria (6), large nuclear bodies (3), mitotic apparatus (2), centrioles (6), dense fine speckles (3), Golgi complex (2), and miscellaneous (3).

These autoantibodies may be targeting important proteins known to be involved in cancer, such as the p53 tumor suppressor protein, but most remain to be fully characterized by further studies. A few patients had autoantibodies that recognized the proliferation-associated proteins NuMA and PCNA, autoantibody specificities never before described in breast cancer patients. Altered molecular pathways in cancer may make some molecules immunogenic and may generate antibodies that, although not effective in combating the disease, might still be useful in detecting novel proteins involved in malignant tumors. Once novel breast cancer proteins that are the target of autoantibodies are determined, characterization of the proteins may eventually lead to a better understanding of the basic cell and molecular biology of breast cancer.

PUBLICATIONS

Hashimoto, S., Setareh, M., Ochs, R.L., Lotz, M. Fas/Fas ligand expression and induction of apoptosis in chondrocytes. Arthritis Rheum. 40:1749, 1997.

Ochs, R.L. Autoantibodies and interstitial cystitis. Clin. Lab. Med., 17:571, 1997.

Ochs, R.L. Methods used to study structure and function of the nucleolus. Methods Cell Biol., in press.

Ochs, R.L., Stein, T.W., Jr., Chan, E.K.L., Ruutu, M., Tan, E.M. cDNA cloning and characterization of a novel nucleolar protein. Mol. Biol. Cell 7:1015, 1996.

Ochs, R.L., Tan, E.M. Autoimmunity and interstitial cystitis. In: Interstitial Cystitis. Sant, G.R. (Ed.). Lippincott-Raven, Philadelphia, 1997, p.47.

Zweyer, M., Riederer, B.M., Ochs, R.L., Fackelmayer, F.O., Kohwi-Shigematsu, T., Bareggi, R., Narducci, P., Martelli, A.M. Association of nuclear matrix proteins with granular and threaded nuclear bodies in cell lines undergoing apoptosis. Exp. Cell Res. 230:325, 1997.

Top of Page

---

Autoimmunity Induced by Xenobiotics

K.M. Pollard, D.H. Kono,^* D.K. Lee, B. Hildebrandt, P. Hultman,^** E.M. Tan

^* Department of Immunology, TSRI
^** Linköping University, Linköping, Sweden

Our work focuses on the systemic autoimmunity elicited in mice by treatment with mercury and the similarities between this murine model of autoimmunity and systemic autoimmune disease in humans.

IFN- AND INDUCTION OF XENOBIOTIC-INDUCED SYSTEMIC AUTOIMMUNITY

Mercury-induced autoimmunity is T-cell dependent. Moreover, evidence suggests that the disease is primarily driven by type 2 CD4⁺ T-helper cells that produce IL-4 rather than by type 1 CD4⁺ T-helper cells that produce IFN-. In support of this possibility, we found that spleens from mice treated with mercury had levels of IL-4 four times higher (P =.003) than control mice treated with phosphate buffered saline (PBS). Levels of IFN- in the two groups did not differ significantly. Knockout mice deficient in IL-4 or IFN- were tested for susceptibility to mercury-induced autoimmunity. Antinucleolar autoantibodies and tissue deposition of immune complexes characteristic of autoimmunity developed in IL-4--deficient mice but not in mice deficient in IFN-. In other studies, we compared IFN- knockout mice with wild-type mice for the response to the T cell--dependent antigen (4-hydroxy-3-nitrophenyl)acetyl. The results indicated that IFN- knockout mice have an essentially normal response to exogenous antigens despite the markedly attenuated response to self-antigens and decreased MHC class II expression after injection of mercury.

These findings suggest that the mice are tolerant to the autoantigens and that an IFN---dependent mechanism, possibly antigen presentation, is required to overcome this tolerance and promote production of autoantibodies. Future studies will define which of the pleiotropic IFN- effects are responsible for the induction of autoimmunity. Such findings may have therapeutic potential for the management of systemic autoimmunity.

XENOBIOTIC-INDUCED ACCELERATION OF SYSTEMIC AUTOIMMUNITY

Susceptibility to mercury-induced autoimmunity is largely determined by MHC haplotype. However, the similarity of this disease to spontaneous systemic lupus erythematosus suggests non-MHC lupus-predisposing genes may also play a role. To test this possibility, we used BXSB mice. Systemic lupus erythematosus develops spontaneously in BXSB mice, and the animals express the H-2^b haplotype that in normal strains confers resistance to mercury-induced autoimmunity. For 4 weeks, three groups of animals were given twice weekly subcutaneous injections of 40 mug of either HgCl₂ or NiCl₂ in PBS or PBS alone. After the 4 weeks of treatment, all mice were examined for indications of systemic autoimmunity.

Compared with NiCl₂- and PBS-treated mice, mice treated with HgCl₂ had significantly greater lymphadenopathy (P < .001) and greater increases in total IgG (P < .0001) and in immunoglobulin isotypes IgG1 (P < .0001) and IgG2a (P < .0001) (Fig. 1). The appearance of humoral autoimmunity (including autoantibodies to chromatin, P < .01; Fig. 2) and immune-complex disease was also accelerated in the mice treated with mercury. In contrast, treatment with nickel did not accelerate any autoimmune features. These observations support the hypothesis that in the presence of non-MHC genetic susceptibility and adequate body burden, xenobiotic exposure can contribute to immunologic aberrations, which can lead to overt autoimmunity.

CHANGES IN THE MOLECULAR AND ANTIGENIC PROPERTIES OF FIBRILLARIN AFTER INTERACTION WITH MERCURY

The mechanisms by which a xenobiotic such as mercury elicits an autoantibody response that predominantly targets a single intracellular protein autoantigen remain uncertain, but access of mercury to the intracellular environment may be required. We found that mercury-induced cell death was associated with loss of fibrillarin antigenicity and with modification of the molecular properties of fibrillarin as shown by electrophoretic studies. The sensitivity of the HgCl₂-induced modification of fibrillarin to 2-mercaptoethanol, iodoacetamide, and hydrogen peroxide suggested interaction of mercury with the two cysteines in the fibrillarin sequence. This hypothesis was confirmed by mutation of the cysteines to alanines, which abolished the aberrant migration of fibrillarin in the presence of HgCl₂. The modification of the molecular structure of fibrillarin by mercury reduced immunoprecipitation by autoantibodies to fibrillarin, pointing to unmodified fibrillarin as the B-cell antigen, and implicating mercury-modified fibrillarin as the source of T-cell antigenicity. These observations help explain the antigenic specificity of mercury-induced murine autoimmunity.

PUBLICATIONS

Chan, E.K.L., Pollard, K.M. Autoantibodies to ribonucleoprotein particles by immunoblotting. In: Manual of Clinical Laboratory Immunology, 5th ed. Rose, N.R., et al. (Eds.). American Society for Microbiology, Washington, DC, 1997, p. 928.

Pollard, K.M., Hultman, P. Effects of mercury on the immune system. In: Metal Ions in Biological Systems. Vol. 34: Mercury and Its Effects on Environment and Biology. Sigel, H., Sigel, A. (Eds.). Marcel Dekker, New York, 1997, p. 421.

Pollard, K.M., Lee, D.K., Casiano, C.A., Blüthner, M., Johnston, M.M., Tan, E.M. The autoimmunity-inducing xenobiotic mercury interacts with the autoantigen fibrillarin and modifies its molecular and antigenic properties. J. Immunol. 158:3521, 1997.

Rubin, R.L., Pollard, K.M. Chemical-induced autoimmunity. In: Handbook of Experimental Immunology, 5th ed. Hertzenberg, L.A., et al. (Eds.). Blackwell Scientific, London, 1997, p. 135.1.

Top of Page

---

Structure and Function Studies of the High-Affinity Receptor for Immunoglobulin E

M.W. Robertson, C. Dargan, P. Nguyen, R.A. Forcada-Lowrie, B. Albrecht

The high-affinity receptor for IgE (FcRI) is expressed by a number of hematopoietic cells, including mast cells and basophils. Activation of these cells occurs when the receptor binds to IgE-antigen complexes. Binding leads to a series of cellular events that culminate in the release of pharmacologically active substances that contribute to the clinical signs and symptoms of allergy. The interaction of cell-bound FcRI with circulating IgE is therefore of central importance to the molecular origin of atopic disorders.

One of the main goals of this research program is to define the structural basis of FcRI-IgE recognition. For these studies, we use both protein engineering and structural studies. In addition, recent evidence suggests that a soluble form of FcRI may be able to downregulate the production of IgE in class-switched B cells, thereby prompting further studies into the biological consequences of the FcR-IgE interaction.

Previously, we showed that the ectodomain of the FcR -chain, which contains the complete IgE-binding site, can be expressed in a functional form on the surface of filamentous bacteriophage. This property provides unique protein engineering opportunities for (1) mapping the structures on the receptor that bind ligand (IgE) and (2) constructing specific libraries of IgE-binding receptor analogs, displayed on the surface of phage, for subsequent screening for IgE-binding characteristics. Because these studies use phage display of chimeric molecules consisting of the -chain of FcRI and FcRIII, a receptor for IgG, these constructs are also useful for mapping the IgG-binding sites on the structurally similar molecule (FcRIII). The goal of the latter studies is to eventually engineer a receptor analog with enhanced ligand-affinity characteristics.

Soluble fragments of the -chain of FcRI, produced by secretion from Escherichia coli, have IgE-binding activity similar to that of the native receptor. One of our goals is to produce a complex of the IgE receptor bound to a part of the IgE Fc region. To this end, we have expressed several IgE Fc fragments from bacteria and are evaluating the properties of the various complexes formed between the Fc fragments and the IgE receptor fragment. Eventually, a suitable complex will serve as material for crystallization trials and subsequent crystallographic structural analysis.

REGULATION OF IgE SYNTHESIS

Soluble preparations of the -chain of FcRI, or a suitable analog, may have therapeutic potential for the treatment of atopic disease. Two approaches of modulating the allergic response by using soluble high-affinity, high-avidity IgE-binding molecules are under investigation. One approach focuses on a search for receptor analogs that may serve as IgE-binding antagonists. The second focuses on the capacity of soluble FcRI-chains to modulate IgE synthesis in B lymphocytes.

In the latter studies, human peripheral blood leukocytes from atopic donors are transferred to mice that have severe combined immunodeficiency. In this model of in vivo production of IgE, high IgE titers are produced within 2--4 weeks of adoptive transfer of cells, and these levels can be further enhanced by treatment with IL-4. In collaboration with G. Mudde, Novartis Forschunginstitut GmbH, Vienna, we are testing the effect of several IgE receptor constructs for the capacity to modulate IgE production in both IL-4--treated and untreated mice.

CHARACTERIZATION OF IgG-FCRIII

Additional studies in the laboratory have focused on characterization of the binding of the recombinant low-affinity IgG receptor (FcRIII, CD16) to various human IgG subclasses. Binding of the Fc region of IgG antibodies to low-affinity Fc receptors triggers important effector functions in the immune system. To better understand the nature of the ligand-receptor interaction, we produced and purified recombinant FcRIII from both eukaryotic and prokaryotic cells. In collaboration with C. Sautès and colleagues at the Curie Institute, Paris, we have determined the affinity and kinetics of the interaction between IgG and its receptor. BIAcore analysis revealed that monomeric human IgG1 and IgG3 bound to immobilized CD16 with low affinity (1.3 x 10⁶ M^-1 and 2.6 x 10⁵ M^-1, respectively) and that the affinity of human IgG3 for aglycosylated CD16, produced and purified from E. coli, was enhanced approximately twofold.

PUBLICATIONS

Galon, J., Robertson, M.W., Galinha, A., Mazières, R., Spagnoli, R., Fridman, W.-H., Sautès, C. Affinity of the interaction between Fcreceptor type III and monomeric human IgG subclasses: Role of Fcreceptor type III glycosylation. Eur. J. Immunol., in press.

Top of Page

---

Drug-Induced Lupus

R.L. Rubin, A. Kretz-Rommel, T.-H. Wu

A common side effect of therapy with several dozen medications currently in use is the appearance of autoantibodies; less often, a syndrome similar to the autoimmune disease systemic lupus erythematosus occurs. Drug-induced lupus typically appears after more than 1 year of continuous therapy with the offending medication and improves within a few weeks after the medication is discontinued. Lupus-inducing drugs have no apparent chemical structure or pharmacologic activity in common, yet the clinical features and the specificity of the autoantibodies they elicit are largely indistinguishable. Drug-induced lupus is a perplexing autoimmune phenomenon that defies conventional explanations, and this situation reflects a general ignorance in our understanding of how the immune system discriminates self-antigens from foreign pathogens.

REACTIVE DRUG METABOLITES PRODUCED BY ACTIVATED NEUTROPHILS

We have shown that a common denominator that links lupus-inducing drugs is their capacity to undergo biotransformation by activated neutrophils. This transformation is an extracellular process that accompanies the oxidative burst and degranulation reactions of neutrophils activated by certain stimuli in which the drug and neutrophil-generated hydrogen peroxide are cooxidized by myeloperoxidase. Some of these drug metabolites have been determined, such as procainamide-hydroxylamine (PAHA), which is derived from procainamide, the most common lupus-inducing drug. Under certain conditions, these drug metabolites are cytotoxic, but PAHA has no capacity to activate T cells or B cells or enhance antigen presentation, findings that suggest that the peripheral immune system is not the site of action for induction of autoimmunity. However, neutrophils are abundant and widely distributed, raising the possibility that labile, reactive drug metabolites could be formed within regions of the central immune system where tolerance to self is initially established.

MOUSE MODEL FOR DRUG-INDUCED LUPUS

Because immune tolerance to self-antigens occurs as T cells are being selected in the thymus, we examined the effect of PAHA in this organ. Two intrathymic injections of PAHA into normal mice given 2 weeks apart resulted in the appearance of IgG antibodies to chromatin approximately 3 weeks after the second injection. Antichromatin activity could be largely accounted for by antibodies against the subnucleosome particle (H2A-H2B)-DNA, which is also the predominant antibody activity in procainamide-induced lupus patients. T-cell reactivity against chromatin was also detected in the periphery after a 2-day ex vivo exposure of thymuses to PAHA, indicating that appearance of autoreactive T cells was a consequence of the effect of PAHA during T-cell maturation in the thymus.

PAHA in the thymus could induce autoimmunity either by blocking deletion (negative selection) of autoreactive T cells or by disrupting positive selection. No gross changes in the quantity of CD4 and CD8 double- or single-positive T cells were detected in the thymus after PAHA treatment, and PAHA had no effect on negative selection of V_ß8-specific T cells by Staphylococcus enterotoxin B. It has been proposed that positive selection of T cells occurs through cognate low-affinity interactions with peptides derived from self-molecules in a process that leaves the cells anergic (unresponsive) to the selecting peptide. In an in vitro model of T-cell anergy, we found that concentrations of PAHA as low as 2 muM completely prevented establishment of anergy.

We suggest that PAHA may induce autoimmunity by preventing anergy during positive selection of chromatin-specific T cells and that peptides involved in positive T-cell selection are commonly derived from chromatin because of the high abundance of this material in the thymus. To examine this hypothesis and further explore the possible involvement of negative selection, we are testing the effect of PAHA in transgenic mice that express the neo--self-antigen pigeon cytochrome c and/or the T-cell receptor for this antigen. These studies are expected to provide further support for the role of the central T-cell tolerance machinery in drug-induced lupus and provide insight into how self and nonself are normally distinguished.

PUBLICATIONS

Burlingame, R.W., Rubin, R.L. Autoantibody to the nucleosome subunit (H2A-H2B)-DNA is an early and ubiquitous feature of lupus-like conditions. Mol. Biol. Rep. 23:159, 1996.

Jiang, X., Wu, T.-H., Rubin, R.L. Bridging of neutrophils to target cells by opsonized zymosan enhances the cytotoxicity of neutrophil-produced H₂O₂. J. Immunol., in press.

Kretz-Rommel, A., Duncan, S.R., Rubin, R.L. Autoimmunity caused by disruption of central T cell tolerance: A murine model of drug-induced lupus. J. Clin. Invest. 99:1888, 1997.

Kretz-Rommel, A., Rubin, R.L. A metabolite of the lupus-inducing drug procainamide prevents anergy induction in T-cell clones. J. Immunol. 158:4465, 1997.

Lefkowith, J.B., Kiehl, M., Rubenstein, J., Di Valerio, R., Bernstein, K., Kahl, L., Rubin, R.L., Gourley, M. Heterogeneity and clinical significance of glomerular binding antibodies in SLE. J. Clin. Invest. 98:1373, 1996.

Rubin, R.L. Drug-induced lupus. In: Dubois' Lupus Erythematosus, 5th ed. Wallace, D.J., Hahn, B.H. (Eds.). Williams & Wilkins, Baltimore, p. 871, 1997.

Rubin, R.L. Environmentally induced systemic lupus. In: Rheumatic Diseases and the Environment. Kaufman, L.D., Varga, J. (Eds.). Chapman and Hall, New York, in press.

Rubin, R.L. Enzyme-linked immunosorbent assays for antibodies to native DNA, histones and H2A-H2B)-DNA. In: Manual of Clinical Laboratory Immunology, 5th ed. Rose, N.R., et al. (Eds.). American Society for Microbiology, Washington, DC, p. 935, 1997.

Rubin, R.L. (H2A-H2B)-DNA autoantibodies. In: Textbook of Autoantibodies. Peter, J.B., Shoenfeld, Y. (Eds.). Elsevier, New York, p. 364, 1996.

Top of Page

---

Regulation of Bradykinin Effects in Inflammatory Diseases

B.L. Zuraw, Z.K. Pan, K.M. Woessner, S.C. Christiansen, R.D. Ye^*

^* Department of Immunology, TSRI

Our laboratory is interested in the regulation of bradykinin generation and actions in inflammatory diseases. Bradykinin mediates a variety of actions associated with inflammation, including pain, vascular permeability, smooth muscle contraction, neuropeptide release, and stimulation of phospholipases. Results from our laboratory showed that bradykinin also stimulates the synthesis of proinflammatory cytokines and chemokines.

C1 INHIBITOR DEFICIENCY

Hereditary angioedema is an autosomal dominant disease characterized by various C1 inhibitor deficiency and recurrent angioedema. We determined a variety of new mutations of the gene for C1 inhibitor in patients with hereditary angioedema. To assess the relationship between the different genomic mutations and the variable clinical phenotypes, we are transfecting cloned mutant cDNAs into COS-7 cells and are analyzing both intracellular and extracellular expression of C1 inhibitor. To study the transcriptional control of synthesis of C1 inhibitor, we constructed promoter-luciferase reporter plasmids.

We are also characterizing the functional effect of a C1 inhibitor polymorphism that may play a disease-modifying role in hereditary angioedema. We showed that another polymorphism, involving the B2 bradykinin receptor, correlated significantly with the occurrence of angioedema in subjects with C1 inhibitor deficiency.

KALLIKREIN-KININ SYSTEM IN ASTHMA

Asthma is a chronic inflammatory disease of the airways in which bradykinin may play a critical regulatory role. Previously, we found that bronchoalveolar fluid from asthmatic patients contains tissue kallikrein and that bronchial challenge with allergen rapidly increases levels of tissue kallikrein and generation of kinin in the airway. Kinins exert their biological effects by binding to specific high-affinity receptors, which belong to the superfamily of seven transmembrane domain receptors coupled to GTPase-binding proteins.

We found that inflammatory cytokines can upregulate expression of B2 bradykinin receptors in cultured cell lines and in primary cultures of epithelial cells obtained from human airways. We recently showed that binding of bradykinin through B2 bradykinin receptors leads to increased activation of the transcription factor NF-B and production of cytokines in human fibroblast and epithelial cell lines and in primary cultures of human bronchial epithelial cells. We showed that bradykinin-induced activation of NF-B requires activation of the small GTPase-binding protein RhoA.

We are using ribonuclease protection assays and in situ hybridization to measure the in vivo effects of challenge with allergen or bradykinin on the synthesis of cytokines and the expression of kallikrein and kinin in airway tissue. We found that nasal epithelial cells from subjects with allergic rhinitis express significantly higher levels of interleukin-8 and intracellular adhesion molecule-1 mRNA than nonallergic subjects do.

PUBLICATIONS

Christiansen, S.C., Martin, S.B., Schleicher, N.C., Koziol, J.A., Hamilton, R.G., Zuraw, B.L. Exposure and sensitization to environmental allergen of predominantly Hispanic asthmatic children in San Diego's inner city. J. Allergy Clin. Immunol. 98:288, 1996.

Christiansen, S.C., Martin, S.B., Schleicher, N.C., Koziol, J.A., Mathews, K.P., Zuraw, B.L. Evaluation of a school-based asthma education program for inner-city children. J. Allergy Clin. Immunol. 100:613, 1997.

Christiansen, S.C., Zuraw, B.L. Allergy. In: Encyclopedia of Human Biology, 2nd ed., Vol. 1. Academic Press, Orlando, 1997, p. 249.

Lung, C.C., Chan, E.K.C., Zuraw, B.L. Analysis of an exon 1 polymorphism of the B2 bradykinin receptor gene and its transcript in normal subjects and C1 inhibitor deficient patients. J. Allergy Clin. Immunol. 99:134, 1997.

Pan, Z., Zuraw, B.L., Lung, C.C., Prossnitz, E.R., Browning, D.D., Ye, R.D. Bradykinin stimulates NF-B activation and interleukin-1b gene expression in cultured human cells. J. Clin. Invest. 98:2042, 1996.

Zuraw, B.L. Urticaria, angioedema and autoimmunity. Med. Clin. North Am. 17:559, 1997.

Zuraw, B.L., Christiansen, S.C. Kinins. In: Inflammatory Mechanisms in Asthma. Claude Lenfant Series Lung Biology in Health and Disease. Holgate, S.T., Busse, W. (Eds.). Marcel Dekker, New York, in press.

Top of Page

---