The Neurobiology Research laboratory (NBRL) of the Departments of Neurology and Pharmacology (Toxicology and Therapeutics), University of Kansas Medical Center (KUMC), was established at the Kansas City Veterans Affairs Medical Center, now part of the Heartland Veterans Integrated Service Network (VISN 15), in 1976. Barry W. Festoff, M.D., the founding Director, NBRL, is Professor of Neurology and Pharmacology, KUMC. For almost thirty years, the NBRL has pursued approaches equally divided between clinical neurology and neurobiology in a research and academic environment. The guiding theme, or thread, of the research studies of the NBRL is that development, plasticity after injury, and disease of the nervous system all involve similar mechanisms in a relatively simple repertoire of means with which this system can react.
I. Proteases and protease inhibitors:
Since coming to KUMC and the KCVAMC Dr. Festoff and the NBRL has largely focused its studies at the neuromuscular synapse, the neuromuscular junction (NMJ), beginning with studies of the isoforms of the enzyme, acetylcholinesterase (AChE), viewed as being markers of neurotrophic interaction. The results of these studies pointed to the role of proteases in the plasticity of the NMJ. Protease studies continue to the present time and constitute one of the major thematic approaches to neural function and dysfunction. One specific disease entity that has been featured in these studies, as a model disease of this system, was amyotrophic lateral sclerosis (ALS or Lou Gehrig's disease). Studies at the NMJ have been applied to concepts of synapse degeneration in ALS.
A. ALS and the NMJ
Beginning in 1980, Dr. Festoff and the NBRL were led to the conclusion that ALS was a syndromic disorder, not a specific disease entity that was caused by various stimuli or agents, and progressed via a common pathogenesis. They proposed one that developed into a unifying hypothesis, operating through activation of cascades of extracellularly acting proteases that disrupted synaptic connections with catastrophic repercussions on the anterior horn cell in the spinal cord. Animal studies in the NBRL along with tissue culture experiments pointed towards one or more serine proteases that might be involved. Subsequent studies invoked plasminogen activation and after a sabbatical in France, Dr. Fest off and colleagues at INSERM 153 (Drs. Michel Fardeau and Daniel Hantaï) along with Drs. Janine and Claudine Soria (Hôtel Dieu and Hôpital Lariboisiere) Paris, France concluded that urokinase-like plasminogen activator (uPA) was the physiologic regulator of remodeling at the NMJ. At the same time, it was also proposed that normally an inhibitor(s) of such proteases exists at the NMJ to prevent the disruption that occurs with ALS or other forms of denervation. This model was suggested to also stand for similar synaptic degenerations in the central nervous system.
Subsequent collaboration with Dr. Joffre Baker (now V.P., Discovery Research, Genentech Inc.), who had just moved from UC, Irvine (Dr. Dennis Cunningham's lab) to KU in Lawrence, ultimately resulted in the demonstration that the serpin, protease nexin I (PNI), was concentrated at the NMJ. PNI was the first of several protease inhibitors to be located at synapses. PNI targets specific serine proteases such as plasmin, trypsin, plasminogen activators and thrombin in solution. However, when bound to the extracellular matrix (ECM), it becomes the most potent tissue-based inhibitor of thrombin. During this time, Dr. Denis Monard's group at the Friedreich Miescher Institute in Basel had found a protein that was produced by glioma cells that promoted neurite outgrowth in neuroblastoma cell culture. This was later shown to also be a functional protease inhibitor. After returning from the Paris sabbatical, Dr. Festoff encouraged Baker and Monard to exchange their proteins and antibodies and PNI was found to be as effective a neurite outgrowth promoter as the glial protein. In 1987, PNI was cloned and found to be identical to the glia-derived neurite promoting factor of Monard (first renamed glia-derived nexin, and then changed back to protease nexin I). Cunningham's group had also discovered another nexin, named PNII, in fibroblasts. This was subsequently found to be the secreted, Kunitz-containing form of the beta amyloid precursor protein (bAPP).
The entire field of serine proteases and inhibitors, in general, progressed rapidly during the mid-'80s with Dr. Robin Carrel of Cambridge University and Dr. James Travis of the University of Georgia coining the term "serpin" (serine protease inhibitor) in 1985. In 1988, bAPP was shown to consist of more than one isoform that contained a Kunitz-type serine protease inhibitor (KPI) domain. In 1989, Dr. Festoff organized and directed, along with Drs. Carrell, Hantaï, Georgia Barlovatz Meimon and Gustave Moonen, the first international meeting dealing with these rapid developments of serine proteases and serpins in the nervous system under N.A.T.O. sponsorship. During this meeting, known as The Maratea Conference, the field changed dramatically with the demonstration by Cunningham's lab and the group at Athena Neurosciences (now Elan) that PNII was, in fact, the KPI-containing form(s) of APP.
In collaboration with Dr. Douglas Brenneman at NIH, Dr. Festoff and the NBRL showed that subnanomolar concentrations of PNI promoted neuronal survival while higher concentrations, those that support neurite outgrowth, do not. This occurred more prominently under the influence of the neuropeptide, vasoactive intestinal neuropeptide (VIP) and when electrical silence of neurons was induced by tetrodotoxin (TTX), making PNI the first serpin to be classed as an activity-dependent neurotrophic factor. We also showed that PNI was acting by inhibiting thrombin, since picomolar thrombin killed these cells. With Drs. Lucien Houenou and Ron Oppenheim at Bowman Gray, we showed that PNI rescued motor neurons from programmed (naturally-occurring) cell death in ovo and sciatic neurectomy-induced apoptotic cell death in neonatal mice.
We since focused on programmed elimination of polyneuronal synapses in neonatal mice, as a model of synapse loss in disease states. In collaboration with Dr. Philip Nelson of NIH, we showed that PNI prevented elimination of polyneuronal innervation in vitro. The specific thrombin inhibitor from the leech, hirudin, as potent as PNI (in the nanomolar range), allowed us to show that thrombin was the responsible protease involved in an activity-dependent in vitro model of synapse elimination. Members of the NBRL, Mikhail Zoubine and Irina Smirnova, actively worked on showing that thrombin was also responsible for in vivo neonatal elimination by infusing hirudin locally by osmotic minipumps. Our model proposed that thrombin induced its effects at developing NMJs by activating protease-activated receptor 1 (PAR-1), the principal thrombin receptor on cells. With Dr. Smirnova, we showed that thrombin killed mouse motor neurons in an apoptotic manner, confirming early studies with total spinal cord cells, in the nanomolar range, by activating PAR-1, mobilizing calcium and activating heterotrimeric G-proteins (Gs) and monomeric ras proteins such as Rho A. Others have shown similar results in hippocampal neurons from the rat. With Drs. Houenou and Turgeon, we showed similar effects in embryonic chick motoneurons. In both mouse and chick motoneurons, we found that PAR-1 activation by thrombin results in subsequent activation of caspases. G-protein effector molecules such as cholera toxin, mastoparan, the C3 exoenzyme of C. botulinum and lovastatin can modulate this.
With Dr. Bruce Citron in the NBRL and collaborators Drs. Carter and Jones of the Eleanor Roosevelt Institute in Denver and Drs. Cerosaletti and Fournier of the Hutchinson Cancer Center in Seatle we assigned the gene for (PI7) to chromosome 2q33-35 in the human and to syntenic regions in both mouse and sheep genomes. This is of interest to disease states since this is the same location for the gene ALS2, for Tunisian autosomal recessive ALS. With Drs. David Patterson at Eleanor Roosevelt Institute and Jackie Beckmann at Généthon in Paris, we are pursuing PNI as a candidate gene for this and other diseases. We are also exploring potential mutations in PNI as possibly being involved in dysregulation of the protease:serpin balance in neurodegeneration.
B. Protease and inhibitors in neurodegeneration
Studies of proteases and serpins at synapses are also ongoing in Alzheimer's disease (AD) and ALS tissue. DeKosky and Scheff, and Terry and colleagues, found significant evidence supporting a concept Dr. Festoff and Dr. Stanley Appel proposed more than 25 years ago that synapse loss might be an initiating event in AD pathogenesis. This evolved into theories of protease dysregulation at the synapse in ALS, as a disease initiated at synapses traveling retrograde to the motor neuron soma. Support for this has come recently from the group of Sasaki that found significant synapse loss in advance of cell loss on ALS spinal cord anterior horn cells. This has evolved into the concept of synaptic apoptosis or "synapoptosis".
One area under intense study in the NBRL is the role of protease inhibition in preventing apoptotic cell death after spinal cord injury (SCI) in a controlled contusion rat model. Recent exciting data in the effects of the caspase-3 specific tetrapeptide inhibitor DEVD-fmk has been submitted for publication.
In addition, as exciting data have recently been achieved using recombinant soluble thrombomodulin (TM; Solulin™) in the same model. TM is an integral membrane protein that binds thrombin at higher affinity than PAR-1 and prevents it from interacting with PAR-1 and other substrates.
Another area the NBRL focuses on is the interaction between thrombin and bAPP, as well as with specific cytokines such as TNF-a and interleukin 1 beta (IL-b) in SCI and in both cellular and animal models of AD with Dr. Zhiming Suo in the NBRL.
II. Growth factors and binding proteins:
A. Human growth hormone and insulin-like growth factors
Earlier studies with ALS patients suggested altered carbohydrate metabolism. This resulted in studies showing decreased insulin receptors, even in circulating monocytes, with Dr. Osvaldo Perurena and impaired insulin sensitivity (insensitivity/resistance), with Drs. Erlinda Reyes and Wayne Moore, in ALS patients. This avenue was pursued and resulted in a treatment approach with a multi-center trial (with Drs. Richard Smith, Theodore Munsat, Shlomo Melmed, along with Dr. Barry Sherman, then of Genentech) of recombinant human growth hormone (rhGH) in ALS. This was the first recombinant molecule treatment trial in a neurologic disease. Although not curative of ALS we demonstrated that the hGH:insulin-like growth hormone I (IGF-I) axis was intact in ALS patients. This work evolved into subsequent multi-center trials of rhIGF-I in over 260 ALS in North America and Europe, funded by Cephalon, Inc. The results of this trial, indicating a dose-dependent improvement in primary and secondary outcome parameters and prolongation of survival, have been published. However, because a 2nd trial in Europe, with a different design, failed to show significance, the FDA demanded a 3rd trial before granting NDA approval. This 3rd trial, conducted through the NIH, is currently in progress. At fundamental levels, focus in the NBRL pursued specific IGF binding proteins (IGFBPs) with Dr. Shi Yang that modulated IGF-I's action at its receptors.
III. Other Projects:
Other studies in the NBRL emphasized protein molecules that are important in the nervous system, both at CNS synapses and at the NMJ. These include the synaptic proteins, a-synuclein, implicated in Parkinson's disease (PD), as well as bAPP and PNI, and how these may be cross-linked by the enzyme tissue transglutaminase (tTG) to ensure synaptic stability during development, but which may also invoke disease in pathologic conditions, contributing to formation of inclusion bodies common in a number of neurodegenerative disorders. This member of the transglutaminase superfamily is a dual function protein, able to bind and hydrolyze GTP (a G-protein known as Gah/tTG) as well as induce gamma-glutamyl epsilon lysine (GGELs) cross-links. How these may be altered during the life of the organism, after CNS injury and in genetic mutants, is the focus of a number of studies. These studies are carried out with Drs. Paul Arnold, and others in the NBRL, in collaboration with Dr. Peter Davies, UT, Houston and Ugra Singh, Texas A&M in Temple, TX. In addition to injury, we are studying a neurotoxic model of Parkinson's disease (PD) induced in mice by a cocktail of MPTP and probenecid. This results in inclusion body (Lewy body-like) formation and apoptotic cell death in substantia nigra neurons. They include studies at molecular biology levels, using cDNA probes with production of anti-sense cRNAs, along with cellular neurobiology and RT-PCR, under the direction of Dr. Bruce Citron in the NBRL. Our recent success in showing that trauma and degeneration are both characterized by changes in tTG transcription, with the production of an alternatively-spliced transcript that lacks the GTP binding domain of the dual function Gah/tTG. This short transcript is preferentially produced in early spinal cord development, but is also rapidly found in neuronal culture models treated with mechanical trauma or tumor necrosis factor alpha (TNF-a) and in brains of AD and PD patients as well as rapidly (4 h) after controlled spinal cord injury (SCI) in the rat.
Current studies are designed to test the essentiality of tTG activation and/or alternative mRNA splicing in CNS injury and/or neurodegeneration. Other studies are directed at understanding the regulation of alternative splicing of tTG mRNA by analyzing transcription and splicing factors. Finally, treatment trials are directed at testing whether specific tTG activity inhibitors, anti-sense oligonucleotides or small interfering NA (siRNA) are of value in preventing inclusion body formation and/or apoptosis.
IV. Current Studies: Mechanisms and treatment of Spinal Cord Injury (SCI)
A. Role of thrombin and thrombin signaling
Currently, the role of thrombin in SCI is being investigated in vivo. Initial studies were done in rats that showed rapid upregulation of prothrombin and proteinase-activated receptor 1 (PAR1), most common thrombin receptor, within hours following contusion SCI. Current studies have moved to mice, with use of two different mutant strains used as models to illustrate the role of thrombin in SCI. We created genomically overexpressing mice for the human protease nexin I (PNI) gene (TghPNI mice). PNI is a potent inhibitor (serpin) of thrombin that theoretically should allow these mice to recover better from SCI if thrombin is a major player. The other strain of mice is a mutant of the murine thrombomodulin (TM) gene. TM is a transmembrane chondroitin sulfate proteoglycan (CSPG) that is able to sequester extracellular thrombin. The binding complex of TM and thrombin then can activate the zymogen protein C (PC) to activated APC. We have published, and patented, use of a soluble, truncated human TM (sTM; Solulin™) in SCI (see below). These mice have a glutamic acid (glu) replaced by a proline (pro) at the 404th amino acid residue of TM, and are termed TmProPro for the homozygote. The TmProPro mice can still bind thrombin with about 100-fold less affinity as the wild type (wt) but the TM:thrombin complex is essentially unable to bind and activate PC (1000-fold less than wt). We hypothesize that TmProPro mice should do worse than wt mice in recovering from SCI, if the TM: APC pathway is significant in endogenous plasticity. The TghPNI and TmProPro mice represent two ends of the spectrum for thrombin’s role in SCI. In other studies with Dr. Mohammed Farooque, we have stereotactically injected 12- mL of 1-10 nM a-thrombin into mouse spinal cords and followed the histological response and recovery patterns to this model of SCI. We found that this injection causes microglial reaction and proliferation with clustering around motor neurons, specifically.
B. Natural anticoagulants in SCI treatment
With Drs. Armelle Pindon and Daniel Hantaï in Paris we studied expression of TM in mouse brain astrocytes and showed it was functionally identical to TM on endothelial cell surfaces. We proposed a brain micro-coagulation network that produced homeostatic balance. Subsequently, those colleagues showed that TM is upregulated in reactive astrocytes after CNS injury but the increase was delayed for 36-48 hours. We postulated that early treatment with Solulin™ would “tie up” activated thrombin in the injury site and both prevent activation of PARs as well as promote APC formation. We then showed that this promoted much greater recovery of function and neuronal preservation, reduced tissue damage and microglial activation after SCI in rats.
Others showed that TM was neuroprotective in cultured cells and APC also prompted recovery after compression SCI. Subsequently, APC has been shown to reduce p53 induced apoptosis in neurons and to reduce ischemic injury in rodent stroke models.
Current studies are exploring mechanistic bases for TM’s beneficial effects using these and other strains of mutant mice. Our collaborator in Milwaukee, Dr. Harmut Weiler, has shown that TMpropro mice are susceptible to coagulation disorders but even more vulnerable to endotoxin or lipopolysaccharide (LPS). LPS at 1/10th the LD50 for wt mice kills 100% TMpropro mice. W also have another mutant strain of mice in which the D1 domain (-NH2 terminal C-type lectin domain) is missing. Our collaborator in Belgium, Dr. Ed Conaway, has shown that TMLed mice are also extremely susceptible to LPS but can adequately activate PC, suggesting that TM possesses more than one mechanism for its anti-inflammatory effects. This mechanism might involve binding and engaging a master cytokine, high mobility growth box 1 (HMGB1) protein, a DNA architectural protein with nerve outgrowth enhancing properties. We perform contusion SCI in both strains of mice, as well as in TghPNI mice.
C. Secondary Damage by Rho
Recent work in the NBRL has been directed towards characterizing the activation of rho and the ability of exogenous effecter molecules to cause neurite retraction in a neuronal cell line and injured spinal cord neurons. Rho is a member of the Ras superfamily of GTPases and is active when GTP is bound and inactive after hydrolysis to GDP. Rho can become activated after upstream cleavage activation of PAR1 by thrombin. Numerous other pathways can also activate rho by signal transmission through a diverse variety of membrane-bound proteins: Nogo/NoGoR, semaphorins, Eph/ephrins, etc. After SCI the blood-spinal cord barrier (BSCB) is disrupted and thrombin is able to come in contact with spinal cord neural cells. When this happens thrombin cleaves PAR1, RhoA is activated downstream, and neurites are subsequently retracted. In efforts to promote recovery in patients who suffer from SCI this secondary damage should be inhibited so that neurite extension may be promoted.
D. Minocycline and 3rd generation tetracyclines in SCI
We were amongst the first groups to show effects of minocycline, a 2nd generation tetracycline in SCI. We are now evaluating newer 3rd generation compounds with industry (Paratek) and other (Daniel Anthony, Ph.D., Oxford University) collaborators. Dr. Anthony and colleagues are also working with us to show TM’s effects on blocking invasion of polymorphonuclear neutrophils (PMNs) in to the cord after SCI.