Surgical Approach and Safety of Spinal Cord Stem Cell Transplantation

The October 2009 edition of Neurosurgery featured the timely report of our team's preclinical work to develop safe techniques for ventral horn spinal cord stem cell transplantation.1 The report coincided with Food and Drug Administration (FDA) approval of the first trial to examine the safety of spinal cord stem cell transplantation for motor neuron disease. We anticipate that this trial will be followed by a series of trials in North America, Europe, and Asia. These trials will coincide with similar approaches applied to traumatic and demyelinating spinal cord disease. The FDA approved protocol is entitled “A Phase I, Open–label, First–in–human, Feasibility and Safety Study of Human Spinal Cord–Derived Cell Transplantation for the Treatment of Amyotrophic Lateral Sclerosis.” As alluded to in the title, the therapeutic product is derived from NIH–banked human fetal spinal cord. Technology developed originally in Ron McKay's laboratory was used as the intellectual property platform of a company called NeuralStem, Inc. (Rockville, Maryland). Unlike many protocols for the propagation of stem cells, NeuralStem's cells are propagated on laminar surface rather than as free–floating neurospheres.2-4 Hopes for amyotrophic lateral sclerosis (ALS) therapy rest on experiments in the SOD1 mutant rodent model of familial ALS conducted published in 2006. The SOD1 gene has been found to have a variety of point mutations in a subset of patients with familial ALS. When mutant human SOD1 is expressed in transgenic animals (rodents and pigs), they develop a phenotype that closely resembles human ALS. Xu et al demonstrated that spinal cord grafts of the NeuralStem cells had the ability to preserve motor neuron numbers in the spinal cords of SOD1 rodents and also prolonged their survival. Several mechanisms are postulated to explain the efficacy of the grafts. First, a fraction of the cells are found to develop a gabaergic neuronal phenotype. These inhibitory cells form synapses with surrounding cells providing a means to suppress excitotoxicity thought to play a role in the etiology of degenerative motor neuron death. The remaining cells develop an astrocytic phenotype. Excitotoxicity in ALS has been ascribed to defects in glial excitatory amino acid scavenging. Thus, it is possible that the remaining cells prevent toxic build up of excitatory transmitters. Finally, the cells are demonstrated to secrete a variety of neural growth factors that may contribute to neural protection. It is clear that these cells do not replace lost motor neurons and neuromuscular junctions. As mentioned, a variety of competing approaches exist for both molecular and cellular spinal cord therapies of motor neuron diseases, spinal cord injury, and demyelinating diseases. Our team began work on the techniques for safe and accurate ventral horn targeting in collaboration with Clive Svendsen, PhD (Cedars–Sinai, Los Angeles, California; and Madison, Wisconsin), coauthor on our October Neurosurgery manuscript in 2004. Dr Svendsen's team had documented the neuroprotective properties of human fetal cortically derived cells. Unlike the NeuralStem's cells, these cells are grown as neurospheres. In addition, these cells do not form neurons on transplantation, but rather all differentiate into astrocytes. To augment the protective capacity of these cells, Dr Svendsen's team used lentiviral vectors to induce the expression and secretion of glial cell-derived neurotrophic factor (GDNF). Thus, the Svendsen cells act as organic minipumps for growth factors in addition to scavenging excitatory amino acids. His team has demonstrated the ability of these grafts to preserve spinal cord motor neurons in the SOD1 rat model.5 In 2005, Dr Svendsen and I submitted a PreIND application for transplantation of these cells into humans. The manuscript published in October 2009 reports on some of the critical preclinical work conducted to support this application. The master cell bank has now been completed, and vector production has been funded by the National Gene Vector Laboratories. We anticipate a final FDA application with Dr Svendsen's cells sometime in the next year. We are also supporting the preclinical development of stem cells intended for use in spinal cord transplantation for the treatment of ALS and transverse myelitis by Q Therapeutics (Salt Lake City, Utah), and the academic team of Angelo Vescovi in Italy. Other teams are pursuing the idea of embryonic stem cell transplantation for spinal muscular atrophy in infants. Extensive work has been done in the Far East with human application of spinal cord grafting. Most notably Dr Huang (Beijing) has reported a large series of olfactory ensheathing cell grafts into the spinal cords of patients with chronic spinal cord injuries.6-8 He has also transplanted these cells into the brains of ALS patients, a limited group of which received free hand cervical injections as well. The results of this work have not, to our knowledge, been published as yet. In my personal correspondence with Dr Huang, it has become clear that he has abandoned this procedure, though it is not clear why. Anecdotal evidence supports efficacy of the transplants for spinal cord injury, and it is clear that this approach bears further scrutiny. However, based on our work in over a hundred pigs, we are emphatically opposed to free hand cord injections. This approach has no reproducible targeting accuracy, and more importantly leaves the patient vulnerable to sheer injuries, pressure injury, and graft reflux. The FDA approved NeuralStem trial is aimed at establishing safety and feasibility. We have developed the concept of “risk escalation” to replace the common “dose escalation” used in pharmacological trials. We envision the ultimate therapy involving staged lumbar and cervical transplants using multiple bilateral lumbar injections to preserve ambulation, and multiple unilateral cervical injections (C3–C5) to preserve diaphragmatic and proximal upper extremity function. To begin with the least risk possible, we will initially enroll non–ambulatory patients for lumbar unilateral injections, proceeding to lumbar bilateral injections in non–ambulatory patients. The main risk of these first cohorts is pain and bowel and bladder dysfunction. We will then proceed to ambulatory patients, starting with unilateral multiple injections and proceeding to bilateral multiple injections. The increased risk to these cohorts involves potential loss of ambulation. Next, we will move to unilateral cervical injection with the risk of quadriplegia, followed finally by staged bilateral lumbar followed by unilateral cervical injection. The device described in the October 2009 manuscript has been modified in a variety of ways to optimize safety and accuracy. Subsequent reports describing this development are in production or review at present. While preclinical studies address the safety of a specific biological product, very little work has been done in large animals that models spinal cord transplantation into humans. That is, when we transplant human cells into pigs, they are, by definition, xenografts. While this is the required safety data requested for an IND, it is not a good model for human allografts proposed for trials. As such, our team at Emory has recently submitted an RO1 application for continued study of surgical techniques, graft rejection, imaging, and graft control in the pig model. It is not clear which cell line will prove the most beneficial for ALS patients. However, we have great hopes that much will be learned from these initial trials about the best way to conduct human spinal cord transplantation. Nicholas Boulis Thais Federici Atlanta, Georgia