Human Marrow Stromal Cell Transplants In A Collagen Matrix Support Axonal Regeneration of Descending Pathways Across Complete Spinal Cord Transections

Introduction: Human marrow stromal cells are similar to stem cells. They are easily harvested from bone marrow by aspiration under local anesthesia. Harvested cells are multi-potential and can differentiate into a variety of different types of cells. They are also easily modifiable through ex-vivo gene therapy to deliver several select therapeutic gene products including neurotrophins. The purpose of the current study was to compare genetically unmodified human marrow stromal cells to unmodified and genetically modified fibroblast transplants, which are known to be permissive to axonal regeneration when transplanted after complete spinal cord transections.

Methods: 42 adult rats underwent mid-thoracic laminectomy and complete spinal cord transection, creating an approximately 3-4 mm defect. The defect was immediately filled with a transplant of a carefully titrated number of cells as follows: 1)Unmodified human marrow stromal cells alone (n=4); 2)Unmodified human marrow stromal cells in a Vitrogen collagen matrix (n=12); 3)Unmodified fibroblasts in a Vitrogen collagen matrix (n=10); 4)Fibroblasts genetically modified to secrete Neurotrophin-3 (NT-3) in a Vitrogen Matrix supplemented by platelet derived growth factors (n=5); 5)Fibroblasts genetically modified to secrete Brain Derived Neurotrophic Factor (BDNF) in a Vitrogen Matrix supplemented by platelet derived growth factors (n=5); 6) Control group (n=6). Rats also received cyclosporin A immunosuppression post-operatively. At six months, rats were sacrificed. Graft site and adjacent spinal cord tissue was harvested and serially sectioned for histologic analysis. Staining for Nissyl-Myelin, RT-97 (axonal neurofilament), CGRP (dorsal root ganglion sensory axons), and Serotonin (Brainstem-Spinal axons) was performed. Additionally fluorogold retro-grade axonal labeling, and some preliminary staining for GAP-43 (a constituent of the axonal growth cone) was carried out.

Results: In all animal groups, transplants were permissive to axonal regeneration into the graft, as demonstrated by immunostaining for RT-97 (axonal neurofilament). This was particularly true for axons stained with CGRP and therefore of dorsal root ganglion origin. Additional staining of the graft for GAP-43 (axonal growth cone) indicated permissiveness to regenerating axons other than those from the dorsal root. Axonal penetration into the transplant was determined as a percent area fraction as follows: 25% for human marrow stromal cells, 15% for fibroblast transplants, and 10% for controls. Whereas control animals or animals that received transplants of human marrow stromal cells without a carrier matrix were permissive to axonal re-growth into the graft, penetration through the graft into distal host spinal cord tissue by brainstem-spinal axons (i.e. serotonergic staining of descending pathways) was not shown. In contrast, 30% of animals that received transplants of human marrow stromal cells with the collagen matrix demonstrated brainstem-spinal axons going into and through the graft (3-4mm), and penetrating the distal host tissue for up to 6 millimeters. This compared to 20% of animals transplanted with unmodified fibroblasts in a collagen matrix, 40% of those with genetically modified fibroblasts secreting NT-3, and 20% of those with genetically modified fibroblasts secreting BDNF.

Conclusions: By providing certain cellular transplants in an appropriate carrier matrix, regeneration in adult rats across a complete spinal cord transection into distal spinal cord tissue by brainstem spinal axons is possible. The ideal cell for transplantation, however, has yet to be determined. Marrow stromal cells have some advantages over fibroblasts or Schwann cell transplants in that they are easily harvested from human bone marrow, they are multi-potential, and they are easily modified by ex-vivo gene therapy. Our data suggests that unmodified human bone marrow stromal cells in a collagen matrix may be as efficacious as fibroblasts that are genetically modified to secrete neurotrophic factors, in altering the typically inhibitory CNS environment and aiding axonal regeneration by brainstem-spinal neurons. Future studies will be directed at human marrow stromal cells that are also genetically modified to secrete specific neurotrophic factors.