Development of a Method for Generation of Spinal Cord Neurons from Embryonic Stem Cells for Treatment of Spinal Cord Injury

Objectives

            Stem cell research has been around for many years and has significantly contributed to the fields of haematopoiesis and embryology.  Recently, remarkable discoveries that stem cells may be generated in different adult tissues throughout life and may differentiate into a variety of specialized cells have brought into focus their possible therapeutic potential (for review see Lovell-Badge, 2001; Gage, 2002).  The research has shown that the vertebrate brain continues to produce new neurons throughout lifespan (Altman, Das, 1965; Barnea, Nottebohm, 1994; Cameron, McKay, 1999).  Moreover, adult stem cells demonstrate surprising pluripotency, for example stem cells isolated from the bone marrow may differentiate into neurons and glia after transplantation into mouse brain (Nakano et al., 2001).  Research on therapeutic application of adult stem cells for treatment of neurological disorders is constrained by ethical considerations, lack of reliable sources and inability to propagate stem cells in sufficient quantities (Liu et al., 2000).  There are reports which show the propagation of stem cells in culture is possibly limited and that their differentiation potential decreases over time (Carpenter et al., 2001).  There are also concerns, that adult stem cells, may have their genome reprogrammed through epigenesis, which may also limit their ability to differentiate, and induce accelerated senescence (for review see Surani, 2001; Rakic, 2002).  Embryonic ES (ES) cells derived from blastocyst may provide a partial solution to these problems.  Experiments showed that ES cells might be maintained in culture over a year (more than 250 passages) retaining their phenotype and ability to differentiate (Amit et al., 2000).  In addition to the ability of almost indefinite replication, ES cells have a normal unaltered genome, which makes them totipotent.  Because ES cells may be easily cultured, propagated and genetically manipulated, they may prove to be a better choice for transplantation applications.  Moreover, there are few cases showing that ES cells can be successfully used for treatment of animal models of disease (McDonald et al., 1999; Kim et al., 2002)

The focus of this multidisciplinary proposal, representing joint research efforts of five established investigators, is to develop a method for targeted differentiation of ES cells into specialized cells of spinal cord and to establish a method for efficient transplantation of new cells which will result in functional recovery.

The specific objectives for this two-year project are:

1. To develop a method for generation of neurons specific for spinal cord.  This will be accomplished by transfection of ES cells with specific factors driving differentiation of ES cells into spinal dorsal horn interneurons.  The differentiation of ES cells into neurons will be monitored using specific antibodies and profiling gene and protein expression. [Year 1].

2. To establish procedure for efficient transplantation of generated new cells into host tissue.  This will be accomplished by using scaffolds at a time of cell transplantation with subsequent monitoring of functional integration of new cells with host tissue and analysis of functional recovery using immunohistology, electrophysiology, and behavioral tests. [Year 2].

B.  Project Plan

Preliminary data (in collaboration with Dr. Brewer)

            Preliminary studies from Dr. Murashov’s lab show that mouse ES cells may be easily grown and propagated in our laboratory and successfully differentiated into neuronal precursors using retinoic acid treatment (Liu et al., 2000) or treatment with NGF.  For tracking transplanted cells, ES cells were stably transfected with green fluorescent protein (GFP).

ES cells in phase-contrast	Fluorescence of ES cells with GFP

After induction ES cells formed floating clusters of cells, named embryoid bodies (EBs).  After treatment with retinoic acid and plating, some cells became precursors of neurons or glial cells.  Immunohistochemistry was performed with antibodies against glial fibrillary acidic protein (GFAP). 

EBs after retinoic acid treatment stained with anti GFAP	GFAP positive cell

       Immunofluorescence on neuronal cells

Immunofluorescence on neuronal cells 5 days culture

Spinal cord injury was induced by injection of quisqualic acid (QUIS) according to protocols form Dr. Brewer' laboratory (Abraham, Brewer , 2001).  QUIS was injected intraspinally into male, C57BL/6 mice (0.6 ml total volume, 500mm below the surface of the cord, level T12-L2) to create a unilateral excitotoxic injury. Control animals received an equal volume of normal saline.  This model produces pathological and behavioral changes consistent with those observed following human spinal cord injury.  The model was previously developed in the rat, and has been proven to be a reliable method by which to study the phenomenon of post-SCI pain states.

Previous studies have shown that simulating a chemical change that occurs during injury leads to pathophysiologic changes similar to those produced by traumatic or ischemic injuries, including cavitation of the spinal cord and altered response properties of surviving neurons (Yezierski et al., 1996). In addition, excitotoxic SCI produces predictable pain behaviors including mechanical and thermal allodynia (Yezierski et al., 1998), as well as overgrooming in approximately 80% of injured animals, a behavior believed to represent the presence of at-level neuropathic pain (Yezierski et al., 1998; Gorman et al., 2001).  A major advantage of the excitotoxic model is that the location and extent of damage is controllable. As a result, specific neuronal populations can be targeted for destruction. This allows for the assessment of the effects of damage to the sensory pathways, without introducing motor or autonomic complications.  Therefore, the sensory component of SCI can be studied independently of the motor component. 

Seven days after injury transplants were injected into the site of QUIS injection - 100,000 cells, 0.6ml.  Only one animal out of five mice which have injections with transplants developed overgrooming behavior.  While all four control animals which have received injection of Quis and no transplants exhibited overgrooming.

Animal showing sign of excessive grooming

Cavitation of the spinal cord is characteristic of QUIS –induced damage.  Histological sections showed formation of a lumen in QUIS-injected animals at the site of injection.  Surprisingly, no lumen was found in spinal cord of animals that received cell transplants.

Lumen in spinal cord 7 days after QUIS injection. Staining with anti-NeuN	No lumen in the area of QUIS injection 7 days after transplantation (anti-NeuN)

Histological sections of spinal cord with transplants were examined for GFP positive cells under fluorescent microscope.  GFP positive cells were observed at the site of the injection 7 days after transplantation.  Some of the GFP positive cells migrated to the ventral horn or even to the contralateral side.  This shows that transplanted cells successfully survive in host spinal cord and possibly contribute to reparation.  More experiments are required to confirm that transplanted cells form functional connections with host cells and survive over long periods of time.  Moreover, behavioral and electrophysiological studies are necessary to access the level of functional recovery in mice.

GFP positive cells in the spinal cord	The same section. Stained anti-NeuN

1.  Research Plan

a.  Experimental design.

Specific Aim 1. 

To develop a method for generation of neurons specific for spinal cord.  This will be accomplished by transfection of ES cells with specific factors driving differentiation of ES cells into spinal dorsal horn interneurons.  The differentiation of ES cells into neurons will be monitored using specific antibodies and profiling gene and protein expression. [Year 1].

Rationale

Few studies have shown that transplantation of predifferentiated ES cells may lead to integration of transplanted cells with the host tissue.  In the spinal cord the majority of transplanted cells differentiate into oligodendrocytes and astrocytes and less than 10% into neurons (McDonald et al., 1999; Liu et al., 2000).  One of the challenges is to direct cells to differentiate into neurons specific for spinal cord, such as dorsal horn interneurons.  The development of spinal cord is guided by several signaling molecules including sonic hedgehog  (SHH) and retinoic acid (RA).  RA induces differentiation of cell along rostral/caudal axis, while SHH together with bone morphogenetic proteins (BMPs) induce differentiation along ventral/dorsal axis (Lee, Pfaff, 2001; Panchision, McKay, 2002).  Homeodomain transcription factor Pax7 plays a critical role in subsequent neuronal differentiation in the dorsal spinal cord (Muller et al., 2002).  Further, Lbx1 is required for correct specification of three early dorsal interneuron populations and late-born neurons that form the substantia gelatinosa (Gross et al., 2002).  Homeodomain transcription factor Pax6 is required for motor neuron differentiation (Arber et al., 1999; Thaler et al., 1999).  Later, regulation of motor neuron development is directed by differential expression of transcription factor HB9 and type II cadherins- cell surface proteins implicated in cell adhesion and recognition (Price et al., 2002).  Expression of type II cadherin, MN-cadherin defines clustering of motor neurons into small groups, termed motor pools soon after their exit from the cell cycle.  Please see fig.1 and 2.

Fig.1. Dorsal Interneuron Differentiation (Modified from Gross et al., 2002)

Fig. 2. Motor Neuron Differentiation (Modified from Goulding, 1998)

We propose to use RA, to initially direct development of ES cell.  Transfection with homeodomain transcription factors Pax7 will be used to further drive differentiation towards interneuron phenotypes.  Cellular decisions concerning differentiation are reflected in altered patterns of gene and protein expression.  Quantitative assessment of gene expression in ES cells is essential for understanding the molecular events underlying differentiation.  To profile gene and protein expression of precursors and postmitotic cells we will use specific markers for different types of spinal cord neurons.  The following figure 3 list specific markers expressed by precursors and postmitotic neurons which can be used for cell identification. 

Fig.3. Summary of Murine Dorsal Spinal Cord Development (Adapted from Muller et al., 2002).  Schematic display of six distinct dorsal neuronal subtypes (dI1–dI6) and of the progenitor domains that produce these neurons. The precursor cells express Pax7, Dbx 1, Dbx2, Mash1.  Postmitotic neuronal subtypes express specific combination of factors including Lbx1, Lim1/2, Lmx1b, Brn3a, Isl1/2, Lh2a/b.

The use of Real-time PCR is critical as a method for rapid and precise quantitation of nucleic acids.  This technique has shown to be superior to endpoint quantitation both in accuracy and reproducibility (Bustin, 2000) and it has been successfully used for precise quantitation of gene expression in small subsets of highly purified ES cell populations (Raaijmakers et al., 2002).  Dr. Katwa, Co-Pi on this proposal, has Real-time PCR machines in his laboratory, and he is an expert in quantitative Real-time PCR.

Protein profiling will be done by using specific antibodies for immunocytochemistry, and immunoprecipitation with subsequent mass-spectrometry.  Protein profiling by mass-spectrometry has been successfully used previously (Petricoin, et al., 2002).  Dr. Jan Teller, Co-PI on this proposal has long track record using mass-spectrometry for protein profiling. 

Specific Aim 2. 

To establish procedure for efficient transplantation of generated new cells into host tissue.  This will be accomplished by using scaffolds at a time of cell transplantation with subsequent monitoring of functional integration of new cells with host tissue and analysis of functional recovery using immunohistology, electrophysiology, and behavioral tests. [Year 2].

Rationale

            Damage to the spinal cord often results in progressive tissue loss and subsequently in cavity formation.  These cavities may be of substantial diameter leaving only a small rim of white matter (Oudega et al., 2001).  The inability of axons to regenerate across the cavity may lead to permanent paralysis.  The potential differentiation of ES cells in the adult brain raises exciting new possibilities for treatment of spinal cord injuries.  However to bridge a large gap in the injured tissue may be difficult if not impossible without tissue engineering.  A scaffold grafted into the site of injury may provide necessary mechanical support for the transplanted cells, guide axonal growth and promote better integration with host tissue.  Different compounds were used previously, including fibronectin (Tate et al., 2002), alginate hydrogel (Novikov et al., 2002), Schwann cell grafts (Bunge et al., 1994), glioma cells C6-R (Hormigo et al., 2001) and different fibers or tubular scaffolds made of multi-component biopolymers (Oudega et al., 2001; Novikov et al., 2002; Teng et al., 2002). All of these scaffolds promoted neuronal survival and regeneration, although none of them resulted in a complete restoration of missing tissue or total functional recovery.  The potential problem may be based on a type of cells used to populate the scaffold as well as on the development of a glial scar around the injury. 

            Recently, an alternative approach utilizing enzyme chondroitinase ABC has been reported (Bradbury et al., 2002).  It is well-known that at the site of the spinal cord injury a glial scar forms containing extracellular matrix molecules including chondroitin sulphate proteoglycans which are inhibitory to axonal growth.  The investigators have used specific enzyme chondroitinase ABC to degrade chondroitin sulphate.   After treatment of injured spinal cord with chondroitinase the glial scar did not develop which stimulated regeneration of both ascending and descending corticospinal tract axons and promoted functional recovery of locomotor and proprioceptive behaviors.

            In our proposal we will test treatment with chondroitinase ABC and different scaffolds to promote recovery.  For correct function, a scaffold for tissue engineering has to be a porous three-dimensional structure which serves as a template for initial cell attachment and subsequent migration, growth and differentiation.  The requirement for scaffold systems is to arrange cells in an appropriate 3D configuration and present molecular signals in an appropriate spatial and temporal fashion so that the individual cells will grow and form the desired tissue structures and do so in a way that can be carried out reproducibly, economically and on a large scale (for review see Hutmacher, 2001; Bottaro et al., 2002).  It has also been well established that cultured neurons did not show nerve fiber growth unless fibronectin, collagen or trophic factors were incorporated into the scaffold.  Because the cavity in the spinal cord after excitotoxic injury may be of irregular shape which may also vary from animal to animal the use of rigid tubular scaffold or fibers is not practicle.  Hence, for our experiments we will tests gels containing fibronectin, collagen and different trophic factors including, NGF, BDNF, and NT-3.  After transplantation the recovery will be monitored using histology and immunohistochemistry at different time points after implantation.  In parallel, inflammatory response will be assessed during spinal cord injury and subsequent recovery, using cytokine profiling on protein array system Luminex-100.

Supported by grant from North Carolina Biotechnology Center