Is the Outgrowth of Transplant-Derived Axons Guided by Host Astrocytes and Myelin Loss?

Abstract

Loss of cortical neurons may lead to sever and sometimes irreversible deficits in motor function in a number of neuropathological conditions. Absence of spontaneous axonal regeneration following trauma in the adult central nervous system (CNS) is attributed to inhibitory factors associated to the CNS white matter and to the non-permissive environment provided by reactive astrocytes that form a physical and biochemical barrier scar. Neural transplantation of embryonic neurons has been widely assessed as a potential approach to overcome the generally limited capacity of the mature CNS to regenerate axons or to generate new neurons in response to cell loss. We have recently shown that embryonic (E14) mouse motor cortical tissue transplanted into the damaged motor cortex of adult mice developed efferent projections to appropriate cortical and subcortical host targets including distant areas such as the spinal cord, with a topographical organization similar to that of intact motor cortex. Several parameters might account for the outgrowth of axonal projections from embryonic neurons within a presumably non-permissive adult brain, among which are astroglial reactions and myelin formation. In the present study, we have examined the role of astrocytes and myelin in the axonal outgrowth of transplanted neurons.

Introduction

Central nervous system (CNS) lesion in the adult mammal is followed by more or less severe and quasi irretrievable functional deficits. The inhibitory nature of the adult mammalian CNS is supposed to prevent spontaneous axonal regeneration following trauma. One way to circumvent the generally limited capacity of the mature CNS to regenerate axons or to generate new neurons in response to degeneration or injury is neural transplantation of embryonic neurons.

Following lesion of the adult motor cortex, we transplanted E14 motor cortical neurons obtained from green fluorescent protein (GFP) transgenic embryos as we previously described (ref. ¹). Two months after grafting, GFP axons displayed significant, appropriate and long distance projections to several motor cortex targets including distant ones such as the thalamus and the spinal cord. We have also shown that the visualized GFP signal is not due to cell fusion between the grafted cells and the host neurons that might have transported the GFP signal along their own axons. The transplant was well integrated within the host as it filled the lesion cavity and received axonal projections as evidenced by electron microscopy. In addition, GFP axons were often myelinated and formed synapses with the host (see Fig. 1 for a summary of these results). However, not all cortical neurons retained the possibility to reconstruct the motor lesioned pathway as grafts from the visual cortex placed into the pre-lesioned motor cortex contacted mainly visual targets.

Figure 1.

Summary of the main findings following grafting of motor cortex GFP embryonic cells into the prelesioned adult motor cortex. (A) Grafted GFP neurons receive synaptic contacts from the host as evidenced by electron microscopy. The GFP signal is revealed with a secondary antibody labeled with gold particles (black dots) and is present here in a GFP cell body (down) contacted by an axonal terminal (up) from the host filled with neurotransmitter vesicles. (B) Grafted cells do not fuse with host cells as evidenced by FISH analysis. The grafts (left) are derived from a female embryo and are not labeled following FISH targeting the Y chromosome while male host cells (right) show a clear nuclear labeling. (C) Large numbers of GFP+ axons (green) exit the transplant and grow in close association with astrocyte processes (red). (D) Following double labeling of GFP (green) and PLP (red), dense patches of GFP fibers were observed within striatal bundles of descending fibers in which myelin levels were reduced. (E) A GFP axonal terminal (middle) contacted by two axons from the host as evidenced by electron microscopy. (F) Several GFP axons were myelinated. (G) GFP axons contacted distant targets such as the spinal cord. Here cell bodies within the GFP transplant are retrogradely labeled (yellow) following tracer injection into host pyramidal decussation. Note that the host cortical neurons (red) are also labeled.

Materials and Methods

Animals.

Housing of the animals and all animal experimental procedures were carried out in accordance with the guidelines of the French Agriculture and Forestry Ministry (decree 87849) and of European Communities Council Directive (86/609/EEC). All efforts were made to reduce the number of animas used and their suffering.

Surgical procedures.

Four to five month-old C57BL/6 mice (n = 21) supplied by R. Janvier (Le Genest-Saint Isles, France) were used as recipients. The transplantation procedure was carried out as described before (ref. ¹). Briefly, the animals were placed in a stereotaxic apparatus under Avertin (250 mg/Kg) anesthesia. Frontal cortex was aspirated from approximately 0.5 to 2.5 mm rostral to the Bregma and from 0.5 to 2.5 mm lateral to the midline, taking care to leave the corpus callusom intact.

The day following mating is designed embryonic day (E0). Transplants were harvested from E14 fetuses removed by caesarean section from transgenic EGFP mice (Dr. M. Okabe,² Osaka University). Fragments of presumptive frontal cortex were dissected out and gently deposited into the host lesion cavity. Care was taken to maintain the original dorso-ventral and antero-posterior orientations of the cortical fragments during the transplantation procedure. The animals were transplanted immediately after the lesion. Two months after transplantation the animals were transcardially perfused with 4% paraformaldehyde 0.1 M phosphate buffer. The brains were removed and cryoprotected with 30% sucrose solution and 40 µm Microtome sections were cut and collected in Tris-buffered saline (TBS) before being processed for immunohistochemical analysis.

Immunohistochemistry.

Sections were washed in TBS, blocked with 3% bovine albumin (Sigma) with 0.3% Triton X-100 in TBS, and incubated overnight with primary antibody in blocking solution followed by secondary step for single and double staining. Primary antibodies against GFP 1/1000 (rabbit or mouse, Molecular Probes, Eugene, Oregon) were used to identify transplanted neurons, anti-GFAP for identification of astrocytes 1/400 (rabbit, DAKO), anti-MBP 1/500 (rat, SIGMA) or anti-PLP 1/300 (rabbit, gift from Dr. Zalc) for identification of myelin. The sections were then exposed to the appropriate fluorophore conjugated secondary antibody. The sections were examined with a BX60 microscope equipped with fluorescein and rhodamine filters. Selected sections were examined and photographed by a confocal laser scanning microscope setting (Bio-Ra MRC 1024).

Results

The adult CNS is generally considered as a non-permissive environment for axonal out-growth. One of the major barriers to axonal regeneration is glial scarring, the main component of which is astrocytic gliosis.³ Normally quiescent astrocytes in the adult show a vigorous response to injury. However, several studies suggest that, after a traumatic injury, the molecular composition of the glial scar evolves with time from non-permissive to axon growth permissive properties.^4,5 To examine the role of astrocytes in the axonal outgrowth of transplanted neurons, we double labelled the sections with antibodies raised against glial fibrillary acidic protein (GFAP) and GFP. Two months after transplantation, GFAP immunoreactive (IR) astrocytes were present in and around the graft and bilaterally in the corpus callosum and cortex. Higher density of GFAP-IR cells and processes were present in those areas where GFP-IR fibers left the grafts (Fig. 2A–D). At the graft-host interface, the astrocytic processes observed were elongated compared to the short, thickened GFAP processes typical of reactive astroglia. Furthermore, GFAP processes were frequently oriented parallel with the longitudinal axis of the GFP fibers. These findings suggest that the astrocytic processes could be the substrate directing the orientation of the growing donor axons. A number of studies have demonstrated that the radial glial cells are the primary substrate for neural migration in the developing neocortex⁶ and that the intracortical growth of developing callosal axons is similarly linked to radial glia.⁷ In addition, alignment of astroglial cells in the developing central nervous system has been shown to be important for correct outgrowth of several axon tracts, such as the corticospinal tract.⁸ Our results suggest that, under lesion and transplantation condition, the adult cortical astrocytes retain the capacity to re-express earlier developmental cues sufficient to direct efficient axon growth from embryonic transplants. This is in agreement with a previous study demonstrating that adult cortical astrocytes retain the capacity to re-express an earlier developmental phenotype that may partially underlie the active migration of transplanted neurons and neural precursors.⁹

Figure 2.

Astroglial reaction (A–D) and myelin loss (E–H) in host 60 days post-grafting of GFP E14 embryonic cells from the motor cortex. (A) GFAP reactivity in the cortex (Cx) (red) was present in and around the transplant (T). (B) Large numbers of GFP+ axons (green) exit the transplant and grow parallel to astrocytes processes (red). (C and D) Higher power images showing astrocytes processes (C) and close association between GFP fibers and these processes (D). (E–H) Double labeling of GFP (green) and PLP (red). (E and F) Dense patches of GFP fibers were observed within striatal bundles of descending fibers in which myelin levels were reduced (arrowheads) and to a lesser extent where host myelin was intact (arrows). (G and H) High power images showing high density of GFP+ fibers extending through demyelinated striatal bundles of fibers (G) and lower density within still myelinated bundles (red) (H). Scale bar: (A and B) 150 µm; (C and D), 30 µm; (E), 133 µm; (F), 67 µm; (G and H), 25 µm.

The general assumption that the adult brain can only permit a weak axonal regeneration has been attributed to the presence of growth-inhibiting factors in the adult CNS white matter.¹⁰ Myelin proteins that inhibit axonal regeneration include Nogo,^11,12 myelin-associated glycoprotein (MAG)^13,14 and oligodendrocytemyelin glycoprotein (OMgp).¹⁵ Major components of central myelin include proteolipid protein (PLP) and myelin basic protein (MBP). In order to examine whether pre-existing degenerated motor pathways promote graft axons development to deafferented targets, we have used antibodies raised against PLP and MBP and investigated myelin loss in the host following cortical lesion. Transplantation in the cortex of an unlesioned host does not allow for GFP+ axons to develop and reach target areas even close targets such as the ipsilateral adjacent cortex. Cases in which lesion and graft did not reach the deep cortical layers showed only a mild decrease in host myelin and only few GFP+ fibers were identified beyond the caudate putamen nucleus (CPu). In animals where lesion and graft reached the deep cortical layers, a significant decrease in host myelin and an increase in GFP+ fiber density were observed. Indeed, following double labeling of GFP and PLP or MBP, dense patches of GFP+ fibers within striatal white matter were found where host myelin was reduced while fewer GFP+ fibers occurred where host myelin was still present (Fig. 2E–H). Taken together, these data indicate that transplant-derived axons preferentially follow demyelinated pathways. Following damage to the neocortex, significant functional recovery may depend critically on whether neurons can reconstruct the specific projection patterns of the damaged host neurons. Graft axons might be guided to the subcortical targets by cues provided by the degenerated axons. One argument in favor of this hypothesis is that GFP+ fibers are frequently found in bundles of host degenerated fibers where myelin was reduced; this was particularly obvious in the CPu. Both the nature and degree of potential axon guidance cues in different regions of adult CNS following lesions remain unclear.

Discussion

In summary, our results suggest that (1) elongated astrocytic processes constitute a supportive substrate for axonal outgrowth of transplanted embryonic neurons and (2) that transplant-derived axons preferentially follow the degenerated, myelin free, pathways. Along with our previous report (ref. ¹), these results indicate that the adult brain is subject to a greater plasticity than previously suspected and that it is still permissive to axonal growth derived from transplanted embryonic neurons. However, our results also indicate that transplanted neurons retain their specificity and that not all neurons are capable of successfully integrating any brain region and emitting significant and appropriate axonal projections. Thus, research with stem cells should focus on signals/factors, that are yet to be identified, and that not only promote the maturation of stem cells into neurons but, perhaps more importantly, into a specific and appropriate neuronal phenotype and subtype. Failing to do so may result into inadequate or insufficient axonal projections to target areas. This is of particular interest given the substantial amount of research that is currently performed in fields such as the Parkinson's disease (PD) field where significant efforts to restore lacking dopamine into the striatum are directed towards transplanting stem cells-derived dopamine neurons. To this regard, our results regarding the possibility for transplanted embryonic neurons to send long distance axonal projections also indicate that restoration of degenerated nigral dopamine pathways by intranigral grafting should be more thoroughly investigated with mesencephalic embryonic neurons and, later on, with dopamine neurons derived from stem cells. Care should be taken however with the later since not all dopamine neurons might have the capacity to reinnervate the denervated striatum. Among the dopamine neurons types used for transplantation, only the substantia nigra pars compacta subtype appears to possess the capacity to innervate the striatum^16,17 which implies that neurons used for cell transplantation in PD should be of the correct nigral dopamine phenotype.