Abstract

Reconstruction of CNS circuitry is a major aim of neural transplantation, and is currently being assessed clinically using foetal striatal tissue in Huntington's disease. Recent work suggests that neuronal precursors derived from foetal striatum may have a greater capacity than primary foetal striatum to project to the usual striatal target areas such as the globus pallidus and substantia nigra, raising the possibility that they have a greater potential for circuit reconstruction. However, comparing the reconstructive capacity of the two donor cells types is confounded by the fact that many precursor experiments have been carried out in a xenogeneic background in order to utilize species-specific markers for tracking the donor cells, whereas most primary foetal transplant studies have utilized an allograft paradigm. Thus, differences in immunogenic background could influence the findings; for example, xenogeneic grafts may not recognize host inhibitory signals, thereby encouraging more profuse and extensive projections. We have addressed this issue directly by comparing foetal neural precursor and primary foetal grafts in both allo- and xenograft environments using several labelling techniques, including GFP-transgenic mice and LacZ-labelled cells as donor tissue and iontophoretic injection of the anterograde tracers BDA, neurobiotin and PHA-L in the host. We present clear evidence that foetal neural precursors produce grafts with richer axonal outgrowth than primary foetal grafts, and that this is independent of the immunogenic background. Furthermore, both neural precursor and primary grafts derived from human foetal tissue produced a significantly richer outgrowth than do grafts of mouse donor tissue, which may relate to their large final graft volume and the greater intrinsic potential of human CNS neurons for greater axon elongation.

Introduction

Grafts of primary foetal (PF) striatum dissected from human foetal material appear to improve function in patients with Huntington's disease (HD) (Bachoud-Levi et al., 2006). However, the scarcity of human foetal tissue demands that alternative donor cells be identified, one potential source being human foetal neural precursor (FNP) cells (Kelly et al., 2006). An essential requirement for graft function is that the donor cells reconstruct circuitry lost due to the disease, and it is generally expected that the closer the reconstruction to normal, the better the functional improvement (Fricker-Gates et al., 2001). Depending on the nature of the pathology, it may be important that a graft can project to distant sites. In HD specifically, medium spiny neurons are lost from the striatum, and cells grafted into this region may be required to project to the globus pallidus, substantia nigra and subthalamic nucleus (Wictorin et al., 1992). PF striatal allografts appear to send axons to distant targets such as the globus pallidus and substantia nigra (Wictorin, 1992). However, these projections are less dense than those seen from human tissue grafts where significant dense projections are seen at long distances (Pundt et al., 1996). Extensive outgrowth, including to striatal-specific targets, has also been reported from 10-day expanded human FNPs following transplantation into the adult lesioned rat brain (Armstrong et al., 2000). This could represent a substantial intrinsic capacity of FNPs to project into the host parenchyma or could be a result of the xenograft environment, i.e. it may be that there are specific guidance cues or ‘stop’ signals that are species-specific and thus produce differential growth patterns depending on the origin of the donor cells (Garcia et al., 1995). A further interpretation is that the profuse outgrowth seen from the human FNPs was due to the high labelling efficiency of the species-specific marker, and that allograft projections have been previously under-reported due to the technical difficulties of reliably identifying allografted cells and their axons. Thus, it is crucial that a systematic comparison of allografted and xenografted cells takes account of this and that labelling systems are of comparable efficiency in both.

Much use has been made of species-specific markers such as NF70, which reliably label human donor neuronal projections but not the rat host CNS. Of course, species-specific markers could not be used for allografts. Furthermore, we found the mouse-specific markers M2 and M6 (which are reported to label mouse neuronal and astrocytic tissue, respectively) to be markedly less reliable or specific in our hands than has generally been reported, necessitating the identification of alternative methods of labelling graft projections. We explored several options, each with its own limitations: the use of donor tissue from transgenic mice expressing GFP under a prion promoter; transfecting the donor cells with a lentiviral vector (LV) carrying the LacZ transgene; and labelling grafts with anterograde tracers neurobiotin (NBT), biotynilated dextran amine (BDA) or Phaseolus vulgaris leucoagglutinin (PHA-L).

This study aimed to determine whether grafted FNPs do indeed project more extensively than grafted PF cells in a lesion model of HD, or whether previous indications that this is the case related to the immunogenic background or is an artefact of labelling techniques.

Materials and methods

The experimental design was structured so as to allow a comparison between allo- and xenograft environments as well as PF versus FNP tissues. For human tissue of course, the allograft arm of the experiment would have required transplantation into humans and so a full four-way systematic comparison of human allo- and xenografted PF and FNPs could only be achieved for rodent tissue. As species differences have been noted for FNPs, it was important to validate the results using human cells (in the xenograft arm alone). Table 1 illustrates the experimental design.

Table 1

Outlines the transplant paradigms in which graft axonal projections were analysed, and presents the number of animals with surviving grafts in each group

Donor tissue Labels No. surviving grafts/no. animals grafted 
  Host lesioned animal 
  Mouse Rat 
Control Lesion only */8 */8 
Mouse PF GFP 8/8 6/8 
 LacZ 8/8 7/8 
 BDA 8/8 6/8 
 Neurobiotin 7/8 5/8 
 PHA-L 7/8 7/8 
Mean (%)  95 77.5 
Mouse FNP GFP 8/8 7/8 
 LacZ 7/8 5/8 
 BDA 7/8 7/8 
 Neurobiotin 7/8 6/8 
 PHA-L 8/8 6/8 
Mean (%)  92.5 77.5 
Human primary pNF70 7/8 
Human FNP pNF70 6/8 
Mean (%)   81 
Donor tissue Labels No. surviving grafts/no. animals grafted 
  Host lesioned animal 
  Mouse Rat 
Control Lesion only */8 */8 
Mouse PF GFP 8/8 6/8 
 LacZ 8/8 7/8 
 BDA 8/8 6/8 
 Neurobiotin 7/8 5/8 
 PHA-L 7/8 7/8 
Mean (%)  95 77.5 
Mouse FNP GFP 8/8 7/8 
 LacZ 7/8 5/8 
 BDA 7/8 7/8 
 Neurobiotin 7/8 6/8 
 PHA-L 8/8 6/8 
Mean (%)  92.5 77.5 
Human primary pNF70 7/8 
Human FNP pNF70 6/8 
Mean (%)   81 

Note: All animals receiving xenografts were immunosuppressed with Cyclosporin A starting 1 day prior to transplantation. Analysis was carried out on surviving grafts only and all animals that presented with rejected grafts were eliminated from the analysis. *The lesion groups (n = 8) did not receive grafts. X, absent condition.

Human and mouse tissue preparation

Collection, dissection and preparation of human CNS tissue

Human foetal tissue was collected by transvaginal ultrasound guided low pressure aspiration at routine surgical termination of pregnancy (TOP), following the guidelines of the Polkinghorne report (Polkinghorne, 1989) and with Local Research Ethics Committee approval and the UK Department of Health (1995). Full consent was obtained from the maternal donor, as part of the MRC-sponsored, South Wales initiative for transplantation in HD (SWIFT-HD) program. Foetal tissue ranged in age from 7 to 11 weeks post conception and foetal age was based on ultrasound measurements and confirmed using foetal morphometric details using a mathematical model (Evtouchenko et al., 1996; Hurelbrink et al., 2003). Striatal eminences were dissected and human cells were treated in one of the two ways: (i) stored as primary cells overnight in hibernation medium (Hibernate E, Gibco) at 4°C at a density of 500 000 cells/ml or (ii) expanded as FNPs for 10 days (see later).

Collection, dissection and preparation of mouse CNS tissue

Mouse tissue was obtained from embryonic day 14 (E14) transgenic mice carrying the green fluorescent protein (GFP) under the prion promoter on a C57/BL6 background (Feil et al., 1996). The stock was maintained in house, and pregnant dams were sacrificed for embryonic tissue donation at E14 days of embryonic age. Embryos were collected in Hanks balanced salt solution over ice. Following decapitation, the striatal eminencies were dissected (Dunnett et al., 1992). Dissected tissue was collected using a Pasteur pipette and left to settle in Hanks solution on ice.

Propagation of mouse (m) and human (h) striatal FNPs

Propagation was the same for both mouse and human tissue except that human tissue culture media contained an extra growth factor, leucaemia inhibitory factor (LIF). Coarse single cell suspensions of whole ganglionic eminence were prepared as described previously (Fricker et al., 1996). Briefly, the medium was removed and 200 μl of trypsin/DNAse was added to the tissue for 20 min at 37°C. Trypsin inhibitor and DNAse were added, mixed and incubated for 5 min at 37°C. The tissue was then washed twice with normal medium, Dulbecco's modified Eagle's medium (DMEM F-12) supplemented with 1% PSF, and then collected by centrifugation at 1000 rpm for 3 min. The medium was aspirated off and the tissue was resuspended in normal medium plus growth factors (FGF-2, 20 ng/ml, R&D Systems and EGF, 20 ng/ml, Sigma), and in the case of human tissue cultures LIF (10 ng/ml, Sigma), this basic media plus growth factors constitutes the proliferation medium. The tissue was resuspended in 200 μl normal medium and then triturated to produce a single-cell suspension. Cells were counted in a haemocytometer using trypan blue exclusion to assess the viability of the cells. The cell concentration was adjusted to culture the cells at a concentration of 200 cells/μl in T25 flasks with 10 ml of proliferation medium (normal media plus growth factors). Cultures were maintained at 37°C in humidified 5% CO2, 95% atmospheric air.

FNPs were fed, by replacing half the medium with fresh medium containing twice the concentration of B27, EGF, FGF-2 and LIF (where appropriate), every 3–4 days.

For transplantation tissue was either prepared from fresh (PF) tissue or expanded in culture (FNP) for 10 days, as described, and the concentration was adjusted to graft 5 × 105 cells per animal in all cases.

Characterization of FNPs in vitro

The characteristics of striatal FNPs from either species following 10 days propagation were assessed in vitro. Briefly, spheres were dissociated to a coarse single-cell suspension and plated onto poly-l-lysine-coated coverslips at a density of 1 × 105 cells in 30 μl differentiation medium. After 4–6 h cells were flooded with 500 μl of differentiation medium and allowed to differentiate for 7 days prior to fixation. Cells were fed by replacing half the medium with fresh medium every 3 days.

Indirect fluorescent immunocytochemistry was performed using standard protocols with primary antibodies directed against β-III tubulin (1:1000, Sigma), β-galactosidase (β-gal, 1:6000, Promega) and glial fibrillary acidic protein (GFAP, 1:1000, DAKO). Fluorescent staining was visualized on a Leitz DRMB microscope, and cell counts performed at ×40 magnification. Pseudocolour fluorescent images were obtained using Openlab 2.1 image analysis software.

LacZ labelling of PF and FNP cells in vitro

NP and 10-day expanded FNP cells were treated with LacZ prior to transplantation. Conditions for infection were optimized by varying the concentration of virus and infection time with optimal infection achieved at MOI (multiplicity of infection) = 2 for 1 h. All experimental cell suspensions were treated similarly.

Transplantation experiments

A total of 192 animals were lesioned (104 rats and 88 mice) and received grafts with an n = 8 for each graft group as laid out in Table 1. All animal experiments were performed in full compliance with local ethical guidelines and approved animal care according to the UK Animals (Scientific Procedures) Act 1986 and its subsequent amendments. Adult female Sprague-Dawley (Harlan UK) rats typically weighing 200–250 g at the start of the experiments were used. They were housed in cages of four in a natural light–dark cycle with access to food and water ad libitum.

All surgery was performed under isoflurane anaesthesia. Anaesthesia was induced in an induction box with isoflurane and oxygen (4 l/min), and maintained by passive inhalation of isoflurane (1–2 l/min) and a mixture of oxygen (0.8 l/min) and nitrous oxide (0.4 l/min). Animals were allowed to recover in a warmed recovery chamber and received analgesia through paracetamol dissolved in drinking water (2 mg/ml) for 3 days subsequent to surgery.

Quinolinic acid lesioning of mouse and rat striatum

All grafts were placed into the unilateral quinolinic acid (QA) lesioned striatum. Animals received unilateral injections of 45 nmol quinolinic acid into the right striatum. An amount of 0.75 μl of QA was infused to two sites over a 4-min period using a 1 μl Hamilton syringe targeted at stereotaxic coordinates: −3.2/−2.4 mm lateral (L) of bregma, +0.4/+1.4 mm anterior (A) of bregma and −5.0/−4.5 mm below dura (vertical, V) with the incisor bar set to 2.3 mm below the interaural line. The syringe needle was left in situ for 3 min before retraction following each injection. After completion of both injections the wound was cleaned and sutured.

For quinolinic acid lesions of the mouse striatum, one infusion site was used with injection of 0.5 μl over a 4-min period targeted at the stereotaxic coordinates: A = +0.9 mm, L = −1.8 mm, V = −2.7 mm with the incisor bar set to 0 mm below the interaural line.

Neural transplantation

All animals that were receiving a xenograft were immunosuppressed with daily intraperitoneal injections of Cyclosporin A (CsA, Sandimmun, 10 mg/kg) for the duration of the experiment, commencing the day prior to transplantation. Antibacterial prophylaxis was administered by addition of aureomycin to the drinking water (5 g/l) with sucrose (5 g/l) and sodium chloride (0.5 g/l) throughout the course of the experiment.

For transplants into rat brain, 5 × 105 cells in 2 µl were injected stereotactically over 2 min via a 10 μl Hamilton microsyringe with a thin-walled widebore needle (diameter = 0.25 mm). All grafts were placed ipsilateral to the side of the lesion and were targeted to the striatum (A = +1.0 mm, L = −2.8 mm, V = −5.0/−4.5 mm, with the incisor bar set 2.3 mm below the interaural line). The needle was left in situ for 3 min following grafting to minimize reflux of grafted cells along the needle tract, following which the needle was removed and the skin incision cleaned, closed and sutured.

For transplants into adult mouse brain, 2 μl of the graft suspension were targeted to the lesioned striatum at the coordinates: A = +0.9 mm, L = −1.8 mm, V = −2.7 mm, with the incisor bar set to 0 mm below the interaural line, and injected as earlier. Animals in the tracer groups received transplants of mouse tissue untreated. One week prior to perfusion, 11 weeks post-transplantation animals received an iontophoretic injection of tracer to the graft site. The tracer was injected with a square wave pulse of 10 µA at a rate of 7 s on 7 s off for 20 min with a micropipette of 20–50 μm diameter. Control animals received grafts of dead cells and in the tracer groups control animals received an injection of tracer to the lesioned striatum to estimate the extent of non-specific labelling within the lesioned striatum when analysing the graft data.

Transplant immunohistochemistry

Twelve weeks following transplantation animals were transcardially perfused and their brains processed for histological analysis. Serial coronal 40 μm frozen sections were prepared, collected and stored appropriately. A 1:12 series of sections were processed for Nissl staining using cresyl violet. A further 1:6 series was processed for indirect single-label immunohistochemistry with the following primary antibodies: mouse anti-HuNu (1:1500, Chemicon) (human-specific nuclei), NF70 (1:500, supplied by Prof. Soriano, e-mail address: soriag@aol.com), rabbit anti-β-Gal (1:6000, Cappel), chicken anti-GFP (1:4000, Chemicon), Neurobiotin, PHA-L and BDA.

The basic protocol was identical in all cases and visualization was via the DAB method. Where double labelling was carried out, the first antibody was completed before commencing the second antibody that was visualized using the Vector SG kit. Staining controls consisted of omission of the primary antibody and these confirmed the specificity of staining in all cases.

Quantification of graft parameters

Graft volume and cell number was based on Nissl staining and determined using the Olympus CASTgrid stereology system on a 1:12 series. The fractionator method was used which samples a fraction of the total cells, i.e. the fraction of the area samples, the fraction of sections stained and the ratio between the section thickness and the dissector from which they were counted (West, 1999). For each labelling technique, 4–8 sections were analysed for each animal as pilot data has shown this to be sufficient. A sampling grid of 545 µm at 100× objective was used with at least 10 sampling areas per section. Graft volume was calculated for each experimental group separately, except for animals in which post-graft anterograde tracing methods were used. As grafted cells were all wild-type CD1 mouse-derived cells for these experiments, the groups were combined for analysis of volume and cell number, which was calculated using a stereological random sampling method. Axonal outgrowth from the graft was calculated by manual counts of axonal projections or, in the case of human grafts where the fibre density was such that accurate counts could not be made, an estimation of fibre density was made.

Results

Cell characterization in vitro

Prior to transplantation, striatal mouse (m) FNPs underwent a 5× increase in absolute cell number after 10 days, whereas human (h) FNPs underwent a 1.6× increase over the same time period. This difference most likely reflects previously reported species differences in the rate of cell turnover (Kelly et al., 2005).

Both PF and FNP cell suspensions prior to transplantation showed viabilities above 95% in all cases, based on trypan blue exclusion assay analysis. Exposure to the LacZ LV in vitro resulted in some cell death and so cell suspensions were adjusted to an appropriate volume to account for this for transplantation purposes. Uptake of the virus was analysed by immunohistochemistry and found to be in the range of 60–70%.

Lesion morphology

Control animals that were lesioned, but received no grafts, showed histological characteristics typical of a QA striatal lesion: shrinkage of the striatum with a compensatory increase in the ventricular volume on the side of the lesion (Fig. 1). To control for the possibility of tracer leakage into host parenchyma, non-grafted quinolinic acid lesioned animals received an identical injection of tracer. Except for very rare fibres around the immediate periphery of the lesion in a small number of animals, no projections were seen and certainly no projections at a distance.

Fig. 1

Nissl staining of (A) QA lesion-only in mouse striatum; (B) mPF grafts into a QA lesioned mouse striatum (allograft); (C) QA lesion-only in rat striatum; (D) mPF graft into QA lesioned rat striatum (xenograft). This demonstrates the characteristics of a QA lesioned brain with enlargement of the ventricles resulting from the cell death in the striatum (A and C). Grafted tissue (B and D) can be identified by the increased density in staining in the right striatum. Scale bar = 500 μm.

Fig. 1

Nissl staining of (A) QA lesion-only in mouse striatum; (B) mPF grafts into a QA lesioned mouse striatum (allograft); (C) QA lesion-only in rat striatum; (D) mPF graft into QA lesioned rat striatum (xenograft). This demonstrates the characteristics of a QA lesioned brain with enlargement of the ventricles resulting from the cell death in the striatum (A and C). Grafted tissue (B and D) can be identified by the increased density in staining in the right striatum. Scale bar = 500 μm.

Graft survival

Graft survival in general was good: 94% of grafts survived in all mouse-to-mouse transplants and 77.5 and 81% for all mouse to rat and all human to rat, respectively. (Table 1). All animals with a rejected graft were removed from the analysis and all results are based on analysis only of the animals with surviving grafts. There was no significant difference in the volume or cell number of PF compared to FNP mouse grafts in either the allo- or xenograft situation. However, despite the transplantation of equal numbers of cells, at 12 weeks post-transplantation, grafts of human cells were significantly larger and contained significantly more cells than those of mouse [Fig. 2A, F(2,192) = 562.31, P < 0.001 and Fig. 2B, F(2,192) = 372.95, P < 0.001, respectively]. Animals that received tracer injections had no significant difference in the number of cells within the graft for each tracer used (Fig. 2C). Lesion only control animals receiving dead cells showed no signs of positive cells within the graft area at 8 or 12 weeks post-transplantation. Lesion-only control animals that received an anterograde tracer injection showed no labelling of cells and no major projections, although there were occasional positive fibres from cells on the edge of the lesion.

Fig. 2

(A) There was no overall difference in graft volume between PF and FNP grafts. Volume of mouse-to-mouse and mouse-to-rat grafts did not differ but there was a significant difference in the volume of human-to-rat grafts compared to all other graft groups. (B) The total number of cells was calculated from Nissl stained sections, animals receiving grafts of human tissue was greater compared to all other grafted animals. (C) There was no significant difference in the number of graft cells labelled with each of the anterograde tracers for each experimental group. (M–M = donor mouse cells to mouse host, M–R = donor mouse cells to rat host, H–R = donor human cells to rat host).

Fig. 2

(A) There was no overall difference in graft volume between PF and FNP grafts. Volume of mouse-to-mouse and mouse-to-rat grafts did not differ but there was a significant difference in the volume of human-to-rat grafts compared to all other graft groups. (B) The total number of cells was calculated from Nissl stained sections, animals receiving grafts of human tissue was greater compared to all other grafted animals. (C) There was no significant difference in the number of graft cells labelled with each of the anterograde tracers for each experimental group. (M–M = donor mouse cells to mouse host, M–R = donor mouse cells to rat host, H–R = donor human cells to rat host).

Axonal projections

Comparison of labelling techniques

By using a number of different labelling techniques, each of which has recognized strengths and weaknesses, we aimed to increase confidence in the results. Following transplantation, GFP transgenic mouse tissue (both PF and FNP) was downregulated, resulting in partial labelling within the graft. LacZ-labelled grafts were identified using an antibody to β-galactosidase. Not all cells in the graft were labelled with the virus, however, a substantially greater proportion of the graft labelled with the LacZ virus than with GFP. The anterograde tracers BDA, Neurobiotin and PHA-L were also used. Each tracer was injected iontophoretically into the grafted striatum and resulted in extensive labelling of the graft, which was similar for all anterograde labels (Fig. 2C). For all methods, but to a lesser extent with GFP, graft-derived projections were observed as far rostrally as the olfactory bulb and caudally in the globus pallidus, substantia nigra, basal nucleus, internal capsule, corpus callosum, ventrolateral thalamic nucleus and the entopeduncular nucleus.

In the following sections we report the findings for each condition as summarized in Table 2.

Table 2

Summary of fibre projections in the host brain using all approaches

  CC IC GP STN SN 
Allograft Primary mouse to mouse +/− ++ +/− 
 Iontophoresis only +/− ++ +/− 
Xenograft Expanded mouse to mouse ++ +++ +/− +/− 
Iontophoresis only ++ +++ +/− +/− 
Primary mouse to rat +/− +++ +/− 
Iontophoresis only +/− +++ +/− 
Expanded mouse to rat ++ +++ +/− 
Iontophoresis only ++ +++ 
Primary human to rat +++ +++ ++ 
 Expanded human to rat +++ +++ +++ 
  CC IC GP STN SN 
Allograft Primary mouse to mouse +/− ++ +/− 
 Iontophoresis only +/− ++ +/− 
Xenograft Expanded mouse to mouse ++ +++ +/− +/− 
Iontophoresis only ++ +++ +/− +/− 
Primary mouse to rat +/− +++ +/− 
Iontophoresis only +/− +++ +/− 
Expanded mouse to rat ++ +++ +/− 
Iontophoresis only ++ +++ 
Primary human to rat +++ +++ ++ 
 Expanded human to rat +++ +++ +++ 

Note: 0 = no fibres, +/− = negligible fibres, + = tens of fibres, ++ = hundreds of fibres and +++ = thousands of fibres, observed in each region, based on estimated or actual counts. Iontophoresis only animals for the corresponding groups are shown in red.

mPF allo- and xenografts

Analysis of mPF grafts derived from the GFP transgenics revealed that not all cells in the defined graft region expressed GFP, as detailed earlier. Furthermore, no migrating cells were recorded. Axonal projections appeared limited; fibres were only seen adjacent to the graft in the striatal neuropil and none were seen traversing other brain regions (Figs 3A–C and 4A–C). In contrast, in animals in which non-transgenic CD1 mouse cells were transplanted, either following infection with LacZ LV or non-infected cells later labelled with anterograde tracers, cells were seen to have migrated out from the graft, for example in the corpus callosum and the internal capsule (Figs. 3D–F and 4D–F), and projections were observed crossing the graft–host border into the corpus callosum, and into the medial and lateral globus pallidus and internal capsule ipsilaterally.

Fig. 3

All sections are taken from mPF striatal xenografts in the QA lesioned rat adult brain. (A) and (E) are low-power magnification showing a few projections protruding from the graft core. (B) Higher power of boxed area in (A) showing GFP positive projections (arrows). (F) Demonstrates the greater proportion of projections that were labelled with LacZ. Some GFP immune positive cells were seen in the corpus callosum and in the globus pallidus (C and D, respectively, arrows). (G) and (H) show LacZ positive fibres in the internal capsule and the subthalamic nucleus, respectively. Scale bar (A and B) = 500 μm, all others = 200 μm. (I) Represents a schematic of the summary of areas where projections (as shown by grey lines) were seen in all primary mouse grafts to the rat host brain. Corpus callosum (cc), globus pallidus (GP), internal capsule (ic), subthalamic nucleus (STN).

Fig. 3

All sections are taken from mPF striatal xenografts in the QA lesioned rat adult brain. (A) and (E) are low-power magnification showing a few projections protruding from the graft core. (B) Higher power of boxed area in (A) showing GFP positive projections (arrows). (F) Demonstrates the greater proportion of projections that were labelled with LacZ. Some GFP immune positive cells were seen in the corpus callosum and in the globus pallidus (C and D, respectively, arrows). (G) and (H) show LacZ positive fibres in the internal capsule and the subthalamic nucleus, respectively. Scale bar (A and B) = 500 μm, all others = 200 μm. (I) Represents a schematic of the summary of areas where projections (as shown by grey lines) were seen in all primary mouse grafts to the rat host brain. Corpus callosum (cc), globus pallidus (GP), internal capsule (ic), subthalamic nucleus (STN).

Fig. 4

All sections are taken from mPF striatal allografts in the QA lesioned mouse brain. (A–C) are GFP-labelled grafts and (D–F) are LacZ-labelled grafts. (A) Shows the graft mass of the largest graft in the GFP group with all others being small thin grafts, similar to those shown in Figs 5 and 6. Positive cells and projections were seen in the globus pallidus (B and F, arrows), internal capsule (C), striatum (D, arrow) and the corpus callosum (E). Insert show higher power magnification images of GFP projections (C) and LacZ-labelled neurons (F). (G) Represents a schematic of the summary of areas where projections (as shown by grey lines) were seen in all primary mouse grafts to the mouse host brain. Scale bar = 500 μm (A–E) and = 200 μm (F). Corpus callosum (cc), globus pallidus (GP), internal capsule (ic).

Fig. 4

All sections are taken from mPF striatal allografts in the QA lesioned mouse brain. (A–C) are GFP-labelled grafts and (D–F) are LacZ-labelled grafts. (A) Shows the graft mass of the largest graft in the GFP group with all others being small thin grafts, similar to those shown in Figs 5 and 6. Positive cells and projections were seen in the globus pallidus (B and F, arrows), internal capsule (C), striatum (D, arrow) and the corpus callosum (E). Insert show higher power magnification images of GFP projections (C) and LacZ-labelled neurons (F). (G) Represents a schematic of the summary of areas where projections (as shown by grey lines) were seen in all primary mouse grafts to the mouse host brain. Scale bar = 500 μm (A–E) and = 200 μm (F). Corpus callosum (cc), globus pallidus (GP), internal capsule (ic).

Fig. 5

All sections are taken from mFNP striatal xenografts in the QA lesioned rat brain. (A–D) are GFP-labelled grafts and (E–H) are LacZ-labelled grafts. Projections can be seen emanating across the graft–host border (A and E), (B and F) are higher power views of the indicated areas in (A and E). GFP positive fibres were seen projecting in the internal capsule (C) and to the globus pallidus (D). Inserts represent higher power images of the graft projections. LacZ-labelled cells and fibres were identified in the globus pallidus and the cortex (arrows), G and H, insert represent higher power images of the graft projections. Scale bar = 500 μm. (I) Represents a schematic of the summary of areas where projections (as shown by grey lines) were seen in all expanded mouse grafts to the rat host brain. Corpus callosum (cc), globus pallidus (GP), internal capsule (ic), cortex (Ctx).

Fig. 5

All sections are taken from mFNP striatal xenografts in the QA lesioned rat brain. (A–D) are GFP-labelled grafts and (E–H) are LacZ-labelled grafts. Projections can be seen emanating across the graft–host border (A and E), (B and F) are higher power views of the indicated areas in (A and E). GFP positive fibres were seen projecting in the internal capsule (C) and to the globus pallidus (D). Inserts represent higher power images of the graft projections. LacZ-labelled cells and fibres were identified in the globus pallidus and the cortex (arrows), G and H, insert represent higher power images of the graft projections. Scale bar = 500 μm. (I) Represents a schematic of the summary of areas where projections (as shown by grey lines) were seen in all expanded mouse grafts to the rat host brain. Corpus callosum (cc), globus pallidus (GP), internal capsule (ic), cortex (Ctx).

Fig. 6

All sections are taken from mFNP striatal allografts in the QA lesioned mouse brain. (A–D) are GFP-labelled grafts and (E–H) are LacZ-labelled grafts. Projections can be seen emanating across the graft–host border (A, B and E). GFP positive fibres were seen projecting in the internal capsule (C) and to the globus pallidus (D and H, arrows). Immune positive fibres (arrow) were observed in the grey matter close to the graft (F). Insert in F shows a high-power image of the graft projections. Dense axonogenesis of the grafted cells was observed on LacZ-labelled cells (G, arrow). Scale bar = 500 (A and E) and 200 μm (all others). (I) Represents a schematic of the summary of areas where projections (as represented by grey lines) were seen in all mFNP mouse grafts to the rat host brain. globus pallidus (GP), internal capsule (ic).

Fig. 6

All sections are taken from mFNP striatal allografts in the QA lesioned mouse brain. (A–D) are GFP-labelled grafts and (E–H) are LacZ-labelled grafts. Projections can be seen emanating across the graft–host border (A, B and E). GFP positive fibres were seen projecting in the internal capsule (C) and to the globus pallidus (D and H, arrows). Immune positive fibres (arrow) were observed in the grey matter close to the graft (F). Insert in F shows a high-power image of the graft projections. Dense axonogenesis of the grafted cells was observed on LacZ-labelled cells (G, arrow). Scale bar = 500 (A and E) and 200 μm (all others). (I) Represents a schematic of the summary of areas where projections (as represented by grey lines) were seen in all mFNP mouse grafts to the rat host brain. globus pallidus (GP), internal capsule (ic).

mFNP allo- and xenografts

mFNP GFP-labelled grafts showed some limitation of GFP expression within both grafts and projections, but not to the extent seen in mPF grafts described earlier. In contrast to mPF grafts, cell migration from the mFNP graft core was more pronounced in both mouse and rat host striatum. Projections were observed in similar areas to that of mPF grafts, such as the corpus callosum, the internal capsule and the globus pallidus. However, mFNP projections were also observed in more distal regions such as the substantia nigra, the ventrolateral thalamic nucleus, the subthalamic nucleus, the entopeduncular nucleus and the basolateral amygdaloid nucleus (Figs 5 and 6). As for the PF grafts described earlier, the use of LacZ-labelled cells and the anterograde tracers allowed the identification of graft projections that were not observed from GFP-labelled cells due to the downregulation of the GFP transgene in vivo (Fig. 7).

Fig. 7

Anterograde tracers were inotophoretically injected into the graft region one week prior to perfusion. Immune positive fibres were identified for each tracer in a similar pattern to that of LacZ-labelled cells and a sample of each tracer is shown here for clarity of presentation. (A) and (B) show BDA positive cells within mPF xenograft. (C) and (D) show neurobiotin positive projections in the corpus callosum and the internal capsule from mFNP xenografts. (E) and (F) show PHA-L positive projections in the globus pallidus from mPF (E) and mFNP (F) allografts. There were no major differences in the pattern of projections between specific labels. Arrows point to immune positive fibres. Scale bar = 500 μm. Corpus callosum (cc), globus pallidus (GP), internal capsule (ic).

Fig. 7

Anterograde tracers were inotophoretically injected into the graft region one week prior to perfusion. Immune positive fibres were identified for each tracer in a similar pattern to that of LacZ-labelled cells and a sample of each tracer is shown here for clarity of presentation. (A) and (B) show BDA positive cells within mPF xenograft. (C) and (D) show neurobiotin positive projections in the corpus callosum and the internal capsule from mFNP xenografts. (E) and (F) show PHA-L positive projections in the globus pallidus from mPF (E) and mFNP (F) allografts. There were no major differences in the pattern of projections between specific labels. Arrows point to immune positive fibres. Scale bar = 500 μm. Corpus callosum (cc), globus pallidus (GP), internal capsule (ic).

hPF xenografts

hPF tissue grafts were identified using the human-specific immune marker NF70 (Fig. 8). Grafted cells were located throughout the striatum as well as along the needle tract as a result of reflux during the grafting procedure. Projections were seen crossing the graft–host border to the corpus callosum, globus pallidus, and caudally to the subthalamic nucleus and entopeduncular nucleus and substantia nigra.

Fig. 8

An antibody to NF70 was used to identify the human-derived grafted PF cells in the host brains: (A) shows part of the graft core, (B) shows a fibre bundle under higher power. (C) Represents a schematic of the summary of areas where projections (as represented by grey lines) were seen in all PF and FNP human grafts to the rat host brain.

Fig. 8

An antibody to NF70 was used to identify the human-derived grafted PF cells in the host brains: (A) shows part of the graft core, (B) shows a fibre bundle under higher power. (C) Represents a schematic of the summary of areas where projections (as represented by grey lines) were seen in all PF and FNP human grafts to the rat host brain.

hF NP xenografts

hFNPs were also identified using NF70. Immune positive graft cells were located in the striatum with projections emanating across the graft–host border. Immune positive projections were identified in the corpus callosum and globus pallidus, and more caudally in the subthalamic nucleus, entopeduncular nucleus and the substantia nigra.

Discussion

Labelling techniques

Not surprisingly, given the lower proportion of labelled graft cells with GFP compared to the other methods, many fewer projections were observed emanating from these grafts (for example see panels C and G for Fig. 5). Compared to Nissl stained sections, only a small proportion of the cells were positive for GFP, suggesting downregulation of the transgene in vivo. This was in accordance with a subsequent study, in which we demonstrated a reduction in the proportion of GFP-labelled cells in the grafts with increasing post-transplantation survival times (C.M Kelly and A.E. Rosser, unpublished observation). Eriksson and colleagues (2003) also reported a similar downregulation of GFP expression, albeit under the control of a different promoter. Thus, although the GFP transgenic grafts have provided information about the relative capacity of PF and FNP cells to project in the host CNS, we conclude that they under-represent projections.

The anterograde tracers Neurobiotin, BDA and PHA-L have been shown to label neurons clearly in vivo (Novikov, 2001). There was no significant difference in the labelling pattern of the three tracers, although in our hands BDA and Neurobiotin gave a cleaner reaction than did PHA-L. Results from the anterograde tracing experiments corresponded to those observed in LacZ-labelled grafts. Other studies incorporating the use of anterograde tracers to label the PF-derived projections in the host brain have also reported fibre outgrowth in the globus pallidus, and in some cases in the entopeduncular nucleus and substantia nigra (Pritzel et al., 1986; Wictorin et al., 1989). Where human tissue was the donor, immune positive fibres were also observed more caudally in the substantia nigra and the cerebral peduncle.

Graft-derived projections

The results presented here and summarized in Table 2, confirm that PF cells project into the host parenchyma, including some distant sites and that expanded FNPs show a more marked ability to do so. Moreover, these results were not affected by the allo- or xenograft background. The projections were present in sites both rostral and caudal to the graft core in all groups.

mPF and mFNP grafts in either the allo- or xenograft environment sent projections to striatal targets, possibly via the internal capsule, including the pallidum, globus pallidus, subthalamic nucleus and basal nucleus. Projections were also observed in host white matter tracts in which branching of the fibres could be seen. Tiny numbers of fibres were observed in the substantia nigra of mFNP grafts only, but numbers were similar for allo- and xenograft environments. Human graft-derived projections were also seen in the same striatal target regions but generally in greater numbers, and were also observed at greater distances from the graft, for example in the olfactory bulb. More dense projections were seen in the subthalamic nucleus from hFNPs, supporting the finding we report using mouse tissue.

There are a number of explanations for the findings presented in this study.

FNPs appear to project more densely than PF cells

It is clear from our results that denser projections were generally seen from FNP grafts than from PF grafts. One possibility could be that FNP cells have a greater proliferation potential following transplantation resulting in increased cell number and consequently more graft cells capable of sending out projections. However, there was no significant difference in either graft volume or graft cell number between PF or FNP grafts suggesting that this is unlikely to be a major factor.

FNP cells had a greater tendency to migrate outside the graft core. Thus, another possibility is that these migrated cells made a significant contribution to apparent projections, in particular having the potential to reach more caudal brain regions. This was addressed by using iontophoretic injection of tracers that label donor cells in the graft core only. These experiments revealed projections in more caudal nuclei such as the substantia nigra only in FNP grafts but also confirmed that for most regions the majority of projections originated in the graft core and not from migrated cells. Nevertheless, it is likely that small numbers of additional projections do originate from cells that have migrated away from the graft core.

Does the xenograft background encourage projections into the host?

As outlined in the introduction, previous studies have suggested that the xenograft background may encourage graft derived projections into the host parenchyma. Indeed, it has been previously reported that hPF tissue transplanted to the rat brain can send out extensive projections to the pallidum and the substantia nigra (Wictorin et al., 1990, 1992). Furthermore, xenograft projections have been reported to be limited to the normal target sites of the donor tissue itself (Isacson et al., 1984; Wictorin et al., 1992; Garcia et al., 1995). Armstrong and colleagues (2000) have reported that hPF striatum, grafted to the lesioned adult striatum, projects only as far as the globus pallidus, whereas the projections of expanded hFNP grafts extended further and were present in non-specific targets of the host tissue. In this study we performed a direct comparison of projections both from PF and expanded FNPs in both allo- and xenograft background and found no evidence that the immunological background influenced the capacity of the cells to project into the host parenchyma nor affected the sites to which they projected.

A further consideration is whether immunosuppression itself could have an effect since all xenografted, but not the allografted, animals in this study were administered CsA throughout the experiment. CsA has been reported to have an effect on neuronal survival and growth in vitro (Steiner et al., 1997). However, this also is unlikely to be a major factor, since there were no significant differences between the allo- and xenograft environments; rather, the major factor affecting outgrowth appears to be in vitro cell expansion. Further studies in which all animals, both allo- and xenografted, are immunosuppressed are required for direct comparison between all graft groups to fully understand more subtle effects of CsA on the graft.

Human grafts project more extensively

Human-derived grafts sent greater numbers of projections into the rat host brain than did mouse-derived grafts. However, both hPF and hFNP grafts were much larger in volume than the equivalent mouse grafts, which can account for the denser projections. The fact that the same numbers of cells were grafted in all experimental groups suggests that human foetal tissue continued to proliferate in vivo following transplantation to a greater extent than did rodent cells.

In addition to more dense projections, the human grafts also appeared to project further in the rat brain than did the mouse cells. This may relate to the relatively larger final size of the human brain: human tissue being derived from a phylogenetically more mature donor presumably has the necessary potential for long-distance growth. Studies of pig tissue xenografted to the rat brain also show the potential for long-distance projections to traverse the host brain (Armstrong et al., 2002, 2003). Thus, caution must be used when interpreting human to rat xenografts in terms of understanding their clinical potential.

We conclude from this study that the capacity for outgrowth from FNPs is greater than that of PF-grafted tissue. Further studies assessing the physiological and behavioural differences between PF and FNP grafts are warranted to understand the functional significance of these projections.

Acknowledgements

The authors would like to thank Miss Jane Heath for her help with the histology and Professor Paul Bolam for his help with the iontophoresis. These studies were supported by a Lister Institute fellowship to AER and a MRC studentship to CMK.

References

Armstrong
RJE
Huelbrink
CB
Tyres
P
Ratcliffe
EL
Richards
A
Dunnett
SB
, et al.  . 
The potentail for circuit reconstruction by expanded neural precursor cells explored through porcine xenografts in a rat model of Parkinson's disease
Exp Neurol
 , 
2002
, vol. 
175
 (pg. 
98
-
111
)
Armstrong
RJE
Tyers
P
Jain
M
Richards
A
Dunnett
SB
Rosser
AE
, et al.  . 
Transplantation of expanded neural precursor cells from the developing pig ventral mesencephalon in a rat model of Parkinson's disease
Exp Brain Res
 , 
2003
, vol. 
151
 (pg. 
204
-
17
)
Armstrong
RJE
Watts
C
Svendsen
CN
Dunnett
SB
Rosser
AE
Survival, neuroanl differentiation and fiber outgrowth of propagated human neural precursor grafts in an animal model of Huntington's Disease
Cell Transplant
 , 
2000
, vol. 
9
 (pg. 
1
-
10
)
Bachoud-Levi
AC
Gaura
V
Brugieres
P
Lefaucheur
JP
Boisse
MF
Maison
P
, et al.  . 
Effect of fetal neural transplants in patients with Huntington's disease 6 years after surgery: a long-term follow-up study
Lancet Neurol
 , 
2006
, vol. 
5
 (pg. 
303
-
9
)
Department of Health
Guidance on the supply of fetal tissue for research diagnosis and therapy
1995
London
Dunnett
SB
Bjorklund
A
Neural transplantation a practical approach
1992
 
Oxford University Press
Eriksson
C
Bjorklund
A
Wictorin
K
Neuronal differentiation following transplantation of expanded mouse neurosphere cultures derived from different embryonic forebrain regions
Exp Neurol
 , 
2003
, vol. 
184
 (pg. 
615
-
35
)
Evtouchenko
L
Studer
L
Spenger
C
Dreher
E
Seiler
RW
A mathematical model for the estimation of human embryonic and fetal age
Cell Transplant
 , 
1996
, vol. 
5
 (pg. 
453
-
64
)
Feil
R
Brocard
J
Mascrez
B
LeMeur
M
Metzger
D
Chambon
P
Ligand-activated site-specific recombination in mice
Proc Natl Acad Sci U S A
 , 
1996
, vol. 
93
 (pg. 
10887
-
90
)
Fricker
RA
Barker
RA
Fawcett
JW
Dunnett
SB
A comparative study of preparation techniques for improving the viability of striatal grafts using vital stains, in vitro cultures, and in vivo grafts
Cell Transplant
 , 
1996
, vol. 
4
 (pg. 
173
-
200
)
Fricker-Gates
RA
Lundberg
C
Dunnett
SB
Neural transplantation: restoring complex circuitry in the striatum
Restor Neurol Neurosci
 , 
2001
, vol. 
19
 (pg. 
119
-
38
)
Garcia
AR
Deacon
TW
Dinsmore
J
Isacson
O
Extensive axonal and glial fiber growth from fetal porcine cortical xenografts in the adult rat cortex
Cell Transplant
 , 
1995
, vol. 
4
 (pg. 
515
-
27
)
Hurelbrink
CB
Tyers
P
Armstrong
RJ
Dunnett
SB
Barker
RA
Rosser
AE
Long-term hibernation of human fetal striatal tissue does not adversely affect its differentiation in vitro or graft survival: implications for clinical trials in Huntington's disease
Cell Transplant
 , 
2003
, vol. 
12
 (pg. 
687
-
95
)
Isacson
O
Brundin
P
Kelly
PA
Gage
FH
Bjorklund
A
Functional neuronal replacement by grafted striatal neurones in the ibotenic acid-lesioned rat striatum
Nature
 , 
1984
, vol. 
311
 (pg. 
458
-
60
)
Kelly
CM
Dunnett
SB
Rosser
AE
Sanberg
PR CDDa
Cell-based therapy for Huntington's disease
Cell therapy for brain repair.
 , 
2006
, vol. 
1
 
Totowa
Humana Press Inc.
(pg. 
83
-
117
)
Kelly
CM
Tyers
P
ter Borg
MG
Svendsen
CN
Dunnett
SB
Rosser
AE
Mouse and rat embryonic neural precursor responsivness to EGF and FGF-2 and the effects of rat gestation
Brain Res Bull
 , 
2005
, vol. 
68
 (pg. 
83
-
94
)
Novikov
LN
Labeling of central projections of primary afferents in adult rats: a comparison between biotinylated dextran amine, neurobiotin and Phaseolus vulgaris-leucoagglutinin
J Neurosci Methods
 , 
2001
, vol. 
112
 (pg. 
145
-
54
)
Polkinghorne
J
Review of the guidance on the research use of fetuses and fetal material.
 , 
1989
London
HMSO
Pritzel
M
Isacson
O
Brundin
P
Wiklund
L
Bjorklund
A
Afferent and efferent connections of striatal grafts implanted into the ibotenic acid lesioned neostriatum in adult rats
Exp Brain Res
 , 
1986
, vol. 
65
 (pg. 
112
-
26
)
Pundt
LL
Kondoh
T
Conrad
JA
Low
WC
Transplantation of human striatal tissue into a rodent model of Huntington's disease: phenotypic expression of transplanted neurons and host-to-graft innervation
Brain Res Bull
 , 
1996
, vol. 
39
 (pg. 
23
-
32
)
Steiner
JP
Connolly
MA
Valentine
HL
Hamilton
GS
Dawson
TM
Hester
L
, et al.  . 
Neurotrophic actions of nonimmunosuppressive analogues of immunosuppressive drugs FK506, rapamycin and cyclosporin A
Nat Med
 , 
1997
, vol. 
3
 (pg. 
421
-
8
)
West
MJ
Stereological methods for estimating the total number of neurons and synapses: issues of precision and bias
Trends Neurosci
 , 
1999
, vol. 
22
 (pg. 
51
-
61
)
Wictorin
K
Anatomy and connectivity of intrastriatal striatal transplants
Prog Neurobiol
 , 
1992
, vol. 
38
 (pg. 
611
-
39
)
Wictorin
K
Brundin
P
Sauer
H
Lindvall
O
Bjorklund
A
Long-distance directed axonal growth from human dopaminergic mesencephalic neuroblasts implanted along the nigrostriatal pathway in 6-hydroxydopamine lesioned adult-rats
J Comp Neurol
 , 
1992
, vol. 
323
 (pg. 
475
-
94
)
Wictorin
K
Clarke
DJ
Bolam
JP
Brundin
P
Gustavii
B
Lindvall
O
, et al.  . 
Extensive efferent projections of intra-striatally transplanted striatal neurons as revealed by a species-specific neurofilament marker and anterograde axonal tracing
Prog Brain Res
 , 
1990
, vol. 
82
 (pg. 
391
-
9
)
Wictorin
K
Simerly
RB
Isacson
O
Swanson
LW
Bjorklund
A
Connectivity of striatal grafts implanted into the ibotenic acid-lesioned striatum—III. Efferent projecting graft neurons and their relation to host afferents within the grafts
Neuroscience
 , 
1989
, vol. 
30
 (pg. 
313
-
30
)

Abbreviations:

    Abbreviations:
  • BDA

    biotynilated dextran amine

  • FNP

    foetal neural precursor

  • NBT

    neurobiotin