Neuronal Replacement as a Tool for Basal Ganglia Circuitry Repair: 40 Years in Perspective

Abstract

The ability of new neurons to promote repair of brain circuitry depends on their capacity to re-establish afferent and efferent connections with the host. In this review article, we give an overview of past and current efforts to restore damaged connectivity in the adult mammalian brain using implants of fetal neuroblasts or stem cell-derived neuronal precursors, with a focus on strategies aimed to repair damaged basal ganglia circuitry induced by lesions that mimic the pathology seen in humans affected by Parkinson's or Huntington's disease. Early work performed in rodents showed that neuroblasts obtained from striatal primordia or fetal ventral mesencephalon can become anatomically and functionally integrated into lesioned striatal and nigral circuitry, establish afferent and efferent connections with the lesioned host, and reverse the lesion-induced behavioral impairments. Recent progress in the generation of striatal and nigral progenitors from pluripotent stem cells have provided compelling evidence that they can survive and mature in the lesioned brain and re-establish afferent and efferent axonal connectivity with a remarkable degree of specificity. The studies of cell-based circuitry repair are now entering a new phase. The introduction of genetic and virus-based techniques for brain connectomics has opened entirely new possibilities for studies of graft-host integration and connectivity, and the access to more refined experimental techniques, such as chemo- and optogenetics, has provided new powerful tools to study the capacity of grafted neurons to impact the function of the host brain. Progress in this field will help to guide the efforts to develop therapeutic strategies for cell-based repair in Huntington's and Parkinson's disease and other neurodegenerative conditions involving damage to basal ganglia circuitry.

Keywords: dopamine; embryonic stem cells; induced pluripotent stem cells; nigrostriatal pathway; regenerative medicine; striatum; substantia nigra.

Figure 1

(A) Striatal connectivity comprises two major neuronal circuits: the cortico-striato-pallido-thalamic circuit and the cortico-striato-nigro-thalamic circuit, which in turn are interlinked by an important regulatory hub, the subthalamic nucleus. The striatal projection neurons (which constitute more than 90% of all neurons in the striatum) are of two types: the D1 receptor-bearing striatonigral neurons that innervate the internal globus and the pars reticulata of the substantia nigra, and the D2 receptor-bearing striatopallidal neurons which innervate the external globus pallidus. (B) Damage to the striatal projection neurons, caused by an intrastriatal injection of ibotenic acid (IBO) or quinolinic acid(QUIN), will disrupt these two circuits and result in a disinhibition of the downstream targets, the pallidum and the substantia nigra, pars reticulata. This lesion mimics the striatal damage seen in patients with Huntington’s disease. (C) Lesion of the nigrostriatal dopamine (DA) pathway, induced by the injection of 6-hydroxydopamine (6-OHDA) or MPTP, removes an important regulatory control, resulting in motor impairments similar to those seen on patients with Parkinson’s disease.

Figure 2

(A,B) The loss of striatal neurons, and the reduction in overall striatal volume, seen in the excitotoxin lesioned striatum (as shown in A), are largely restored by the transplanted fetal striatal primordium (B; AchE stain, 6 months survival. T = transplant). (C–E) Injection of the retrograde axonal tracer FluoroGold into the Globus pallidus (hatched area in E) results in the labeling of large numbers of DARPP-32+ striatal projection neurons in the graft (bright spots in C). The vast majority of these are located within the DARPP-32+ patches in the grafts (striped area in D). (F) Injection of a retrograde axonal tracer into the striatal graft (hatched area, T) reveal extensive afferent inputs from the same regions of the host brain that innervate the striatum in the intact brain, including the frontoparietal cortex (FCX), the intralaminar thalamic nuclei (CL, CM, PF, Pc, Po, VL, VM), and the substantia nigra (SNC), with a distribution that is closely similar to that seen in the intact animal. (G–I) The DARPP-32+ areas of the grafts (G) are densely innervated by TH+ axons from the host nigrostriatal pathway (H), as well as from the host frontal cortex (labeled with the anterograde tracer PHA-L in I) adapted from Wictorin et al. (1989b) and Wictorin and Björklund (1989).

Figure 3

(A,B) Cartoons illustrating the disinhibitory effect of the excitotoxic striatal lesion on the downstream targets, globus pallidus (GPe and GPi), and substantia nigra pars reticulata (SNc), and the reversal of this effect induced by the striatal graft. (C) Recovery of skilled motor performance is the paw reaching test, as seen in two groups of lesioned and grafted rats, using grafts derived from the lateral ganglionic eminence (LGE) only (modified from Nakao et al., 1996). (D) Recovery of the use of the paw contralateral to the lesion and grafted side (open bars) is well correlated to the volume of the DARPP-32+ portion of the fetal GE grafts, obtained from the whole GE at different donor ages (green bars; modified from Fricker et al., 1997). (E) Graft-induced recovery in the performance of delayed alternation in the classic T-maze task in rats with bilateral striatal lesions and transplants (modified from Isacson et al., 1986). (F) Graft-induced recovery of habit learning in rats with unilateral striatal lesions and transplants. In this test, the grafted animals had to relearn the task over a similar period, 6–8 weeks, as seen in intact rats learning the same task for the first time (modified from Brasted et al., 1999). *p < 0.05, **p < 0.01.

Figure 4

Re-establishment of the nigrostriatal pathway from a transplant of mouse fetal ventral mesencephalic (VM), implanted as a cell suspension in substantia nigra of a 6-OHDA lesioned mouse, 16 weeks post-grafting. The VM tissue was obtained from a transgenic Pitx3-green fluorescent protein (GFP) mouse allowing the outgrowing axons to be visualized using GFP immunostaining, as illustrated in the computer-assisted drawings derived from three horizontal sections in panel (A). The micrographs in panels (B–D) are taken from the areas marked in panel (A). Amy, amygdala; CPu, caudate-putamen; H, hippocampus; NAc, nc. Accumbens; NSP, nigrostriatal pathway; Pir, piriform cortex. Modified and redrawn from Thompson et al. (2009).

Figure 5

The functional effect depends on the area of the striatum reinnervated by the fetal DA neuron transplants. (A) The reinnervation obtained from VM cell suspension grafts placed in the central or lateral part of the striatum is restricted to the area surrounding the graft deposits. (B) Grafts reinnervating the central vs. lateral striatum, as shown in panel (A), have markedly different effects on behavior: the centrally placed grafts abolish amphetamine-induced rotation but have little effect on sensorimotor behavior. The laterally placed grafts, by contrast, has little effect on amphetamine rotation but is highly efficient in restoring sensorimotor behavior. The more complex version of the task, called diseagage behavior, remains unaffected by these transplants but is well restored in animals with more wide-spread reinnervation of the striatal complex, as shown in Figure 6. Modified and redrawn from Mandel et al. (1990). *p < 0.01.

Figure 6

More complete recovery can be obtained by spreading the graft tissue over multiple implantation sites. (A) Extent of graft induced reinnervation obtained with fetal VM grafts spread over seven injection sites distributed over the entire striatum, including the nc. accumbens, 10 months post-grafting. (B) In these animals, significant functional recovery is seen in a broad range of drug-induced and spontaneous motor tests, but remains incomplete in most of the tests Data compiled from Winkler et al. (1999). *Different from Control at p < 0.05; ^†different from Lesion only at p < 0.05.

Figure 7

Performance of human fetal and human embryonic stem cell (hESC)-derived DA neurons grafted to the striatum in the rat 6-OHDA lesion model. (A–C) Time-course of functional recovery in the amphetamine rotation test obtained with fetal rat VM transplants (A), human fetal VM transplants (B), and hESC-derived DA neuron transplants (C). The time-course of recovery is notably similar for the fetal and hESC-derived human DA neurons, but much slower than that seen with rat DA neurons. In the experiment shown in panel (A), the original functional deficit returned within a week after the graft had been removed with a second 6-OHDA lesion. (D) A single deposit of hESC-derived DA neurons is sufficient to reinnervate the entire striatum in the rat PD model, as visualized using a human-specific NCAM antibody. The graft-derived innervation pattern in (D) is notably similar to the distribution of the endogenous TH-positive innervation, as shown in (E; adapted from Nolbrant et al., 2017). Data compiled from Dunnett et al. (; A), Lelos et al. (; B), and Kirkeby et al. (; C).

Figure 8

The functional impact of hESC-derived DA neuron grafts is tunable using chemo-and optogenetics. (A) The recovery of paw use in the cylinder test (contralateral to the 6-OHDA lesion) is abolished when the activity of the grafted DA neurons is inhibited by CNO (hM4Di), and it is further potentiated when the activity is increased by CNO (hM3Dq). (B) The effect of inhibition and activation of the grafted human DA neurons is similar in the amphetamine rotation test. The 6-OHDA-induced ipsilateral turning bias is abolished by the grafted DA neurons in all three groups (open bars). Activation of the inhibitory DREADD blocks the graft effect, seen as induction of an ipsilateral turning bias (similar to what is seen in lesioned controls), and activation of the excitatory DREADD potentiates the graft effect, seen as induction of turning in the direction away from the transplant. (C) The graft-induced recovery in sensorimotor performance seen in the corridor test (gray bars) is completely blocked when the activity of the grafted DA neurons is inhibited by light (green bars). Data in panels (A,B) are redrawn from Chen et al. (2016), data in panel (C) redrawn from Steinbeck et al. (2015). *p < 0.05, **p < 0.01.