Is forebrain neurogenesis a potential repair mechanism after stroke?

Abstract

The use of adult subventricular zone (SVZ) neurogenesis as brain repair strategy after stroke represents a hot topic in neurologic research. Recent radiocarbon-14 dating has revealed a lack of poststroke neurogenesis in the adult human neocortex; however, adult neurogenesis has been shown to occur, even under physiologic conditions, in the human striatum. Here, these results are contrasted with experimental poststroke neurogenesis in the murine brain. Both in humans and in rodents, the SVZ generates predominantly calretinin (CR)-expressing GABAergic interneurons, which cannot replace the broad spectrum of neuronal subtypes damaged by stroke. Therefore, SVZ neurogenesis may represent a repair mechanism only after genetic manipulation redirecting its differentiation.

Stimulation of adult neurogenesis in the subventricular zone (SVZ) may represent an important brain repair mechanism and future therapeutic strategy to replace neurons depleted after stroke. Early data obtained in the experimental middle cerebral artery occlusion model in adult rats showed that stroke markedly increased SVZ neurogenesis, followed by sustained migration of newborn neuroblasts into the damaged striatum.¹ More than 40% of newborn neurons differentiated to medium-sized spiny projection neurons that represent 95% of all neurons in the striatum.¹ The differentiation of newborn neurons to the phenotype of most of the neurons destroyed by the ischemic lesion suggested that the adult brain may possess the capacity for self-repair after insults causing extensive neuronal death. These data indicated that SVZ neurogenesis may represent per se, without external intervention, a repair mechanism and therapeutic option of cell replacement after stroke.

However, more recent studies did not confirm these initially promising results. Indeed, newborn neurons were even detected under physiologic conditions outside the classic sites of neurogenesis (SVZ/olfactory bulb and hippocampal dentate gyrus), i.e., in the neocortex and striatum² of adult rats. However, these neurons belonged exclusively to few subclasses of small inhibitory GABAergic interneurons with a morphology similar to adult-generated granule cells, expressing either calretinin (CR) in the striatum or CR or calbindin in the neocortex.² In the latter region, these neurons were located in the deep layers, in the medial striatum they were in the immediate vicinity of the SVZ.² Although their function remains unknown, it is plausible that they are integrated in local neuronal circuits. In any case, it seems obvious that these specific newborn interneurons cannot replace the majority of neocortical and striatal neurons (pyramidal neurons and medium-sized spiny neurons, respectively).

Thus, in contrast to earlier expectations, brain injury does not seem to alter the intrinsically restricted differentiation potential of adult SVZ neuroblasts to CR-positive interneurons, as shown both in neonatal³ and in adult⁴ rat models. In the striatum, exclusively newborn CR-positive interneurons, morphologically similar to those described before by Dayer et al,² were detected using the classic Vannucci model of neonatal hypoxic/ischemic (H/I) brain damage.³ This proliferative process appeared quantitatively robust, resulting in a twofold increase in the density of these neurons in the ipsilateral striatum, and sustained, with long survival rates of newborn neurons (5 months).³ In addition, only the same subtype of CR-positive interneurons was found to be generated in the middle cerebral artery occlusion model in adult rats.⁴ Retroviral fate-mapping studies confirmed that newly born striatal CR-positive interneurons were descendants of SVZ progenitors.^{3, 4} Surprisingly, considering the previous results of Arvidsson et al,¹ no other striatal neuronal subtype (DARPP32-, calbindin-, parvalbumin-, somatostatin-, or choline acetyltransferase-positive neurons) was detected in the striatum after stroke.^{3, 4} Of note, similar newborn interneurons were also found in the rodent neocortex after stroke. In mice, poststroke-generated SVZ-derived neuroblasts migrated long distances into the damaged cortex and survived for at least 35 days in the peri-infarct area where they differentiated into mature neurons.⁵ In this case, all newly born neurons belonged as well to the same class of CR-positive GABAergic interneurons.⁵ The restricted capacity of the SVZ to generate this neuronal subtype only is further strengthened by results obtained in other experimental paradigms, e.g., after electroconvulsive stimulation that triggered massive migration of CR-positive interneurons into the striatum.⁶

In humans, methodological difficulties hampered the analysis of adult neurogenesis for a long time. This has changed due to a new technique of retrospective birthdating by radiocarbon-14 dating. This method takes advantage of atomic weapons tests that resulted in detectable levels of the isotope carbon-14 after 1963. Using this approach, it has recently been shown that humans lack any detectable neurogenesis in the neocortex after stroke.⁷ This was corroborated by immunohistochemical analyses of postmortem brain samples (e.g., lack of cells expressing neurogenesis markers like Pax6 or Tbr2, and lack of neurons devoid of lipofuscin which postmitotic cells accumulate with age).⁷ Such results appear to negate any hope for neurogenesis as brain repair mechanism and strategy. In contrast to these data in the neocortex and after stroke, another recent radiocarbon-14 dating study showed ongoing adult neurogenesis in the human striatum under physiologic conditions.⁸ Importantly, the renewal rate of human striatal neurons was 2.7% per year, with an estimated total turnover rate of 25%, which is comparable to the renewal rate in the adult human hippocampus.⁸ Newly born neurons were identified as belonging to subpopulations of striatal CR-positive and, more rarely, neuropeptide Y-expressing GABAergic interneurons.⁸ Although the human striatum has not yet been investigated after stroke, one could speculate that even more newborn GABAergic interneurons may be detected postischemia. These results are of real importance, since ischemic stroke in humans mostly results from occlusion of the middle cerebral artery, which feeds the striatum.

What could be the quantitative and qualitative impact of poststroke neurogenesis? Because CR-positive interneurons represent <1% of all neurons in the adult murine striatum, the potential regenerative capacity of SVZ neurogenesis appears very limited in rodents. However, the proportion of CR-positive interneurons increases sharply to about 10% in humans, CR-positive cells representing the largest population of GABAergic interneurons in the human striatum. Therefore, this process could be quantitatively significant in patients. However, even if SVZ/striatal neurogenesis would be triggered by ischemia in humans, this will not influence/change the phenotype of newborn neurons from mainly one neuronal subtype (CR-positive GABAergic interneurons), suggesting a rather low qualitative effect on cell replacement. Therefore, the general contribution of neurogenesis as brain repair mechanism remains questionable, since a plethora of different types of neurons are damaged and have to be replaced after stroke.

Recent evidence, however, indicates that such cell proliferative processes could become useful for brain repair if it would be possible to influence and redirect the differentiation of cell precursors. Thus, retroviral transduction of molecular cell fate determinants—causing repression of the gliogenic factor Olig2 and overexpression of the neurogenic factor Pax6—was shown to be effective in reprogramming resident striatal glia toward neuronal differentiation after transient brain ischemia.⁹ Promising results were also obtained in the injured cerebral cortex by lentiviral expression of the neurogenic transcription factor Sox2.¹⁰ Similar strategies could be potentially used to increase the proportion of medium-sized spiny neurons developing from SVZ progenitors in the striatum. However, additional intense research is needed to identify the fate determinants modulating the differentiation of progenitor cells toward specific neuronal subtypes. Moreover, the integration of newborn neurons into local circuits and the net functional recovery effect should be studied and improved in the future.

In conclusion, the fact that neurogenesis is not induced in the human neocortex after stroke does not imply that it cannot occur at a significant level in other brain regions. Particularly the striatum represents such a candidate region, where neurogenesis persists in humans throughout life at levels similar to the hippocampus even in the absence of any brain pathology. Future studies in humans and in animal models have to deepen our understanding of neurogenesis and its functional significance in pathologic situations such as ischemia, hypoxia, and trauma, toward the development of clinically effective brain repair strategies.

The authors declare no conflict of interest.