Introduction

The human brain is at constant danger of sustaining structural damage, either due to trauma or neurodegenerative diseases. Although the functional consequences of brain lesions can be somewhat compensated (Nithianantharajah and Hannan 2011), and some symptoms can be treated pharmacologically (Bartus 2000; Connolly and Lang 2014), so far no cure for structural brain damage exists.

However, it is known that the adult mammalian brain produces new neuronal precursor cells in distinct brain areas, the dentate gyrus of the hippocampus (Altman and Das 1965; Gage et al. 1995) and the subventricular zone (SVZ) of the lateral ventricle (Altman 1969; Reynolds and Weiss 1992). In these regions, neural progenitor cells give rise to a special type of neuronal precursor cells, the so-called neuroblasts. These cells are capable of undergoing mitosis but show a limited proliferative potential, since they only differentiate to neurons under physiological conditions (Jablonska et al. 2010).

While neuroblasts generated in the adult dentate gyrus do not migrate to targets outside the hippocampal formation, those that originate from the SVZ travel through the forebrain towards the olfactory bulb where they differentiate into granule and periglomerular cells (Mouret et al. 2009). Therefore, high numbers of these neuroblasts continuously move through the rostral migratory stream (RMS) (first described in Altman 1969). This path of migration is densely surrounded by astrocytes, which form a glial tube that prevents the migrating cells from dispersing into surrounding tissue (Lois et al. 1996; Peretto et al. 1997). The SVZ-derived cells form longitudinal aggregates and migrate along the RMS independently of the guidance of radial glia or axonal processes (Wichterle et al. 1997). During this so-called chain migration the cells interact with each other and migrate at a relatively high speed of more than 70 µm/h (Nam et al. 2007). However, this process is not uniform, since cells may stop migrating or even move caudally through the RMS (Nam et al. 2007).

The regulatory processes involved in the generation and guidance of SVZ-derived neuronal progenitor cells are complex and not yet fully understood. While the number of newly generated neuroblasts entering the RMS depends on connectivity (Jankovski et al. 1998) and activity of the olfactory bulb (OB) (Pothayee et al. 2017), the general process of cell migration in the RMS seems to be OB independent (Jankovski et al. 1998; Kirschenbaum et al. 1999). The direction of migration is rather controlled locally by the microenvironment within the migratory path, which in turn is possibly controlled by the ensheathing astrocytes of the RMS (García-Marqués et al. 2010). A number of chemoattractive and chemorepulsive substances, such as various neurotrophic factors and extracellular matrix proteins, have been identified so far (for reviews see Sun et al. 2010; Leong and Turnley 2011).

An earlier study by Emsley and Hagg (2003) described the role of the interaction between laminin and its receptor integrin on neuroblast migration. Selective antibody blockade revealed that both the α6 and β1 subunit of integrin are required for RMS migration. Additionally, an antibody against the β1 subunit led to a disruption of the typical neuroblast chains and a dispersal of cells from their migration path into the surrounding tissue.

Furthermore, the authors demonstrated that the chemoattractive properties of laminin, or a shorter peptide, representing the α6β1-integrin binding site of laminin, can be utilized to redirect neuroblasts from the RMS into the surrounding tissue. Injection of this laminin subunit dorsal to the RMS led to an accumulation of cells, which are positive for the cell proliferation marker BrdU as well as PSA-NCAM and Tuj1, two markers for immature neurons. Furthermore, the authors injected laminin dorsally to the RMS, while retracting the injection cannula, resulting in a narrow, 2 mm long tract of the protein. 7 days later they found cells positive to the above-mentioned markers at the injection tract and dorsal to it, demonstrating that migrating neuroblasts were diverted from the RMS into the injected tract of laminin.

In the present study, we apply this method in a brain lesion model and test the neuroblasts’ ability to (a) migrate to the lesion along the injected tract of laminin, (b) disperse from the injection tract into the lesioned tissue, and (c) assess the number of neuroblasts within the lesion. Furthermore, we opted for a relatively long survival time after the laminin application to evaluate the long-term effects of the treatment in terms of survival of the cells, possible differentiation, and a sustained flow of neuroblasts into the lesion site. Overall, this study was designed as a proof-of-concept of redirection of neuroblasts into a lesion as a possible future treatment of structural brain damage that only requires a single, relatively simple, minimally invasive surgery.

Materials and methods

Subjects

A total of 25 adult (age 3.5–6.8 months) male Wistar rats (Charles River, Germany) were used in this study. The animals were kept under standard housing conditions (12 h light/dark cycle, lights on at 7 a.m., water ad libitum, standard lab chow 12 g/rat/day) in groups of 4 to 6 animals per cage. The experiments were performed in accordance with the National Institutes of Health ethical guidelines for the care and use of laboratory animals for experiments and were approved by the local animal care committee (Senatorische Behörde, Bremen, Germany).

Treatment

All animals underwent two stereotactic microinjections. First, they were lesioned bilaterally by injection of ibotenic acid (Cayman Chemical Company, MI, USA; 6.7 mg/ml saline; injection volume of 0.4 µl at 0.3 µl/min infusion rate) into the medial prefrontal cortex (3.2 mm rostral, ± 0.5 mm lateral, − 4.4 mm ventral, all relative to bregma).

Five days later the laminin tract (L2020, Sigma–Aldrich Chemie GmbH, Germany) was applied bilaterally. For this purpose, the injection cannula (custom made from 30 G epidermal stainless steel cannula) was placed directly dorsal and slightly medial to the RMS (rc. + 3.0 mm; l ± 1 mm; vd. − 6.3 mm, relative to bregma; see Fig. 1). After a dwell time of 3 min, the laminin solution (0.25 µg/ml; Lam group) or vehicle (phosphate buffered saline, PBS; C group) was injected at a constant flow rate (0.16 µl/min) by means of an injection pump while the injection cannula was slowly retracted dorsally (0.6 mm/min) into the lesion site (vd. − 4.6 mm). The injection volume of 0.48 µl was chosen to match the volume of the resulting tract. After another 3 min the cannula was fully retracted before the trepanations were closed with bone wax (SMI AG, Belgium). After application of an antiseptic ointment (Betaisodona, Mundipharma, Germany) and a local anesthetic (Xylocaine 2%, AstraZeneca GmbH, Germany), the skin was sutured. An additional group was treated with a higher concentration of laminin (0.25 µg/µl; LamHi group).

Fig. 1
figure 1

(modified from Paxinos and Watson 1998)

a Sagittal view of the physiological path of neuroblasts in adult rats: Neuroblasts are generated at the walls of the lateral ventricle (green) and migrate along the RMS (red) through the forebrain towards the OB where they disperse and differentiate (position of the RMS illustrated according to Altman 1969). b Experimental procedure: Laminin was injected while retracting the cannula dorsally, resulting in a narrow tract of laminin (blue) between the RMS (red) and the excitotoxic lesion (− 6.3 to − 4.6 mm ventral to bregma). The approximate position of lesion is represented by the dashed line. The images show the sagittal section 1.4 mm lateral to bregma (a) and the coronal brain section 3.2 mm rostral to bregma (b)

Full size image

All surgical procedures were performed under isoflurane (CP-Pharma, Germany) anesthesia (vaporizer in circle system). The spontaneous breathing rate (BR) was constantly monitored and the vaporizer setting was adjusted to keep BR at 40–60 breaths per minute.

Immunohistochemistry

After a survival time of 6 weeks after the second intracranial injection, the animals were killed by intraperitoneal injection of sodium pentobarbital (200 mg/kg bodyweight at a concentration of 20 mg/ml PBS; Sigma–Aldrich Chemie GmbH, Germany), before being transcardially perfused with PBS followed by 4% paraformaldehyde solution. Immunohistochemistry was performed on 40 µm thick coronal cryosections (distance of 240 µm). Sections were blocked free floating in a 10% normal donkey serum (Jackson ImmunoResearch, PA, USA) in PBS and 0.05% Triton-X. The same solution was used for incubation with NeuN (1:1000; Millipore Rabbit anti-NeuN (RRID: AB_10807945)) and doublecortin (DCX) [1:1000; Santa Cruz Doublecortin Antibody (C-18) (RRID: AB_2088494)] antibodies for 72 h at 4 °C. After blocking in 10% bovine serum albumin (Sigma–Aldrich Chemie GmbH, Germany) for 1 h, the sections were incubated in secondary antibody solution [CruzFluor 488 Donkey anti-Rabbit IgG (RRID: AB_10989100), 1:2000 and Santa Cruz Biotinylated Donkey anti-Goat (RRID: AB_631726), 1:1000] for 48 h at 4 °C. The streptavidin-conjugated fluorescent dye [Jackson Streptavidin-Alexa Fluor 568 (RRID: AB_2337250), 1:2000] was allowed to incubate for another 24 h at room temperature. Subsequently, the sections were counterstained in 0.9% w/v Sudan Black (Acros Organics, Belgium) in 70% ethanol and then mounted, air dried, and coverslipped. Additionally, exemplary brain slices were stained against DCX and the astrocyte marker glial fibrillary acidic protein (GFAP) to reassure the specificity of the DCX antibody for neuroblasts. The staining procedure followed the same protocol described above, but the primary antibody for NeuN was replaced by Dako Rabbit anti-GFAP (RRID: AB_10013382) at the same concentration.

Image acquisition and analysis

Images were acquired on a fluorescent microscope (Axioscope 100, Carl Zeiss AG, Germany) and a monochrome digital camera (Spot, Visitron Systems GmbH, Germany). Brains with severe misplacement of the laminin tract or the lesion were excluded from the analysis, resulting in group sizes of 12 hemispheres for the sham laminin injected (C) group, 19 for the laminin injected (Lam) and 8 for the high dose laminin (LamHi) group. Overview images were taken at the rostro-caudal position of the laminin tract and later stitched in Microsoft Image composition editor (Microsoft Cooperation, version 1.4.4.0) using the planar motion 1 (rigid scale) setting. Adjustment of brightness and contrast as well as creation of scale bars were done in the ImageJ (version 1.49 h) (Schindelin et al. 2015; Schneider et al. 2012) based software FIJI (Schindelin et al. 2012).

The brain lesions were identified in the NeuN staining, whereas the RMS showed strong immunoreactivity against DCX. The laminin/vehicle injection tract was clearly visible in both the NeuN and DCX staining; either as a fissure in the slice or due to strong background staining at the site of the injection (see Fig. 3). The hemispheres of the animals were analyzed separately. Sections with insufficient staining or severe damages at the area of interest were excluded from the analysis. In the section closest to the laminin tract or the corresponding vehicle injection, the lesion was then surrounded with the polygon tool in FIJI to measure its size. Within this region, DCX-positive cells were counted. Only cells with a luminance similar to that of the RMS and with the typical morphology of neuroblasts were counted (see Fig. 2). This analysis was performed by an observer blind to the treatment. To account for the thickness of the brain slices, the number of cells was assessed per region of interest volume (i.e., area of interest multiplied by the slice thickness). The statistical analyses were conducted in IBM SPSS Statistics (version 20 for Windows).

Fig. 2
figure 2

a–c Low magnification photomicrographs of the gliotic response 6 weeks after the excitotoxic lesion, coronal section, 3.0 mm rostral to bregma. GFAP-positive cells have formed a glial scar around the lesion. A smaller number of GFAP-positive cells are also found around the laminin tract (arrowheads in a and c; scale bars: 500 µm). d–i Two different cell populations found in brain lesions are immunoreactive for DCX. d–f Example of a polar neuroblast immunoreactive for DCX (arrowheads in e and f). These cells show one long process extending from the soma towards the migration direction (leading process) and often another, shorter process oriented towards the opposite direction (trailing process). These cells do not show any overlap with GFAP staining (d and f). g–i Example of a stellate cell positive for DCX (h). The stellate cells show a round morphology with a comparatively large soma and multiple processes in all directions. They are also positive for the astrocyte marker GFAP (g), although only a small portion of GFAP-positive cells are also positive for DCX (cp. overlay in i). Note that here contrast and brightness are adjusted for maximum visibility of the cells. Side by side the polar neuroblasts show markedly brighter fluorescence than the stellate cells in the DCX assays. Scale bars in d–i 100 µm

Full size image
Fig. 3
figure 3

DCX-positive neuroblasts following the laminin tract dorsally towards the lesion site. a–c Low magnification of coronal section showing the lesion (circumscribed by dotted line in a) and the RMS (arrowheads in b and c). The laminin tract is visible by the line of stronger background staining in vertical direction (marked with an asterisk in c). Scale bars: 500 µm. d Detail of the DCX staining (position marked by dashed box in b). The leading processes of the neuroblasts are oriented parallel to the path of the laminin injection and directed dorsally, towards the lesion. Scale bar: 50 µm

Full size image

Results

Ibotenic acid-induced brain lesion

Microinjections of ibotenic acid reliably caused brain lesions, characterized by the absence of NeuN-positive cell bodies. The lesion sites showed some variability in size and shape but were mostly limited to the prelimbic and infralimbic cortex (see Fig. 1 for the approximate position). In some cases, parts of the cingulate cortex, area 1 and the dorsal peduncular cortex were affected. At the coronal plane of the laminin injection (3.0 mm rostral to bregma), the lesions stretched over an area of 1.24 ±0.15 mm² (standard error of the mean; n = 37 hemispheres).

Two different populations of cells are positive for DCX

Two different populations of cells were identified in the lesions: One type of cells showed a distinctly weaker fluorescence and astrocyte-like morphology with multiple processes [stellate cells; cf. (Kunze et al. 2015)]. The other cell type showed only one or two processes originating from a smaller soma and a much stronger staining (polar cells). In exemplary brain slices, we reproduced the finding that the DCX-immunoreactive stellate cells were also positive for the astrocytic marker GFAP, but none of the polar cells were (Fig. 2). Furthermore, the stellate cell type was only found within the lesions of laminin treated animals, whereas the polar cell type was also observed in the SVZ, the RMS, and the olfactory bulb. Moreover, these cells were found between the RMS and the lesion in laminin treated animals. Overall, the distribution and morphology of GFAP-positive cells within the lesion corresponded to that described earlier for excitotoxic lesions (Dusart et al. 1991), with the addition that a subpopulation of these cells was also immunoreactive for DCX. For the quantification of neuroblasts, only cells that were clearly characterized as polar type were counted.

Neuroblasts follow the laminin tract, and then disperse into the lesion site in a non-uniform orientation

Neuroblasts were found in 73.7% of the lesions of the Lam group (n = 19 hemispheres), whereas all of the hemispheres of the LamHi group (8 hemispheres) showed DCX-positive cells within the lesion site. Two of the 12 control hemispheres (16.7%) had DCX-positive cells in the lesion site.

All neuroblasts found between the lesion site and the RMS were located in close vicinity to or directly at the laminin tract with their processes oriented in the direction of the injection path (Fig. 3). In contrast to the physiological migration through the RMS, these cells did not show any signs of chain formation. The RMS of laminin-treated animals did not show any visible alterations from that of the control animals, indicating that the physiological SVZ-OB migration was largely unaffected.

Neuroblasts in the lesion were not restricted to the laminin tract and did not show a coherent orientation (Fig. 4a–c). Furthermore, many of these cells showed the canonical morphology of orientation changing neuroblasts (Martinez-Molina et al. 2011). This means that the cells kept migrating within the lesion probably following different orientation cues. In six hemispheres of the laminin-treated groups single cells, both stained for DCX and NeuN were observed (Fig. 4i–k). These cells were all located at the border of the lesions and most of them resembled the morphology of adult neurons with long processes towards the intact surrounding tissue.

Fig. 4
figure 4

a–c Neuroblasts within the lesion; dotted line represents the border of the lesion. After following the laminin-guided migration from the RMS the neuroblasts (DCX in b) disperse within the lesion site (area without NeuN-positive cells in a). They no longer show a uniform orientation and many of the cells resemble typical direction changing neuroblasts (Martinez-Molina et al. 2011). d–h Show higher magnifications of neuroblasts depicted in b (magnification is identical for d–h). i–k Example of a cell within the confines of the lesion, positive for the neuroblast marker DCX (j) and the marker for adult neurons, NeuN (i). The cell shows long processes towards the surrounding intact neuronal tissue. All scale bars: 100 µm

Full size image

Laminin dose affects the reliability of the method but not the density of relocated cells

DCX-positive neuroblasts were counted within the lesion site at the coronal plane of the laminin or vehicle injection (Fig. 5). A Kruskal–Wallis H test showed a significant difference in cell density between the treatment groups (Χ² = 16.29, p < 0.001, mean ranks: C: 9.75; Lam: 22.89; LamHi: 28.50). The pairwise comparison demonstrated a significant difference between the mean cell density of the C group and both the Lam (p = 0.003) and LamHi (p < 0.001) group, but not between the Lam and LamHi group (p = 0.687). However, the higher dose of laminin caused a more reliable relocation of neuroblasts towards the lesion, which resulted in a lower variability within the LamHi group [interquartile range (IQR): 0.61–20.86 cells/mm³] compared to that of the Lam group (IQR: 13.54–18.60 cells/mm³).

Fig. 5
figure 5

Neuroblasts within the brain lesion site presented as cells per mm³ of lesioned tissue. Both treatments led to a significantly higher density of neuroblasts than the vehicle application (Kruskal–Wallis H test, Χ² = 16.29, p < 0.001, mean ranks: C: 9.75; Lam: 22.89; LamHi: 28.50). According to the pairwise comparison, the mean cell density differed significantly between the C group and both the Lam (p = 0.003) and LamHi (p < 0.001) group, but not between the Lam and LamHi group (p = 0.687). However, the higher dose of laminin caused a relocation of neuroblasts towards the lesion more reliably, which resulted in a lower variability within the LamHi group (IQR: 20.25 cells/mm³) compared to that of the Lam group (IQR: 5.06 cells/mm³). Boxes represent IQR; whiskers mark the highest and lowest numbers within 1.5 × IQR; dots represent extreme outliers

Full size image

Discussion

In the present study, we demonstrate that the previously described chemoattractive properties of laminin can be utilized to redirect the migration of neuroblasts into lesions in rat brains. Our data suggest that a part of the cells traveling through the forebrain keep migrating along the surgically introduced migration path for as long as at least 6 weeks without disrupting the physiological migration into the olfactory bulb of the animals. Furthermore, the neuroblasts leave the laminin tract once they reach the lesion and disperse within the damaged tissue, where at least a part of them seems to mature into adult neurons.

Astrocytes in the lesions of laminin-treated animals can be immunoreactive for DCX

Physiologically, DCX, a marker for migrating neuronal precursors, is scarcely observed outside of the neurogenic zones of the brain of rats (Omori et al. 1998; Brown et al. 2003) and human (Verwer et al. 2007). Nonetheless, there have been reports of astrocytes, expressing DCX under pathological conditions (Kunze et al. 2015; Verwer et al. 2007). The reason for the occurrence of this microtubule-associated protein in glia cells is not yet fully understood. Possible explanations include the involvement of DCX in the migration of astrocytes, but also the transdifferentiation of astrocytes towards a stem cell-like phenotype (Kunze et al. 2015). Furthermore, neuroblasts can differentiate into GFAP-positive astrocytes in vitro (Reynolds and Weiss 1992), so that the co-expression of GFAP and DCX for some time is conceivable. Although our data does not provide a sufficient explanation to this phenomenon, it is noteworthy that we only observed DCX-positive cells with astrocyte-like morphology and immunoreactivity for GFAP in the lesions of laminin-treated animals where also DCX-positive polar neuroblasts were found. This finding goes along with the results of Kunze et al. (2015), who identified DCX-positive astrocytes alongside DCX-positive neuroblasts that probably migrated through the striatum towards a cortical lesion. Therefore, we favor the hypothesis that DCX expression in astrocytes is related to an interaction between migrating neuronal precursor cells rather than a spontaneous reaction of astrocytes to brain damage or degeneration.

This observation suggests that if astrocytes are indeed capable of dedifferentiating towards a more stem cell-like phenotype (Leavitt et al. 1999; Steindler and Laywell 2003), this process might be promoted by the presence of neuroblasts. A similar conversion of astrocytes to DCX-positive neuroblasts has been demonstrated after in situ reprogramming of spinal cord astrocytes by introduction of the transcription factor SOX2 (Su et al. 2014).

However, although our data confirm that the immunohistochemical detection of DCX was not exclusive to neuroblasts, we consider the combined assessment of staining intensity and cell morphology as a reliable tool to identify neuroblasts to analyze their number and location within the lesioned tissue.

Neuroblasts follow migratory cues towards brain lesions

The SVZ and the RMS are considered distinct regions from which neuronal precursors rarely disperse into surrounding tissue under physiological conditions. However, there are a number of reports about precursor cells migrating towards damaged brain areas. For example, migration of DCX-positive neuroblasts towards the striatum in a mouse model for stroke was observed (Lee et al. 2006). Similarly, after focal ischemia, DCX-positive cells were found to migrate into the striatum and the cortical regions adjacent to the infarct (Jin et al. 2003; Kunze et al. 2015). Following traumatic brain injury, DCX- and PSA-NCAM-positive cells migrated towards a cortical lesion in the parietotemporal cortex in mice (Dixon et al. 2015). Taken together, these studies demonstrate that neuroblasts migrate towards brain lesions, independent of the nature of brain damage. However, the migratory potential deviating from the physiological route seems to be limited to regions in the direct vicinity of the neurogenic regions or the RMS. In the case of neuroblast migration towards lesions of the cortex, the interface between cortex and corpus callosum appears to provide an alternative path of migration (Jin et al. 2003). In our study, the animals received excitotoxic lesions that did not touch the ventricle wall or the RMS. Hence, neuroblasts did only rarely travel towards the lesion site within the control group, which received a vehicle injection instead of the laminin tract (2 out of 8 hemispheres, both at a lower cell density than the laminin treated groups; cp. Fig. 5).

In the original study that demonstrated the capability of laminin to divert neuroblasts from the RMS into the surrounding tissue (Emsley and Hagg 2003), the authors also reported single cases of neuroblasts entering the vehicle tracts. We support their original assumption that these migratory processes are triggered by substances released due to the injection-induced microlesion in the vicinity of the RMS, which concurs with the above mentioned more recent studies on spontaneous migration of neuroblasts towards brain lesions.

Temporal aspects of neuroblast relocation

In the present study, we preferred to use the neuronal precursor and neuroblast marker DCX over the often used proliferation markers such as BrdU. In the context of this proof-of-concept study, the advantages of this method outweigh the restrictions with respect to a limited insight of the temporal aspects of laminin-based neuroblast redirection. Besides demonstrating that neuroblasts could be diverted from the RMS into a brain lesion we could validate that the relocated cells are present in the lesion site even after a survival time of 6 weeks. Meanwhile, a proliferation marker only marks cells that undergo reproduction while the substance is present systemically so that cells produced either prior to or after the application cannot be detected. Furthermore, the application of BrdU during formation of the glial scar would yield unclear results, since during this phase enhanced proliferation of glia cells is expected (Wanner et al. 2013).

However, the data at hand allow some interpretation of the underlying processes. First, in many cases, neuroblasts were found close to the laminin tract between the RMS and the lesion site. This could be either interpreted as a sign of disrupted migration with cells stuck on their way towards the lesion or seen as an indicator for neuroblast migration towards the lesion still going on by the end of the experiment. Since we did not observe a single case of neuroblasts accumulating ventral to the damaged tissue, we favor the assumption that cell migration towards the lesion has not yet ceased by the time the animals were killed.

Furthermore, cells resembling the morphology of adult neurons immunoreactive for both DCX as well as NeuN were found at the border of some lesion sites. This is especially remarkable since co-expression of these markers in developing neurons only occurs during a few days under physiological conditions (Brown et al. 2003). This finding implies that at least some of the relocated cells are capable of differentiating into adult neurons. This hypothesis is further supported by the results of Reynolds and Weiss (1992), who isolated and grafted cells from the adult rodent brain that were capable of generating new neurons. Although these authors used different cellular markers, the results show striking parallels to the outcome of our study: After 21 days in culture, a number of cells migrated away from the previously built cell spheres and differentiated into either a neuronal or astrocytic phenotype. The morphology of the migrating neuroblast, the astrocytes, as well as the differentiated neurons closely resembles our results [cp. (Reynolds and Weiss 1992), Fig. 3].

It is possible that more neuroblast have differentiated than deduced from our data, since neuroblasts that had completed maturation earlier might blend in with the NeuN-positive cells surrounding the lesion. However, this hypothesis needs to be evaluated in follow up experiments.

Another aspect that could be addressed by targeted use of proliferation markers is the long-term fate of newly generated neurons. Under physiological conditions, newly generated neuroblasts travel from the SVZ towards the OB within about 7 days in rats (Brown et al. 2003; Peretto et al. 1997). At about 15–30 days after proliferation most of these cells differentiate into mature neurons (Petreanu and Alvarez-Buylla 2002), but only a part of these cells survive for a prolonged time (Mizrahi et al. 2006; Mouret et al. 2009). It remains to be determined if, and to what extent the fate of replaced neurons in brain lesions differs from the development of replaced neurons in the olfactory bulb.

Potential clinical significance of neuroblast relocation

One major objective of this study was to evaluate the method originally described by Emsley and Hagg (2003) with regard to its potential for a future therapy of structural brain damage. Although our data provide further evidence that relocated neuronal precursor cells might contribute to a clinical application, a number of open questions need to be addressed.

First, the details of neuroblast migration in the human forebrain remain controversial. While neurogenesis in the hippocampus and SVZ of the adult human brain are undisputed (Curtis et al. 2011), the existence of a human RMS is topic of an ongoing debate. Initially, the lack of specific markers commonly found in the RMS of rodents and nonhuman primates (Kornack and Rakic 2001) has led to doubts, if neuroblast migration occurs in the human brain at all (Sanai et al. 2004). In later studies, however, an RMS was identified, though it showed a different anatomy and a markedly lower number of migrating neuroblasts compared to the RMS of rodents (Curtis et al. 2007; Wang et al. 2011). It has to be noted that the data on human neuroblast migration is based on brains of older subjects, since only post mortem analysis is applicable. More recent studies reported the human RMS activity to decline drastically during infancy (Bergmann et al. 2012; Sanai et al. 2011) and that neuroblasts in adult humans might rather migrate towards the striatum instead of the OB (Ernst et al. 2014). However, it has to be determined if neuronal progenitor cells located within the adult human SVZ are susceptible to migratory cues. If these cells do not enter the RMS physiologically but can be stimulated to migrate towards signal molecules, the application of a substance tract directly towards the SVZ might provide an advantageous alternative to the method used in this study.

Another important issue of neuroblast relocation presented here is the relatively low number of new cells found in the lesion site. Although even a lower number of neuronal precursor cells in damaged brain tissue have beneficial effects on functional recovery (Dixon et al. 2015; Li et al. 2010), the quantity of relocated neuroblasts needs to be significantly higher to compensate for the lost neuronal tissue. Nonetheless, our data indicates a long-term effect of the single laminin injection which might lead to a higher number of replaced neurons over time. Furthermore, the adult neurogenic niche is dynamically regulated by various systems (Ming and Song 2011) and the rate of neurogenesis is upregulated after neurological disorders such as stroke (Arvidsson et al. 2002) or seizures (Parent et al. 1997). A better insight into the underlying mechanisms and possible methods to influence these regulatory processes might increase the value of neuroblast redirection.

Conclusion

The method described here might eventually lead to a therapy of structural brain damage. It offers a comparatively simple and minimally invasive approach and the limited proliferative potential of the neuroblasts increases the safety of this method over methods based on embryonic stem cells (Meyer et al. 2010). Moreover, transplantation studies provide evidence that redirected neuroblasts repopulate damaged brain tissue and integrate into the complex networks successfully (Gallina et al. 2010; Shin et al. 2000).