Subventricular zone cytogenesis provides trophic support for neural repair

  1. Michael R. Williamson
  2. Stephanie P. Le
  3. Ronald L. Franzen
  4. Nicole A. Donlan
  5. Jill L. Rosow
  6. Andrew K. Dunn
  7. Theresa A. Jones
  8. Michael R. Drew

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    The authors use a mouse stroke model to address a potential cellular source of functional recovery. Using multiple lineage tracing paradigms, they show that undifferentiated progenitor cells that migrate from the subventricular zone produce trophic factors including VEGF that promote functional and cellular recovery. These findings will be of interest to the neuroscience community, and those who study neural repair.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer #2 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Stroke enhances proliferation of neural precursor cells within the subventricular zone (SVZ) and induces ectopic migration of newborn cells towards the site of injury. Here we characterize the identity of cells arising from the SVZ after stroke and provide insight into their function by uncovering a mechanism through which they facilitate neural repair and functional recovery. Using genetic lineage tracing, we show that SVZ-derived cells that migrate towards stroke- induced cortical lesions in mice are predominantly undifferentiated precursors, suggesting that the main function of post-injury cytogenesis is not cell replacement. We find that SVZ-derived cells are a unique cellular source of trophic factors that instruct neural repair. Chemogenetic ablation of neural precursor cells or conditional knockout of VEGF in the adult neural stem cell lineage impairs neuronal and vascular reparative responses and worsens functional recovery after stroke. In addition, normal aging markedly diminishes the cytogenic response to stroke, resulting in worse functional recovery. Therapeutic replacement of VEGF in peri-infarct cortex is sufficient to induce neural repair and functional recovery in mice with arrested cytogenesis. These findings indicate that the SVZ cytogenic response following brain injury is a source of trophic support that drives neural repair and recovery.

Article activity feed

  1. eLife

    Author Response

    Reviewer #1 (Public Review):

    In this article, the authors are trying to ascertain how emigrated SVZ cells can be beneficial - via neuroreplacement or neuroprotection. They provide evidence for the latter and also show that it is primarily precursors and not differentiated cells that migrate to photo-thrombotic cortical models of stroke.

    The writing is lucid and the flow of the experiments logical. The images and quality of data are high and the depth of investigation appropriate (eg 100 cells examined per marker in Figure 1). The methods are clearly described. They appropriately control for changes in cortical lesion size. The photo-thrombotic lesion is a good choice in terms of controlling lesion placement and size.

    A distinctive advantage of this paper is they show that reducing SVZ cytogenesis in the stroke model diminishes recovery, especially behavioural (single seed reaching behavior). This essential experiment has been remarkably under-utilized in the field.

    The 2-photon imaging of dendric spines after stroke combined with multi-exposure speckle imaging is a technical tour-de-force especially since they combine it with ganciclovir-induced loss of cytogenesis and behavioural assays. Importantly, they show that SVZ cells are needed for full spine plasticity.

    They are correct to examine the SVZ response in aging as it diminishes dramatically in animal models but in humans is associated with more strokes. As expected, they show reduced SVZ proliferation after stroke. This was associated with significantly worse performance in the seed-reaching task and depleting SVZ precursors with ganciclovir did not make it worse.

    The viral VEGF delivery rescue experiment is fantastic. Behavior, blood vessel growth, and spine density are all rescued.

    The idea that SVZ cells are beneficial via mechanisms other than cell replacement is not really that new. For example, neural stem cells from the SVZ have been shown to reduce inflammation and thereby be neuroprotective as the authors themselves acknowledge and cite (Pluchino et al., 2005).

    The fact that it is primarily precursor cells that migrate towards the stroke does not mean that cell replacement does not occur. The precursors could gradually differentiate (even after 6 weeks post-injury) into more mature cells that do replace cells lost to injury. Also, the two events are not mutually exclusive.

    Our findings indicate that there is no appreciable differentiation of SVZ-derived cells up to 6 weeks after stroke. By this time, we find complete recovery of behavioral deficits. While it is conceivable that cells may differentiate after this timepoint, such a phenomenon would not be contributing to recovery.

    Overall this is an interesting addition to the literature and methodologically it is quite strong. It is sure to generate follow on studies showing how different growth factors may be secreted by SVZ cells in various models of neurological disease.

    Reviewer #3 (Public Review):

    Williamson et al. have investigated the role of cells derived from a neural stem cell (NSC) region of the adult mouse brain called the subventricular zone (SVZ) in a model of stroke. The authors labeled SVZ cells with Nestin-CreER and the Ai14 (tdTomato) reporter, induced cortical infarcts 4 weeks later, then analyzed brains 2 weeks thereafter. Most of the tdTomato+ cells in the peri-infarct regions were not neurons but less differentiated neural precursor cells. They then ablated proliferating NSCs in the SVZ with GFAP-TK mice and ganciclovir (GCV) administration, and this reduced SVZ-derived peri-stroke cells and impaired motor recovery. Older mice have less proliferation in the SVZ, and these older mice have fewer peri-infarct SVZ-derived cells and worse recovery than younger mice. Using multi-exposure speckle imaging (MESI) and 2 photon imaging, the authors found that ablation of proliferating SVZ cells reduced vascular remodeling and synaptic turnover in peri-infarct areas. Immunohistochemical analysis revealed the expression of VEGF, BDNF, GDNF, and FGF2. The authors selected VEGF for functional studies, conditionally knocking out VEGF in SVZ cells and finding that this reduced recovery and neuronal spine density. Finally, the authors expressed VEGF by AAV vectors in mice with ablated SVZ, finding that VEGF could improve repair and recovery after stroke.

    The results presented in the paper support some of the authors' general conclusions and may be of interest to investigators of adult mouse SVZ. The use of genetic labels for lineage analysis and studies of VEGF conditional knockout in SVZ cells are technical strengths of the study. The results support the idea that VEGF in SVZ cells is important for recovery from stroke in younger adult mice. However, the impact of the work may be somewhat limited, as outlined below.

    1. It is already well known that VEGF is an important aspect of stroke recovery (at least in rodent models), and that ectopic expression of VEGF can be beneficial. Showing that some of the VEGF in peri-stroke regions might come from SVZ-derived cells would be a relatively incremental discovery.

    We disagree. In our view, the identification of SVZ-derived cells as a major cellular source of an important trophic factor for recovery is itself an important finding. We also demonstrate that VEGF produced by this cell population is necessary for effective neural repair and recovery, while replacement of VEGF is sufficient to induce repair and recovery in mice lacking this cell population. Moreover, these findings provide a compelling explanation for the worsening of recovery and diminishment of repair that occurs with age (i.e., loss of VEGF signaling from the neural stem cell lineage). Finally, the demonstration that replacing VEGF rescues deficits that accompany loss of SVZ stem cells provides rationale for the replacement of neural stem cell lineage factors as a potential treatment.

    The molecular mechanism aside, our main goal was to understand the function of SVZ cytogenesis in stroke recovery. Our findings that 1) the majority of cells arising from the SVZ after stroke remain in an undifferentiated state, 2) these cells facilitate neuronal and vascular reparative processes in order to promote recovery, and 3) very few new neurons are produced during the recovery phase, provide a new and unexpected understanding of the purpose of post-injury cytogenesis. The dogma of past literature is that neural stem cells produce new neurons that mature and integrate into damaged circuits after injury. An implication of this dogma is that neuronal replacement after stroke is an important treatment target. Accordingly, substantial effort has been devoted to developing cell transplantation and conversion treatment strategies to create new neurons. Our study reframes the function of newborn cells in stroke recovery and provides a compelling rationale against treatment strategies aimed at replacing neurons, instead demonstrating that trophic factor mediated repair and remodeling of spared tissue is sufficient for profound recovery of function. To emphasize these important findings, we have expanded our discussion (lines 338-378).

    1. Furthermore, while it seems clear that the VEGF conditional knockout (VEGF-cKO) in SVZ cells reduces behavioral recovery and certain histological measures, it is not clear that these impairments are due to a lack of VEGF delivery from the SVZ cells. It is possible that VEGF-cKO changed the proportion of SVZ cells that arrive in the peri-stroke region. It is also possible that VEGF-cKO makes these cells impaired in the expression of other trophic factors.

    We disagree with this interpretation. We used a cell type-specific, inducible knockout of VEGF to examine the function of VEGF produced by the adult neural stem cell lineage after stroke. We show in Figure 6E that numbers of lineage traced cells are not different between control and cKO mice. We have added new data, as Figure 6 – figure supplement 2, showing that the proportion of Ascl1-expressing cells is not different between groups, indicating that there is no change in the amount of differentiation. We have also added staining demonstrating that VEGF cKO cells still express GDNF, BDNF, and FGF-2 (Figure 6 – figure supplement 2). Notably, we did not detect a decrease in these three proteins in peri-infarct regions in mice in which neural stem cells were ablated, suggesting that while SVZ-derived cells do produce them, their production is small relative to other cell types. VEGF is thus unique in that large quantities of it are produced by SVZ-derived cells. We also provide evidence for direct effects of VEGF on other cell types (rather than cell-autonomous effects of VEGF regulating other factors in SVZ-derived cells that then act on other cell types) since restoring VEGF in mice with ablated neural stem cells rescued repair and recovery. Importantly, even if VEGF cKO led to perturbed expression of some other proteins, our conclusion would still be that VEGF produced by SVZ-derived cells is crucial for promoting repair and recovery.

    1. The cytogenic response to stroke was not characterized in much detail at the cellular level. Essentially only one time point (2 weeks) was selected for immunohistochemistry (Fig. 1), and so the dynamics of this response cannot be evaluated. Does the proportion of cell types change over time? Are migratory cells more homogeneous and then diversify after arrival to the peri-stroke region? At longer time points, do these SVZ-derived cells still exist? Such an analysis is important to the story since the behavior was evaluated at a range of time points (3-28 days after stroke), and recovery was noted as early as 7 days. Are SVZ-derived cells already at the peri-stroke area after 7 days? If they are not already there, then how would the recovery be explained? The behavioral recovery also continues to improve at 28 days; are SVZ-derived cells still present in large numbers at that time? How would the authors explain continued recovery if the SVZ-derived cell population drops away after 2 weeks?

    Thank you for these suggestions – we agree that these are important points to address. In our original submission we provided evidence that there is no significant differentiation into neurons even 6 weeks after stroke. We have added additional data (Figure 1-figure supplement 2) in which we assessed expression of a range of markers at 6 weeks post-stroke, a time by which recovery is typically complete in this model. These new data show that cell type distribution of lineage traced cells at 6 weeks is highly similar to the 2-week timepoint and show that little differentiation occurs during the course of recovery. We have also included new data quantifying numbers of SVZ-derived cells in peri-infarct cortex at 1, 2, and 6 weeks post-stroke. These data show that lineage traced cells are abundant by day 7 and numbers increase progressively to 6 weeks, suggesting that these cells are present when we initially detect functional improvement, survive, and are continuously produced during the course of recovery. Finally, we have added data showing that the proportion of Ascl1-expressing cells does not change from 1 to 6 weeks, which is consistent with the idea that there are no dynamic changes in cell phenotype during recovery.

    1. The SVZ-derived peri-stroke cells were not characterized in much detail at the molecular/transcriptomic level. The authors studied 4 trophic factors by antibody staining, but there are many other potential genes that may contribute to the effect. Transcriptomic analyses of SVZ-derived peri-stroke cells (e.g., by single-cell RNA-seq) may provide deeper insights into potential mechanisms.

    We acknowledge that single-cell RNA sequencing of SVZ-derived cells may reveal other interesting molecular mechanisms, but such a study could easily stand on its own. There are also technical limitations, such as the population of cells being relatively small, that would create difficulty in generating such a dataset. Instead, we focused on protein-level expression of a small cohort of factors that could be potentially involved based on our initial findings and past literature. In particular, we focused on examining proteins known to potently drive angiogenesis, axon outgrowth, and/or synapse formation given our findings of deficient vascular and synaptic repair in mice lacking cytogenesis. Even if single cell sequencing provided new molecular targets, the ensuing workflow would mirror what we have done in our study (validation of protein level expression, loss-of-function manipulation, gain-of-function manipulation). The magnitude of deficits in VEGF cKO mice did not completely match that seen in mice in which neural stem cells were ablated, making it likely that SVZ-derived cells also contribute to recovery by other mechanisms. We have added to the discussion: “Importantly, these past studies have identified an array of factors produced by precursor cells depending on context. It is possible that multiple factors produced by SVZ-derived cells promote recovery after stroke. This is suggested by our finding that recovery is worse in mice with ablated neural stem cells compared with VEGF cKO mice. Thus, future studies could examine other molecular targets. These efforts could be aided by techniques such as single-cell RNA sequencing.” (lines 338 - 343).

    1. The significance of this work for understanding stroke in human patients is unclear since the adult human brain SVZ is essentially devoid of neurogenic stem cells. Thus, although some of the observations in this paper are interesting, the cytogenic response to stroke described here may not occur in human patients.

    We disagree for several reasons. First, while there is an ongoing debate on whether neurogenesis or neural stem cells persist in adult humans, this debate has not been resolved. At least in our opinion, the preponderance of evidence is in support of persistent cytogenesis and neural stem cells in the adult human SVZ due to convincing data across many studies (e.g., PMID: 10328940, 9809557, 14973487, 11333968, 10870078, 24561062). While it appears that neural stem cell proliferation declines with aging, as in rodents, there is evidence of increased SVZ proliferation and cytogenesis in response to stroke in adult and elderly humans (e.g., PMID: 20054008, 16924107, 17167100, 20104652). Thus, although it is exceedingly difficult to study in humans, it is likely that neural stem cells persist in the adult human brain and can respond to injury by producing new cells. One reason for the somewhat sparse evidence of post-stroke cytogenesis in humans may be that the focus of past studies has been on finding new neurons. Importantly, our study demonstrates that other cells types, especially undifferentiated precursors, arise from the SVZ after stroke in far greater numbers than neurons, which may spur further examination of the phenomenon in humans.

    Second, while stroke incidence increases with age, stroke is not uncommon in the young. Moreover, incidence of stroke in the young is increasing (PMID: 32015089). It is generally accepted that young humans have intact neural stem cells, and the phenomenon we describe in our study shows clear benefits of SVZ cytogenesis in young mice.

    Third, our study provides evidence that “neurogenic” capacity of neural stem cells may not be important for the beneficial functions of cytogenesis after injury. The overwhelming majority studies in humans and mice have focused on “neurogenesis”. Our study demonstrates that undifferentiated precursors constitute the majority of SVZ-derived cells after stroke and identifies Ascl1 as a marker of them, which may be useful for identifying these cells in humans.

    Fourth, if neural stem cell numbers are substantially reduced in aged humans, as in rodents, our study provides clear rationale for the development of treatments to restore stem cell numbers/activation or limit their decline with aging.

    Fifth, our study not only identifies VEGF as a mechanism by which SVZ-derived cells promote repair and recovery after stroke, but also demonstrates that replacing VEGF is sufficient to improve repair and recovery in mice lacking neural stem cells. Thus, even if the argument that cytogenesis does not occur in adult or elderly humans is true, our study shows that identification and replacement of factors produced by the neural stem cell lineage, such as VEGF, could be a reasonable treatment strategy with clear translational potential.

    In order to more clearly state these points in the manuscript, we have expanded our discussion of them.

  2. eLife

    Evaluation Summary:

    The authors use a mouse stroke model to address a potential cellular source of functional recovery. Using multiple lineage tracing paradigms, they show that undifferentiated progenitor cells that migrate from the subventricular zone produce trophic factors including VEGF that promote functional and cellular recovery. These findings will be of interest to the neuroscience community, and those who study neural repair.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer #2 agreed to share their name with the authors.)

  3. eLife

    Reviewer #1 (Public Review):

    In this article, the authors are trying to ascertain how emigrated SVZ cells can be beneficial - via neuroreplacement or neuroprotection. They provide evidence for the latter and also show that it is primarily precursors and not differentiated cells that migrate to photo-thrombotic cortical models of stroke.

    The writing is lucid and the flow of the experiments logical. The images and quality of data are high and the depth of investigation appropriate (eg 100 cells examined per marker in Figure 1). The methods are clearly described. They appropriately control for changes in cortical lesion size. The photo-thrombotic lesion is a good choice in terms of controlling lesion placement and size.

    A distinctive advantage of this paper is they show that reducing SVZ cytogenesis in the stroke model diminishes recovery, especially behavioural (single seed reaching behavior). This essential experiment has been remarkably under-utilized in the field.

    The 2-photon imaging of dendric spines after stroke combined with multi-exposure speckle imaging is a technical tour-de-force especially since they combine it with ganciclovir-induced loss of cytogenesis and behavioural assays. Importantly, they show that SVZ cells are needed for full spine plasticity.

    They are correct to examine the SVZ response in aging as it diminishes dramatically in animal models but in humans is associated with more strokes. As expected, they show reduced SVZ proliferation after stroke. This was associated with significantly worse performance in the seed-reaching task and depleting SVZ precursors with ganciclovir did not make it worse.

    The viral VEGF delivery rescue experiment is fantastic. Behavior, blood vessel growth, and spine density are all rescued.

    The idea that SVZ cells are beneficial via mechanisms other than cell replacement is not really that new. For example, neural stem cells from the SVZ have been shown to reduce inflammation and thereby be neuroprotective as the authors themselves acknowledge and cite (Pluchino et al., 2005).

    The fact that it is primarily precursor cells that migrate towards the stroke does not mean that cell replacement does not occur. The precursors could gradually differentiate (even after 6 weeks post-injury) into more mature cells that do replace cells lost to injury. Also, the two events are not mutually exclusive.

    Overall this is an interesting addition to the literature and methodologically it is quite strong. It is sure to generate follow on studies showing how different growth factors may be secreted by SVZ cells in various models of neurological disease.

  4. eLife

    Reviewer #2 (Public Review):

    These studies investigated the identity of cells that migrate in response to stroke from the stem cell niche, the subventricular zone (SVZ). They also showed that these cells are important in the repair processes following cortical ischemia as mice who had stem cells ablated or had age-associated reduced progenitor number had less improvement in a motor task. Finally, they identify the mechanism for this progenitor-driven repair as both synaptic plasticity and angiogenesis following ischemia that is driven by the production of trophic factors most notably VEGF. The major strengths of the paper are the use of multiple promoters to drive the lineage tracing fluorescent marker. In addition to the traditional NesinCre-ER mice with a tdTomato tag, they use an Ascl-1Cre-ER mice which is in fewer progenitors but is more specific to neural progenitors and not upregulated in activated astrocytes to support their findings that the majority of migrating cells are progenitors. To further support this finding they also show the majority of cells do not express the mature astrocyte marker S100beta. The neural stem cell ablation model is the well-established GFAP-TK mouse model which uses ganciclovir to ablate neural progenitors and most importantly they show that it is working for them which increases the rigor of the study. The mechanistic studies are convincing because not only do they use a cortical window and two-photon microscopy to measure changes in the synapsis and vasculature over time but they also do gain and loss of function studies to support that VEGF is a major driver of the reparative response.

  5. eLife

    Reviewer #3 (Public Review):

    Williamson et al. have investigated the role of cells derived from a neural stem cell (NSC) region of the adult mouse brain called the subventricular zone (SVZ) in a model of stroke. The authors labeled SVZ cells with Nestin-CreER and the Ai14 (tdTomato) reporter, induced cortical infarcts 4 weeks later, then analyzed brains 2 weeks thereafter. Most of the tdTomato+ cells in the peri-infarct regions were not neurons but less differentiated neural precursor cells. They then ablated proliferating NSCs in the SVZ with GFAP-TK mice and ganciclovir (GCV) administration, and this reduced SVZ-derived peri-stroke cells and impaired motor recovery. Older mice have less proliferation in the SVZ, and these older mice have fewer peri-infarct SVZ-derived cells and worse recovery than younger mice. Using multi-exposure speckle imaging (MESI) and 2 photon imaging, the authors found that ablation of proliferating SVZ cells reduced vascular remodeling and synaptic turnover in peri-infarct areas. Immunohistochemical analysis revealed the expression of VEGF, BDNF, GDNF, and FGF2. The authors selected VEGF for functional studies, conditionally knocking out VEGF in SVZ cells and finding that this reduced recovery and neuronal spine density. Finally, the authors expressed VEGF by AAV vectors in mice with ablated SVZ, finding that VEGF could improve repair and recovery after stroke.

    The results presented in the paper support some of the authors' general conclusions and may be of interest to investigators of adult mouse SVZ. The use of genetic labels for lineage analysis and studies of VEGF conditional knockout in SVZ cells are technical strengths of the study. The results support the idea that VEGF in SVZ cells is important for recovery from stroke in younger adult mice. However, the impact of the work may be somewhat limited, as outlined below.

    1. It is already well known that VEGF is an important aspect of stroke recovery (at least in rodent models), and that ectopic expression of VEGF can be beneficial. Showing that some of the VEGF in peri-stroke regions might come from SVZ-derived cells would be a relatively incremental discovery.
    2. Furthermore, while it seems clear that the VEGF conditional knockout (VEGF-cKO) in SVZ cells reduces behavioral recovery and certain histological measures, it is not clear that these impairments are due to a lack of VEGF delivery from the SVZ cells. It is possible that VEGF-cKO changed the proportion of SVZ cells that arrive in the peri-stroke region. It is also possible that VEGF-cKO makes these cells impaired in the expression of other trophic factors.
    3. The cytogenic response to stroke was not characterized in much detail at the cellular level. Essentially only one time point (2 weeks) was selected for immunohistochemistry (Fig. 1), and so the dynamics of this response cannot be evaluated. Does the proportion of cell types change over time? Are migratory cells more homogeneous and then diversify after arrival to the peri-stroke region? At longer time points, do these SVZ-derived cells still exist? Such an analysis is important to the story since the behavior was evaluated at a range of time points (3-28 days after stroke), and recovery was noted as early as 7 days. Are SVZ-derived cells already at the peri-stroke area after 7 days? If they are not already there, then how would the recovery be explained? The behavioral recovery also continues to improve at 28 days; are SVZ-derived cells still present in large numbers at that time? How would the authors explain continued recovery if the SVZ-derived cell population drops away after 2 weeks?
    4. The SVZ-derived peri-stroke cells were not characterized in much detail at the molecular/transcriptomic level. The authors studied 4 trophic factors by antibody staining, but there are many other potential genes that may contribute to the effect. Transcriptomic analyses of SVZ-derived peri-stroke cells (e.g., by single-cell RNA-seq) may provide deeper insights into potential mechanisms.
    5. The significance of this work for understanding stroke in human patients is unclear since the adult human brain SVZ is essentially devoid of neurogenic stem cells. Thus, although some of the observations in this paper are interesting, the cytogenic response to stroke described here may not occur in human patients.

  6. Version published to 10.1101/2022.06.14.496078v1 on bioRxiv