Acute Inflammation Initiates the Regenerative Response in the Adult Zebrafish Brain

See allHide authors and affiliations

Science  07 Dec 2012:
Vol. 338, Issue 6112, pp. 1353-1356
DOI: 10.1126/science.1228773


The zebrafish regenerates its brain after injury and hence is a useful model organism to study the mechanisms enabling regenerative neurogenesis, which is poorly manifested in mammals. Yet the signaling mechanisms initiating such a regenerative response in fish are unknown. Using cerebroventricular microinjection of immunogenic particles and immunosuppression assays, we showed that inflammation is required and sufficient for enhancing the proliferation of neural progenitors and subsequent neurogenesis by activating injury-induced molecular programs that can be observed after traumatic brain injury. We also identified cysteinyl leukotriene signaling as an essential component of inflammation in the regenerative process of the adult zebrafish brain. Thus, our results demonstrate that in zebrafish, in contrast to mammals, inflammation is a positive regulator of neuronal regeneration in the central nervous system.

After a traumatic brain injury in the telencephalon, zebrafish can efficiently restore the tissue architecture and replace the lost neurons upon neurogenic activity of the ventricularly located radial glial cells (14). However, the signaling mechanisms involved in the activation of those stem cell niches after injury (reactive proliferation) and the production of new neurons (regenerative neurogenesis) from neurogenic progenitors are largely unknown.

Immune cell activation is among the first responses detected in the tissue after a severe central nervous system injury in both mammals and zebrafish (59). Various types of immune cells from the bloodstream (leukocytes) or resident in the brain tissue (microglia) accumulate in the injured tissue, secrete cytokines and chemokines to modulate the environment, and are responsible for the removal of cell debris and accumulated metabolites, as has been shown in mammals (10). Acute neuroinflammation in mammals has generally been considered to have a negative effect on neurogenesis and regeneration by promoting the formation of a glial scar and hampering the proliferation of the precursor cells and the migration, survival, maturation, and integration of the newborn neurons into the existing circuitry (1115). In contrast, the inflammatory response observed at the lesion site after injury in the zebrafish brain does not hamper the regeneration of the neurons (3). Thus, we hypothesized that an acute inflammatory reaction in fish might provide a context in which the molecular programs for regenerative neurogenesis could be initiated.

First, we characterized the inflammatory response upon traumatic brain injury, which leads to the activation of microglia and leukocytes that can be detected with L-plastin immunostaining (Fig. 1A) (16). After the brain lesion, the number of microglia and leukocytes in the injured telencephalic hemisphere increased significantly, and those elevated levels persisted for several days (Fig. 1, A to C). Based on the morphological transformation of the L-plastin–positive cells, as well as the elevated expression levels of the proinflammatory cytokines interleukin-8 (IL-8), IL-1β, and tumor necrosis factor–α, we concluded that an active inflammatory response takes place rapidly after traumatic brain injury (fig. S1). These results suggest that lesion in the adult zebrafish brain induces a bona fide inflammatory response characterized by proliferation (3) and a change of microglia morphology, which resolves within a few days without indications of chronic inflammation.

Fig. 1

Leukocytes infiltrate the brain rapidly after traumatic injury and sterile infection. (A and B) Immunohistochemistry (IHC) for L-plastin in unlesioned and lesioned telencephalons. (C) Quantification of the L-plastin cells in sham-operated, lesioned, and unlesioned telencephalic hemispheres in time course. (D and E) IHC for L-plastin in phosphate-buffered saline (PBS)– and zymosan A–injected brains. (F) Quantification of the L-plastin cells in PBS- and zymosan A–injected brains in time course. ns, not significant; **P < 0.01, ***P < 0.001; scale bars, 100 μm; n = 3 brains for every experiment.

Because traumatic brain injury induces active inflammation yet leads to efficient neural regeneration in zebrafish, we hypothesized that the inflammatory response might directly activate radial glial cells. In order to test this hypothesis, we sought to experimentally dissect the traumatic brain injury from the inflammatory response by injecting immunogenic zymosan A BioParticles (15) conjugated with fluorophores, using cerebroventricular microinjection (fig. S2, A to D) (17). In this way, the processes of the radial glial cells were not injured, and only sterile inflammation was induced. Similar to the response after traumatic brain injury, zymosan A injections led to a significant increase in the number of L-plastin–positive cells at 24 hours after injection, and the response was resolved 2 days later (Fig. 1, D to F). We also found that leukocytes engulf the yeast particles and undergo extensive shape changes from ramified to amoeboid morphology, and the levels of proinflammatory cytokines increase similarly as in the lesioned brains (fig. S2, E to L). These results indicate that zymosan A–induced sterile inflammation can mimic aspects of the inflammatory conditions after traumatic brain injury, whereas at the same time there is no effect on cell viability, as shown by TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling) and hematoxylin and eosin staining (figs. S3 and S4).

After traumatic brain injury, radial glial cells are activated, increase their proliferation levels significantly, and produce neurons to compensate for the neuronal loss (3, 4). To test whether induced inflammation would lead to similar outcomes without a lesion, we performed proliferating cell nuclear antigen (PCNA) immunostaining (marking proliferative cells) and S100β immunostaining (marking radial glial cells) on zymosan A–injected brains 1 day after the injection (Fig. 2, A and B). When we quantified the double-positive cells, we observed a significant increase of 49.7 ± 12.0% in the zymosan A–injected fish when compared to vehicle-injected fish (Fig. 2C), indicating that acute inflammation can enhance progenitor cell proliferation in the adult zebrafish brain. To test whether the increase in the proliferation of the progenitors leads to elevated levels of neurogenesis, after injecting zymosan A into the brain ventricle, coupled with bromodeoxyuridine (BrdU) pulses, we found that the number of newborn neurons was increased significantly by 39.7 ± 1.3% as compared to vehicle-injected fish (Fig. 2, D to F).We performed similar experiments in the cerebellum, where the same effect of increased progenitor proliferation and neurogenesis was observed (fig. S5).

Fig. 2

Inflammation is sufficient and necessary for the initiation of reactive proliferation and reactive neurogenesis. (A and B) IHC for S100β and PCNA in PBS- or zymosan A–injected zebrafish brains. (C) Quantification of S100β/PCNA-positive cells in PBS and zymosan A–injected brains. (D and E) IHC for HuC/D and BrdU. (F) Quantification of the HuC/D/BrdU double-positive cells between PBS and zymosan A injections. (G and H) IHC for S100β and PCNA. (I) Quantification of proliferating radial glia in both lesioned and unlesioned hemispheres between Dex- and vehicle-treated animals. (J and K) IHC for HuC/D and BrdU. (L) Quantification of the newborn neurons (HuC/D/BrdU double-positive cells). *P < 0.05, **P < 0.01, ***P < 0.001; scale bars, 100 μm; n = 4 brains for every experiment.

Our data suggest that the inflammatory response might be one of the molecular cues that imminently precede the activation of radial glial cells. Thus, to analyze whether the inflammatory response was necessary for the initiation of the reactive proliferation, we immunosuppressed the zebrafish by administering the anti-inflammatory drug dexamethasone (Dex). After determining the efficiency of the immunosuppression assay (fig. S6), we concluded that Dex treatment did partially suppress the immune system of the adult zebrafish, and hence we could proceed to test the effect during regeneration. We killed the lesioned (control and immunosuppressed) fish at 3 days after lesion and analyzed the proliferation response of the radial glial cells (Fig. 2, G to I). We observed that compared to the control brains, the reactive proliferation response in the lesioned brains was significantly reduced after immunosuppression (61.2 ± 2.2%). We subsequently tested the effect of immunosuppression on reactive neurogenesis.The number of newborn neurons was significantly lower (55.8 ± 1.8%), as shown by HuC/D and BrdU double labeling (Fig. 2, J to L). To examine whether Dex had any effect on the constitutive proliferation and neurogenesis, we performed the immunosuppression assay also in unlesioned animals, where no difference was detected between control and immunosuppressed animals (fig. S7). This finding shows clearly that Dex impedes only the regenerative response, without any additional effect on brain homeostasis. Moreover, we examined the effect of immunosuppression in the caudal fin, and we observed that immunosuppressed animals could not regenerate their caudal fins (fig. S8).

Because inflammation is involved in the activation of reactive proliferation and neurogenesis, we performed a transcriptome screen to identify possible molecular players involved in the regenerative response. We determined that cysteinyl leukotriene receptor 1 (cysltr1) was induced after traumatic brain injury, especially in the lesioned hemisphere (Fig. 3A). By using the transgenic line Tg(her4.1:GFP), which marks radial glial cells, we found that CysLT1 is predominantly expressed in those cells (Fig. 3B and fig. S9A). After zymosan A injection, cysltr1 expression was also up-regulated in the ventricular zone (Fig. 3, C and D), suggesting that the inflammatory response after lesion might enhance leukotriene signaling in radial glial cells. Next, we analyzed the need for CysLT1 signaling in the reactive proliferation of the radial glial cells and the subsequent neurogenesis by injecting Pranlukast, an antagonist of CysLT1 (18, 19), into the ventricle after lesion. When we analyzed the proliferation response of the glial cells, we observed that compared to vehicle-injected fish, injury-induced proliferation levels were reduced significantly, by 54.9 ± 5.5% in lesioned and Pranlukast-injected telencephalons (Fig. 3, E to G). We subsequently found that the number of newborn neurons was significantly reduced by 33.5 ± 4.6% in the Pranlukast-injected telencephalons as compared to the vehicle-injected ones (Fig. 3, H to J).

Fig. 3

The CysLT1–LTC4 pathway is required and sufficient for enhanced proliferation and neurogenesis. (A) cysltr1 expression is up-regulated in the lesioned hemisphere at the ventricular zone at 3 days after lesion. (B) IHC shows that CysLT1 is expressed in radial glial cells. (C and D) cysltr1 is up-regulated significantly at the ventricular zone after zymosan A injections. (E and F) IHC for S100β and PCNA. (G) Quantification of S100β/PCNA-positive cells in dimethyl sulfoxide (DMSO)– and Pranlukast-injected brains. (H and I) IHC for HuC/D and BrdU. (J) Quantification of HuC/D/BrdU-positive cells in DMSO- and Pranlukast-injected brains. (K and L) IHC for S100β and PCNA in MetOH- and LTC4-injected brains (M) LTC4 injections initiate the reactive proliferation response. (N and O) IHC for HuC/D and BrdU. (P) LTC4 injections induce reactive neurogenesis. *P < 0.05, **P < 0.01, ***P < 0.001; scale bar in (A), 200 μm; scale bar in (B), 10 μm; scale bars in (C), (D), (K), (L), (N), and (O), 100 μm; dashed circle in (A) indicates the lesion site; n = 3 brains for every experiment.

Then, we hypothesized that the leukotriene signaling pathway itself might increase the proliferation of radial glial cells and the overall neurogenesis response. To test this hypothesis, we injected leukotriene C4 (LTC4), one of the ligands for CysLT1, which had been sufficient to enhance its receptor expression through a feedback mechanism (fig. S9, B and C) and had not caused an inflammatory response itself (fig. S10), into the ventricle of the zebrafish brain without lesion. We analyzed the proliferation of the glial cells at 24 hours after injection. Compared to control injections, LTC4-injected brains displayed significantly elevated levels of progenitor proliferation (46.7 ± 0.8%) (Fig. 3, K to M). Additionally, we observed that LTC4 injection also significantly increased the consequent neurogenesis at 21 days after injection (44.4 ± 4.7%) (Fig. 3, N to P).

Our data suggest a role of the inflammatory response and immune cells in reactive proliferation of the radial glial cells and consequently enhanced neurogenesis. However, it is known that zebrafish radial glial cells proliferate constitutively and that newborn neurons are formed continuously (4, 2024). This poses the question of whether the inflammatory outcome enhances cell proliferation and neurogenesis by amplifying the already-existing signals of adult neurogenesis or initiates an injury-induced molecular program specific to the regenerative response. To address this question, we took advantage of the injury-induced expression of the zinc-finger transcription factor Gata3 as a marker (25). We have previously shown that gata3 is induced only after traumatic brain injuries in the adult zebrafish telencephalon, whereas it is undetectable under homeostatic conditions. We found that zymosan A injections were sufficient to induce gata3 expression at the ventricular zone at 24 hours after injection, similar to the lesion model (Fig. 4, A to C). Additionally, by immunosuppression studies we found that inflammation is required for the induction of gata3 expression, because after zymosan A injection, gata3 expression is reduced significantly upon immunosuppression (Fig. 4D). Moreover, in a Gata3 knockdown background, reactive proliferation ceased upon zymosan A injection (fig. S11).

Fig. 4

Inflammation is sufficient and necessary for inducing regeneration-specific molecular programs. (A and B) gata3 is induced at the ventricular zone 24 hours after zymosan A injections in comparison to vehicle injections. (C) Relative fold change of gata3 expression between vehicle and zymosan A injections. (D) Relative expression levels of gata3 at 3 days after lesion in immunosuppressed and control animals. (E and F) gata3 is induced significantly in the ventricular region 24 hours after LTC4 injection, in comparison to vehicle-injected brains. (G) Relative expression levels of gata3 in vehicle- and LTC4-injected brains. (H) Quantification graph indicates that gata3 expression reduces significantly after Pranlukast injection, in comparison to vehicle-injected telencephalons. *P < 0.05, **P < 0.01, ***P < 0.001; scale bars, 100 μm; n = 3 brains for every experiment).

LTC4 injection without the lesion was able to activate gata3 expression in the adult zebrafish telencephalon (Fig. 4, E to G). When we blocked the CysLT1 activity using Pranlukast, gata3 expression levels were significantly diminished after lesions (Fig. 4H). Together these data indicate that the inflammatory response, including the LTC4 signaling activity, initiates regeneration programs that are distinct from the constitutive neurogenesis in the adult zebrafish brain.

In mammals, which cannot efficiently regenerate neurons after injuries, several studies have shown that acute inflammation impedes adult neurogenesis and regenerative capacity (13, 26). Our results suggest that acute inflammation can promote central nervous system regeneration, because it provides cues necessary for the initiation of reactive proliferation and regenerative neurogenesis in the adult zebrafish brain (fig. S12). Our findings reveal a signaling pathway in zebrafish that couples the inflammatory response to efficient enhancement of stem cell activity and the initiation of neural regeneration. Thus, the zebrafish offers an opportunity to study the crosstalk between the inflammatory response and successful regeneration programs in the central nervous system of a regeneration-capable vertebrate, with the aim of translating such findings into potential therapeutic applications to treat traumatic brain injuries and neurodegenerative disorders.

Supplementary Materials

Materials and Methods

Figs. S1 to S12

Tables S1 and S2

Reference (27)

References and Notes

  1. Acknowledgments: We thank A. Menge for cryosections and paraffin sections, G. Kempermann and C. L. Antos for useful discussions, T. Chavakis for discussions and critical reading on the manuscript, N. Trede for discussion and assistance with the immunosuppression assay, and M. Redd for kindly providing us the L-plastin antibody. This work was supported by research grants from the Deutsche Forschungsgemeinschaft (SFB-655), European Union (ZF Health, EC Grant Agreement HEALTH-F4-2010-242048), and Technical University of Dresden.
View Abstract

Navigate This Article