Fate of Newborn Dentate Granule Cells after Early Life Status Epilepticus

Authors

  • Brenda E. Porter,

    1. Division of Pediatric Neurology, The Children's Hospital of Philadelphia, and Departments of Neurology and Pediatrics at the University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, U.S.A.
    Search for more papers by this author
  • Margaret Maronski,

    1. Division of Pediatric Neurology, The Children's Hospital of Philadelphia, and Departments of Neurology and Pediatrics at the University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, U.S.A.
    Search for more papers by this author
  • Amy R. Brooks-Kayal

    1. Division of Pediatric Neurology, The Children's Hospital of Philadelphia, and Departments of Neurology and Pediatrics at the University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, U.S.A.
    Search for more papers by this author

Address correspondence and reprint requests to Dr. B.E. Porter at Division of Child Neurology, 6 Wood, The Children's Hospital of Philadelphia, 34th and Civic Center Boulevard, Philadelphia, PA 19104, U.S.A. E-mail: Porterb@email.chop.edu

Abstract

Summary: Purpose: To determine the fate of newborn dentate granule cells (DGCs) after lithium-pilocarpine–induced status epilepticus (SE) in an immature rat.

Methods: Postnatal day 20 (P20) rats were injected with lithium and pilocarpine to induce SE, and then with bromodeoxyuridine (BrdU) 4, 6, and 8 days later (P24, 26, and 28), and killed 1 day (P29), 1 week (P34), and 3 weeks (P50) after the last dose of BrdU for cell counts. Immunohistochemistry and TUNEL staining were performed to assess the fate of newborn DGCs.

Results: Pilocarpine-treated animals had significantly more BrdU-labeled DGCs than did littermate controls at all times. The day after the final BrdU injection (P29), sixfold more cells were found in pilocarpine-treated animals than in controls, which was reduced to threefold, 3 weeks later. A decrease in the BrdU-labeled cell density was noted from P29 to P50 in the control and pilocarpine-treated animals. Evidence of DGC cell death was seen in pilocarpine and control animals, with threefold more TUNEL-positive cells in the pilocarpine-treated than in the control animals at P29. The surviving newborn DGCs became mature neurons; expressing the neuronal marker NeuN in both control and pilocarpine-treated animals.

Conclusions: These findings suggest that SE during postnatal development increases the birth and death of DGCs. A subset of the newborn DGCs survive and mature into dentate granule neurons, resulting in an increased population of immature DGCs after SE that may affect hippocampal physiology.

The generation of a large number of new neurons throughout life is a distinctive feature of the hippocampal dentate gyrus (DG) (1,2). Limited studies have suggested that immature dentate granule cells (DGCs) have unique physiologic properties, and the birth of new DGCs is necessary for the establishment of certain types of memory in the rodent (3–5). As the total numbers of DGCs do not increase throughout life, a concomitant loss of DGCs must occur. The gradual loss of DGCs over weeks to months has been demonstrated in several species; however, why certain cells die is not well understood (6,7).

Several growth factors have been implicated in DG neurogenesis (8–10), and alterations in the environment, drugs, hormones, age, and seizures all influence the rate of neurogenesis (11–14). Multiple types of seizures have been shown to modify the birth rate of DGCs, increasing the rate from one to as much as tenfold (14–17), although also see (18). Along with the increase in the birth of newborn DGCs, an increase of injured and dying DGCs also occurs after status epilepticus (SE) (19,20). Many of the dying cells are found at the dentate/hilar border, the area of DGC birth, suggesting that immature DGCs may be particularly susceptible to dying, although the number and phenotype of the dying cells has not been extensively characterized. Because SE increases both the birth and death rate of DGCs, it remains unclear how many newborn DGCs survive long after SE and thus could contribute to alterations in dentate physiology. To better understand the role of newborn neurons in the DG after an episode of SE during postnatal development, we characterized the fate of newborn DGCs after lithium-pilocarpine–induced SE in the immature rat.

METHODS

The Animal Care and Use Committee at The Children's Hospital of Philadelphia approved this study.

Lithium-pilocarpine injections

All animals were housed in standard cages with ad libitum feed and 12-h light/dark cycle. Male rat pups on P19 were injected first with lithium chloride (Sigma, St. Louis, MO, U.S.A.), 3 mEq/kg intraperitoneally (i.p.). Twelve to 16 h later, pups were injected with pilocarpine, 60 mg/kg i.p. (Sigma) or an equivalent amount of saline (lithium controls). Pilocarpine injections typically trigger long-duration (>30 min) seizures within 10–30 min of the injection. Rats that did not exhibit class three behavioral seizures within 1 h of pilocarpine injection were not included (21). A 75% survival was found at 3 days after SE, with no animals dying after this time. We also included a group of naïve control animals that did not receive lithium or saline injections and were kept with their mothers until weaning at P21. Naïve animal results were combined with the lithium-saline treatment results, as no difference was seen in bromodeoxyuridine (BrdU) or TUNEL labeling between these control groups.

BrdU injections

On P24, 26, and 28, the rats received two i.p. injections of 150 mg/kg BrdU (Sigma), 10 mg/ml in 0.007N NaOH, and 0.9 M NaCl, 12 h apart (2).

Immunohistochemistry

All animals received ketamine, 80 mg/kg, and/xylazine, 12 mg/kg (Sigma) i.p., before undergoing intracardiac perfusion of saline followed by 4% paraformaldehyde (EM Sciences, West Grove, PA, U.S.A.) in phosphate-buffered saline (PBS), pH 7.2. Animals were killed on P29, P34, and P50 for BrdU staining and P29 for TUNEL staining. After the brains were dissected, they were drop-fixed overnight in 4% paraformaldehyde at 4°C and then treated with sequential days of dehydration in 10%, 20%, and 30% sucrose. They were frozen on dry ice and cryostat sectioned at 10 μm for BrdU staining or 30 μm for TUNEL staining. Animals had the BrdU and TUNEL density assessed from sections that are morphologically similar to those found between 2.5 and 5.0 mm from bregma in the rat brain atlas (22).

BrdU staining

Sections were incubated with anti-BrdU antibody, 1:100 dilution (Roche Diagnostics, Indianapolis, IN, U.S.A.), developed by using an ABC alkaline phosphatase kit and NovaRed kit from Vector Laboratories (Vector Labs, Burlingame, CA, U.S.A.), and counterstained with cresyl-violet. The area of dentate was assessed by using a freehand drawing and quantified by using a Metamorph Imaging System (Downingtown, PA, U.S.A.). Manual profile cell counts for density of BrdU-labeled cells were carried out on a minimum of three matched sections per animal, separated by 60 μm, and were counted by an examiner blinded to the animal's treatment group. The total number of animals assessed for BrdU cell counts at each time was P29, n = 6, lithium/pilocarpine; n = 5, lithium/saline; n = 4 naïve; P34, n = 4, lithium/pilocarpine; n = 2, lithium/saline; n = 4, naive; P50, n = 4, lithium/pilocarpine; n = 2, lithium/saline; n = 2 naïve. BrdU-labeled cells both within and at the edge of the dentate/hilar border were included in density counts.

TUNEL staining

Animals used for TUNEL assay were killed on P29 and did not receive BrdU injections. The assay was carried out by using a previously published protocol (23). In brief, 30-μm sections were incubated with biotin-labeled deoxyuridine triphosphate (dUTP)16 (Roche Diagnostics) and terminal deoxytransferase (Roche) for 1 h. A Vector ABC peroxidase kit (Vector Labs), and 3–3'-diaminobenzadine (Sigma) was used for isolated TUNEL staining. The sections were counterstained with cresyl-violet and mounted with Vectashield containing DAPI-nuclear stain (Vector Labs) or Crystal Mount (Sigma). Blinded cell counts were performed by using the technique of manual area measurements and manual cell counts. The DG was identified as the neuron-dense region counterstained with cresyl-violet. The distance of the TUNEL-positive cells across the width of the dentate was measured by using a line drawn by hand, and its length measured by using the Metamorph Imaging System. The distance of each TUNEL-positive cell from the dentate/hilar border was measured, and at the same location, the entire width of the DG was measured. Three naïve, four lithium/saline, and six lithium/pilocarpine–treated animals were assessed for TUNEL cell density by using two to four sections per animal.

Co-staining

All primary antibodies, unless otherwise stated, were diluted in PBS with 2% horse serum and incubated overnight at 4°C. We used a rat anti-BrdU antibody, fluorescein tagged (1–25; Accurate Scientific, Westbury, NY, U.S.A.) and NeuN (1–100; Chemicon, Temecula, CA, U.S.A.) with a goat–anti-mouse TRITC (Sigma) secondary antibody. The TUNEL assay was carried out as mentioned earlier, substituting a streptavidin-Cy3 (1–500; Jackson Immunolabs, West Grove, PA, U.S.A.) secondary, and for the NeuN, a goat–anti-mouse fluorescein isothiocyanate (FITC; Sigma) secondary antibody was used. The tissue sections were incubated with PSA-NCAM antibody (1–200; Chemicon) in 5% dry milk at 4°C for 48 h. A goat–anti-mouse immunoglobulin M (IgM)-Cy-2 (Jackson Immunolabs) was used as a secondary antibody. Cell counts of co-labeled, NeuN-TUNEL, PSA-NCAM-TUNEL, and NeuN and BrdU were carried out by using a conventional fluorescent Zeiss microscope with superimposed digital pictures analyzed on Image Pro software; a limited number of co-labeled cells were confirmed by using a Leica laser confocal microscope (TUNEL, n = 8; and BrdU, n = 4).

RESULTS

The treatment paradigm for the pilocarpine and BrdU treatment is shown in Fig. 1. Animals underwent pilocarpine-induced SE at P20 and BrdU injection to label dividing cells 4, 6, and 8 days later. We chose to inject during these times to capture the period of robust DGC division after pilocarpine-induced seizures (14,17). We then killed animals over the next 3 weeks to assess the survival of the newborn DGCs during the latent period, but before the development of spontaneous seizures (24). The pilocarpine-treated animals killed the day after the final dose of BrdU, P29, had 5.5-fold more newborn cells per area of DG than did control animals (Fig. 2A and B; pilocarpine, mean, 4.53 × 10−4 cells/μm2 area of dentate; SE, 0.10 × 10-4; control, mean, 8.25 × 10−5, SE, 0.40 × 10-5, p < 0.0001, unpaired t test). The pilocarpine animals killed on P34 and P50 had 2.2- and 3.6-fold more BrdU-labeled cells than did controls (P34 pilocarpine, mean, 2.25 × 10-4cells/μm2; SE, 0.17 × 10-4; P34 control, mean, 1.01 × 10-4; SE, 0.01 × 10-4, p < 0.0001; P50 pilocarpine, mean, 7.26 × 10-5; SE, 0.96 × 10-5; P50 control, mean, 1.68 × 10-5; SE, 0.30 × 10-5; p < 0.0001). Thus at all times studied, substantially more newborn DGCs were found in the pilocarpine-treated animals than in controls.

Figure 1.

Experimental timelines. Top: Bromodeoxyuridine (BrdU) protocol. Bottom: TUNEL protocol.

Figure 2.

Figure 2.

An overall increase occurs in the number of newborn dentate granule cells in the weeks after status epilepticus (SE). A: Photomicrograph showing increased numbers of bromodeoxyuridine (BrdU)-labeled cells in the dentate after SE at P20. In both control (left), and pilocarpine treated (right) animals, the BrdU-labeled cells reside at the border of the hilus/dentate granule neuron layer. Both sections are from P29 animals, 9 days after SE (×10). B: Density measurements of BrdU-labeled cells in the dentate from control and pilocarpine-treated animals killed on P29, 34, or 50. All animals received two doses of BrdU per day on P24, 26, and 28. Significantly more BrdU-labeled cells are seen per area in the pilocarpine-treated animals than in the controls at all times (*p < 0.0005, ANOVA). A significant decrease occurs in the density of BrdU-labeled cells in the pilocarpine-treated animals at all times (#p < 0.0001). No difference was found in the density of BrdU-labeled cells in the control animals at P29 compared with P34; however, at P50, a significant decrease in the density of the BrdU-labeled cells was seen in the control animals compared with that on P29 and P34 (^p < 0.001).

Figure 2.

Figure 2.

An overall increase occurs in the number of newborn dentate granule cells in the weeks after status epilepticus (SE). A: Photomicrograph showing increased numbers of bromodeoxyuridine (BrdU)-labeled cells in the dentate after SE at P20. In both control (left), and pilocarpine treated (right) animals, the BrdU-labeled cells reside at the border of the hilus/dentate granule neuron layer. Both sections are from P29 animals, 9 days after SE (×10). B: Density measurements of BrdU-labeled cells in the dentate from control and pilocarpine-treated animals killed on P29, 34, or 50. All animals received two doses of BrdU per day on P24, 26, and 28. Significantly more BrdU-labeled cells are seen per area in the pilocarpine-treated animals than in the controls at all times (*p < 0.0005, ANOVA). A significant decrease occurs in the density of BrdU-labeled cells in the pilocarpine-treated animals at all times (#p < 0.0001). No difference was found in the density of BrdU-labeled cells in the control animals at P29 compared with P34; however, at P50, a significant decrease in the density of the BrdU-labeled cells was seen in the control animals compared with that on P29 and P34 (^p < 0.001).

BrdU-labeled cells in the pilocarpine-treated animals at P50 were 16% of those present at P29 (p < 0.001, ANOVA, with Tukey-Kramer for multiple comparisons). In control animals, differences were seen in the density of the BrdU-labeled cells in the control animals at P34 versus P29; however, at P50, 24% of the BrdU-labeled cells were found compared with those at P29 (p < 0.001). Thus over a 3-weeks period, in both pilocarpine and control animals, the density of BrdU-labeled cells in the dentate decreased.

To determine the fate of the newborn DGCs, we performed double-labeled immunohistochemistry of the remaining BrdU-labeled cells and NeuN, a marker of mature DGCs at P50 (Fig. 4D). In control animals, 45 (89%) of 51 of the BrdU-labeled cells co-expressed NeuN, similar to the 57 (95%) of 60 co-labeled in the pilocarpine-treated animals; thus SE does not appear to alter the neuronal fate of the surviving BrdU-labeled cells.

Figure 4.

Fate of dentate granule cells (DGCs) born after status epilepticus (SE) in postnatal development. A: The newly-born DGCs are susceptible to dying as immature neurons. Two arrows, TUNEL-positive cells at the hilar/dentate border in a SE-treated animal at P29. Arrowhead, A TUNEL-positive cell at the dentate/inner molecular layer border. The nuclei of all cells are stained with Dapi, showing the tightly packed dentate granule layer (×10). B: The immature neuronal marker, PSA-NCAM (green, Cy2) stains the surface of immature dentate granule neurons at the hilar border, a region in which most of the TUNEL-positive cells (red, Cy3) were found (20X-P29 pilocarpine-treated animal). C: The majority of TUNEL-positive cells (red, Cy3) lay along the hilar edge of the dentate and below the mature NeuN-expressing cell layer (green, FITC) (40X- P29 pilocarpine-treated animal). D: P50, 30 days after SE, in both the control (not shown) and pilocarpine-treated animals (shown), the majority of BrdU-positive (green fluorescein) cells coexpress NeuN (red, TRITC), a mature neuronal marker, and appear yellow. Arrows, the two green BrdU-labeled cells, one in the hilus and one at the dentate hilar border that do not coexpress NeuN (×20). H, the hilar region in each of the photomicrographs.

The loss of BrdU-labeled cells in the dentate could be due to death of the newborn cells or dilution of BrdU label by continued cell division. We looked for nicked DNA in dentate cells as a marker of cell death by using the TUNEL assay and found TUNEL-positive cells in both control and pilocarpine-treated animals. The pilocarpine-treated animals had 3.8-fold more TUNEL-positive cells than the control animals (Fig. 3A; pilocarpine mean, 3.05 × 10−5 cells/μm2 area of dentate; SE, 0.55 × 10-5; control, mean, 8.08 × 10−6; SE, 1.2 × 10-6; p < 0.0001, unpaired t test). The cells were distributed unevenly within the DGC layer. The majority of TUNEL-positive cells in both the control and pilocarpine-treated animals were located at the dentate/hilar border as seen in Fig. 3B, where the distance of TUNEL-positive cells as a ratio of the total width of the DGC layer is graphed (see also Fig. 4A, arrows). Note that a subset of TUNEL-positive cells in the pilocarpine-treated animals [9% or (11 of 117)] is located at the dentate inner-molecular border (arrowhead, Fig. 4A).

Figure 3.

Figure 3.

An increased number of TUNEL-positive neurons are seen in the dentate of animals subjected to seizures at P20. A: In the dentate gyrus of the pilocarpine-treated animals, 2.6-fold more TUNEL-positive cells are found than in controls 9 days after SE. (**p < 0.0001, t test). B: The TUNEL-positive cells within the dentate are concentrated at the hilar/granular layer border in both the seizure and control animals. The relative positions of the TUNEL-positive cells across the width of the dentate are graphed, with the dentate/hilar border labeled zero, and the dentate/inner-molecular layer (IML) labeled one. Inset 3B: The width of the dentate granule cell layer (W) was measured from the hilus to the IML, and then the distance of each TUNEL-positive cell was measured from the hilus (D). D/W is the relative distance graphed (×40). Although the vast majority of cells in both the SE and control animals are at the hilar/granular border, a small population of TUNEL-positive cells is seen at the granular/IML in the SE-treated animals.

Figure 3.

Figure 3.

An increased number of TUNEL-positive neurons are seen in the dentate of animals subjected to seizures at P20. A: In the dentate gyrus of the pilocarpine-treated animals, 2.6-fold more TUNEL-positive cells are found than in controls 9 days after SE. (**p < 0.0001, t test). B: The TUNEL-positive cells within the dentate are concentrated at the hilar/granular layer border in both the seizure and control animals. The relative positions of the TUNEL-positive cells across the width of the dentate are graphed, with the dentate/hilar border labeled zero, and the dentate/inner-molecular layer (IML) labeled one. Inset 3B: The width of the dentate granule cell layer (W) was measured from the hilus to the IML, and then the distance of each TUNEL-positive cell was measured from the hilus (D). D/W is the relative distance graphed (×40). Although the vast majority of cells in both the SE and control animals are at the hilar/granular border, a small population of TUNEL-positive cells is seen at the granular/IML in the SE-treated animals.

To confirm that the newborn cells are preferentially dying and to determine the phenotype of the cells undergoing DNA degradation, we did TUNEL staining and immunohistochemistry for the mature and immature DGC markers, NeuN and PSA-NCAM in the pilocarpine-treated animals. The majority, 27 (79%) of 34 TUNEL-positive cells reside in the PSA-NCAM–expressing layer of immature DGCs (Fig. 4B). This layer is below the mature NeuN-expressing DGCs, and only five (21%) of 23 TUNEL-positive cells co-express NeuN (Fig. 4C). Thus DGCs expressing the immature neuronal marker PSA-NCAM are most susceptible to dying after SE.

DISCUSSION

Dramatic increase in the birth of DGCs, seen after an episode of pilocarpine-induced SE, has been demonstrated by others in pilocarpine, kainate, and kindling models, in both adult and immature animals (14,15,17,25). We previously found a ∼20% increase in DGC numbers in adult animals that developed epilepsy after SE at P20, suggesting a possible persistence of DGCs born in the weeks after SE (24). Ekdahl et al. (20) reported an 85% loss of BrdU-labeled cells over the month after SE in adult animals. These findings raised the question; how many DGCs born after SE persist in the long term and contribute to dentate function? The current report is the first to quantify the survival and the resulting phenotype of DGCs born after SE in an immature animal and to compare the survival of DGCs with that of littermate controls. We find that the majority of new DGCs born after SE in the developing brain do not persist over the ensuing weeks. Despite this decrease, threefold more newborn DGCs remain after SE compared with those in control animals, and the surviving neurons develop a mature neuronal phenotype.

Immature DGCs are more susceptible to long-term potentiation (LTP) and paired-pulse facilitation, suggesting that increased numbers of immature DGCs seen after SE may influence dentate synaptic plasticity (4,5). The physiology, morphology, and neurotransmitter-receptor expression of the DGCs is altered after pilocarpine-induced SE, and these changes may contribute to the development of spontaneous seizures (26–28). GABA physiology and GABA-receptor subunit expression is heterogeneous among DGCs, and at least two distinct populations of DGCs have been identified from humans with temporal lobe epilepsy (29,30). Here we have shown that in the weeks after SE, newborn neurons are increased in number and might be candidates for one of the physiologically distinct subpopulations of DGCs identified in the previously mentioned studies.

The role of dentate neurogenesis in epileptogenesis is not well understood. Some lines of evidence suggest that newborn DGCs may contribute to epileptogenesis. Scharfman et al. (31) showed that a population of DGCs, possibly newborn, migrate abnormally and contribute to a hyperexcitable population of hilar DGCs. However, blockade of neurogenesis after SE did not inhibit mossy fiber sprouting, suggesting that mature neurons are capable of aberrant sprouting (32). Further, increased DGC neurogenesis due to environmental enrichment has been correlated with a decrease in seizure susceptibility, suggesting that neurogenesis might be protective for some seizures (33). The increased numbers of newborn DGC found in the pilocarpine-treated animals after SE but before the development of spontaneous seizures warrants further study to determine the role of these cells in epileptogenesis.

We believe that much of the loss of BrdU-labeled cells is due to death of immature DGCs, as we find an increase in TUNEL-positive cells at P29 in the pilocarpine-treated animals concentrated in the dentate/hilar border, the region of immature DGCs. How long a dying DGC remains TUNEL positive is not known, making it difficult to show a direct correlation between decreased numbers of BrdU-labeled DGCs and TUNEL staining density. We do show that most of the TUNEL-positive cells are residing in the immature PSA-NCAM–expressing layer and do not coexpress the mature marker NeuN, suggesting that immature neurons are particularly susceptible to dying after SE. Several other groups have seen evidence either by TUNEL, Fluoro-Jade, or morphology of increased numbers of dying cells at the dentate/hilar region after pilocarpine-induced SE (19,20), but also see (34).

An alternative cause for the loss of BrdU-labeled cells is dilution of the label to imperceptible levels after multiple cell divisions. As we did not see any decrease in staining intensity of the labeled cells at P50, it would suggest that if dilution was responsible, it was present in only a subpopulation of cells. Although we did not perform long-term EEG monitoring of the animals, none had witnessed spontaneous seizures or handling-induced seizures during the course of the study (until P50). This finding is consistent with our prior monitoring studies of P20 pilocarpine-treated animals, which found that spontaneous seizures began on average at P65, with the earliest onset of seizures occurring at P54 (24). Thus spontaneous seizures should not have influenced the number of BrdU- or TUNEL-labeled cells at the times studied. Our cell-counting data do not compensate for increased dispersion of DGCs seen in temporal lobe epilepsy, which may have caused us to underestimate the density of BrdU- and TUNEL-stained cells in the pilocarpine-treated animals (35).

The presence of the small subset of TUNEL-positive cells at the dentate/inner-molecular border in the pilocarpine-treated animals suggests that a subset of mature DGCs are susceptible to SE-induced injury. Why these cells are uniquely susceptible is unclear, but their concentration at the dentate/inner-molecular border suggests that regional factors within the mature dentate control DGC susceptibility to SE-provoked death.

As the number of DGCs does not continue to increase throughout life, the birth and death of DGCs appear to be coordinated, although the molecular mechanisms are not well understood. Although the majority of dying cells in the pilocarpine-treated animals appear to be immature DGCs, replacement of the small increased number of dying mature DGCs after SE may be related to the increased survival of newborn DGCs after SE. Further studies are needed to understand how SE regulates the fate of DGCs and to determine the importance of a threefold increase in immature DGCs that persist during the latent period after SE.

Acknowledgments

ACKNOWLEDGMENT:  Support for this study was provided by the NINDS to A.B.K., and by the Child Neurology Foundation to A.B.K. and B.E.P.

Ancillary