A behavioral task with more opportunities for memory acquisition promotes the survival of new neurons in the adult dentate gyrus

Abstract

It has been suggested that the dentate gyrus, particularly its new neurons generated via adult neurogenesis, is involved in memory acquisition and recall. Here, we trained rats in two types of Morris water maze tasks that are differentially associated with these two memory processes, and examined whether new neurons are differently affected by the two tasks performed during the second week of neuronal birth. Our results indicate that the task involving more opportunities to acquire new information better supports the survival of new neurons. Further, we assessed whether the two tasks differentially induce the expression of an immediate early gene, Zif268, which is known to be induced by neuronal activation. While the two tasks differentially induce Zif268 expression in the dentate gyrus, the proportions of new neurons activated were similar between the two tasks. Thus, we conclude that while the two tasks differentially activate the dentate gyrus, the task involving more opportunities for memory acquisition during the second week of the birth of new neurons better promotes the survival of the new neurons.

Figure 1

Experimental design. (A) Experimental timeline. (B) Reference memory and working memory versions of the Morris water maze task. The platform location was kept constant across all training days for the reference memory group, whereas the location was changed every day for the working memory group. (C) Within training days, the platform location was kept constant.

Figure 2

Performance of the reference memory group in the water maze task. (A) Latencies to reach the platform in individual trials during days 1–7 of the training period. (B) Latencies to reach the platform in individual trials on the retraining day. (C) Percentages of time spent in each quadrant during trials 1–4 on the retraining day. The platform was in the target quadrant. *p < 0.05, ***p < 0.005 in a one-way ANOVA. All data are shown as the mean ± s.e.m.

Figure 3

Performance of the working memory group in the water maze task. (A) Latencies to reach the platform in individual trials during days 1–7 of the training period. (B) Latencies to reach the platform in trials 1–4 averaged over the 7-day training period. *p < 0.05 in post hoc LSD test. (C) Latencies to reach the platform in individual trials on the retraining day. (D) Percentages of time spent in each quadrant in trials 1–4 on the retraining day. The platform was in the target quadrant. ***p < 0.005 in a one-way ANOVA. All data are shown as the mean ± s.e.m.

Figure 4

Immunofluorescence detection of BrdU-, NeuN-, and Zif268-positive cells. (A) Representative epifluorescence images from four sections used for the analyses. The sections were immunostained for NeuN. One in every 12 of the 40-µm-thick sections was selected, and the selected sections were designated #1–4. Scale bar: 500 µm. (B) An example of confocal image tiles, consisting of 12 images (512 × 512 pixels each) covering the whole granule cell layer. Blue: NeuN, red: Zif268, green: BrdU. Scale bar: 200 µm. (C) A confocal image showing BrdU/NeuN/Zif268-triple-positive cells (arrowheads). Scale bar: 30 µm.

Figure 5

Density of BrdU-positive cells. (A,C,E) Overall densities of BrdU-positive (A); BrdU/NeuN-double-positive (C); and BrdU-positive, NeuN-negative (E) cells from all sections. **p < 0.01, ***p < 0.005 in a two-tailed t-test. (B,D,F) Mean densities in individual sections of BrdU-positive (B); BrdU/NeuN-double-positive (D); and BrdU-positive, NeuN-negative (F) cells. * in B: p < 0.05 in post hoc two-tailed t-test. * in D: p < 0.05 in Group effect, repeated measures ANOVA. All data are shown as the mean ± s.e.m.

Figure 6

Correlation of reference memory task performance with the densities of BrdU-positive and BrdU/NeuN-double-positive cells. (A) Relationship between BrdU-positive cell density and summed latency over the 7-day training period. Top: reference memory group; bottom: working memory group. Each dot represents an individual rat. r: Pearson’s correlation coefficient; p: p-value for correlation (same for D and E). (B) Summed latency over the 7-day training period. Data are shown as the mean ± s.e.m. (C) Summed latency for all training days. Each line represents an individual rat. Orange arrows indicate the initial improvement of performance of individual rats. Rats are numbered/color coded according to the rank of their summed latency over the 7-day training period. (D) Correlation between summed latency on day 4 and the densities of BrdU-positive (top) and BrdU/NeuN-double-positive (bottom) cells. (E) Correlation between the day before a large reduction in latency occurred and the densities of BrdU-positive (top) and BrdU/NeuN-double-positive (bottom) cells.

Figure 7

Zif268 expression in BrdU/NeuN-double-positive cells. (A) Mean overall percentages of BrdU/NeuN-double-positive cells expressing Zif268 from all sections. (B) Mean percentages of BrdU/NeuN-double-positive cells expressing Zif268 in the individual sections. All data are shown as the mean ± s.e.m.

Figure 8

Zif268 expression in NeuN-positive cells. (A) Mean overall percentages of NeuN-positive cells expressing Zif268 from all sections. (B) Mean percentages of BrdU/NeuN-double-positive cells expressing Zif268 in the individual sections. *p < 0.05 in a post hoc two-tailed t-test. All data are shown as the mean ± s.e.m.