Hippocampal functions in memory and emotion are sensitive to stress. Stress and corticosteroids modulate neuronal activity and plasticity in the hippocampus and regulate its contributions to behavior. One feature of the hippocampus that is highly sensitive to stress is adult neurogenesis. Stress alters the proliferation and survival of adult-born neurons in the hippocampal dentate gyrus (DG). Furthermore, adult-born neurons modulate the behavioral response to stress. How newborn neurons respond to stress at a cellular level still remains relatively unexplored, but is important for understanding how they regulate behavioral responses. Here we examined whether two experiences that differ in the degree of stress, novel context exposure and physical restraint, alter levels of immediate-early genes (IEGs), mineralocorticoid receptors (MR) and glucocorticoid receptors (GR) in immature adult-born neurons and putative older neurons in the DG. We found that experience, particularly restraint stress, reduces levels of zif268 and MR in immature neurons but not older neurons. In contrast, both context exposure and restraint stress increased GR levels in immature neurons and mature neurons. These divergent cellular responses suggest that these two neuronal populations may have distinct functions in regulating the stress response.

The dentate gyrus subregion of the hippocampus is comprised of a heterogeneous population of cells due to ongoing neurogenesis throughout adulthood. Adult-born neurons transition through defined developmental stages and, during a period of immaturity, they have enhanced synaptic plasticity and different patterns of activity compared to more mature neurons^{[1][2][3][4][5][6]}. One commonly-used measure of activity is the immediate-early gene (IEG) response to behavioral or physiological stimuli. IEGs are rapidly-induced following synaptic activity and promote plasticity that is required to store and process information^[7]. Adult-born neurons show increased expression of IEGs as they mature and integrate into hippocampal circuits. For example, adult-born neurons gradually increase Fos expression until they reach similar rates as mature neurons^[8]. However, zif268 expression shows a pronounced peak during cellular immaturity^[2] and elevated expression in older adult-born neurons following learning tasks, compared to developmentally-born mature neurons^[9].

The factors that recruit adult-born neurons during behavior remain unclear but may depend on stress. In response to a stressor, glucocorticoids are released and bind to GRs and MRs, which act via genomic and non-genomic mechanisms to alter neuronal function^[10][11]. Notably, even acute stressors are capable of producing long-lasting changes in memory and behavior^[12][13][14]. It is likely that glucocorticoid-dependent signalling regulates adult-born neurons since they express GRs and MRs^[15][16], show enhanced survival in response to chronic stress^[17], and require GRs for their functional maturation^[18]. Since corticosteroid receptor levels vary across hippocampal subregions at baseline and in response to corticosterone treatment^[19][20], this raises the possibility that MR and GR levels may also differ between young and old neurons within the DG and endow them with distinct cellular functions. This could arise from many possible divergent transcriptional targets of MRs vs GRs^[21], distinct functions in LTP where MRs promote post-stress LTP in the ventral hippocampus and GRs impair post-stress LTP in the dorsal hippocampus^[22], or different roles in spine formation/turnover^[23][24]. Differences in MR levels might also be expected to impact probability of synaptic glutamate release^[25] and changes in GR levels could specifically impact IEG expression and memory consolidation^{[26][27][28][29][30]}.

Behavioral evidence also implicates adult-born neurons in the stress response. Preventing the addition of adult-born neurons leads to altered emotional responses and HPA output in response to stress^[31][32]. Furthermore, new neuron functions in context fear conditioning depend on the number of footshocks, suggesting stress-dependent recruitment^[33]. However, it is unclear how stress impacts the IEG response in newborn neurons. In some instances, stressful experiences lead to Fos expression in immature neurons^[34][35]. In other cases, Fos expression appears selective to putative older neurons^[36]. Interestingly, initial reductions in zif268 in immature DG neurons after water maze testing dissipate after extended training^[4][6]. This raises the question of whether the stress of a new, challenging experience might ,in fact, reduce IEG expression in newborn neurons. In short, further characterization is needed to clarify how cellular responses in the DG contribute to the stress-related changes in behavior.

To determine whether stressful experiences alter the expression of IEGs and nuclear levels of stress hormone receptors in immature adult-born neurons and older neurons in the DG.

Young adult male rats were either left in their home cage (no stress), exposed to a novel context (mild stressor), or subjected to acute restraint stress (a more severe stressor) to examine how stressors alter nuclear immunostaining for the immediate-early genes zif268 and Fos, and the corticosteroid receptors MR and GR. Corticosterone levels tended to be higher in the restrained animals but differences were not statistically significant due to the return to baseline over the post-stress interval (controls: 114±16 ng/ml, context: 82±27 ng/ml, restraint: 186±76 ng/ml).

All gene products were observed in the DG of all groups, but levels were graded and patterns varied significantly across groups. The majority of DG neurons, including immature BrdU+ neurons, displayed graded levels of MR and GR, with the exception of cells bordering the hilus, where precursor cells reside (Fig. 1B, 1C). This is consistent with previous reports that both MR and GR are highly expressed in the DG except in very immature neurons and precursor cells^[16][15]. Generally, a small proportion of DG neurons expressed graded levels of zif268 and Fos. While levels of the IEGs zif268 and Fos likely reflect recent transcriptional history, MR and GR levels could reflect both expression and translocation^[19][37]. We, therefore, use the generic term “signal” when referring to immunostaining levels for each antigen.

To capture graded, experience-dependent changes we measured the immunostaining signals of each immature BrdU+ adult-born neuron and each mature non-BrdU+ cell, located in the superficial granule cell layer. In BrdU+ cells, zif268 signal was greatest in cells from control rats; it was reduced by both behavioral manipulations and was lowest in BrdU+ cells after restraint (44% reduction compared to controls). These data are consistent with previous findings that zif268 expression in immature neurons is reduced following water maze training and exploration of a novel environment^[4][6]. That zif268 reduction was greatest in restrained rats suggests that stress may be responsible. In contrast to the adult-born neuronal population, behavioral manipulations did not alter zif268 signal in mature, non-BrdU+ cells (Fig. 1D). In both BrdU+ and non-BrdU+ cells, Fos signal was not altered by behavioral manipulations (Fig. 1E).

The lack of difference in zif268 and Fos signal in mature non-BrdU+ cells might seem at odds with previous studies showing experience-dependent increases in IEGs^{[34][8][4][6]}. This apparent discrepancy is likely due to methodological differences in quantification. Since there is only sparse activation of DG neurons during behavior, it is possible that only a small proportion of DG neurons show IEG upregulation. Indeed, in the scatterplots it is clear that a handful of non-BrdU+ cells show high levels of Fos but only in the context-exposed and restraint groups. Thus, IEG upregulation in a small proportion of neurons is masked by a lack of upregulation in the majority of DG neurons. A similar effect may be occurring with zif268 in non-BrdU+ cells. In contrast, zif268 appears to be reduced in a large proportion of immature BrdU neurons. The reason for the different patterns is unclear but, since zif268 levels are highest in immature neurons from control rats, this may suggest that immature neurons are active during "offline" processes and actively suppressed during stressful experiences.

In BrdU+ cells, MR signal was reduced following experience in a stepwise fashion, with a significant ~20% reduction following restraint stress relative to controls. In non-BrdU+ cells, neither context exposure nor restraint stress altered MR signal (Fig. 1F). Finally, GR signal was increased by both context exposure and restraint stress, in both BrdU+ and non-BrdU+ cells (Fig. 1G). While the experience-dependent increase in GR signal in adult-born cells was statistically significant, the effect was small in comparison to non-BrdU+ cells (10% vs 35%).

Since MRs and GRs both regulate cellular activity and plasticity, including the expression of IEGs, we examined co-expression of IEGs and corticosteroid receptors at the single cell level. Across all groups and both cell types, there were significant correlations between zif268 and MR, and between Fos and GR (Suppl. Material). Due to antibody and fluorophore limitations, we did not examine zif268-GR and Fos-MR relationships. Nonetheless, these data provide preliminary support for the possibility that IEGs and corticosteroid receptors cooperate in both immature and mature DG neurons cells to promote plasticity or other physiological processes.

Behavioral experience reduces zif268 and MR levels in immature adult-born DG neurons, but not older neurons. That this reduction is greatest in restrained rats suggests that stress may be an important regulator of cellular functions of immature DG neurons. Immature and mature DG neurons both show experience-dependent increases in GR levels, though this effect may be weaker in immature neurons. Collectively, immature adult-born neurons display distinct cellular responses to stressful experiences.

While our IEG data indicates that zif268 changes are widespread amongst immature neurons, it is possible that a sparse population of neurons showed dramatic increases in IEG expression that were obscured by pooling measurements from a large number of DG neurons (as discussed, above). Larger sample sizes would permit analyses of strong vs weak IEG-expressing populations. Since IEG expression, MR/GR expression and translocation vary with time after an experience, it is also possible that we failed to capture stress-related changes in these signals by sampling at only a single timepoint. Finally, our data do not distinguish whether changes in MR and GR signal are due to changes in receptor expression vs. localization; mRNA analyses or measurements of cytoplasmic vs nuclear signal could distinguish these possibilities in the future.

While we observed correlations between IEGs and MR/GR levels, it is unclear whether MRs and GRs directly regulate IEG expression, such as the stress-related decrease in zif268 expression. To test a causal relationship, future studies could manipulate MRs and GRs and examine effects on IEG expression.

Subjects and treatments

Male Long Evans rats were generated in-house, doubly-housed on a 12 h light-dark schedule and were injected with the mitotic marker BrdU to label adult-born neurons at 2 months of age (200 mg/kg/injection). Injections were given on days 1, 3, 5, 8, 10, 12. Rats were handled for 5 min/day on days 19, 22–26 to familiarize them with the experimenter. On day 29 rats were divided into 3 groups that were subjected to different treatments: the Control Group (n=2) was perfused with 4% paraformaldehyde directly from their home cage with no manipulations; the Context Group (n=4) was exposed to a novel open field environment in an unfamiliar room for 30 min, transferred to a new empty cage in another unfamiliar room for 30 min and then perfused; the Restraint Group (n=4) was subjected to 30 min of immobilization restraint stress in plastic tubes in an unfamiliar room for 30 min, transferred to a new empty cage in another unfamiliar room for 30 min and then perfused. Testing on the last day occurred 7–9 h into the light phase. Blood was rapidly collected immediately after behavioral manipulations and prior to perfusion for subsequent corticosterone measurements by radioimmunoassay. After perfusion, brains were collected, post-fixed for 48 h in 4% paraformaldehyde, cryoprotected in glycerol, and sectioned coronally at 40 µm with a freezing sliding microtome.

Immunohistochemistry

Immunohistochemical analyses of neuronal phenotype were performed on free-floating sections with fluorescent detection. Sections were treated with 2N HCl for 30 min, incubated at 4°C for 3 days in PBS with 10% triton-x, 3% horse serum and combinations of the following antibodies: rat anti-BrdU (1:200; AbD Serotec, OBT0030G), rabbit anti-zif268 (1:1000; Santa Cruz, sc-189), goat anti-c-fos (1:250; Santa Cruz, sc-52G), rabbit anti-GR H-300 (1:200, Santa Cruz, sc-8992, validated by Sarabdjitsingh et al.^[38]), mouse anti-MR (1:200, Developmental Studies Hybridoma Bank, rMR1-18 1D5, validated by Gomez-Sanchez et al.^[39]). Due to limitations in the available numbers of antibody species and fluorophore channels, separate sections were stained for BrdU/zif268/MR and BrdU/Fos/GR. Visualization was performed with Alexa488/555/647-conjugated donkey secondary antibodies (Invitrogen/Thermofisher) diluted 1:250 in PBS with 10% triton-x and 3% horse serum for 60 min at room temperature. Sections were counterstained with DAPI, mounted onto slides and coverslipped with PVA-DABCO.

Sampling and Image Analysis

Sections containing dorsal hippocampus were examined using sequential scanning on a confocal microscope (Leica SP8) and a 40X oil immersion objective (NA 1.3). Approximately 50 BrdU+ cells/rat were analyzed for zif268 and MR, and ~30 BrdU+ cells/rat were examined for Fos and GR. Mature and developmentally-born neurons are located in the superficial portion of the granule cell layer^[40][41]. To analyze these cells we examined 50 BrdU-DAPI+ cells (we refer to as "non-BrdU+ cells") for zif268, Fos, MR and GR in each animal. BrdU+ cells were sampled across the full mediolateral extent of the suprapyramidal blade since they were relatively sparse and therefore required sampling across entire sections. BrdU-DAPI+ cells were sampled from the middle of the suprapyramidal blade. Cells were selected from the BrdU and/or DAPI channels, to avoid biased selection of cells based on marker expression. For MR and GR, we reasoned that, since both are widely expressed in the DG, merely counting a cell as positive or negative is unlikely to reveal changes in nuclear levels; the majority will remain MR or GR-positive and therefore to capture experience-dependent changes it is necessary to quantify the signal, as others have shown^[19]. Likewise, while IEG levels are typically measured by calculating the percentage of cells that express above some arbitrary (and often undefined) threshold, this binary approach also cannot effectively reveal variability in expression levels that may occur. To quantify immunostaining levels in individual cells an intensity measure (“signal”) for each marker in each cell was determined: the average fluorescence intensity in the middle z-plane of the cell nucleus was measured and divided by the intensity of a nearby region in the subgranular zone or hilus that contained no DAPI+ nuclei, using ImageJ.

Statistical analyses

Since immunostaining signal intensities for the various markers were not normally distributed, the effect of treatment on cellular signal intensity was analyzed by separate non-parametric Kruskal-Wallis tests for BrdU+ and non-BrdU+ cells (cells pooled across animals). Specific treatment effects were examined with Dunn’s test, corrected for multiple comparisons. Significance was set at p=0.05.

This work was supported by startup funds from the Psychology Department & Faculty of Arts at the University of British Columbia.

The authors declare no conflicts of interest.

All procedures were approved by the Animal Care Committee at the University of British Columbia and conducted in accordance with the Canadian Council on Animal Care guidelines regarding humane and ethical treatment of animals.

No fraudulence is committed in performing these experiments or during processing of the data. We understand that in the case of fraudulence, the study can be retracted by ScienceMatters.