Dek Loss Induces Sex-Dependent, Task-Specific Cognitive Deficits and Reprograms the Hippocampal Transcriptome in Mice

Abstract

Cognitive decline with aging, and some neurodegenerative conditions like Alzheimer's disease, disproportionately affects females yet few mechanisms beyond steroid hormone signaling fully explain this sex-specific vulnerability. The chromatin-remodeling DEK protein, upregulated by estrogen and progesterone and broadly expressed in the brain, including the hippocampus, may be one such mechanism. We have previously linked DEK loss with indices of neuronal dysfunction, including increased DNA damage, impaired neurite development, and apoptosis, suggesting a potential neuroprotective role. Here, we investigated the molecular and behavioral consequences of Dek loss in vivo. Female Dek constitutive knockout (cKO) mice exhibited a sex-specific behavioral phenotype, with impairments in sensorimotor gating, as measured by pre-pulse inhibition, and in reversal learning in the Morris Water Maze. These findings are suggestive of deficits in pre-attentive sensory processing and cognitive flexibility, respectively. Notably, these cognitive deficits were not observed in male Dek cKO mice and were not attributable to differences in general learning ability, locomotor activity, or anxiety-like behavior. The absence of impairment in object recognition and conditioned fear learning and memory in females suggests that the effects of DEK loss are task-specific and likely brain region-specific. Transcriptomic analysis of hippocampal tissue revealed differentially expressed genes related to inflammation, metabolism, and neuropeptide signaling in all Dek-cKO mice, along with a distinct female-specific transcriptomic profile indicative of impaired neuronal function. Combined, we report for the first time that DEK supports certain aspects of cognitive function, particularly in females. These data may be relevant for understanding sex differences in some cognitive disorders.

Keywords: DEK; cognitive dysfunction; hippocampus; memory; sex differences.

Figure 1.. Loss of DEK expression in the brain causes sex-specific memory impairment using the novel object recognition test in adult mice.

(A) Immunohistochemical staining for DEK reveals prominent staining in the dentate gyrus in wild-type mouse brain while expression is absent in brain tissue from Dek conditional knockout (Dek cKO) mice. (B) All mice, regardless of sex or genotype, performed similarly in the open field test for locomotor activity. (C-D). The novel object recognition (NOR) task, at the 1 h retention interval for females, there was a trend of genotype (t(16.689) = 1.77, p = 0.0949) (C), but no difference in males (panel D). █p<0.1 All mice were adults; N=12 WT females, 10 Dek cKO females, 14 WT males, 7 Dek cKO males.

Figure 2.. Dek loss causes impaired spatial memory in females, but not males.

There were no differences in performance during cued trials in the Morris water maze on either latency (A) or speed (B). On hidden platform acquisition trials for females, there was a genotype × day interaction for latency (F(4,52.7) = 2.87, p = 0.0319). On the probe trial 24 h after the last acquisition trial, there were no effects on entries for females and males. For average distance to the former platform site there was no effect in females, but there was a trend in cKO males to be further away (t(19) = 1.85, p = 0.0793).. On reversal trials for females for latency there was a genotype × day interaction (F(4,61.2) = 3.54, p = 0.0115; female cKO mice had longer latencies than WT females on day 5 with a trend on day 3 (panel C). There were no differences in males. Example path tracings are shown in panel D. On the reversal probe trial for females, there was an effect of genotype on platform zone entries (t(11.384) = −2.82, p = 0.0161), but no effect in males. For females and males there were no differences in average distance from the former platform location. Group sizes: N=12 WT females, 10 Dek cKO females, 14 WT males, 7 Dek cKO males.

Figure 3.. Dek loss results in sex-specific impairment in tactile startle response but not acoustic startle responses.

(A,C) There is no difference in acoustic startle V_max. Acoustic stimulus is a 20 ms, 120 dB (SPL) mixed frequency sound burst with 1.5 ms rise time. (B,D) There are no statistically significant differences between Dek cKO and WT mice for V_max on acoustic prepulse inhibition trials. Of note, there is a trend for increased V_max in Dek cKO males and females compared with WT males and females. (E,G) Tactile startle showed no differences between WT and Dek cKO males on either day 1 or day 2. However, there were more pronounced and consistent increases in startle in female Dek cKO mice compared with WT females, differences supported by an effect of genotype (F(1,22.7) = 10.48, p = 0.0037) in which the cKO females were hyperreactive. On Day 2, females again had a significant genotype effect (F(1,21.4) = 17.19, p = 0.0004) (panel G, right). The tactile stimulus was a 20 ms 60 psi air puff, delivered to the dorsal surface of the mouse through a tube inserted in the top of the animal holder. (F, H) There were no differences between WT and Dek cKO males for tactile PPI. Compared with WT females, female Dek cKO showed reduced prepulse induced suppression (panel (H): F(1, 20.6) = 10.69, p =0.0037) that was not observed in males (panel F). N=12 WT females, 10 Dek cKO females, 14 WT males, 7 Dek cKO males

Figure 4.. Dek loss leads to transcriptional deregulation of genes across multiple cellular processes.

(A) Dek mRNA levels in the hippocampus of WT and Dek cKO mice using qRT-PCR shows loss of Dek expression in the Dek cKO mice. N=6 per group. (B) Principal component analysis (PCA) shows the separation of groups based on sex and genotype due to variation in the transcriptome using hippocampal mRNA. The ellipses represent 95% confidence intervals. N=3/genotype/sex. (C) The volcano plot depicts genes upregulated (red) and down-regulated (blue) in Dek cKO hippocampus compared with WT mice. (D-E) Differentially expressed genes from Dek cKO mice compared with WT mice were utilized to create cellular networks with Cytoscape, with cellular functions as central nodes and associated genes attached. Males and females were pooled for each genotype to find common deregulated genes. (D) Blue circles are down-regulated. Relevant down-regulated processes are nucleic acid binding, behavior, calcium homeostasis, and biosynthetic processes. and (E) Red circles are up-regulated processes. Up-regulated processes included immune response nodes, suggestive of neuro-inflammation.

Figure 5.. Dek loss leads to sex-specific transcriptomes in the mouse hippocampus.

(A, C) The volcano plot depicts genes upregulated (red) and down-regulated (blue) in male (A) and female (C) Dek cKO hippocampus compared with WT mice. (B, D) Differentially expressed genes from Dek cKO mice compared with WT mice were utilized to create cellular networks with Cytoscape, with molecular functions as central nodes and associated genes attached. Molecular functions in males (B) were distinct from the molecular functions of DEGs in females (D). Blue circles are down-regulated and red are up-regulated processes. (E) A Venn diagram depicts the overlap of differentially expressed genes based on sex and Dek genotype. (F) The 76 shared DEGs from male and female Dek cKO mice were analyzed for gene ontologies using Enrichr (https://maayanlab.cloud/Enrichr/) N=3/sex/genotype

Figure 6.. Identification of sex-specific signaling pathways in the hippocampus of WT mice, and identification of a unique transcriptional profile for female Dek cKO mice that highlights differences in dementia-relevant pathways.

(A) A Heatmap depicts regulated pathways from the Pathway Commons database that are different in pairwise comparisons of all four sex X genotype groups, including sex differences in WT mice. Differentially expressed genes were identified using AltAnalyze with Kallisto and differentially expressed genes assess for pathway analysis using GO-Elite. The first column identifies sex-specific differences of the transcriptome from the hippocampal tissue of WT C57Bl/6 mice (Female WT vs. Male WT). Key upregulated pathways in female hippocampus from WT include DNA repair and stress pathways like ATM, ATR, and p53, in addition to mitogenic, pro-proliferative pathways like mTor, EGFR, FGF, and E2F transcription factor targets. In comparison, as shown in the second column, female Dek cKO mice compared with Dek WT females have down-regulation of these same pro-proliferative and DNA damage/stress pathways. This suggests that Dek cKO female mice have lost sex-specific differences and neuroprotective pathways. (B) Differentially expressed genes from female DekcKO mice compared with female WT were utilized to create cellular networks with Cytoscape, with cellular functions as central nodes and associated genes attached. Blue circles are down-regulated and red are up-regulated processes. Relevant down-regulated processes are limited but include genes relevant to dendrites and synaptic membranes/synapses including the glutamate receptor Grik2and protein kinase C-gamma (Prkcg). Up-regulated processes included extracellular factors like paracrine signaling proteins, collagen, and basement membrane. Of interest, the sex-determining factor Rspo1 is upregulated as well as several inflammation-relevant signaling factors like interleukins and pro-inflammatory Ccl3. The neurodegeneration and inflammation marker Lrg1 is also upregulated in female Dek cKO mice. (C) Marker discovery analysis to identify genes that could uniquely identify each sex/genotype group was performed with MarkerFinder and AltAnalyze. A heatmap shows the differentially expressed genes that identify each group (yellow are up-regulated, and blue are down-regulated). Relevant gene ontologies and p-values are shown on the left. (D) Cytoscape network analysis was used to visualize the marker gene set for female Dek cKO mice, which identified Wnt signaling, lipid metabolism, and calcium ion binding as relevant pathways.