Sex-specific hippocampus-dependent cognitive deficits and increased neuronal autophagy in DEspR haploinsufficiency in mice

Abstract

Aside from abnormal angiogenesis, dual endothelin-1/VEGF signal peptide-activated receptor deficiency (DEspR^−/−) results in aberrant neuroepithelium and neural tube differentiation, thus elucidating DEspR's role in neurogenesis. With the emerging importance of neurogenesis in adulthood, we tested the hypothesis that nonembryonic-lethal DEspR haploinsufficiency (DEspR^+/−) perturbs neuronal homeostasis, thereby facilitating aging-associated neurodegeneration. Here we show that, in male mice only, DEspR-haploinsufficiency impaired hippocampus-dependent visuospatial and associative learning and induced noninflammatory spongiform changes, neuronal vacuolation, and loss in the hippocampus, cerebral cortex, and subcortical regions, consistent with autophagic cell death. In contrast, DEspR^+/− females exhibited better cognitive performance than wild-type females and showed absence of neuropathological changes. Signaling pathway analysis revealed DEspR-mediated phosphorylation of activators of autophagy inhibitor mammalian target of rapamycin (mTOR) and dephosphorylation of known autophagy inducers. Altogether, the data demonstrate DEspR-mediated diametrical, sex-specific modulation of cognitive performance and autophagy, highlight cerebral neuronal vulnerability to autophagic dysregulation, and causally link DEspR haploinsufficiency with increased neuronal autophagy, spongiosis, and cognitive decline in mice.

the dual endothelin-1/vegf signal peptide-activated receptor (DEspR), formerly Dear (44), was originally cloned from a Dahl salt-sensitive hypertensive rat brain cDNA library and was shown to be a single transmembrane receptor coupled to a Ca²⁺-mobilizing transduction pathway binding endothelin-1 (ET-1) and angiotensin II (ANG II) with equivalent affinities (44). Subsequent molecular studies elucidated that the mouse ortholog does not interact with ANG II but binds ET-1 and the vascular endothelial growth factor signal peptide (VEGFsp) with equal affinities instead (16). DEspR^−/− deficiency in mice resulted in embryonic lethality due to impaired vasculogenesis, abnormal angiogenesis, and vascular network formation (16). DEspR^−/− embryos also showed abnormal neurogenesis marked by a hyperconvoluted neuroepithelium and dysregulated neural tube differentiation from telencephalon to myelencephalon (16). This phenotype is strikingly opposite to the proapoptotic effects observed in the developing neural tube in VEGF^+/− deficient mice (5, 9), although abnormalities in vasculogenesis and angiogenesis are similar (16). Given two embryonic lethal abnormal angiogenesis phenotypes with differential neurogenic phenotype, the data suggest a DEspR-specific role in neurogenesis. Furthermore, on the basis of noncongruence with phenotypes exhibited by ET-1-null mice (27) and both ET_A and ET_B receptor-null mice (7, 19), observed angiogenic and neurogenic phenotypes in DEspR^−/− mice appear to be mediated by VEGFsp activation of DEspR (16).

Interestingly, while developmental angiogenic effects are observed in both male and female DEspR^−/− mice (16), several observations suggest that DEspR acts within a sex-specific context in adults. In rats, the DEspR/Dear locus is genetically linked to hypertension susceptibility in female but not male F2 (Dahl R × Dahl S) intercross rats (21) and is genetically associated with hypertension in males in a human case control hypertensive cohort (11). Likewise, DEspR-haploinsufficiency results in decrease in blood pressure in female (21.7 mmHg, P < 0.01) but not male DEspR^+/− mice compared with wild-type control mice (16). Furthermore, in a subcutaneous melanoma model DEspR haploinsufficiency results in decreased tumor angiogenesis and tumor growth in female but not male mice (16). These cumulative observations identify sex-specific modulation of DEspR-mediated functions in the adult.

Given these observations, we tested the hypothesis that DEspR's developmental angiogenesis-neurogenesis roles contribute to adult neuronal homeostasis and function. Partial deficiencies, as in DEspR^+/− haploinsufficient adult mice, will therefore impact neuronal health and hence cognitive performance. We therefore tested cognitive performance with established behavioral paradigms and analyzed histopathological changes in the brains of mice used in behavioral analysis for cognitive performance.

MATERIALS AND METHODS

Animals.

All animal procedures were approved by the Boston University School of Medicine institutional animal care and use committee. Originally, homologous recombination and production of chimeric mice were done in the SVJ genetic background (16). We proceeded to inbreed the DEspR^+/− line onto the C57BL/6 genetic background. A speed congenic strategy was implemented in order to establish a >99% congenic line (32). Sixteen backcross 1 (BC¹) DEspR^+/− heterozygous males were genotyped with 88 markers informative for the SVJ × C57BL/6 intercross. We identified one “best breeder” containing 92% of C57BL/6 genetic background, which was utilized to continue the inbreeding program (16). Animals utilized in the experiments described below are backcross 10 (BC¹⁰) from the “best breeder” to ascertain >99.98% of C57BL/6 genetic background. Wild-type and DEspR^+/− littermate mice were produced for the study from the same BC¹⁰ (+/+) × (+/−) intercross. The female cohorts consisted of 16 (+/−) and 16 (+/+) mice and the male cohorts of 11 (+/−) and 14 (+/+) mice. All mice were individually housed 1 wk before the beginning of behavioral testing. Testing was done at 12–16 wk of age in the following consecutive order: Morris water maze (MWM), social recognition (SR), and social transmission of food preference (STFP) tasks. A 1-wk resting period was observed between the different tests.

Quantitative real-time PCR.

Brains were harvested from four wild-type males, three wild-type females, five DEspR^+/− heterozygous males, and three DEspR^+/− heterozygous females. RNA from mouse brains was extracted with TRIzol reagent (Invitrogen, CA) as described previously (16), according to the manufacturer's instructions. Total RNA concentration was determined by absorbance at 260 nm, and nondegradation of RNA samples was assessed by 1% agarose formaldehyde gel electrophoresis. A 50-μg portion of each sample was further purified with the RNeasy Kit (Qiagen) with DNase I treatment per manufacturer's specifications. cDNA synthesis and amplification was performed by one-step real-time PCR with the StepOnePlus PCR System (Applied Biosciences). Equal amounts of RNA (25 ng for DEspR and 2.5 ng for cyclophilin A) were used in duplicates for quantitative real-time PCR (QRT-PCR) using the iScript one-step QRT-PCR kit with SYBR Green (Bio-Rad) and two primer sets: for DEspR (forward: 5′-TCATCTAAAGCCAGCAACTT-3′, reverse: 5′-GGCTTGCGGTCTTTTGTAT-3′) and for endogenous control cyclophilin A (forward: 5′-GCGTCTSCTTCGAGCTGTT-3′, reverse: 5′-RAAGTCACCACCCTGGCA-3′). QRT-PCR and cycling parameters were optimized: cDNA synthesis for 30 min at 50°C, thermal inactivation of reverse transcriptase for 10 min at 95°C, and 40 cycles of PCR (melting for 15 s at 95°C, annealing/synthesis for 1 min at 55°C). A standard curve for absolute quantification was generated for DEspR with six points defined by serially diluting 1:2 from 100 ng of wild-type RNA down to 3.125 ng. For the endogenous control (cyclophilin A), the range utilized was from 10 ng to 0.3125 ng of wild-type RNA. Data acquisition and analysis were carried out with StepOne software v2.0 (Applied Biosciences). We utilized cyclophilin A as normalizer for the target gene DEspR since it has been shown to have low variations between animals and stable expression in experimental brain trauma in mice (54).

Morris water maze test.

The MWM task was performed as described previously (42), comprising of training with visible platform (cued learning) and acquisition testing of spatial learning, followed by probe trial testing for retention of spatial memory. Training with a visible platform comprised four blocks, with three trials per block or test and two blocks per day. Acquisition of spatial learning was conducted with eight blocks of three trials each and two blocks per day. This schedule was done in order not to confound testing with fatigue. The mouse's path was monitored and recorded with the SMART computer tracking system (San Diego Instruments). The water maze was divided into four imaginary quadrants. Each animal was subject to 2 days of visible platform training followed by 4 days of hidden platform test.

Visible platform version.

The water maze system was located in a small observation room with blank walls. The platform was raised 1 cm above the water (26 ± 1°C, rendered opaque by the addition of 1 quart of 2% fat reduced milk) and was cued by attaching two 8.3-cm green cylinders to two opposing corners. A trial was initiated by placing the mouse in the water facing the pool perimeter in one of the chosen quadrants. The platform and start positions were randomly changed for each trial to avoid habituation to a particular quadrant. Animals were counterbalanced with respect to subject groups. A 60-s limit was imposed on each trial. The mouse was guided to the platform and allowed to rest for 10 s if it failed to locate it in the specified time.

Hidden platform version.

The experiment was done in the same room as the visible platform version but with numerous salient visual cues placed in predefined locations in the vicinity (0.7–1.0 m) of the pool. The pool (70 cm in diameter, platform size 8 cm × 8 cm) was filled until the platform was submerged 0.5 cm below the surface of the water, so that the mouse could rest on it once located. This allowed the mouse to acquire spatial learning by using extra maze cues to locate the platform. The trials were initiated in the same manner as the visible platform version, but this time the platform was maintained in one position for all the trials and the mouse was randomly dropped in one of the three quadrants that did not contain the platform. Six trials were done per day and structured in the same manner as the visible platform version. On each trial, a maximum swim time of 60 s was imposed. Between trials, a 10-s interval was imposed with the mouse on the platform. Spatial learning was measured as the mean distance traveled until the mouse found the platform. A shorter or more direct route to the platform represents better spatial learning.

Probe trial.

At the end of the sixth trial of day 4, a probe trial was done to test for retention of spatial memory. This was done with the platform removed and the mouse placed onto the quadrant diagonally opposite to the quadrant where the platform was located. During this probe trial the mouse was allowed to swim for 60 s. The greater amount of distance traveled in the target quadrant compared with all other quadrants represents retention of spatial memory.

Assessment of inherited food preference.

Separate cohorts of 12 male and 11 female C57BL/6 mice were used. Subjects were familiarized with powdered food and deprived of food for 22 h. Two flavored foods were then presented in the subjects’ home cage in separate jars. After 2 h, consumption was measured. This procedure was repeated on subsequent consecutive days with one flavored food pair per day. Food pairs consumed equivalently were validated for STFP testing; food pairs exhibiting preference for one over the other were not included in STFP testing. Several flavored food pairs were tested for each sex. For the males we tested clove (0.25% wt/wt) vs. garlic (0.2%), marjoram (2%) vs. thyme (1%), ground anise (1%) vs. cumin (0.4%), sage (0.25%) vs. clove (0.25%), curry (0.25%) vs. ginger (0.25%), and sage (0.25%) vs. onion (0.25%). The flavored food pairs tested for the females were ground anise (1%) vs. cumin (0.4%), ginger (0.25%) vs. turmeric (0.75%), onion (0.25%) vs. garlic (0.2%), curry (0.25%) vs. thyme (1%), sage (0.25%) vs. clove (0.25%), and marjoram (2%) vs. nutmeg (0.5%). Each type of scented food was encountered only once.

Social transmission of food preference task.

This task was performed essentially as described previously (10, 45, 46). Mice were trained or shaped to consume powdered food (Harlan 2018 rodent chow) for 3 days before testing. During the shaping period, mice were subjected to a 23.5-h food deprivation schedule and offered powdered food in their home cage for 30 min per day for 3 consecutive days. Food consumption was measured during the 3 shaping days, and an animal was considered “shaped” when it consumed at least 0.5 g of powdered food. Shaped subjects were then deprived of food for 24 h. An anesthetized 6-wk-old CD-1 mouse (Harlan) of the same sex as the test subject was used as demonstrator. A specific scented food was sprinkled on top of the head of the demonstrator. The demonstrator was placed in the test subject's cage and left there for 20 min for the subject to inspect. At the end of the exploratory period, the demonstrator was removed and a 5-min and 24-h time delay was enforced. At the end of the time delay, the test subject was offered two scented foods of a predetermined odor pairing in separate jars (5 cm apart), one of which was identical to the scented food presented by the demonstrator. The food was presented in a counterbalanced manner in the absence of water. Preassigned odor pairings previously validated for equivalent species preference for male mice were 1% anise vs. 0.4% cumin for the 5-min retention time and 0.25% clove vs. 0.25% sage for the 24-h retention time. Odor pairings for female mice were 1% thyme vs. 0.25% curry for the 5-min retention time and 0.25% clove vs. 0.25% sage for the 24-h retention time.

Social recognition test.

This task was performed essentially as described previously (45, 46). All tests were conducted in dim red light during the dark phase of a 12:12-h light-dark cycle and after a 2-h acclimation period. All observations were recorded onto videotape with a video camera equipped with an infrared light source. Six-week-old CD-1 mice (Harlan) of the same sex as the test subject were used as social stimuli. Testing began when a juvenile was introduced into the subject's home cage for a 5-min initial exposure. At the end of the exposure period, the juvenile was removed and returned to its cage. A second 5-min exposure to the same juvenile was conducted after a certain interval of delay. Thirty- and ninety-minute delay intervals were used in this experiment. Social investigatory behavior was previously defined (45, 46) and included direct contact while sniffing, close following, grooming or generally inspecting any body surface of the juvenile, as well as the tip of the subject's nose being proximally oriented within ∼1 cm of the juvenile. Observational data were scored and analyzed with the Observer Base Package for Windows version 3.0 (Noldus Information Technology).

Histological analysis of mouse brains.

One week after behavioral analysis, mouse brains were isolated and immersion fixed in PBS-buffered 4% paraformaldehyde, paraffin embedded, and processed, and 5-μm serial sections were obtained through the hippocampus. Hematoxylin and eosin (H & E) sections were obtained every 20 sections. Photomicroscopy was done with a Zeiss Axioskop2-Plus light microscope and Zeiss AxioVision 4.5 software. Exposure settings were identical among sections per magnification using ×5, ×20, and ×100 oil immersion objectives. Once linear best-fit setting was identified, this was used for all sections per magnification and stain type. Minimal image adjustments were applied to match the actual image when viewed directly through the eyepiece. Differential interference contrast (DIC) microscopy was used in the analysis of hematoxylin-counterstained, immunostained sections with peroxidase-diaminobenzidine (DAB) as chromogen. Nestin immunostaining was performed with nestin sc-33677 antibody (Santa Cruz Biotechnology) as described previously (16). Images were then sized and assembled for figures with AdobePhotoshop, with no color, contrast, or level corrections.

The number of hippocampal subgranular zone (SGZ) progenitor cells in male brains were counted on ×40 high-power fields of dentate gyrus hilus (n = 6 sections/group, representative of 3 mice/group). Two-tailed nonparametric test was done for statistical significance.

Signaling pathway analysis.

Analysis of ligand-specific DEspR signaling pathways was custom performed by Kinexus, utilizing the Kinex Antibody Microarray System. We compared the effects of ET-1-DEspR activation, VEGFsp-DEspR activation, and control nonstimulated DEspR (reference) in permanent transfectant Cos1 cells expressing human DEspR (Cos1-hDEspR cells). Antibody array (ab-array) analysis used phosphorylation-specific and panspecific antibodies spanning different signal transduction pathways implicated in multiple biological processes. For the present study, a single time point of 30 min, a single optimal dose for DEspR ligands based on binding affinity results (ET-1 10 nM; VEGFsp 10 nM), and an ab-array that queried signaling pathways with phosphorylation-specific and panspecific antibodies in duplicates, some in quadruplicates, or six replicates, were used. Data were normalized to background; only values exhibiting signal-to-noise ratios of ≥1.5-fold or greater and with percent error <10% were accepted. Changes in pAkt1^*S473 levels detected in the ab-array analysis were corroborated by Western blot analysis as described previously (16), with 0, 10, 15, 30, 60, and 180 min time points. Quantitative analysis of ET-1 (10 nM)- and VEGFsp (10 nM)-induced pAkt1^*S473 phosphorylation was done by densitometry of corresponding Western blots adjusted for Akt1 levels detected with pan-Akt1 antibody.

RESULTS

Sex-specific diametrical effects on hippocampus-dependent memory.

To ascertain true phenotypic differences due to DEspR haploinsufficiency, and eliminate confounders from genetic background variation, we utilized DEspR^+/− mice back-crossed for 10 generations (BC¹⁰) containing ≥99.98% C57BL/6 genetic background. The C57BL/6 genetic background was selected since this mouse strain performs competently in a variety of behavioral assays for reference memory (18, 50). The effect of DEspR haploinsufficiency on brain DEspR mRNA expression was investigated by QRT-PCR. As shown in Fig. 1, DEspR haploinsufficiency resulted in 54% and 51% reduction in DEspR mRNA levels in DEspR^+/− males (P < 0.01) and DEspR^+/− females (P < 0.05), respectively, compared with wild-type controls (Fig. 1D). This ascertains that any phenotypic differences detected are most likely due to DEspR haploinsufficiency.

Fig. 1.

Dual endothelin-1/VEGF signal peptide-activated receptor (DEspR) mRNA expression in brains of wild-type and DEspR^+/− mice. Quantitative real-time PCR (QRT-PCR) performed on serial dilutions of RNAs for cyclophilin A (CypA, green) and DEspR (magenta) are shown in A, plotting threshold cycle (C_T) values against log transformations of PCR baseline-subtracted relative fluorescence units (RFU). The resulting equations of the standard curves for CypA (green) and DEspR (magenta) are depicted in B. C: amplification plots (linear view) for CypA and DEspR for all samples tested: 7 wild-type (blue; 4 males, 3 females) and 8 DEspR^+/− (red; 5 males, 3 females) mice. D: relative mRNA expression of DEspR normalized to CypA in wild-type (WT) and DEspR^+/− (KO) male (black) and female (gray) mice. *P < 0.05, **P < 0.01, Tukey's pairwise multiple comparison test after 1-way ANOVA P < 0.001.

To investigate the effect of DEspR haploinsufficiency on representative cognitive paradigms, we evaluated age- and sex-matched BC¹⁰ male and female mice, comparing DEspR^+/− and wild-type littermates. Since the hippocampus is particularly vulnerable to varied stress states spanning aging, ischemia, neurodegeneration, and Alzheimer disease, we explored the effects of DEspR haploinsufficiency on two established experimental paradigms of hippocampus-dependent cognitive performance: the MWM, which evaluates hippocampus-dependent visuospatial cognition (34), and the STFP task, which measures hippocampus-dependent cognition for a natural odor-odor association (4).

Using an established MWM protocol, validated for mice with hippocampal pathology that is comprised of cued learning (4 sessions), acquisition of spatial learning (8 sessions), followed by probe trial for spatial memory (42), we observed the following. All mice, DEspR^+/− and wild-type male [Fig. 2B, training, sessions 1–4; F_1,30 = 0.06, not significant (ns)] and female (Fig. 2C, training, sessions 1–4; F_1,45 = 0.029, ns) mice demonstrated comparable ability for cued learning, indicating the absence of sensory motor deficit and validating the use of the MWM test to evaluate hippocampus-dependent spatial learning and memory (42, 57).

Fig. 2.

Morris water maze (MWM) testing comprising hippocampus-dependent spatial learning and memory in DEspR^+/− and wild-type mice. A: key aspects of MWM testing: 1) representative mouse track (•→) in a cued task done to ascertain equivalent abilities to perform in MWM with a visible platform (▪); 2) representative probe trial showing a 60-s mouse track (•→) indicative of poor or no spatial memory for the target quadrant (T), where the platform was previously located during acquisition (dashed box), resulting in the mouse spending time exploring all quadrants randomly; and 3) representative probe trial showing 60-s mouse track (•→) of a mouse that has spatial memory for the target quadrant, spending most time exploring said quadrant for the platform. A_R, adjacent right quadrant; A_L, adjacent left quadrant; O, opposite quadrant. B and C: results of MWM testing for male (B) and female (C) DEspR^+/− (⧫) and wild-type (○) mice. Cognitive performance in cued task and acquisition is measured as mean distance (±SE) traveled to locate the escape platform. Performance in the probe trial is measured as mean % distance (± SE) traveled in each quadrant (search quadrant %) during the probe trial. *P < 0.001, Tukey's pairwise multiple comparison test comparing % search distance in target quadrant vs. each of the other quadrants after 1-way ANOVA P < 0.05.

Acquisition of spatial learning was next assessed with distance traveled to locate the platform rather than escape latency duration in order to eliminate potential confounders from different swim speeds between contrasting strains. In males, acquisition of spatial learning trended to be slower than wild-type males, but differences were not statistically significant (Fig. 2B, sessions 5–12; F_1,70 = 1.00, ns).

To assess spatial memory, the probe trial was done after the last training session, in which impaired performance represents spatial memory deficits (42). On probe trial for spatial memory, DEspR^+/− male mice did not exhibit enhanced preference for the target quadrant over other quadrants, in contrast to wild-type male control mice, which exhibited target selectivity (Fig. 2B, probe trial; DEspR^+/− mice: P > 0.7, wild-type mice: P < 0.001), indicating that DEspR^+/− males exhibit impaired spatial memory.

In contrast, DEspR haploinsufficiency did not impair hippocampus-dependent spatial learning and memory in females. MWM testing of DEspR^+/− and wild-type females showed comparable acquisition of spatial learning (Fig. 2C, acquisition, sessions 5–12; F_1,105 = 0.002, ns). However, in the probe trial, DEspR^+/− female mice showed target selectivity with “spatial bias” toward the target quadrant, whereas wild-type female mice did not (Fig. 2C, probe trial; DEspR^+/− female mice: P < 0.001, wild-type female mice: P > 0.2). Direct comparison of MWM performance by genotypes based on sex shows diametrical sex-specific effects of DEspR haploinsufficiency on hippocampus-dependent spatial memory: impaired in males but improved hippocampus-dependent spatial memory in females.

To corroborate findings on MWM, we used another test for hippocampus-dependent cognition that is natural and nonspatial, the STFP test (4). We also employed testing of retention for short term (after 5 min) and long term (after 24 h). The test subject (observer) learns a preference for a particular scented food as “safe” after said observer mouse interacts with a demonstrator mouse that has recently consumed a distinctively scented food, thus demonstrating the scented food's safety pertinent to survival. Thus determining amount of food consumption of demonstrated-preference and nondemonstrated scented foods presented in a pair-choice paradigm at different retention times can be used to measure STFP long-term memory (Fig. 3A).

Fig. 3.

Testing of hippocampus-dependent social transmission of food preference (STFP) in DEspR^+/− and wild-type mice. A: diagram of STFP. 1: Interaction of an “observer mouse” (dark gray) with a “demonstrator mouse” (light gray) exposed to scented food 1 (white pellets). After a designated time interval (5 min and 24 h), testing of STFP retention or memory is done, determining whether the “observer mouse” remembers scented food 1 compared with “never-encountered” scented food 2. 2: Representative diagram of a mouse that consumes scented food 1 (white pellets) much more than scented food 2 (black pellets), consistent with STFP memory. 3: Representative diagram showing a mouse with no or less STFP memory consuming both scented foods (black and white pellets) equivalently. B–K: STFP results for male mice (B, D–G) and female (C, H–K) DEspR^+/− (filled bars) and wild-type (open bars) mice. B and C: mean % (±SE) of food eaten in a choice test (to assess inherited food preferences) by male (B) and female (C) C57BL/6 mice presented with the following scented food pairings: for males: 1,1′ = cumin-anise, 2,2′ = sage-clove, 3,3′ = thyme-marjoram, 4,4′ = ginger-curry, 5,5′ = garlic-clove, 6,6′ = onion-sage; for females: 1,1′ = cumin-anise, 2,2′ = ginger-turmeric, 3,3′ = onion-garlic, 4,4′ = curry-thyme, 5,5′ = sage-clove, 6,6′ = marjoram-nutmeg. D–K: mean % odor preference (±SE) in male (D–G) and female (H–K) DEspR^+/− and wild-type observers at 5-min (D, E, H, I) and 24-h (F, G, J, K) retention times. Tr, trained odor; Non-tr, nontrained odor. *P < 0.05, **P < 0.01, ***P < 0.001, P values calculated with an unpaired Student's t-test.

Since this test is valid only if paired odors used for STFP testing are ascertained for equivalent baseline preference among mice per sex, we first identified equivalent scent preference of foods with an independent group of male and female C57BL/6 mice. Male mice showed inherent preference for thyme > marjoram, ginger > curry, garlic > clove, and onion > sage (Fig. 3B). Female mice demonstrated inherent preference for ginger > turmeric and marjoram > nutmeg (Fig. 3C). Thus these pairings were not utilized in the STFP experiments. Preassigned odor pairings with ascertained equivalent preference in male mice were anise vs. cumin for the 5-min retention time and clove vs. sage for the 24-h retention time. Odor pairings for female mice were thyme vs. curry for the 5-min retention time and clove vs. sage for the 24-h retention time.

On STFP testing, wild-type male mice showed efficient performance at 5-min (Fig. 3E; t = 3.26, P < 0.004) and 24-h (Fig. 3G; t = 4.15, P < 0.001) retention times. In contrast, DEspR^+/− male mice exhibited difficulties in performing this task. They were able to show preference for the trained odor at 5-min (Fig. 3D; t = 6.56, P < 0.001) retention time; however, they did not recognize the trained odor after 24 h of retention time (Fig. 3F, t = 2.03, ns). In contrast, DEspR^+/− female mice were able to perform this task at 5-min (Fig. 3H; t = 3.25, P < 0.01) and 24-h (Fig. 3J; t = 2.12, P < 0.05) retention times, while wild-type female mice did not demonstrate preference for the trained odor at 24-h (Fig. 3K; t = 1.03, ns) retention time. Concordantly, hippocampus-dependent STFP results parallel observations made in the MWM task, thus demonstrating that DEspR haploinsufficiency reduces hippocampus-dependent cognitive performance in male, but not female, mice. In fact, DEspR^+/− females performed better than wild-type DEspR^+/+ females in both hippocampus-dependent cognitive performance tasks (Figs. 2 and 3). Diametrical effects of DEspR haploinsufficiency confirm sex-specific functional roles for DEspR, as well as suggesting complexities with estrogen-mediated modulation of DEspR function necessitating further study.

Differential effects on social recognition.

To assess DEspR haploinsufficiency effects on other cognitive functions, we tested performance in a SR task, which measures the natural ability to recognize conspecifics through social investigation, mainly through olfactory investigation (48, 55), relevant to species survival (53). A mouse that remembers previous exposure to a conspecific will spend less time exploring said conspecific, whereas poor learning/memory will manifest as a mouse not remembering a conspecific to which it was previously exposed and thus exploring the subject equivalently as in the first encounter (Fig. 4A). Thus difference in duration of social investigation of conspecifics presented at different retention times is used as a measure of social recognition (48, 55).

Fig. 4.

Social recognition (SR) in DEspR^+/− and wild-type mice. A: diagram depicting SR testing. 1: Conspecific juvenile mouse (light gray) is exposed to test mouse (dark gray); test mouse investigates juvenile. After a specific experimental interval (30 or 90 min), test mice are exposed again to the same juvenile. 2: Test mice that remember (check mark) spend less time investigating the juvenile to which they were previously exposed since they remember it (conspecific). 3: Test mice that do not remember (?) investigate the juvenile for the equivalent amount of time as in the first exposure (as though they were not previously exposed to it). Memory is inversely related to duration of investigation. B and C: mean (±SE) duration of investigation of the juvenile by male (B) and female (C) DEspR^+/− (filled bars) and wild-type (open bars) mice on the first exposure (0 min) and the second exposure performed later (clock) after 30 or 90 min. *P < 0.01, **P < 0.0001, compared with first exposure; P values calculated with a paired Student's t-test.

In contrast to hippocampus-dependent MWM and STFP, both DEspR^+/− and wild-type male mice performed equivalently in SR tests. Both were able to recognize conspecifics at the 30-min (Fig. 4B; DEspR^+/−: t = 4.22, P < 0.001; wild-type: t = 3.42, P < 0.01) but not the 90-min (Fig. 4B; DEspR^+/−: t = 1.53, ns; wild-type: t = −0.41, ns) retention time. In contrast, DEspR^+/− females demonstrated a robust SR performance recognizing a conspecific after 30-min (Fig. 4C; DEspR^+/−: t = 6.3, P < 0.0001) and 90-min (Fig. 4C; DEspR^+/−: t = 3.69, P < 0.01) retention times, while wild-type DEspR^+/+ females were unable to perform this task at both retention times (Fig. 4C; 30 min: t = 2.08, ns; 90 min: t = 0.75, ns). While DEspR haploinsufficiency is associated with improved function in females rather than the deficits seen in males, improvements are detected in all cognitive tasks tested, suggesting a more global effect. Although estrous cycle stages were randomly represented in both DEspR^+/− and DEspR^+/+ female mice, the possible differential effects of estrogen cannot be eliminated and remain to be investigated in future. Nevertheless, observations in SR indicate specificity of deficits in hippocampus-dependent MWM and STFP in DEspR^+/− male mice, and again reiterate sex-specific effects of DEspR haploinsufficiency.

Spongiform changes and increased autophagy in hippocampal and cerebral cortical neurons in DEspR^+/− male mice.

To investigate possible neuropathological correlates of decreased hippocampus-dependent spatial and associative learning and memory in DEspR^+/− male mice, histopathological analyses of brains were completed on a representative group of mice after behavioral studies (n = 6–8 per DEspR^+/− group, n = 4 for wild-type controls), analyzing five serial sections across the hippocampus. We detected spongiform changes in the neuropil and granule cell layers (CA1-3, dentate gyrus) of the hippocampus of DEspR^+/− male brains but not in age-matched, littermate DEspR^+/− female and wild-type male and female mouse brains (Figs. 5–7) at 6–7 mo of age. Spongiform changes with neuronal cytoplasmic swelling, vacuolation, and neuronal loss were also detected in subcortical basal ganglia in DEspR^+/− male (Fig. 8A) but not in wild-type male (Fig. 8B) brain, as well as in focal areas of cerebral cortical regions of DEspR^+/− male brain (Fig. 8C) but not in littermate wild-type male brain (Fig. 8D). The increased, irregular cytoplasmic vacuolation with cytoplasmic swelling and dense nuclei are consistent with autophagic stress, defined as the imbalance between autophagy and autophagosome clearance, since increased autophagy is associated with decreased clearance (6). Notably, there were no inflammatory or hemorrhagic changes or astro- or microgliosis surrounding observed spongiform changes, neuronal cytoplasmic vacuolation, and neuronal loss (Figs. 5–8). These observations reveal a key role of DEspR in the regulation of autophagic homeostasis, the perturbation of which leads to increased autophagy, autophagic stress, and programmed cell death of neurons in the hippocampus and cerebral cortex.

Fig. 5.

Spongiform changes in the hippocampus and overlying cerebral cortex in male DEspR^+/− mouse brain compared with control wild-type and female mice. Representative hematoxylin and eosin (H & E)-stained sections of male wild-type (A) and DEspR^+/− (B) mouse brains and female wild-type (C) and DEspR^+/− (D) mouse brains show noninflammatory spongiosis changes in male DEspR^+/− brain section (B) but none in age-matched wild-type mouse brain section (A) or in both female wild-type (C) and DEspR^+/− (D) mouse brains. Spongiform changes are also detected in the male DEspR^+/− cerebral cortex (B). Boxed areas in CA1 (solid box) and in the dentate gyrus (dashed box) are shown in higher magnification in Fig. 6.

Fig. 6.

Representative ×100 oil-immersion differential interference contrast (DIC) imaging of H & E sections of hippocampal CA1 (A and B) and dentate gyrus (C and D) regions boxed in Fig. 5 show neuronal cytoplasmic vacuolation, nuclear dysmorphology, and neuronal loss consistent with autophagy in male DEspR^+/− (A and C) in contrast to control male wild-type (B and D) mouse brain sections.

Fig. 7.

Focal areas of neuronal autophagy and decreased progenitor subgranular zone (SGZ) cells in male DEspR^+/− mouse hippocampus. A and B: representative high-magnification images of hippocampal CA1 region in Fig. 6 show autophagic vacuolations in neurons (yellow arrows), neuronal loss, and dysmorphology in male DEspR^+/− (A) brain in contrast to normal neuronal (green arrows) and neuropil morphology in age-matched, littermate male wild-type mouse brain (B). C and D: representative nestin-immunostained, hematoxylin-counterstained section viewed with DIC optics showing decreased SGZ cells and sparse nestin+ SGZ cells (red arrows) in male DEspR^+/− mouse brain dentate gyrus (C) in contrast to age-matched, littermate male wild-type mouse brain (D) with 2-fold more SGZ cells and nestin+ SGZ cells in wild-type (filled bar) than in DEspR^+/− (open bar) male mice (E). *P < 0.005, unpaired Student's t-test. Bar, 10 μm (A–D).

Fig. 8.

Detection of autophagic vacuolations in the basal ganglia and cerebral cortex in male DEspR^+/− mouse brains. A and B: representative ×100 oil immersion DIC H & E sections of basal ganglia demonstrating spongiosis changes affecting neuropil, neurons (arrows), and fiber tracks (encircled in yellow) in DEspR^+/− brain (A) but none in age-matched littermate male wild-type mouse brain (B) (encircled in green). Neuronal and fiber track vacuolations are consistent with neuronal and axonal autophagy, respectively. C and D: representative ×100 oil immersion DIC H & E sections of cerebral cortex. High magnification reveals neuropil microvacuolations, neuronal cytoplasmic vacuolation and swelling, and nuclear dysmorphology, consistent with autophagy in male DEspR^+/− brain section (C) but none in age-matched, littermate male wild-type control (D). Bars, 10 μm.

Furthermore, we detected a decrease in number of hippocampal SGZ progenitor cells in DEspR^+/− male brains (Fig. 7C) compared with wild-type male brains (Fig. 7D) (average ± SE: 17.2 ± 1.5 vs. 36.0 ± 5.0, respectively, P < 0.005; Fig. 7E). We did not, however, detect any autophagic vacuoles in said SGZ cells, albeit reduced in number. These observations are concordant with the hypothesis that postmitotic neurons are particularly susceptible to developing autophagic stress (6), suggesting that the regulation of autophagy is different in proliferative cells.

Furthermore, we note that neurological deficits such as tremors, ataxia, spasticity, paresis, paralysis, incontinence, or premature death associated with other mutant mouse models exhibiting autophagy (29) were not exhibited by DEspR^+/− males (n = 3) with observation periods up to 15 mo of age (data not shown), thus demonstrating a DEspR-specific role in autophagy.

VEGFsp-stimulated DEspR role in autophagy inhibition.

To investigate the putative mechanism by which DEspR haploinsufficiency results in a net increase of autophagy, we investigated whether upstream signaling pathways regulating autophagy are activated/deactivated by DEspR on stimulation with VEGFsp. VEGFsp is the ligand deduced to be mediating DEspR embryonic roles in neurogenesis since ET-1-knockout mice do not produce a phenotype similar to DEspR-null mice (16). We investigated upstream activators PDK, Akt1, RSK1, and RSK1/2 of the major inhibitory signal of autophagy mammalian target of rapamycin (mTOR) (28, 38), as well as upstream stimulators of autophagy JNK, Jun, and ERK1/2 (3, 13), using a human ab-array containing a duplicate-to-quadruplicate panel of antibodies specific for nonphosphorylated and phosphorylated signaling molecules. We investigated whether VEGFsp-treated Cos1 transfectant cells expressing human DEspR would phosphorylate or dephosphorylate upstream inhibitory or stimulatory regulators of autophagy.

Consistent with evidence for induced autophagy in neurons, 30-min VEGFsp treatment of stable Cos1-hDEspR transfectants resulted in the phosphorylation of several upstream activators of mTOR, the major inhibitory gatekeeper of autophagy: PDK, Akt1, RSK1, and RSK1/2 (Fig. 9). Since Akt1 elicited the smallest percent change from control, we tested Akt1 phosphorylation by Western blot analysis. As shown in Fig. 9, B and C, Western blot analysis confirmed phosphorylation of Akt1 on VEGFsp activation of DEspR in Cos1-hDEspR transfectants, in contrast to no effect by ET-1 stimulation of DEspR.

Fig. 9.

DEspR-mediated effects on signaling pathways that regulate autophagy. A: antibody-array analysis of signaling pathways involved in the stimulation and inhibition of autophagy reveals that VEGF signal peptide (VEGFsp)-DEspR activation dephosphorylates [< −15% change from control (t₃₀/t₀)] autophagy stimulators (each bar = average of duplicate spots): a: 6 replicates of JNK^*T183/Y185. b: 2 replicates of jun^*S63. c: 6 replicates of jun^*S73. d: Erk1/2^*T202+Y204; e: 4 replicates of Erk1/2^{*T202+Y204;*T185/Y187}. VEGFsp-DEspR activation also phosphorylates (> +15% change from control) upstream activators of mammalian target of rapamycin (mTOR), the key inhibitor of autophagy: f: PDK1. g: pPDK1^*S244. h: 4 replicates of phosphatidylinositol 3-kinase. i: 4 replicates of pAkt1^*S473 and at least 4 different phosphorylation sites of RSK1/2: pRSK1/2^*S221/S277, pRSK1/2^*S363/S369, pRSK1/2^*T359/T365, pRSK1/2^*T573/T577. j: RSK1. k: pRSK1^*S380/S386. Values represent % change from controls; control is time 0 or nontreated Cos1-hDEspR stable transfectants. B: Western blot analysis of phosphorylated pAkt1^*S473 (60 kDa) in VEGFsp-stimulated Cos1-hDEspR transfectants at time 0–180 min. Bottom: Akt1 protein levels detected with the pan-Akt1 antibody. C: Western blot analysis of phosphorylated pAkt1^*S473 (60 kDa) in endothelin 1 (ET-1)-stimulated Cos1-hDEspR transfectants at time 0-180 min. Bottom: Akt1 protein levels detected with the pan-Akt1 antibody. D: quantitative analysis of Western blots in B and C with pAkt1^*S473 normalized to Akt1 protein from time 0 to 180 min. VEGFsp-DEspR activation phosphorylates Akt1^*S473, whereas ET-1-DEspR activation does not phosphorylate Akt1^*S473. % pAkt1*S473 (t_x′/t_0′), % change of pAkt1^*S473 from time 0 of normalized values. Molecular mass markers on left in B and C from top to bottom: 150, 100, 75, and 50 kDa.

Concordantly, we also detected dephosphorylation of several stimulators of autophagy: JNK, Jun, and ERK1/2 (Fig. 9A). These data demonstrate that VEGFsp-DEspR interaction inhibits autophagy through a dual mechanism: phosphorylation/activation of upstream activators of mTOR-mediated autophagy inhibition, as well as dephosphorylation/deactivation of autophagy stimulators. These data support the observations that DEspR haploinsufficiency results in the reduction of DEspR-mediated inhibition of autophagy and hence net induction of autophagy as observed in DEspR^+/− male mouse brains.

DISCUSSION

Sex-specific, diametrical effects on hippocampus-dependent cognitive performance.

The observed diametrical effects in males versus females of DEspR haploinsufficiency in hippocampus-dependent visuospatial learning and memory and hippocampus-dependent associative learning provide robust evidence supporting the concept of sex-specific dimorphic/diergic determinants of cognition. Moreover, sex-specific differences in learning have been reported in chimpanzees (31), in rats and mice characterized in the radial and water maze tasks for spatial learning (20), as well as in humans, where studies on brain activation during human navigation have shown sex-specific neural substrates for spatial cognition performance (12). Multispecies to human concordance adds significance to animal model observations of sex-specific differences in cognitive performance.

Concordantly, sex specificity of DEspR-mediated functions also affects other physiological parameters, such as hypertension susceptibility quantitative trait locus in Dahl rats (21) and blood pressure and tumor progression in adult DEspR^+/− mice (16), although sex-specific effects are detected involving both sexes. These observations contribute cumulative evidence supporting the paradigm that genetic determinants contribute to, and hence modify, said sex-specific differences in a variety of physiological traits including cardiovascular (15, 17, 58), depression-like behavior (52), and emotionality-related behavior (41).

Sex-specific noninflammatory spongiform changes, neuronal autophagic stress, and neuronal loss.

The association of reduced performance in hippocampus-dependent MWM and STFP with morphological, noninflammatory, and nonischemic spongiform changes in the neuropil and CA1, CA2, CA3, and dentate gyrus granule cell layers of the hippocampus collectively supports hippocampus-dependent cognitive dysfunction in DEspR^+/− male mice. The absence of abnormalities in hippocampus-dependent MWM and STFP and the spongiform changes in 6- to 7-mo-old DEspR^+/− female mouse brains demonstrate robust sex-specific effects of DEspR haploinsufficiency. The detection of increased neuronal autophagic stress, neuronal loss, and neuronal and neuropil spongiform changes links neuronal autophagic stress with programmed cell death and supports the hypothesis that excessive neuronal autophagic stress causes neuronal cellular injury and destruction (6, 60)—as seen in neuronal development (49) and in Huntington disease (39). We note, however, that studies also show that suppression of basal autophagy in neurons causes neurodegeneration in mice (14, 25), speaking to the critical importance of autophagy in preventing neurons from undergoing protein aggregation-induced degeneration (26, 29, 56). These seemingly paradoxical observations, associating neurodegeneration with both autophagy deficiency (14, 25, 33) and autophagy excess (Ref. 56; this study), strongly indicate that the homeostatic balance of autophagy is critical to the maintenance of constitutive basal levels for cell survival, to responsive modulation to keep up with changes in organelle turnover or misfolded protein accumulation, and to avoidance of excessive autophagy or autophagic stress, which leads to programmed cell death. The relationship of autophagy and cell death does not always need excess, since autophagy has also been implicated in exocytotoxic necrotic-like cell death (56).

Since other known etiologies of increased spongiform changes and/or increased autophagy in the brain such as prion disease or scrapie (51), spongiform encephalopathic neurovirulent viruses such as amphotropic murine leukemia virus (35) and FrCas(E) retrovirus (8), and chromosome 15 Gray tremor mutation (23) are not present, the causal role of DEspR haploinsufficiency in increased neuronal autophagy is unequivocal. The detection of autophagic stress and cognitive deficits only in male DEspR^+/− mice elucidates a sex-specific regulatory mechanism for neuronal autophagy balance, with DEspR as a key mediator. Partial loss of DEspR function, as in DEspR haploinsufficiency, results in sex-specific dysregulation of neuronal autophagy balance leading to neuronal autophagic programmed cell death.

Diametrical effects of DEspR deficiency/haploinsufficiency on angiogenesis and autophagy.

Decreased DEspR levels have diametrical effects on angiogenesis and autophagy: DEspR-null mutation results in abnormal angiogenesis (16), while DEspR haploinsufficiency results in increased autophagy (shown here). This is similar to endostatin and kringle 5, which also inhibit angiogenesis but induce autophagy (36, 40). The commonality in diametrically opposed antiangiogenesis/proautophagy effects mediated by different pathways, kringle 5-, endostatin-, and now DEspR mediated, is intriguing and intuitively assigns the coordination of angiogenesis and autophagy as a paradigm whose significance requires multipathway modulation.

Conclusions.

Altogether the data 1) demonstrate the role of DEspR in sex-specific regulation of neuronal autophagy, 2) link increased neuronal autophagy, programmed cell death, and cognitive dysfunction, and 3) highlight the susceptibility of hippocampal granular neurons as well as cerebral cortical neurons to autophagic programmed cell death. More importantly, these observations elucidate a clinical translational paradigm: that the modulation of DEspR functional levels is critical and could therefore comprise a therapeutic target for vectorial modulation of autophagy's paradoxical roles in cell survival and death, as dictated by disease context. On one hand, decreased DEspR functions to induce autophagy, thus harnessing autophagy-mediated neuroprotection for protein aggregate diseases (33, 43, 59), such as Alzheimer disease (37), Huntington disease (47), or lysosomal storage disease neurodegeneration (24). On the other hand, DEspR stimulation could decrease excessive autophagy, in order to attenuate autophagic neuronal programmed cell death as observed in Parkinson disease (1), aging-associated neurodegeneration (2, 22), prion disease (30, 51), or hypoxia/ischemia (24).

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant HL-69937 to N. Ruiz-Opazo.

DISCLOSURES

Boston University School of Medicine has filed a patent (Application No. PCT/US2005/041594) for DEAR/DEspR.