Impact of CRFR1 Ablation on Amyloid-β Production and Accumulation in a Mouse Model of Alzheimer’s Disease

Abstract

Stress exposure and the corticotropin-releasing factor (CRF) system have been implicated as mechanistically involved in both Alzheimer’s disease (AD) and associated rodent models. In particular, the major stress receptor, CRF receptor type 1 (CRFR1), modulates cellular activity in many AD-relevant brain areas, and has been demonstrated to impact both tau phosphorylation and amyloid-β (Aβ) pathways. The overarching goal of our laboratory is to develop and characterize agents that impact the CRF signaling system as disease-modifying treatments for AD. In the present study, we developed a novel transgenic mouse to determine whether partial or complete ablation of CRFR1 was feasible in an AD transgenic model and whether this type of treatment could impact Aβ pathology. Double transgenic AD mice (PSAPP) were crossed to mice null for CRFR1; resultant CRFR1 heterozygous (PSAPP-R1^+/−) and homozygous (PSAPP-R1^−/−) female offspring were used at 12 months of age to examine the impact of CRFR1 disruption on the severity of AD Aβ levels and pathology. We found that both PSAPP-R1^+/− and PSAPP-R1^−/− had significantly reduced Aβ burden in the hippocampus, insular, rhinal, and retrosplenial cortices. Accordingly, we observed dramatic reductions in Aβ peptides and AβPP-CTFs, providing support for a direct relationship between CRFR1 and Aβ production pathways. In summary, our results suggest that interference of CRFR1 in an AD model is tolerable and is efficacious in impacting Aβ neuropathology.

INTRODUCTION

Alzheimer’s disease (AD) is the most common form of dementia, characterized neuropathologically by the accumulation of senile plaques and neurofibrillary tangles. Amyloid-β (Aβ), a cleavage product of the amyloid-β protein precursor (AβPP), aggregates in extracellular spaces as senile plaques. Tangles are formed by the fibrillization of hyperphosphorylated tau (tau-P). Less than 2% of AD cases are linked to known genetic mutations, with the vast majority of AD cases considered sporadic and possibly resulting from an interaction between other genetic and environmental risk factors that increase susceptibility [1, 2]. Stress is a risk factor highly correlated with AD pathology in rodents and humans [3, 4]. Individuals predisposed to experience anxiety are more likely to develop AD, and the prevalence of the disease is increased among patients diagnosed with early-onset anxiety disorders [5, 6].

The physiological impact of stress is mediated by the hypothalamic-pituitary-adrenal axis. In response to stress, neurons of the hypothalamic paraventricular nucleus secrete corticotropin releasing factor (CRF), which binds to CRF receptor 1 (CRFR1) on the pituitary gland. Although CRFR1 is the pituitary receptor that mediates the neuroendocrine stress response [7], it is also expressed widely in brain, including AD-relevant regions such as the isocortex, hippocampus, and amygdala [8]. CRF signaling through CRFR1 in brain acts as a neuromodulator that modulates cellular activity [9, 10]. Supporting a role of CRF in AD neuropathology, work from our laboratory has demonstrated that both CRF overexpression and acute and repeated restraint stress induced hippocampal tau-P, a process that is dependent on CRFR1 ([11–13] versus [14]). While tau-P resulting from acute stress is transient and potentially involved in stress-related neuroplasticity, tau-P induced by repeated stress or overexpression of CRF is present for extended periods after stress exposure, has a definable structure and is localized to detergent-soluble cellular fractions [12, 13].

Mounting evidence has accumulated showing that CRF signaling is involved in the pathophysiology of AD (e.g., [15–17]) and a substantial number of recent studies demonstrate a role for CRF and CRFR1 in AD-related endpoints [12–14, 17–23]. In the present study, we expanded our investigation of the role of CRFR1 in AD to determine whether chronic interruption of CRFR1 was feasible in an AD mouse model and could impact Aβ pathology. We used double transgenic mice that express mutant forms of the human genes for AβPP and presenilin-1 (PSAPP), and reliably produce Aβ plaques beginning at 3-4 months of age [24]. Aβ accumulation in these animals can be exacerbated by behavioral and physiological stress [19]. To examine the impact of chronic loss of CRFR1 on development of Aβ pathology, we developed a novel transgenic model by crossing the PSAPP line to mice null for CRFR1 [25]. Resultant female offspring were used at 12 months of age to examine the impact of CRFR1 disruption on Aβ peptide levels and severity of Aβ plaque pathology.

MATERIALS AND METHODS

Generation of PSAPP-CRFR1^−/− mice

CRFR1 knockout mice (CRFR1^−/−) were bred from heterozygote breeder pairs of established lines backcrossed to founder mice to achieve a pure C57BL/6 background and genotyped using established PCR protocols [25]. Pregnant females used for generating CRFR1-deficient mice received drinking water supplemented with corticosterone (10 µg/ml; Sigma-Aldrich, St. Louis, MO) from embryonic day 12 to postnatal day 14 to prevent early mortality as a result of pulmonary dysplasia [25]. Female CRFR1^−/− mice were crossed with AD transgenic male offspring of Borchelt PSAPP mouse line 85 (APP-Swe_K595N,M596L and PS1_dE9; Jackson Laboratories, Bar Harbor, ME) [24]. Offspring of this cross were determined by PCR genotyping and named based on genotype status: PSAPP, PSAPP-CRFR1^+/−, PSAPP-CRFR1^−/−. PSAPP CRFR mutant mice were maintained on corticosterone-supplemented drinking water during testing to allow the normal nocturnal bias in appetitive behavior to approximate the circadian fluctuation in circulating hormone levels as described previously [11, 12]. To assess effectiveness of the replacement regimen, plasma corticosterone levels were determined by radioimmunoassay from blood samples collected at sacrifice. The UCSD Institutional Animal Care and Use Committee approved all experimental protocols.

Aβ ELISA

Whole brains from 12-month-old female PSAPP, PSAPP-R1^+/−, and PSAPP-R1^−/− (n = 4 each) mice were homogenized in 5 M guanidine HCL buffer (4 ml per 1.5 ml of whole brain extract lysate previously homogenized in a PBS/5mM EDTA homogenization buffer), sonicated for 15–30 s, and incubated for 3–4 h at room temperature. 500 µl of the guanidine extract was added to an equal volume of PBS-T, and the samples were spun at 4°C for 20 min at 16,000 × g, and levels of Aβ₄₀ and Aβ₄₂ were analyzed by enzyme-linked immunosorbent assay (ELISA; Life Technologies) according to the manufacturer instructions. Plates were read using a BioRad iMark Microplate Reader. 12-month-old female wild-type (WT) and CRFR1^−/− animals (n = 4 each) were used as controls.

Western analysis

Groups of 12-month-old female WT (n = 5), PSAPP (n = 4), PSAPP-R1^+/− (n = 4), PSAPP-R1^−/− (n = 4), and CRFR1^−/− (n = 4) mice were killed by cervical dislocation, decapitated and the hippocampus was rapidly dissected and frozen on dry ice. Tissues were homogenized in RIPA buffer containing protease inhibitors and prepared for western analysis as described previously [11, 12]. 6 µg of protein was loaded and electrophoretically separated on a polyacrylamide gel. Samples for CT-20 analysis were prepared in tricine sample buffer and separated on 10–20% tricine gel. Background subtraction was performed and quantitative band intensity readings were obtained using NIH ImageJ software as described previously [11, 12, 14]. 22C11 (Millipore, 1:5000) was used to detect total levels of AβPP, insulin-degrading enzyme (IDE, EMD Millipore, 1:1000) was used to detect total levels of insulin degrading enzyme, and CT-20, provided by Dr. M. P. Murphy (University of Kentucky, 1:1000), was used to detect C-terminal fragments of AβPP. β-actin (Sigma, 1:1000) was used as a loading control on all blots.

Fluorescent amyloid plaque localization

Groups of 12-month-old female PSAPP, PSAPP-R1^+/−, and PSAPP-R1^−/− (n = 5, each) mice were killed by cervical dislocation, decapitated and the brains were drop fixed in 4% paraformaldehyde for 3 h and then cryprotected overnight in 2% sucrose-PBS. The next morning, 30 µm-thick sections were cut on a freezing-sliding microtome and stored at −20°C in cryoprotectant solution (20% glycerol and 30% ethylene glycol in 0.1 M phosphate buffer). Sections were mounted on Fisherbrand Superfrost®/Plus Microscope Slides and treated with 70% formic acid before immunohistochemical staining for Aβ. Sections were immunolabeled for Aβ using the N-terminal specific anti-human Aβ antibody 82E1 (1:500 from 50 µg/ml stock solution; IBL, Hamburg, Germany) and visualized with a goat anti-mouse secondary antibody conjugated to Alexa Fluor® 568 fluorescent dye (1:575; Invitrogen, Carlsbad, CA). Slides were coverslipped using Fluoromount-G™ mounting medium (Southern Biotech, Birmingham, AL). WT and CRFR1^−/− mice were used as controls (n = 5, each).

Aβ quantification

The extent of the brain occupied by Aβ plaques was assayed from fluorescently labeled sections, with regions identified according to anatomical landmarks local to each region of interest (ROI) as depicted in the Franklin and Paxinos atlas. Images were acquired with a Leica DFC310FX camera on a Leica DM5500B microscope at 5× optical magnification using Leica Application Suite software. Individual, adjacent images were stitched together in ImageJ using a plug-in developed by S. Preibisch. Fourteen grayscale, 12-bit images of coronal sections, no less than 200 µm apart, were collected from each animal and analyzed using Metamorph image analysis software licensed to Leica Microsystems. No two sections taken from the same mouse were less than 200 µm apart. Equivalent brain regions between subjects were identified according to anatomical landmarks established in the third edition of Franklin and Paxinos’ atlas, The Mouse Brain. The piriform, as well as the rostral-most areas of the amygdala and hippocampus were quantified from equivalent sections between ~3mm rostral to the interaural plane to ~2.6mm rostral to the interaural plane. The amygdala and the hippocampus from the most caudal sections continued to be quantified distinctly. However, between ~2.6mm rostral to the interaural plane to the interaural plane itself, other regions ventral to the rhinal fissure: the ectorhinal, perirhinal, and entorhinal cortices, were included with quantification of the piriform. Sections corresponded between subjects, beginning at approximately +1.10mm (relative to bregma), where the decussation of the corpus callosum is first apparent, and ending at approximately −3.88 mm, where the hilar region of the hippocampus can still be seen. ROIs were outlined with a light pen on a Wacom Cintiq 12WX interactive LCD monitor.

Statistics

Optical density readings from ELISA and western blots were analyzed by two-way ANOVA using SPS software. Individual group means were compared using Tukey’s multiple-comparison test. Bar graphs were made using Prism 6 software. Data are presented as mean±SEM. For plaque quantification, to adjust plaque with amyloid expressed as a percentage of the area of the sampling region. Density readings from ROI were analyzed by two-way ANOVA (genotype × coronal level) for genotypic differences in plaque load across serial coronal levels, with coronal level used as the within subjects variable. Likewise, total plaque load area for each brain region was estimated by summing the thresholded area of each region from every section it was sampled within, and expressing it as a percentage of the total region area across sections.

RESULTS

Impact of CRFR1 ablation on Aβ plaque load at 12 months of age

As found in previous studies with AD models, our initial trials with antibodies 4G8 and 6E10 yielded heavy AβPP and/or intracellular Aβ labeling (e.g., [26]), which was problematic for developing adequate threshold programming to solely identify Aβ plaques using our Leica Metamorph software (data not shown). For this reason, we chose antibody 82E1 for detection of Aβ plaques because our tests suggest it provided the most favorable detection of plaques with virtually no cellular labeling (Fig. 1).

Figure 1.

PSAPP mice lacking either both (CRFR1^−/−) or one (CRFR1^+/−) allele for CRFR1 exhibited consistently lower mean values of plaque load per µm² in all brain areas. Statistically significant reductions were seen in the insular, retrosplenial, ectorhinal, and perirhinal cortices (p < 0.05, p < 0.01, p < 0.01, p < 0.05, respectively). The hilar region of the hippocampus and the molecular layer of the dentate gyrus also showed significant decrease in Aβ accumulation (both p < 0.05 for both CRFR1^−/− and CRFR1^+/−) as a function of CRFR1 ablation. PSAPP mice, panels A, D; PSAPP-CRFR1^+/−, panels B, E; PSAPP-CRFR1^−/−, panels C, F. Retrosplenial (RS) cortex; hilar layer of the hippocampus, CA4; the molecular layer of the dentate gyrus (mDG); ectorhinal cortex (Ect); perirhinal cortex (PRh); corpus callosum (cc); rhinal fissure (rf); lateral nucleus of the amygdala (La). Scale bar 400 µm.

Plaque load in cortical regions surrounding the medial longitudinal and rhinal fissures of each animal was measured at fourteen coronal levels, as described in Materials and Methods. We found large reductions in plaque accumulation through hilar region of the hippocampus (both p < 0.05) and the insular, retrosplenial, ectorhinal, and perirhinal cortices (p < 0.05, p < 0.01, p < 0.01, respectively) in both PSAPP-R1^+/− and PSAPP-R1^−/− mice compared to PSAPP cohorts (Fig. 1). Percent differences in total plaque area percentages per region are shown in Table 1.

Table 1.

Percent of total plaque area per µm² in regions showing significant change

	PSAPP	PSAPP-R1+/−	PSAPP-R1−/−	p value versus PSAPP
Insular	3.46	2.15	2.02	<0.05
Retrosplenial	4.87	2.89	2.77	<0.01
Ectorhinal/perirhinal	6.75	4.23	3.82	<0.01
Dentate gyrus mol	6.25	4.16	3.65	<0.05
Hilus	4.72	2.13	1.48	<0.05

Impact of CRFR1 ablation on Aβ peptide levels in brain

To determine whether levels of Aβ were reduced with CRFR1 ablation, we examined total brain levels of Aβ₄₀ and Aβ₄₂ in whole brain lysates by ELISA in PSAPP mice as a function of CRFR1 status at 12 months of age (n = 4/group) (Fig. 2). We found that both PSAPP-R1^+/− and PSAPP-R1^−/− mice had significantly reduced levels of both Aβ₄₀ and Aβ₄₂, compared to PSAPP counterparts (for Aβ₄₀: PSAPP-R1^−/− and PSAPP-R1^+/− versus PSAPP, both p < 0.001; for Aβ₄₂: PSAPP-R1^−/− and PSAPP-R1^+/− versus PSAPP, both p < 0.05) (Fig. 2). Importantly, levels of both Aβ₄₀ and Aβ₄₂ in PSAPP-R1^−/− and PSAPP-R1^+/− were not significantly different from that seen in WT or CRFR1 knockout mice (all p > 0.05).

Figure 2.

Changes in Aβ peptides in PSAPP mice with CRFR1 ablation. Analysis of whole brain lysates by ELISA revealed that (A) levels of Aβ₄₀ and (B) Aβ₄₂ were significantly reduced in PSAPP mice null for one (PSAPP-R1^+/−) or both (PSAPP-R1^−/−) alleles for CRFR1 compared to PSAPP counterparts (for Aβ₄₀, PSAPP-R1^−/− and PSAPP-R1^+/− versus PSAPP, p < 0.001; for Aβ₄₂, PSAPP-R1^−/− and PSAPP-R1^+/− versus PSAPP, p < 0.05). Levels of both Aβ₄₀ and Aβ₄₂ in PSAPP-R1^+/− and PSAPP-R1^−/− were not significantly different than that of wild-type (wt) animals or CRFR1 knockout mice (p > 0.05). All data are expressed as Mean±SEM.

Upstream mechanisms

Although our data suggest an important interaction between CRFR1 signaling and Aβ, the mechanisms underlying this relationship are uncertain. To obtain a preliminary assessment of players involved, we used western blot analysis to assess whether overall levels of AβPP were altered by total or partial loss of CRFR1. We found no change in transgenic overexpression of AβPP in our CRFR1 mutant cohorts (PSAPP-R1^−/− and PSAPP-R1^+/− versus PSAPP, all p > 0.05) (Fig. 3). Control mice (WT, R1^−/− lanes) had expected basal levels of AβPP. In terms of Aβ production pathways, we observed dramatic reduction in AβPP CTFs in PSAPP-R1 mutant mice. In particular, a significant reduction in CTF-β in PSAPP-R1^−/− and PSAPP-R1^+/− (both p < 0.0001 compared to PSAPP) was observed but not CTF-α (both p > 0.05 compared to PSAPP), suggesting a CRF-dependent mechanism of Aβ production. CTFs are generated by secretase cleavage of AβPP. To probe whether CRFR1 is involved in modulation of degradation pathways of Aβ, we tested whether the expression of IDE, involved in the degradation of Aβ, was altered in PSAPP mice lacking CRFR1 (PSAPP versus PSAPP-R1^−/− and PSAPP-R1^+/− compared to PSAPP cohorts, both p < 0.01) (Fig. 3). Collectively, our data demonstrate that PSAPP-R1 mutant mice have reduced IDE and AβPP CTFs, suggesting the mechanisms underlying the observed change may involve Aβ production rather than clearance.

Figure 3.

Upstream mediators of the CRFR1-Aβ relationship. We found no change in transgenic overexpression of AβPP in our CRFR1 mutant cohorts (PSAPP-R1^−/− and PSAPP-R1^+/− versus PSAPP, all p > 0.05), indicating that observed reductions in Aβ in PSAPP-R1 mutant mice are not due to a decrease in transgenic AβPP expression. β-secretase activity on AβPP results in the generation of AβPP C-terminal fragments (CTFs) detected with the CT20 antibody, which are highly present at high levels in PSAPP mice. We found a significant reduction in CTFβ in PSAPP-R1^−/− and PSAPP-R1^+/− (both p < 0.0001 compared to PSAPP) but not CTFα (both p > 0.05 compared to PSAPP), suggesting a CRF-dependent mechanism of Aβ production. The Aβ degrading enzyme, IDE, was significantly reduced in PSAPP-R1^−/− and PSAPP-R1^+/− compared to PSAPP cohorts (p < 0.01, each), potentially indicating reduced activity due to lower levels of Aβ.

DISCUSSION

In this study, we developed a novel triple transgenic model to determine whether ablation of the major receptor involved in stress response, CRFR1, was feasible in an AD mouse model and determined the impact of this ablation on Aβ neuropathology. Our anatomical and biochemical findings demonstrate that accumulation of Aβ is significantly reduced in many important regions involved in learning and memory in animals lacking CRFR1. Importantly, our data provides an important mechanistic link between stress signaling intermediates in the CNS and the neuropathology of AD; CRFR1 ablation leads to significant and selective reduction of CTFβ, which suggests that the reduction in Aβ observed may involve modulation of secretase activity or trafficking. In terms of Aβ degradation, although confirmation with enzyme assays are needed, the reduction in resting levels of IDE signal that is observed with CRFR1 ablation may be due to reduced activity as a result of reduced presence of Aβ.

The role of CRF signaling in AD

Stress sensitivity and disorders are very common in AD patients, which are thought to mechanistically involve alterations in the hypothalamic-pituitary-adrenal axis (reviewed in [27]). Furthermore, anatomical and biochemical data indicate the involvement of CRF in the development of AD, with reduced cortical CRF immunoreactivity (in the face of increased hypothalamic expression) being a prominent neurochemical change [15, 28], a process which occurs early in disease progression in areas vulnerable to AD neuropathology [29–34]. CRF-positive dystrophic neurites have been found associated with Aβ plaques [30] and marked increases in CRF binding have been described in specific cortical regions of AD patients, suggesting upregulation of CRFRs in impacted areas [16]. It has been hypothesized that these changes result from changes in a specific CRF binding protein [35], which is expressed in brain [36] and can reversibly neutralize CRF bioactivity. Furthermore, studies in rodent models demonstrate that CRF overexpression can lead to tau phosphorylation and aggregation, brain atrophy, and cognitive impairment [13, 20–22].

In addition to stress and CRF, considerable attention has been focused on effectors of the stress cascade, such as glucocorticoids, as mediators of neuronal vulnerability in AD. Glucocorticoids are dominant stress hormones, whose increased levels with age have been linked to an enhanced neuronal vulnerability, notably in the hippocampus [37], a key learning and memory-related structure and a major target of both Aβ and tau pathology in AD [38]. Glucocorticoid (cortisol) levels in AD patients are elevated relative to age-matched controls, and several lines of evidence support a role of stress steroids in Aβ regulation (reviewed in [39]). Although glucocorticoids would appear to be an important target in AD, administration of steroids in human AD trials does not alter cognition (ADCS Prednisone Trial [40]) and results of studies from AD rodent models have been mixed [17, 23, 37, 41–46].

The role of stress CRF signaling regulation on Aβ production and processing pathways

Expression of β-secretase has been shown to be upregulated by several forms of cell stress such as hypoxia, ischemia, impaired glucose metabolism, and oxidative stress [47]. Such forms of physiologic stressors can be induced in vivo by exposure to chronic behavioral stressors. For example, increases in lipid peroxidation have been demonstrated in TG2576 mice that have endured repeated restraint [42]. These mice also exhibited enhanced aggregation of Aβ, an effect that could be blocked with acute administration of a CRFR1 antagonist [42].

Aβ is derived from AβPP, by the action of two aspartyl proteases, β- and γ-secretases [47–50]. Resultant fragments of AβPP are left behind in the membrane after proteolysis, called C-terminal fragments (CTFs). β-secretase (BACE-1) cleaves AβPP, generating the C99 fragment, which is also called CTF-β and comprises the N-terminus of Aβ [48, 50]. When γ-secretase cleaves CTF-β, Aβ is released and is found in several forms, the most abundant of which consist of 38, 40, or 42 amino acids [50]. An alternative, non-amyloidogenic pathway also exists and involves a third protease, α-secretase. Unlike BACE-1, α-secretase cuts within the Aβ domain of AβPP, generating the CTF-α fragment (aka C83). γ-secretase cleavage of CTF-α does not lead to Aµ generation [48, 50]. In terms of upstream mechanisms, we found a dramatic reduction in CTF-β but not CTF-α seen in PSAPP-R1 mutant mice, which implicates CRFR1 as a mediator of β or γ-secretase activity, expression and/or trafficking. Our future/ongoing studies are poised to address these issues. In particular, knowing whether CRFR1 is involved in modulation of β-secretase trafficking, such that it would no longer be localized to the locus of γ-secretase in endosomes, is of primary interest [51, 52].

CRFR1 regulation of both tau and Aβ pathways?

As summarized in the introduction, much of our work has been focused on stress-induced tau-P, its mechanistic regulation and relevance to AD [11, 12, 14]. Still unsettled is the precise role of stress-induced tau-P and its relevance to tau pathology in AD; studies propose that stress-induced tau-P is a functional component of neuroplasticity, conferring adaptation to stressful stimuli [53, 54], while our group and others suggest that the relationship between stress and tau-P may also define a potential means by which chronic stress exposure may translate into neuropathology [11, 55–58]. We hypothesize that stress-induced tau-P is an essential process required for stress adaptation that becomes dysfunctional with advancing age and/or with chronic overstimulation. Although the phenomenology may be distinct, a similar case can be made for toxicity associated with Aβ. Aβ is produced in the brain of throughout life, but with advancing age perhaps the ability to clear or process Aβ may become altered or lost, which causes accumulation. Although the upstream mechanisms involved remain to be precisely isolated, our data suggest that CRFR1 interruption impacts production pathways involved in Aβ, which may facilitate the development of new therapeutic strategies for AD.

Implications

In animal models, and in postmortem analysis of brain tissue from human patients, synaptic loss, apoptosis, and tau pathology reliably coincide with cognitive impairment and premature death [59–61]. A prevailing hypothesis in the field postulates that overall plaque burden may not correlate highly with memory loss associated with AD due to the accumulation of Aβ occurring early in the preclinical (asymptomatic) period. In support of this hypothesis, analysis of total Aβ (soluble oligomers and insoluble aggregates) in human brains has revealed that severity of Aβ-burden in the rhinal cortex correlates with severity of synaptic loss in AD [62]. Furthermore, data from studies employing commonly used fluorescent dyes that bind to β-pleated sheets in Aβ plaques are limited by their lack of specificity to Aβ itself and to non-neuritic plaques. The use of antibodies directed against Aβ, which recognize both dense core plaques and diffuse plaques, such 82E1 used in our study, have revealed correlations between plaque load and the behavioral sequelae of AD [63]. This suggests that diffuse or non-fibrillar aggregates of Aβ, which predominate in the areas of the brain where we found genotypic changes in plaque load, may be more toxic than neuritic plaques.

Although the hippocampus is well recognized as an area of learning and memory, the retrosplenial and rhinal cortices are also involved in learning and memory processes. Certain models of memory are negatively impacted by damage to areas of the rhinal cortex and the perirhinal cortex has shown to be important in novel-object recognition tasks (for review, [64]). The perirhinal cortex is also among the first cortical regions affected in AD [38, 65]. Similar pathology is found in other regions of the medial temporal lobe that are intimately connected with the perirhinal cortex, such as the entorhinal cortex [38], which sends projections to the hippocampus along the perforant pathway, which is known to be severely impacted by Aβ . The degree of severity of both Aβ and NFT lesions in this area correlate well with one another, as well as with cognitive decline [38, 66]. Some tests of memory may be more sensitive than others to the impact of Aβ along the rhinal and longitudinal fissures. For example, stress insults to the perirhinal cortex have been shown to impact tests of novel object recognition [67, 68]. Likewise, conditioned place aversion has been shown to be dependent on the insular and cingulate cortices [69, 70].

Future directions

Our future studies aim to determine the upstream mechanisms behind the CRFR1-Aβ relationship and the development of pharmacologic treatment for AD that targets central CRF pathways. Of particular interest here is our observation that the impact on Aβ did not differ significantly (i.e., no gene dose-dependent effect) between animals null for one or both alleles of CRFR1. These data suggest that perhaps total blockade of CRFR1 is not needed to achieve efficacy in this context, which is an important point to consider for future pharmacological testing. In terms of upstream mechanisms, we find the dramatic reduction in CTF-β but not CTF-α seen in PSAPP-R1 mutant mice particularly striking and consider a determination of whether CRFR1 is a mediator of β or γ-secretase activity, expression, and/or trafficking a priority. As mentioned earlier, the data presented here leave open the possibility of multiple upstream steps being involved, with the activity and trafficking of secretase enzymes being important to determine. Lastly, with our study demonstrating that CRFR1 ablation was feasible (i.e. not lethal), our upcoming studies will focus on the testing of small molecule selective antagonists to CRFR1 in the AD model.