The co-chaperone Fkbp5 shapes the acute stress response in the paraventricular nucleus of the hypothalamus of male mice

Disturbed activation or regulation of the stress response through the hypothalamic-pituitary-adrenal (HPA) axis is a fundamental component of multiple stress-related diseases, including psychiatric, metabolic, and immune disorders. The FK506 binding protein 51 (FKBP5) is a negative regulator of the glucocorticoid receptor (GR), the main driver of HPA axis regulation, and FKBP5 polymorphisms have been repeatedly linked to stress-related disorders in humans. However, the specific role of Fkbp5 in the paraventricular nucleus of the hypothalamus (PVN) in shaping HPA axis (re)activity remains to be elucidated. We here demonstrate that the deletion of Fkbp5 in Sim1+ neurons dampens the acute stress response and increases GR sensitivity. In contrast, Fkbp5 overexpression in the PVN results in a chronic HPA axis over-activation, and a PVN-specific rescue of Fkbp5 expression in full Fkbp5 KO mice normalizes the HPA axis phenotype. Single-cell RNA sequencing revealed the cell-type-specific expression pattern of Fkbp5 in the PVN and showed that Fkbp5 expression is specifically upregulated in Crh+ neurons after stress. Finally, Crh-specific Fkbp5 overexpression alters Crh neuron activity, but only partially recapitulates the PVN-specific Fkbp5 overexpression phenotype. Together, the data establish the central and cell-type-specific importance of Fkbp5 in the PVN in shaping HPA axis regulation and the acute stress response.


Introduction
Life is full of challenges and appropriate coping with such events implies proper activation and termination of the stress response. The hypothalamic-pituitary-adrenal (HPA) axis is the central orchestrator of the stress response and its end product glucocorticoids (cortisol in humans, corticosterone (CORT) in rodents) mediates the adaptation to acute and chronic stressors in peripheral tissues as well as in the brain [1]. A hallmark of HPA axis regulation is the negative feedback on the secretion of stress hormones to terminate the stress response which is controlled by glucocorticoids via the glucocorticoid receptor (GR).
A critical regulator of GR and therefore key to a successful termination of the stress response is FK506 binding protein 51 (FKBP5), which is encoded by the FKBP5 gene [2]. FKBP5 is an Hsp90-associated co-chaperone that restricts GR function by reducing ligand binding, delaying nuclear translocation, and decreasing GR-dependent transcriptional activity [3,4]. Hence, higher levels of FKBP5 mRNA are associated with higher levels of circulating cortisol and reduced negative feedback inhibition of the stress response [5][6][7][8][9]. Consequently, GR-induced FKBP5 levels reflect the environmental stress condition, and as such, FKBP5 expression has been used as a stress-responsive gene marker [10]. Importantly, FKBP5 polymorphisms have been consistently associated with stress-related psychiatric disorders such as major depression and PTSD [11][12][13], where a demethylation-mediated increase in FKBP5 expression was identified as causal in risk-allele carriers [14].
Despite the central importance of FKBP5 in stress system biology and stress-related disorders, detailed functional and mechanistic studies are still largely missing. Only a few human post-mortem studies focus on central tissue in order to dissect FKBP5 mechanisms [15,16], while the majority of studies use peripheral blood mononuclear cells as a correlate of FKBP5 brain activity [17]. Most of the animal data were obtained from wild-type (WT) or conventional Fkbp5 knockout mice, thereby lacking cell-type-specific insights of Fkbp5 function [8,18]. To tackle this paucity of information, we here investigate the specific role of Fkbp5 in the paraventricular nucleus of the hypothalamus (PVN) the key brain region orchestrating the stress response [19]. Interestingly, previous data in rats suggested a potentially decisive role of Fkbp5 expression in the PVN to drive HPA axis hyperactivity achieved by a selective breeding approach [20]. Using site-specific manipulations of Fkbp5, single-cell RNA expression profiling, and functional downstream pathway analyzes, our data unravel a key role of PVN Fkbp5 in shaping the body's stress system (re) activity, with important implications for its contribution to stress-related disorders.

Animals and animal housing
All experiments were performed in accordance with the European Communities' Council Directive 2010/63/EU. The protocols were approved by the committee for the care and use of laboratory animals of the Government of Upper Bavaria, Germany. The mouse lines Fkbp5 lox/lox , Fkbp5 PVN−/− , Fkbp5 Frt/Frt , and CRH-ires-CRE/Ai9 were generated in house. Experiments were performed on male mice aged between 3 and 5 months unless stated otherwise in the results section. During the experimental time, animals were kept singly housed in individually ventilated cages (IVC; 30 cm × 16 cm × 16 cm; 501 cm 2 ), serviced by a central airflow system (Tecniplast, IVC Green Line-GM500). Animals were maintained on a 12:12 h light/dark cycle, with constant temperature (23 ± 2°C) and humidity of 55% during all times. Experimental animals received ad libitum access to water and a standard research diet (Altromin 1318, Altromin GmbH, Germany) and IVCs had sufficient bedding and nesting material as well as a wooden tunnel for environmental enrichment. Animals were allocated to experimental groups in a semi-randomized fashion and data analysis was performed blinded to the group allocation.

Generation of Fkbp5 mice
Conditional Fkbp5 knockout mice are derived from embryonic stem cell clone EPD0741_3_H03 which was targeted by the knockout mouse project (KOMP). Frozen sperm obtained from the KOMP repository at UC Davis was used to generate knockout mice (Fkbp5 tm1a(KOMP)Wtsi ) by in vitro fertilization. These mice designated as Fkbp5 Frt/Frt are capable to re-express functional Fkbp5 upon Flp recombinase-mediated excision of an frt-flanked reporter-selection cassette integrated into the Fkbp5 gene. Mice with a floxed Fkbp5 gene designated as Fkbp5 lox/lox (Fkbp5 tm1c(KOMP)Wtsi ) were obtained by breeding Fkbp5 Frt/Frt mice to Deleter-Flpe mice [21]. Finally, mice lacking Fkbp5 in PVN neurons (Fkbp5 PVN−/− ) were obtained by breeding Fkbp5 lox/lox mice to Sim1-Cre mice [22]. The CRH-ires-CRE/Ai9 mouse line was generated previously [23]. Genotyping details are available upon request.

Viral overexpression and rescue of Fkbp5
For overexpression and rescue experiments, stereotactic injections were performed as described previously [24]. In brief, mice aged between 10 and 12 weeks were anesthetized with isoflurane, and 0.2 μl of the below-mentioned viruses (titers: 1.6 × 10 12-13 genomic particles/ml) were bilaterally injected in the PVN at 0.05 μl/min by glass capillaries with tip resistance of 2-4 MΩ in a stereotactic apparatus. The following coordinates were used: −0.45 mm anterior to bregma, 0.3 mm lateral from midline, and 4.8 mm below the surface of the skull, targeting the PVN. After surgery, mice were treated for 3 days with Metacam via i.p. injections and were allowed to recover for 3-4 weeks before the start of the testing. Successful overexpression or reinstatement of Fkbp5 was verified by ISH and RNAscope. For in vivo experiments, we used adeno-associated bicistronic AAV1/2 vectors. In the overexpression experiments, the vector contained a CAG-HA-tagged-FKBP51-WPRE-BGH-polyA expression cassette (containing the coding sequence of human Fkbp51 NCBI CCDS ID CCDS4808.1). The same vector constructs without expression of Fkbp5 (CAG-Null/Empty-WPRE-BGH-polyA) was used as a control. For the Crh specific overexpression of Fkbp5, we used a viral vector containing an AAV1/2-Cre-dept-HA-FKBP51 (pAAV-Cre-dependent-CAG-HA-human wildtype FKBP51 WPRE-BGH-polyA) and a control virus containing AAV1/2-Cre-dept-GFP (pAAV-Cre-dependent-CAG-GFP-WPRGE-BGH-polyA). Virus production, amplification, and purification were performed by GeneDetect. For the rescue experiment, a viral vector containing a flippase expressing cassette (AAV2-eSYN-eGFP-T2A-FLPo, Vector Biolabs; VB1093) was used to induce endogenous Fkbp5 expression in Fkbp5 FRT/FRT mice. Control animals were injected with a control virus (AAV2-eSYN-eGFP; Vector Biolabs; VB1107).

Acute stress paradigm
For acute stress exposure, the restraint stress paradigm was used, as it was shown to be a reliable and robust stressor in rodents [25]. One to 2 h after lights on each animal was placed in a restrainer (50 ml falcon tube with holes at the bottom and the lid to provide enough oxygen and space for tail movement) for 15 min in their individual home cage. After 15 min, animals were removed from the tube and the first blood sample was collected by tail cut. Until the following tail cuts at 30, 60, and 90 min after stress onset, the animals remained in their home cage to recover. Basal CORT levels (morning CORT) were collected one week prior to the acute stress paradigm at 8 a.m.

Combined Dex/CRH test
To investigate the HPA axis function we performed a combined Dex/CRH test as described previously [8]. On an experimental day, mice were injected with a low dose of dexamethasone (0.05 mg/kg, Dex-Ratiopharm, 7633932) via i.p injections at 9 a.m. in the morning. At this dose Dex does not cross the blood-brain barrier and acts predominantly in the periphery [26] Here, the most important site of action in relation to HPA axis function, especially when combined with a challenge with the neuropeptide CRH, is the pituitary. Thus, 6 h after Dex injection, a blood sample was collected via tail cut (after Dex value), followed by an injection of CRH (0.15 mg/kg, CRH Ferrin Amp). Thirty-minute after CRH injection, another blood sample was obtained (after CRH value). All samples from the acute stress experiments and the Dex/CRH test were collected in 1.5 ml EDTA-coated microcentrifuge tubes (Sarstedt, Germany). All blood samples were kept on ice and were centrifuged for 15 min at 8000 rpm and 4°C. Plasma was transferred to new, labeled microcentrifuge tubes and stored at −20°C until further processing.

Sampling procedure
On the day of sacrifice, animals were deeply anesthetized with isoflurane and sacrificed by decapitation. Trunk blood was collected in labeled 1.5 ml EDTA-coated microcentrifuge tubes (Sarstedt, Germany) and kept on ice until centrifugation. After centrifugation (4°C, 8000 rpm for 1 min) the plasma was removed and transferred to new, labeled tubes and stored at −20°C until hormone quantification. For mRNA analysis, brains were removed, snapfrozen in isopentane at −40°C, and stored at −80°C for ISH. For protein analysis, brains were removed and placed inside a brain matrix with the hypothalamus facing upwards (spacing 1 mm, World Precision Instruments, Berlin, Germany). Starting from the middle of the chiasma opticum, a 1 mm thick brain slice was removed. The PVN was further isolated by cutting the slice on both sides of the PVN along the fornices (parallel to the 3rd ventricle) as a landmark and a horizontal cut between the reuniens and the thalamic nucleus. Finally, the slice containing the third ventricle was bisected, and the distal part discarded. The remaining part (containing the PVN) was immediately shocked frozen and stored at −80°C until protein analysis [27]. The adrenals and thymus glands were dissected from fat and weighed.

Hormone assessment
CORT and ACTH concentrations were determined by radioimmunoassay using a corticosterone double-antibody 125 I RIA kit (sensitivity: 12.5 ng/ml, MP Biomedicals Inc.) and adrenocorticotropic double-antibody hormone 125 I RIA kit (sensitivity: 10 pg/ml, MP Biomedicals Inc.) and were used following the manufacturers' instructions. Radioactivity of the pellet was measured with a gamma counter (Packard Cobra II Auto Gamma; Perkin-Elmer). Fifteenminute post-stress analysis of ACTH concentrations was analyzed by using ACTH ELISA (IBL international GmBH, RE53081) and the ELISA was performed as recommended by the manufacturer. Final CORT and ACTH levels were derived from the standard curve.

In situ hybridization (ISH)
ISH was used to analyze mRNA expression of the major stress markers, Fkbp5, Gr, Crh, and Avp. Therefore, frozen brains were sectioned at −20°C in a cryostat microtome at 20 μm, thaw mounted on Super Frost Plus slides, dried, and stored at −80°C. The ISH using 35 S UTP labeled ribonucleotide probes was performed as described previously [28,29]. All primer details are available upon request. Briefly, sections were fixed in 4% paraformaldehyde and acetylated in 0.25% acetic anhydride in 0.1 M triethanolamine/HCl. Subsequently, brain sections were dehydrated in increasing concentrations of ethanol. The antisense cRNA probes were transcribed from a linearized plasmid. Tissue sections were saturated with 100 μl of hybridization buffer containing approximately 1.5 × 10 6 cpm 35 S labeled riboprobe. Brain sections were coverslipped and incubated overnight at 55°C. On the next day, the sections were rinsed in 2 × SSC (standard saline citrate), treated with RNAse A (20 mg/l). After several washing steps with SSC solutions at room temperature, the sections were washed in 0.1 × SSC for 1 h at 65°C and dehydrated through increasing concentrations of ethanol. Finally, the slides were air-dried and exposed to Kodak Biomax MR films (Eastman Kodak Co., Rochester, NY) and developed. Autoradiographs were digitized, and expression was determined by optical densitometry utilizing the freely available NIH ImageJ software. The mean of four measurements of two different brain slices was calculated for each animal. The data were analyzed blindly, always subtracting the background signal of a nearby structure not expressing the gene of interest from the measurements. For Fkbp5, exemplary slides were dipped in Kodak NTB2 emulsion (Eastman Kodak Co., Rochester, NY) and exposed at 4°C for representative pictures; exposure time was adjusted to average expression level. Slides were developed and examined with a light microscope with darkfield condensers to show mRNA expression.

RNAscope analysis and cell counting
For the RNAscope experiments, C57Bl/6 male mice were obtained from The Jackson Laboratory (Bar Harbor, ME, USA). All procedures conformed to National Institutes of Health guidelines and were approved by McLean Hospital Institutional Animal Care and Use Committee. Mice were housed in a temperature-controlled colony in the animal facilities of McLean Hospital in Belmont, MA, USA. All mice were group-housed and maintained on a 12:12 h light/ dark cycle (lights on at 7 a.m.). Food and water were available ad libitum unless specified otherwise. Mice were 12 weeks at the time of tissue collection. Animals were allowed to acclimate to the room for 1 week before the beginning of the experiment. During the experiment, mice were either left undisturbed (ctrl) or subjected to 14 h (overnight) of food deprivation prior to sacrifice. During the stress procedure, animals were kept in their home cages and had free access to tap water. All mice were sacrificed by decapitation in the morning (08:00 to 08:30 a.m.) following quick anesthesia by isoflurane. Brains were removed, snapfrozen in isopentane at −40°C, and stored at −80°C. Frozen brains were sectioned in the coronal plane at −20°C in a cryostat microtome at 18 μm, mounted on Super Frost Plus slides, and stored at −80°C. The RNA Scope Fluorescent Multiplex Reagent kit (cat. no. 320850, Advanced Cell Diagnostics, Newark, CA, USA) was used for mRNA staining. Probes used for staining were; mm-Avp-C3, mm-Crh-C3, mm-Fkbp5-C2, mm-Oxt-C3, mm-Sst-C3, and mm-Trh-C3. The staining procedure was performed according to the manufacturer's specifications. Briefly, sections were fixed in 4% paraformaldehyde for 15 min at 4°C. Subsequently, brain sections were dehydrated in increasing concentrations of ethanol. Next, tissue sections were incubated with protease IV for 30 min at room temperature. Probes (probe diluent (cat. no. 300041 used instead of C1-probe), Fkbp5-C2, and one of the above C3-probes) were hybridized for 2 h at 40°C followed by four hybridization steps of the amplification reagents 1-4. Next, sections were counterstained with DAPI, cover-slipped, and stored at 4°C until image acquisition. Images of the PVN (left and right side) were acquired by an experimenter blinded to the condition of the animals. Sixteen-bit images of each section were acquired on a Leica SP8 confocal microscope using a 40× objective (n = 3 animals per marker and condition). For every individual marker, all images were acquired using identical settings for laser power, detector gain, and amplifier offset. Images of both sides were acquired as a zstack of 3 steps of 1.0 µm each. Fkbp5 mRNA expression and co-expression were analyzed using ImageJ with the experimenter blinded to the condition of the animals. Fkbp5 mRNA was counted manually and each cell containing 1 mRNA dot was counted as positive.

Single-cell sequencing. Tissue dissociation
Single-cell sequencing was performed on PVN tissue dissected from male C57Bl/6 mice aged between 8 and 12 weeks. Therefore, mice were anesthetized lethally using isoflurane and perfused with cold PBS. Brains were quickly dissected, transferred to ice-cold oxygenated artificial cerebral spinal fluid (aCSF), and kept in the same solution during dissection. Sectioning was performed using a 0.5 mm stainless steel adult mouse brain matrice (Kent Scientific) and a Personna double edge prep razor blade. A slide (approximately −0.58 mm Bregma to −1.22 mm Bregma) was obtained from each brain and the extended PVN was manually dissected under the microscope. The PVN from five different mice was pooled and dissociated using the Papain dissociation system (Worthington) following the manufacturer's instructions. All solutions were oxygenated with a mixture of 5% CO 2 in O 2 . After this, the cell suspension was filtered with a 30 μm filter (Partec) and kept in cold and oxygenated aCSF.
Cell capture, library preparation, and highthroughput sequencing Cell suspensions of PVN with~1,000,000 cells/μL were used. Cells were loaded onto two lanes of a 10X Genomics Chromium chip per factory recommendations. Reverse transcription and library preparation was performed using the 10X genomics single-cell v2.0 kit following the 10X genomics protocol. The library molar concentration and fragment length were quantified by qPCR using KAPA Library Quant (Kapa Biosystems) and Bioanalyzer (Agilent high sensitivity DNA kit), respectively. The library was sequenced on a single lane of an Illumina HiSeq4000 system generating 100 bp paired-end reads at a depth of~340 million reads per sample.

Quality control and identification of cell clusters
Pre-processing of the data was done using the 10X genomics cell ranger software version 2.1.1 in default mode. The 10X genomics supplied reference data for the mm10 assembly and corresponding gene annotation was used for alignment and quantification. All further analysis was performed using SCANPY version 1.3.7 [30]. A total of 5.113 cells were included after filtering gene counts (<750 and >6.000), UMI counts (>25.000), and the fraction of mitochondrial counts (>0.2). Combat [31] was used to remove chromium channels as batch effects from normalized data. The 4.000 most variable genes were subsequently used as input for Louvain cluster detection. Cell types were determined using a combination of marker genes identified from the literature and gene ontology for cell types using the web-based tool: mousebrain.org (http://mousebrain.org/ genesearch.html).
Subsequently, blots were washed and probed with the respective horseradish peroxidase or fluorophoreconjugated secondary antibody for 1 h at room temperature. The immuno-reactive bands were visualized either using ECL detection reagent (Millipore, Billerica, MA, USA) or directly by excitation of the respective fluorophore. Determination of the band intensities was performed with BioRad, ChemiDoc MP. For quantification of phosphorylated GR, the intensity of phosphor-GR was always referred to as the signal intensity of the corresponding total GR.

For ChIP analysis
Aliquots of chromatin were diluted ten-times in ice-cold Dilution Buffer [ , and DNA was purified using a Qiagen PCR purification kit. All samples (bounds and inputs) were diluted to a standardized concentration with nuclease-free water and analyzed by qPCR as described below using primers/probes (forward: 5′-TGTCAATGGACAAGTCATAAGAAACC; reverse: 5′-GAATCTCACATCCAATTATATCAACAGAT; probe: 5′-TTCCATTTTCGGGCTCGTTGACGTC). The binding of GR was expressed as a percentage of input DNA, i.e., % Input, which is a measure of the enrichment of steroid receptor bound to specific genomic sequences.

qPCR analysis
Mastermix for qPCR was prepared to contain 900 nM forward and reverse primers, 200 nM probe, 1X TaqMan fast mastermix (Life Technologies, Paisley, UK), and nucleasefree water. Primers and dual-labeled probes with 6-FAM as the fluorescent dye and TAMRA are the quencher were designed using Primer Express software (Version 3.0.1, Life Technologies). Standard curves were performed for each primer pair and the qPCR efficiency was calculated using the equation: E = ((10 − 1/slope) − 1) × 100 (where E is qPCR efficiency and the slope is the gradient of the standard curve). Only primer pairs with efficiencies greater than 90% were used. Quantitative PCR was performed using a StepOne Plus machine (Life Technologies, Paisley, UK). Taq enzymes were activated at 95°C for 20 s, then 40 cycles of 95°C (1 s) to 60°C (20 s) were performed to amplify samples.

Statistical analysis
The data presented are shown as means ± SEM and sample sizes are indicated in the figure legends. All data were analyzed by the commercially available software SPSS 17.0 and GraphPad 8.0. When two groups were compared, the unpaired student's t-test was applied. If data were not normally distributed the non-parametric Mann-Whitney test (MW-test) was used. For four group comparisons, a twoway analysis of variance (ANOVA) was performed, followed by a posthoc test, as appropriate. P values of less than 0.05 were considered statistically significant. The sample size was chosen such that with a type 1 error of 0.05 and a type 2 error of 0.2 the effect size should be at least 1.2-fold of the pooled standard deviation. All data were tested for outliers using the Grubbs test.

Loss of Fkbp5 in the PVN alters HPA axis physiology
To study the effects of Fkbp5 in the PVN, we first generated PVN-specific conditional Fkbp5 knockout mice (Fkbp5 PVN−/− ) by crossing Fkbp5 lox/lox mice (generated in-house; for details see methods) with the Sim1-Cre mouse line, which expresses Cre recombinase in Sim1 + neurons mostly concentrated within the PVN (Fig. 1A). The successful deletion of Fkbp5 in the PVN was assessed by mRNA and protein analysis ( Fig. 1B and Supplementary Fig. 1). Under basal conditions, adult mice (16-20 weeks of age) lacking Fkbp5 in the PVN showed significantly lower body-, and adrenal weights and higher thymus weights compared to their WT littermates (Fig. 1C). Interestingly, Fkbp5 PVN−/− mice in young adolescence (8 weeks of age) showed no significant differences in body weight, adrenal, or thymus weights ( Supplementary Fig. 2), indicating an agedependent phenotype of Fkbp5 in the HPA-axis' response.
As PVN Fkbp5 mRNA levels are highly responsive to an acute stressor [10], we hypothesized that Fkbp5 PVN−/− mice have an altered stress response following an acute challenge. Under basal conditions during the circadian trough,  Supplementary Fig. 1A). C Fkbp5 PVN-/mice (n = 16) presented reduced body weight, lowered adrenal weights, and increased thymus weights under non-stressed conditions compared to their WT littermates (n = 15). D Corticosterone levels were significantly reduced following a 15 min restraint stress until at least 60 min after stress onset. E A combined Dex/CRH test showed a significantly pronounced response to a low dose of dexamethasone as well as a dampened response to CRH injection. Data are received from mice between 16 and 20 weeks of age and are presented as mean ± SEM. All data were analyzed with a student's t-test. *p < 0.05, **p < 0.01, and ***p < 0.001. no differences in corticosterone secretion were detected in young and adult mice (Supplementary Figs. 1 and 2). However, already after 15 min of restraint stress adult and young Fkbp5 PVN−/− mice displayed significantly reduced plasma corticosterone levels compared to the control group (Fig. 1D). The dampened stress response was persistent for 60 min in adult mice, while the effect has vanished in young mice already 30 min after stress onset (Fig. 1D and Supplementary Fig. 2). Levels of the adrenocorticotropic hormone (ACTH) were not altered under basal or acute stressed conditions (Supplementary Fig. 1).
To further investigate the effect of Fkbp5 on GR sensitivity, we performed combined dexamethasone (Dex, a synthetic GC)-corticotropin-releasing hormone (CRH) test on adults Fkbp5 PVN−/− mice. The combined Dex/CRH test is a method to analyze HPA axis (dys)function in depressed individuals or animals, measuring the responsiveness of the body's stress response system through suppression (by Dex injection) and stimulation (by CRH injection) of the HPA axis [33]. The injection of a low dose of Dex (0.05 mg/kg) resulted in a reduction in blood corticosterone levels compared to the evening corticosterone levels in both groups. Interestingly, mice lacking Fkbp5 in the PVN showed 1.5fold lower levels of corticosterone compared to their WT littermates. Following CRH stimulation (0.15 mg/kg) Fkbp5 PVN−/− mice showed a significantly lower reaction to CRH than control mice (Fig. 1E). These data indicate that specific deletion of Fkbp5 in the PVN dampens HPA axis response and enhances GR sensitivity.

Overexpression of Fkbp5 in the PVN induces a stress-like phenotype in C57Bl/6 mice under basal conditions
Chronic or acute stress upregulates Fkbp5 in distinct brain regions, such as the PVN [10]. Therefore, we were interested to explore whether selective overexpression of Fkbp5 in the PVN would be sufficient to affect body physiology and stress system regulation. To do so, we bilaterally injected 200 nl of an adeno-associated virus (AAV) containing an Fkbp5 overexpression vector into the PVN of 10-12 weeks old C57Bl/6 mice (Fkbp5 PVN OE , Fig. 2A, B). The AAV-mediated overexpression resulted in a (fourfold) increase in Fkbp5 mRNA and protein levels (Fkbp5) in the PVN (Supplementary Fig. 3).
Intriguingly, Fkbp5 overexpression altered the physiology of stress-responsive organs. Fkbp5 PVN OE animals showed a significantly reduced thymus weight and increased adrenal weights compared to their littermates (Fig. 2C), the hallmark of chronically stressed animals [34,35]. Furthermore, overexpression of Fkbp5 affected the circadian rhythm of corticosterone secretion, indicated by increased blood corticosterone levels in the morning as well as the evening (Fig. 2D). Consequently, ACTH levels of Fkbp5 PVN OE animals were also increased under nonstressed conditions (Supplementary Fig. 3). Next, we analyzed distinguished stress markers under basal conditions in order to determine whether consequences of PVN-specific Fkbp5 overexpression are also detectable at the molecular level. Nr1c3 and Crh mRNA expression in the PVN were increased in Fkbp5 PVN OE animals under basal conditions; however, to a lesser extent than the increase of Fkbp5 mRNA due to viral overexpression. Avp mRNA levels were not altered (Supplementary Fig. 3). Together, these results are comparable to chronically stressed animals and demonstrate that local overexpression of Fkbp5 in the PVN is sufficient to mimic a stress-like phenotype without physically challenging the animals.
In accordance with the knock-out studies of the Fkbp5 PVN−/− animals, we investigated the endocrinology of Fkbp5 PVN OE animals after an acute challenge. As expected, we detected higher blood corticosterone levels in Fkbp5 PVN OE mice 15 and 30 min after stress onset compared to the stressed controls (Fig. 2E). However, we could not detect any difference in ACTH release after stress ( Supplementary  Fig. 3). No differences between both groups were observed 60 and 90 min after restraint stress ( Fig. 2E and Supplementary Fig. 3). These data show that Fkbp5 PVN OE mice have a hyperactive HPA axis response and are more vulnerable to acute stress exposure.
To further assess GR sensitivity in Fkbp5 overexpressing animals, we again tested the response to a combined Dex/ CRH test. While control animals showed a decline (<5 ng/ ml) in blood corticosterone levels 6 h after Dex injection, Fkbp5 PVN OE mice showed almost no response to Dex treatment (Fig. 2F). Interestingly, the subsequent CRH injection resulted in a higher corticosterone release in Fkbp5 PVN OE mice compared to controls (Fig. 2F). These results suggest that excess levels of Fkbp5 in the PVN lead to a decreased GR sensitivity and thereby to an altered HPA axis response. Taken together, animals overexpressing Fkbp5 in the PVN show a hyperactive function of the HPA axis under basal and acute stress conditions, thereby mimicking the physiological hallmarks of chronic stress exposure and HPA axis hyperactivity, as observed in multiple stress-related diseases [36].

Reinstatement of endogenous Fkbp5 in the PVN of global Fkbp5 knock-out animals normalizes the body's stress response
Global loss of Fkbp5 results in a more sensitive GR and better-coping behavior of mice after stress [8,17,28,34], and our results demonstrated that Fkbp5 in the PVN is necessary for an undisturbed stress system function. Thus, we were encouraged to test whether the reinstatement of native Fkbp5 expression only in the PVN is sufficient to push the HPA axis activity of global Fkbp5 knock-out animals to a wildtype level. Therefore, we injected an Flp recombinase expressing virus into 12-14 weeks old Fkbp5 Frt/Frt mice. These mice carry an FRT flanked reporter selection (stop) cassette within the Fkbp5 locus, leading to a disruption of the Fkbp5 function. We compared Fkbp5 Frt/Frt mice to WT littermates and observed a similar HPA-axis phenotype to the well-established Fkbp5 full KO lines ( [8] and Supplementary Fig. 4). An Flp removes the stop cassette from the Fkbp5 locus, resulting in endogenous Fkbp5 re-expression (Fig. 3A, B, Supplementary Fig. 5). In parallel to the two previous mouse models, we assessed body physiology, basal corticosterone levels, and the acute stress response. Interestingly, mice with re-instated Fkbp5 expression (Fkbp5 PVN Rescue ) showed significantly higher adrenal weights as compared to their control littermates (Fig. 3C). Furthermore, we observed that the reinstatement of Fkbp5 in the PVN resulted in significantly increased blood CORT levels in the morning under basal conditions (Fig. 3D), with no effect on thymus weights, evening CORT, and ACTH levels ( Supplementary Fig. 5). ISH analysis revealed significantly higher levels of Crh mRNA, but no changes in Nr1c3 and Avp mRNA expression in the onset, Fkbp5 PVN OE mice displayed significantly higher corticosterone levels. F Fkbp5 PVN OE mice showed significantly elevated corticosterone 6 h after dexamethasone treatment. The following CRH injection further significantly increased the corticosterone release compared to controls. Data are presented as mean ± SEM. All data were received from mice between 14 and 20 weeks of age and analyzed a with student's t-test. *p < 0.05, **p < 0.01, ***p < 0.001, n.s. not significant. Fig. 5). Next, we monitored blood corticosterone levels after 15 min of restraint stress. Here, we observed significantly higher corticosterone levels 15 and 60 min after stress onset (Fig. 3E). No differences were detected in the combined Dex/CRH test ( Supplementary Fig. 5), which suggests that a PVN-driven over-activation of the HPA axis might be necessary for desensitization of GRs in the PVN and the pituitary. These rescue experiments underline the importance of Fkbp5 in the acute stress response and demonstrate that Fkbp5 in the PVN is necessary and sufficient to regulate HPA axis (re)activity.

Fkbp5 manipulation directly affects GR phosphorylation
It is well known that ligand-binding induced phosphorylation of GR plays an important role in response to hormone signaling [37]. The main phosphorylation sites involved in hormone signaling of GR are Serine(Ser) 203 (mouse S 212 ), Ser 211 (mouse Ser 220 ), and Ser 226 (mouse Ser 234 ) and are associated with GR activity [37,38]. Here, we tested the hypothesis that the co-chaperone Fkbp5 regulates phosphorylation of GR in Under basal conditions, animals lacking Fkbp5 in the PVN showed significantly less GR phosphorylation at Ser 203 (Fig. 4A). Furthermore, Fkbp5 PVN−/− animals displayed higher phosphorylation of GR at Ser 234 and Ser 211 in comparison to their WT littermates (Fig. 4B, C). Under stressed conditions, deletion of Fkbp5 had the same effects on pGR Ser211 and pGR Ser234 as we observed under basal conditions (Fig. 4B, C). Levels of pGR Ser203 were found to be unchanged in the Fkbp5 PVN−/− after acute stress compared to the basal levels. However, pGR Ser203 levels of the control group increased after stress (Fig. 4A).
Intriguingly, Fkbp5 PVN OE animals showed exactly the opposing phenotype at all three phosphorylation sites with less pGr Ser234 and pGr Ser211 and higher phosphorylation at Ser 203 under basal conditions (Fig. 4E-G). In parallel to the unstressed condition, the overexpression of Fkbp5 resulted in less GR phosphorylation at Ser 211 and higher levels of pGR Ser203 compared to their control group after stress (Fig. 4E, F). Interestingly, levels of pGR Ser234 were unchanged after stress (Fig. 4G). Notably, total GR levels were not significantly altered in both experimental groups and conditions (Fig. 4D, H). Despite the altered GR phosphorylation, we could not detect any significant changes in  Supplementary Fig. 4. Data are presented as mean ± SEM. All data were received from mice between 16 and 20 weeks of age and analyzed a with student's t-test. *p < 0.05. GR enrichment at the glucocorticoid response element (GRE) in the Crh gene after acute stress ( Supplementary  Fig. 6), which may be due to the use of an antibody that recognizes all GR molecules irrespective of its phosphorylation state.
Overall, our data demonstrate that Fkbp5 manipulation in the PVN affects GR phosphorylation at all three major phosphorylation sites and thereby affects GR activity.

Fkbp5 in the PVN acts in a complex cellular context
To further unravel the expression profile of Fkbp5 in the PVN and to detect cellular populations that might be mediating the effects of Fkbp5 on HPA axis control, we used a single-cell RNA sequencing dataset consisting of 5113 single cells isolated from the PVN of C57Bl/6 male mice [39]. The single-cell expression data reveal a complex cellular composition, with the majority of cells identified as neurons (38%), ependymal cells (25%), and astrocytes (14%) (Fig. 5A-C). Fkbp5 was found to be differentially and cell-type specifically expressed, with the biggest Fkbp5 + cell population found in GABAergic neurons (42%). A significant expression of Fkbp5 was also detected in neuronal populations known to be directly involved in HPA axis regulation, most prominently in Crh positive neurons (Fig. 5D). However, it is important to point out that the expression levels of Fkbp5 are relatively low. Unfortunately, lowly expressed genes may not be detected using this technique [40] and therefore many Fkbp5 positive cells may have remained undetected in this dataset. To circumvent this problem, we next performed a targeted coexpression study of Fkbp5 with five major markers that are characteristic of the stress response oxytocin (Oxt), somatostatin (Sst), vasopressin (Avp), thyrotropin-releasing hormone (Trh), and Crh under basal and stress conditions (Fig. 5E, Supplementary Fig. 7). We observed a strong but not complete co-localization of Fkbp5 expression with these neuropeptide-expressing cellular populations in the PVN under basal conditions. Interestingly, a detailed quantification of the change in co-expression following stress revealed that there was a significant increase in Crh-Fkbp5 co-localization only in the Crh-expressing neurons (Fig. 5E, F and Supplementary Fig. 8).
These data reveal the complex cell-type-specific expression pattern of Fkbp5 under stress and basal conditions in the PVN. The significant increase of FKBP51 in Fig. 4 Fkbp5 manipulation affects phosphorylation of the glucocorticoid receptor (GR). A Animals lacking Fkbp5 in the PVN showed significantly lower phosphorylation at pGR Ser203 and higher levels of (B) pGR Ser211 and (C) pGR Ser234 compared to the control animals. D Fkbp5 PVN−/− mice showed no differences in total GR. E Fkbp5 PVN OE animals showed the opposite effect on GR phosphorylation with higher phosphorylation on pGr Ser203 . F In addition, we observed significantly lower phosphorylation at the GR sites Ser 211 and (G) Ser 234 . H Fkbp5 overexpression had no effect on total GR protein level. Representative blots are shown in (D) and (H). Group size for (A)-(H): 6 vs. 6. Data were received from animals between 16 and 20 weeks of age and are presented as relative fold change to control condition, mean ± SEM, and were analyzed with a two-way ANOVA. *significant genotype effect, (*p < 0.05, **p < 0.01, ***p < 0.001). +significant genotype × stress interaction ( + p < 0.05), # significant stress effect ( # p < 0.05, ## p < 0.01).
Crh positive neurons after an acute stress challenge encouraged us to specifically manipulated FKBP51 in Crh positive neurons.
Crh-specific overexpression of Fkbp5 in the PVN alters HPA axis physiology and CRH neuronal activity Based on the observed increase in Crh-Fkbp5 colocalization post-stress (Fig. 5E, F and Supplementary   Fig. 8), we were interested whether a Crh-specific Fkbp5 overexpression in the PVN could drive the stress-like phenotype observed in the unspecific PVN overexpression (Fig. 2) and whether this neuron-specific manipulation alters CRH neuronal activity. Therefore, we bilaterally injected 200 nl of an AAV containing a Cre-dependent Fkpb5 overexpression vector into the PVN of adult (26 weeks) CRH-ires-CRE/Ai9 mice expressing tdTomato specifically in CRH neurons (Fkbp5 CRH OE ) (Fig. 6A). Credependent AAV mediated overexpression resulted in a , and astrocytes (14%). Fkbp5 + cells are highlighted. B Diversity of Fkbp5 + cell population. C Neurons could be divided mostly into GABAergic (66%) and glutamatergic (Glut, 11%) cells. Furthermore, the well-known stress markers corticotropinreleasing hormone (Crh, 6%), somatostatin (Sst, 5%), oxytocin (Oxt, 2%), and vasopressin (Avp, 1%) could be detected under basal conditions. D Fkbp5 + cells of selected neuronal subpopulations under basal conditions. E RNAscope validation of Fkbp5 mRNA expression in Crh + neurons. F Quantity of Fkbp5 + cells, as well as Fkbp5 mRNA levels, are significantly increased after stress. Data are were received from animals between 8 and 12 weeks of age and are presented as mean ± SEM. For (E): data were analyzed with student's ttest. ***p < 0.001. fourfold increase of Fkbp5 mRNA level in the PVN and Crh specificity of Fkbp5 overexpression was successfully confirmed by RNAscope (Fig. 6B and Supplementary  Fig. 9E). Under basal conditions, Fkbp5 CRH OE mice displayed significantly increased adrenal weights and reduced thymus weights compared to the control group (Fig. 6C), indicative of chronic hyperactivity of the HPA axis. Surprisingly, the circadian rhythm of corticosterone secretion ( Supplementary Fig. 9A), as well as stress-induced corticosterone level (Fig. 6D), were unaffected by the Crhspecific Fkbp5 overexpression. Further, basal and 15 min post-stress ACTH level and DEX/CRH corticosterone level  Supplementary Fig. 9D). C Fkbp5 CRH OE mice showed significantly increased adrenal weights (n = 10) and reduced thymus weights (n = 9) under non-stressed conditions compared to the controls (n = 10). D Corticosterone levels were mildly, but not significantly increased in Fkbp5 CRH OE at 15, 30, and 60 min post-stress (Fkbp5 CRH OE n = 10; Control n = 9). E Fkbp5 CRH OE mice displayed a decrease in CRH neuronal mEPSC frequency accompanied by an increased amplitude (Fkbp5 CRH OE n mouse = 4; n neuron = 29; Control n mouse = 4; n neuron = 25), which is reflected in two representative recording traces. Data are presented as mean ± SEM. All data were received from animals between 30 and 34 weeks of age and analyzed with the student's t-test. *p < 0.05, **p < 0.01, ***p < 0.001, T = 0.05 < p < 0.1, n.s. not significant. remained unaffected in Fkbp5 CRH OE mice ( Supplementary  Fig. 9B-D). To further assess the impact of Fkbp5 overexpression on neuronal activity in Crh positive (Crh + ) neurons in the PVN, we recorded AMPA receptor-mediated mEPSCs in these cells using cell patch-clamp recordings in a separate cohort. Fkbp5 overexpression decreased the frequency of mEPSC frequency while increasing amplitude (Fig. 6E). This data shows that the Fkbp5 expression level in Crh + neurons can steer CRH neuronal activity within the PVN, whereas the baseline-and stress phenotype of Fkbp5 CRH OE mice is mostly unaffected by the manipulation, identifying Crh + neurons as one important but not the only driver of the observed stress-like Fkbp5 PVN OE phenotype.

Discussion
FKBP5 was first associated with stress-related disorders in 2004 [13] and has been studied extensively over the past 15 years with regard to stress regulation and sensitivity. However, detailed cell-type and region-specific manipulations of Fkbp5 in the brain are still lacking. In this study, we highlight the importance of this co-chaperone in the regulation of the acute stress response through the combined analysis of deletion, overexpression, and rescue of Fkbp5 exclusively in the PVN.
Fkbp5 is a stress-responsive gene and past research has shown that its main effects occur after chronic or acute stress [8,18,28,32,41]. Given that deletion of Fkbp5 in the PVN mimics the previously described phenotype of Fkbp5 KO mice [28,41], with regard to their basal neuroendocrine profile and HPA axis function, our data illustrate that the functional contribution of Fkbp5 to HPA axis activity is centered in the PVN. In addition, reinstatement of native basal Fkbp5 expression in the PVN of Fkbp5 KO mice was sufficient to normalize HPA axis function. Interestingly, the phenotype of intensified HPA axis suppression due to the loss of Fkbp5 in the PVN emerges only in adult animals, excluding developmental effects and underlining the previously reported importance of Fkbp5 in aging [42,43].
Further, our results highlight the essential role of Fkbp5 in stress adaptation, as PVN-specific Fkbp5 excess is sufficient to reproduce all physiological and endocrinological hallmarks of a chronic stress situation [29,44]. Interestingly, the results of our Fkbp5 PVN OE cohort are comparable to the neuroendocrine effect of GR deletion in the PVN [45] and are in line with the high PVN-specific FKBP5 expression in rats with a hyperactive HPA axis [20]. The consequence of a heightened Fkbp5 expression in the PVN is twofold. Firstly, it leads to direct changes in GR sensitivity and downstream GR signaling directly in the PVN. Secondly, PVN Fkbp5 overexpression dramatically affects GR sensitivity and feedback at the level of the pituitary, as demonstrated by the inability of a low Dex dose (that does not cross the blood-brain barrier and acts predominantly at the level of the pituitary [26]) to suppress corticosterone secretion. This secondary effect is likely due to the constant overproduction of CRH in the PVN and very similar to the effects of HPA hyperactivity observed in many depressed patients [33,46].
Mechanistically, we explored the role of Fkbp5 in modulating GR phosphorylation. The status of GR phosphorylation at Ser 211 , Ser 203 , and Ser 234 is associated with transcriptional activity, nuclear localization, and ability to associate with GRE containing promoters [37]. Whereas higher levels of phosphorylation at Ser 211 are associated with full transcriptional activity and localization in the nucleus, increased phosphorylation of Ser 203 is linked to a transcriptionally inactive form of GR within the cytoplasm and thereby less active GR [38,47,48]. In our experiments, overexpression of Fkbp5 resulted in dephosphorylation at Ser 211 and higher phosphorylation at Ser 203 , suggesting that the GR is mostly located in the cytoplasm and less active. Deletion of Fkbp5 showed the opposing effect, indicating a more active GR in Fkbp5 PVN−/− animals. Unfortunately, we were not able to analyze the GR phosphorylation in our Fkbp5 PVN Rescue animals due to technical and breeding issues and therefore can only speculate that FKBP5 PVN Rescue animals might have elevated levels of pGRSer203 and decreased levels of pGRSer211 and pGRSer234. It has previously been reported that GR phosphorylation is regulated by several kinases, including CDK5 and ERK [37,49], and Fkbp5 has also been shown to be associated with CDK5 in the brain [50]. Therefore, we hypothesize that Fkbp5 also interacts with CDK5 to phosphorylate GR at multiple phosphorylation sites, thereby directly affecting ligand-dependent GR activity.
Given the complexity of the different cell types with highly specialized functions in the brain, it is essential to gain a deeper understanding of the cellular architecture of the PVN and the specific function of Fkbp5 in this context. Previously, it was assumed that Fkbp5 is quite widely expressed in most cell types of the nervous system [51]. However, our current data suggest that while Fkbp5 is indeed expressed in the PVN, it is enriched in specific subpopulations, including for example GABAergic neurons, Crh + neurons, and microglia, but largely absent in others, such as astrocytes and endothelial cells. Interestingly, when quantifying Fkbp5 regulation, we identified a highly selective regulation of Fkbp5 in Crh + neurons, further supporting the central role of Fkbp5-controlled GR feedback in this neuronal subpopulation.
To further disentangle the role of Fkbp5 within Crh positive neurons in the acute stress response and HPA-axis feedback regulation, we selectively overexpressed Fkbp5 in Crh + cells within the PVN and assessed baseline-and stress-induced phenotypes paralleled by selective patchclamp recordings from Crh + neurons. Fkbp5 overexpression enhanced the amplitude of mEPSCs in Crh + cells indicating a postsynaptic and, thus, the direct effect of our Fkbp5 manipulation on excitatory neurotransmission onto these neurons. This effect, which was accompanied by a diminished rate of mEPSC, potentially arises from accelerated recycling of internalized AMPA receptors to the postsynaptic density [52] and suggests an increased activity of Crh + -cells. However, Fkbp5 overexpression only partially recapitulating the HPA-axis phenotype observed in the Fkbp5 PVN OE mice. The unselective Fkbp5 overexpression in Fkbp5 PVN OE animals targeted a broad range of Fkbp5 expressing cell populations within the PVN, amongst which are oxytocin and vasopressin. Their essential contribution to the initiation and termination of the HPA-axis is well established, with vasopressin (together with CRH) inducing ACTH release from the pituitary [53] and oxytocin enhancing the negative feedback helping to dampen the stress response [54]. Hence, our data suggest that Fkbp5 might play an essential role in at least one further neuronal subpopulation of the PVN driving the observed stress-like phenotype of Fkbp5 PVN OE mice in concert with CRH neurons.
The current study also comes with a number of limitations. Importantly, only male animals were used and the conclusions should therefore only be drawn with respect to male HPA axis regulation. Furthermore, given the fact that we did not observe differences in ACTH levels after stress, we cannot exclude that Fkbp5 manipulations in the PVN also drive changes in adrenal sensitivity, e.g., via alterations of sympathetic adrenal innervation [55]. Furthermore, the time course of CRH and ACTH release is quite different from the CORT response, so our measures may have missed a differential regulation. In fact, a modest increase in circulating ACTH at certain times of the circadian rhythm can indeed lead to adrenal hypertrophy and increased sensitivity, thereby contributing to increased CORT secretion even following comparable ACTH levels. Finally, the lack of ACTH data in the Dex/CRH test was technically unavoidable but also limits the conclusions with regard to the involved mechanism.
In summary, this study is the first to specifically manipulate Fkbp5 in the PVN and underlines its central importance in shaping HPA axis regulation and the acute stress response. The results have far-reaching implications for our understanding of stress physiology and stress-related disorders.