An 840 kb distant upstream enhancer is a crucial regulator of catecholamine‐dependent expression of the Bdnf gene in astrocytes

Abstract Brain‐derived neurotrophic factor (BDNF) plays a fundamental role in the developing and adult nervous system, contributing to neuronal survival, differentiation, and synaptic plasticity. Dysregulation of BDNF synthesis, secretion or signaling has been associated with many neurodevelopmental, neuropsychiatric, and neurodegenerative disorders. Although the transcriptional regulation of the Bdnf gene has been extensively studied in neurons, less is known about the regulation and function of BDNF in non‐neuronal cells. The most abundant type of non‐neuronal cells in the brain, astrocytes, express BDNF in response to catecholamines. However, genetic elements responsible for this regulation have not been identified. Here, we investigated four potential Bdnf enhancer regions and based on reporter gene assays, CRISPR/Cas9 engineering and CAPTURE‐3C‐sequencing we conclude that a region 840 kb upstream of the Bdnf gene regulates catecholamine‐dependent expression of Bdnf in rodent astrocytes. We also provide evidence that this regulation is mediated by CREB and AP1 family transcription factors. This is the first report of an enhancer coordinating the transcription of Bdnf gene in non‐neuronal cells.

Brain-derived neurotrophic factor (BDNF) is a member of the neurotrophin family with important functions in the developing and adult nervous system.The expression of the Bdnf gene is a result of complex transcriptional regulation: rodent Bdnf gene has eight 5 0 noncoding exons (exons I-VIII) and one 3 0 coding exon (exon IX), each of which is controlled by a separate promoter.To produce Bdnf transcripts, one of the 5 0 non-coding exons is spliced together with the common 3 0 coding exon (Aid et al., 2007;Timmusk et al., 1993).Multiple promoters allow stimulus-and tissue-specific expression of Bdnf.
While the proximal regulatory regions of the Bdnf gene have been well described (West et al., 2014), only few studies have investigated the role of different enhancer regions regulating Bdnf gene expression in neuronal cells (Beagan et al., 2020;Brookes et al., 2023;Calderon et al., 2022;Flavell et al., 2008;Lyons et al., 2012;Tuvikene et al., 2021).This is a critical knowledge gap considering that enhancer regions refine the spatial, temporal, stimulus and cell type-specific transcriptional regulation (Nord & West, 2020;Yap & Greenberg, 2018) and alterations in enhancer regions are increasingly associated with diseases (Maurano et al., 2012).Enhancers can be intra-or extragenic, located upstream or downstream of a gene, and can be located close or far away from the transcription start sites (TSS) of a gene (Perenthaler et al., 2019).To date, three enhancer regions for Bdnf gene have been reported: (1) a MEF2-binding region located 4.8 kb upstream of Bdnf exon I that enhances neuronal activity-dependent transcription from Bdnf promoter I in hippocampal neurons (Flavell et al., 2008) but not in cortical neurons (Lyons et al., 2012); (2) an intronic enhancer region that regulates basal and stimulus-dependent expression of first cluster of Bdnf transcripts in neurons (Tuvikene et al., 2021); and (3) an enhancer located $237 kb downstream of Bdnf exon I that is involved in Bdnf expression during neuronal differentiation (Brookes et al., 2023).
We have previously shown that Bdnf proximal promoters IV and VI need distal regulatory element(s) for catecholamine-dependent induction in astrocytes (Koppel et al., 2018).Here, we aimed to identify the regulatory region(s) involved in catecholamine-dependent Bdnf gene transcription in astrocytes.To this end, we first screened several in silico predicted putative Bdnf enhancer regions in cortical astrocytes using reporter assays.We characterized a region $840 kb upstream of Bdnf that showed strong dopamine-dependent bidirectional transcription and potentiated the activity of Bdnf promoters in rat cortical astrocytes.We demonstrated that the À840 kb enhancer forms long-range chromatin interactions with Bdnf gene, and further validated its activity in the endogenous context using CRISPR interference and activator systems as well as CRISPR-mediated deletion of the region.Finally, we show that the activity of the enhancer is regulated by CREB and AP1 family transcription factors.Together, our results provide a novel mechanism for the regulation of Bdnf gene in astrocytes.

| Cultures of rat cortical astrocytes
All animal procedures were performed in accordance with European Directive 2010/63/EU.Animals were maintained under 12 h light/ dark cycle in a humidity (50 ± 10%) and temperature (22 ± 1 C) controlled room.Rats were group-housed (2-4 animals per cage) in conventional polycarbonate or H-TEMP polysulfone cages with ad libitum access to water and food.Rat cortical astrocyte culture was generated from embryonic (E20/21) Sprague Dawley rats as described previously (Tuvikene et al., 2021).
Briefly, rat cortical astrocytes were grown in 75 cm 2 cell culture flasks in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (Pan Biotech) and 100 U/mL penicillin and 0.1 mg/mL streptomycin (Gibco) at 37 C and 5% CO 2 .At 1 DIV, the whole medium was replaced with fresh supplemented DMEM.At 6 DIV, cell flasks were shaken in a 37 C shaker Certomat ® BS-1 (Sartorius Group) for 18-20 h at 180 rpm.For splitting, medium was removed along with unattached non-astrocytic cells.Astrocyte monolayer was washed with phosphate-buffered saline (PBS) and detached from the flask with trypsin-EDTA solution (0.0625% trypsin, 0.25 mM EDTA, Gibco) in 1Â PBS at 37 C for 5 min.Astrocytes were collected in supplemented DMEM and centrifuged at 200 g for 6 min.Supernatant was aspirated, astrocytes were resuspended in supplemented DMEM, and plated on cell culture plates coated with 0.2 mg/mL poly-Llysine (Sigma-Aldrich) in Milli-Q.At 9 DIV, the whole medium was changed.At 15 DIV > 95% of the cells were GFAP-positive by immunocytochemical analysis and quantification (Figure S1).

| Deletion of the À840 kb enhancer region in mouse embryonic stem cells
A2Lox mouse embryonic stem cells (mESCs) were a kind gift from Michael Kyba.Doxycycline-inducible Neurogenin2 was inserted to A2Lox mESC genome using recombination-mediated cassette exchange procedure (Iacovino et al., 2011).mESCs were grown and passaged as described in Kainov and Makeyev (2020).The enhancer region located À840 kb from Bdnf gene was deleted in the mESCs as described in Tuvikene et al. (2021).Briefly, CRISPR/Cas9 system was used to delete $500 bp of À840 kb enhancer core region.For that, a total of 6 gRNAs (listed in Supplementary file 1) targeting either upstream or downstream of the mouse À840 kb enhancer core region were designed using the Benchling CRISPR tool, and chosen so that the gRNAs had no predicted off-targets (maximum of 2 mismatches, allowing both 5 0 NGG and 5 0 NAG PAM-sites) in the protein-coding regions.The gRNAs were cloned into the pX330 vector (Addgene plasmid #42230).A2Lox-Neurogenin-2 mESCs were cotransfected with a mixture of gRNA-encoding plasmids together with a plasmid containing a blasticidin S deaminase expression cassette.To create control cell lines without any deletion, the mESCs were transfected with pX330 vector containing no gRNA targeting sequence and blasticidin S deaminase-encoding plasmid.One day after transfection, 8 μg/mL blasticidin S (Sigma-Aldrich) was added to the media for 3 days to select for transfected cells.The cells were grown for $2 weeks in 2i media containing LIF and 1% FBS (HyClone™ Fetal Bovine Sera) before selecting single colonies, passaging and genotyping.Deletion of the À840 kb region was determined by PCR and verified using qPCR-based copy number analysis as described in Tuvikene et al. (2021).PCR amplicons were confirmed with sequencing.All used primers are listed in Supplementary file 1.
On 7th day of suspension culture, the cells were plated to produce adherent monolayer of NPCs.For that, the neurospheres were washed from the well and collected in a tube, centrifuged at 72 g for 30 s, resuspended in NSA media with growth factors and plated onto 0.1% gelatin-coated 6-well plates.In the next days, the neurospheres sedimented and produced adherent monolayer cultures of NPCs.
Routinely, the cell media was changed every other day and NPCs were passaged when >90% confluency was reached (3rd or 4th day depending on the clonal cell line).For passaging, cell media was collected to a tube, and cells were trypzinised with 0.025% trypsin-EDTA (Gibco, #15400054) at room temperature for 1 min, diluted in 10 times volume of prewarmed DMEM, and collected to the same tube.The cells were centrifuged at 300 g for 5 min and resuspended in NSA media with freshly added growth factors.Routinely, 50,000 NPCs/cm 2 were plated onto 0.1% gelatin-coated cell culture multi-well plates.
The identity of the NPCs was verified using RT-qPCR for Nes mRNA (primers are listed in Supplementary file 1) and NESTIN immunocytochemistry using anti-NES antibody (Developmental Studies Hybridoma Bank, #rat-401) at 0.2 μg/mL.96% of the cells were NESpositive (Figure S6B).To determine Bdnf mRNA levels in the NPCs, the clonal cell lines were independently plated at different times and lysed for RNA 2 days after plating.

| DNA constructs
The selected putative enhancer regions were amplified from genomic DNA and cloned in both forward and reverse orientation in front of the Firefly luciferase reporter gene into pGL4.15luciferase reporter vector (Promega) without an additional promoter as described previously (Tuvikene et al., 2021).Highly conserved $500 bp regions from rat, mouse and human À840 kb region were cloned in forward or reverse orientation in front of Firefly luciferase coding sequence in pGL4.15vector.The coordinates of the cloned regions are listed in Supplementary file 1.
For CAPTURE-3C-sequencing, the FLAG-tagged biotin acceptor site-containing nuclease-deficient Cas9 (FB-dCas9) and biotin ligase BirA plasmids were obtained from Addgene (plasmid #10054 and #100548) and Liu et al. (2017).The FB-dCas9 was cloned to pLV backbone under the control of human UbC promoter and BirA was cloned to pRRL backbone under the control of human PGK promoter.

| Transfection of cells and luciferase reporter assay
Rat cortical astrocytes were transfected at 13 DIV using 190 ng of luciferase reporter plasmid (or 380 ng of DNA when cotransfected with effector plasmids) and 10 ng of pGL4.83-SRα-hRLucnormalizer plasmid (or 20 ng when cotransfected with effector plasmids) in unsupplemented DMEM with a DNA to Lipofectamine 2000 (Invitrogen) ratio of 1:3 (or 1:2 for cotransfection of reporter plasmid and effector plasmid).The transfection was terminated by changing the media to fresh supplemented DMEM.
After treatments at 15 DIV the cortical astrocytes were lysed in 1Â Passive Lysis Buffer (Promega) and luciferase assay was performed using Dual-Glo ® Luciferase Assay (Promega) system.Luminescence signal was measured using GENios Pro Multifunction Microplate Reader (Tecan).For data analysis, background corrected Firefly luciferase signals were normalized with background corrected Renilla luciferase signals and the averages of duplicates were calculated.

| Lentiviral transduction
The production of lentiviral particles was performed as described previously (Koppel et al., 2018).qPCR was used to determine relative titers of lentiviral particles based on provirus incorporation (primers are listed in Supplementary file 1).For functional experiments, equal amounts of lentiviral particles with >95% transduction efficiency were used to transduce astrocytes.Rat cortical astrocytes were transduced after splitting the astrocytes to 6-well cell culture plates at 7 DIV.

| Drug treatments
Rat cortical astrocytes were treated at 15 DIV and mouse embryonic stem cell-derived astrocytes were treated on the 21st day of differentiation with 0.15% DMSO as a vehicle control, 150 μM dopamine (Tocris Bioscience) or 25 μM norepinephrine (Tocris Bioscience) in unsupplemented DMEM.

| RNA isolation, cDNA synthesis and qPCR
Total RNA from rat cortical astrocytes or mESC-derived cells was isolated using RNeasy Mini Kit (Qiagen) with on-column DNase digestion using RNase-free DNase set (Qiagen) according to the manufacturer's instructions.RNA concentration was measured with BioSpec-nano spectrophotometer (Shimadzu) or with Nanodrop 2000c spectrophotometer (Thermo Scientific).cDNA was synthesized from equal amounts of total RNA using Superscript III or IV Reverse Transcriptase (Thermo Fisher Scientific) with 1:1 mixture of oligo(dT) 20 (Microsynth) and random hexamer primers (Microsynth).qPCR was performed in triplicates using 1Â HOT FIREpol EvaGreen qPCR Mix Plus (Solis Biodyne) or 1Â LightCycler 480 SYBR Green I Master (Roche) on Light-Cycler 480 II Real Time PCR instrument (Roche).All used qPCR primers are shown in Supplementary file 1. Ppib mRNA levels in rat cortical astrocytes and Cnot4 mRNA levels in mESC-derived NPCs and astrocytes were used to normalize mRNA and enhancer RNA expression.
Chromatin was sonicated with sonication beads (Diagenode) using 20 cycles (30 s on, 30 s off) on a Bioruptor Pico sonicator (Diagenode) at 4 C.After sonication, lysates were cleared with centrifugation at 4 C for 5 min 10,500 g.Protein concentration was measured using BCA Protein Assay Kit (Pierce).Equal amount of protein was taken for ChIP and 10% of the ChIP volume was taken as input samples.The input samples were kept on ice until decrosslinking.1% of final Triton X-100 was added to the ChIP samples, and then the samples were diluted 2-fold in dilution buffer (1% Tri- DNA-protein complexes were eluted three times with elution buffer (1% SDS, 100 mM NaHCO 3 , 1 mM EDTA) and DNA cross-links were reversed with 250 mM NaCl at 65 C overnight.The next day, the samples were treated with RNase A (125 μg/mL) for at least 1 h, and then 6 mM EDTA was added and the samples were treated for at least 1 h with Proteinase K (240 μg/mL).The ChIP and input DNA was purified using QIAquick PCR Purification Kit (Qiagen) and DNA enrichment was analyzed with qPCR using 1Â LightCycler 480 SYBR Green I Master (Roche).All used qPCR primers are shown in Supplementary file 1.The enrichment of binding was calculated relative to a non-conserved control region (chr3:99,937,231-99,937,310, Rn6 genome) located $5 kbp downstream of the À840 kb enhancer region and selected specifically for this study.
The whole media was changed every 2-3 days.
The previously published CAPTURE-3C-sequencing protocol (Botten et al., 2023;Liu et al., 2017Liu et al., , 2018Liu et al., , 2020) ) was adapted for cultured astrocytes.At 15 DIV, after 2 h of treatment, astrocytes grown on 145 mm dishes were washed two times with 1Â PBS and chromatin was fixed with 2 mM ethylene glycol-bis(succinic acid N-hydroxysuccinimide ester) (EGS, Santa Cruz Biotechnology, #sc-252807) in PBS with gentle agitation for 45 min.Then 1% formaldehyde (Cell Signaling Technology, #12606S) was added and incubated for additional 10 min.The fixation was quenched with 250 mM glycine for 10 min at room temperature also on an orbital shaker.Next, the astrocytes were washed two times with ice-cold PBS and scraped in 1 mL ice-cold cell lysis buffer (25 mM Tris-HCl [pH 7.5], 85 mM KCl, 0.1% Triton X-100 [Triton™ X-100 Surfact-Amps™, Thermo Scientific, #85111], and freshly added 1Â cOmplete™ EDTA-free protease inhibitor cocktail [Roche] and 1 mM DTT [Invitrogen]).The cells were lysed for 30 min while rotating at 4 C.The nuclei were pelleted by centrifugation at 2300 g for 8 min at 4 C and the supernatant was removed with a pipette and the pellet was frozen in liquid nitrogen.
Freezing the pellet at this stage is recommended as it appears to be an efficient way to help inactivate endonucleases that would otherwise interfere with the protocol.
Each pellet was gently resuspended in 500 μL 1Â buffer O (Orange buffer, Thermo Fisher Scientific) and then centrifuged at 2300 g for 8 min at 4 C.The pellet was then resuspended in 0.5% SDS and heated at 62 C for 10 min to inactivate endonucleases.
Finally, the lysate was cooled for 5 min on ice and SDS was sequestered by adding a final concentration of 1% Triton X-100 and incubated for 30 min at 37 C. Finally, buffer O (Orange buffer, Thermo Fisher Scientific) and 300 U of DpnII (NEB, #R0543M) were added (as a result, the SDS was also diluted down to 0.1%).Note that the use of buffer O instead of DpnII buffer is critical when performing CAPTURE-3C-sequencing from cultured astrocytes, to inhibit the endonucleases present in the cultured cells.DpnII restriction was performed for 4 h while rotating at 37 C. DpnII was inactivated at 65 C for 20 min.Next, the solution was diluted in a Falcon tube 5Â up to 3 mL by adding ligase buffer (NEB, final 1Â), Triton X-100 (final concentration 1%), 1Â cOmplete™ EDTA-free protease inhibitor cocktail (Roche), and Milli-Q.Finally, 2 million CELU units/mL of DNA ligase (NEB, #M0202M) was added per 1 mL.The solution was divided into three DNA LoBind tubes (Eppendorf) and rotated overnight ($16 h) at 16 C.
The next day the ligated chromatin was collected by centrifugation at 3000 g for 10 min at 4 C.The pellet was resuspended in 500 μL RIPA 0 buffer (10 mM Tris-HCl [pH 7.5], 1 mM EDTA [pH 8.0], 150 mM NaCl, 0.1% SDS, 1% Triton X-100, 0.1% sodium deoxycholate, freshly added protease inhibitors, 1 mM DTT), and final concentration of 0.25% N-lauroylsarcosine.The mixture was sonicated using sonication beads (0.2 g of beads per 0.5 mL solution and prewashed 3Â with PBS) for 10 cycles (30 s on and 30 s off) at 4 C with Bioruptor Pico (Diagenode).Finally, the solution was transferred to a new DNA LoBind tube (Eppendorf), centrifuged at 10,500 g for 10 min, transferred to another DNA LoBind tube, and final concentration of 300 mM NaCl was added.
The magnetic beads were prewashed two times with 0.1% Tween-20 in tris-buffered saline (TBS, pH 7.4).The samples were incubated with streptavidin beads for $5 h while rotating at 4 C.The resultant read pairs were further processed using a model-based analysis as previously described in Chen et al. ( 2020) with modifications.For statistical analysis, only data of chromosome 3 was kept.
Based on the sequencing read alignment, the region of Bdnf promoter VI, coordinates chr3:100,787,200-100,788,550 on the Rn6 genome, with the FB-dCas9 peak summit at 100,787,921, was selected as the anchor region.The region ±3 Mb of the anchor region was divided into 20 kb bins with a 5 kb step.For each bin, the number of reads with one end inside the anchor region and the other end in the respective bin was calculated.The distance of each bin from the anchor region was calculated as the distance between the bin midpoint and the FB-dCas9 peak summit ±10 kb.Read pairs with both ends outside the anchor region were used to define the background interaction distribution for each sample by subsampling n anchor read pair distances from chromosome 3 for 10,000 times, where n anchor is the number of read pairs with at least one end in the anchor region in the respective sample.Separate negative binomial distributions were fitted for each bin using maximum likelihood estimation, and p-values were calculated from the cumulative distribution function using the count of reads with one end mapped to the anchor region and the other end mapping to the respective bin region.For bins with zero reads interacting with the anchor region, the p-value was set as

| Selection of putative enhancer regions of Bdnf gene in astrocytes
The enhancer regions of Bdnf have not been studied in non-neuronal cells.In addition, previous research from our group suggests that catecholamines induce Bdnf expression through a yet undescribed enhancer region(s) in cortical astrocytes (Koppel et al., 2018).In search for the possible enhancers of the BDNF gene, we used the GeneHancer subsection of the Genecards database.GeneHancer gathers information about genomic regulatory elements from different human genome-wide analysis derived from numerous tissues (Fishilevich et al., 2017).Altogether 22 potential regions were predicted as BDNF enhancer regions based on the integrated data from several databases (Supplementary file 1).We selected four putative enhancer regions (Table 1, À840, À450, À40, +37 kb) based on the following criteria: (1) high BDNF gene-enhancer score, which indicates the likelihood of a region being an enhancer of the BDNF gene; (2) the region is predicted to act as an enhancer in astrocytes; (3) high conservation in mammals; (4) not previously studied as a potential enhancer candidate (Tuvikene et al., 2021); and (5) not located within a promoter region of another gene.All the selected putative enhancer regions were named after their approximate distance from the rat Bdnf exon I TSS.The metrics of the GeneHancer database are described in the legend of Table 1.3.2 | Identification of the region 840 kb upstream of Bdnf as a putative dopamine-regulated Bdnf enhancer in cortical astrocytes Kim et al. (2010) described bidirectional transcription of short enhancer-RNAs (eRNAs) from active enhancer regions and showed that eRNA expression levels are correlated with the expression of target genes, meaning that the transcription from an enhancer region can be used to determine the activity of an enhancer (Kim et al., 2010).First, we used luciferase reporter assay to evaluate whether transcription can start from the putative enhancer regions upon stimuli.We cloned the selected potential rat enhancer regions (À840, À450, À40, +37 kb, Figure 1a,b) as promoters in front of the luciferase reporter gene (without an additional promoter) in either forward or reverse orientation.The cloned enhancer constructs were transfected into rat cortical astrocytes.To see whether these enhancers respond to dopaminergic stimulation, astrocytes were treated with dopamine or DMSO as a vehicle control for 8 h.The À840 kb region showed the highest orientation-independent induction in response to dopamine treatment ($6-fold) (Figure 1b).
Although the À40 kb region showed the highest bidirectional transcription in basal conditions (Figure S2), the À450, À40 and +37 kb regions did not show dopamine-dependent induction (Figure 1b), implying that these regions are not dopamine-dependent enhancer regions in astrocytes.
Functionally important regulatory regions, including enhancers, are known to be evolutionarily conserved (Dickel et al., 2018;Pennacchio et al., 2013).As the À840 kb region showed the strongest transcriptional activation upon stimuli out of the screened candidates, we next inserted the highly conserved $500 bp long core sequence of the À840 kb region from rat, mouse or human upstream of luciferase coding sequence in either forward or reverse orientation (Figure 1c).In unstimulated astrocytes, mouse and human core enhancer regions upregulated luciferase expression up to $6-fold (Figure 1c).Robust upregulation was seen in cortical astrocytes (Figure 1c), where an 8 h dopamine treatment strongly ($14-56-fold) induced the expression of the reporter gene from all tested À840 kb enhancer core regions.These results demonstrate that the À840 kb region exhibits evolutionarily conserved dopaminedependent activity in cortical astrocytes.
Next, we investigated whether the rat À840 kb region could potentiate the activity of Bdnf promoters in a heterologous context.
For investigation, we chose Bdnf promoters IV and VI because Bdnf exon IV-and exon VI-containing mRNAs show the highest basal and catecholamine-induced expression among Bdnf transcripts in cortical astrocytes (Koppel et al., 2018).Therefore, we cloned rat Bdnf promoters IV or VI in front of the luciferase coding sequence and rat À840 kb enhancer region in forward or reverse orientation downstream of the polyadenylation signal (PolyA) in the same plasmid.
Addition of the À840 kb region did not potentiate the activity of Bdnf promoter IV or VI (Figure 1d) in unstimulated cells, but strongly increased the stimulus-dependent activity $5-6 and $3-4-fold, respectively (Figure 1d).Collectively, our results show that the À840 kb region potentiates the dopamine-dependent activity of Bdnf promoters in rat cortical astrocytes.
Finally, to study whether the À840 kb region is an active enhancer region in endogenous context, we interrogated the chromatin state of the region in rat cultured astrocytes.For that, we performed ChIP-qPCR analysis to investigate the presence of H3K27Ac, a hallmark of active regulatory regions, and H3K27me3, a mark of repressed and condensed chromatin regions (Creyghton et al., 2010;Zhu et al., 2013).Analysis of the À840 kb region showed a $4-fold T A B L E 1 Potential human BDNF enhancer regions predicted by GeneHancer subsection in the GeneCards database and selected for investigation.Luciferase vector without an enhancer region was used as a negative control.At 15 DIV, transfected astrocytes were treated for 8 h with 0.15% DMSO as a control or 150 μM dopamine (DA).The results are depicted relative to the luciferase signal measured from DMSO-treated astrocytes transfected with a no enhancer construct (b-d).(e) Histone modifications were analyzed with ChIP-qPCR assay using anti-H3K27Ac or anti-H3K27me3 antibodies.At 15 DIV, astrocytes were treated for 3 h with 0.15% DMSO as a control or 150 μM DA.The region $5 kb downstream of the À840 kb enhancer was used as a negative control region.Data is depicted as enrichment relative to the control region (dashed line equals to 1).(b-e) Error bars indicate SEM (n = 3 independent experiments).Statistical significance was calculated with two-tailed paired t-test relative to the luciferase level measured from DMSO-treated cells transfected with luciferase vector without an enhancer (b, c), relative to the luciferase level measured from cells transfected with respective promoter-containing vector without an enhancer at respective treatment (d), relative to the enrichment of DNA at the negative control region at respective treatment or relative to the respective region in DMSO-treated astrocytes (e).NS, not significant, *p < .05;**p < .01;***p < .001,p-values were corrected for multiple comparisons with Holm-Šidak method.
higher enrichment of H3K27Ac relative to control region in unstimulated astrocytes, and a robust $7-fold increase in the H3K27Ac mark after 3 h dopamine treatment (Figure 1e).The levels of H3K27me3 were $2-fold lower compared to negative control region in both DMSO-and dopamine-treated astrocytes, although this effect failed to reach statistical significance (Figure 1e).To conclude, our results show that the À840 kb region is enriched with an active enhancer chromatin mark in cortical astrocytes especially upon dopamine treatment.

| The À840 kb enhancer region is in spatial proximity to Bdnf promoter VI in cortical astrocytes
The physical proximity between enhancer and its target promoter has been proposed to be necessary for the activation of transcription (Beagan et al., 2020;Chen et al., 2018).To identify the enhancer regions interacting with the Bdnf locus in astrocytes, we used the CAPTURE-3C-sequencing method (Botten et al., 2023;Liu et al., 2017Liu et al., , 2018Liu et al., , 2020)).Briefly, we used lentivirus-mediated expression of FLAG-tagged biotin acceptor site-containing nuclease deficient Cas9 (FB-dCas9), biotin ligase BirA, and gRNAs targeting Bdnf promoter VI region.We verified that gRNAs targeting Bdnf promoter VI did not have a major effect on the Bdnf mRNA expression (Figure S3A).The transduced astrocytes were treated with dopamine or DMSO as a vehicle control for 2 h and analyzed by CAPTURE-3Csequencing (Figure 2a).The enrichment of Bdnf promoter VI was verified using qPCR before next-generation sequencing (Figure S3B).Our CAPTURE-3C-sequencing results also showed a strong enrichment of the Bdnf pVI region (Figure S3C).In total, we determined 10,559 and 9066 sequencing reads with at least one end in the Bdnf pVI anchor region, for DMSO-and DA-treated cells, respectively and of these, 220 and 249 reads, respectively showed long-range intrachromosomal interaction outside the pVI anchor region.
The CAPTURE-3C-sequencing experiment revealed that the Bdnf promoter VI region exhibits a few statistically significant interactions with distant upstream regions.Specifically, we identified a stimulusindependent interaction with a region located 1.82 Mb upstream of Bdnf gene.Following dopamine treatment, we observed alterations in the three-dimensional structure of chromatin.Namely, an interaction with the À790 kb region was exclusively detected in DMSO-treated cells, while DA-induced interactions were noted with the À840 kb, À990 kb and À2.65 Mb regions.We did not detect any interactions near the Bdnf gene locus, even when looking with a smaller bin size of 5 kb (data not shown).Interestingly, we did not find any interactions downstream of the Bdnf gene in astrocytes.Collectively, our findings demonstrate that the À840 kb region physically associates with the Bdnf promoter VI in astrocytes in a catecholamine signalingdependent manner, supporting the role of the À840 kb region as a stimulus-dependent enhancer of the Bdnf gene.

| The À840 kb enhancer regulates catecholamine-induced Bdnf expression in cortical astrocytes
Next, we investigated the functionality of the À840 kb enhancer region to potentiate Bdnf expression in the endogenous context.We hypothesized that if the region under investigation enhances the transcription of Bdnf gene, modulating the activity of the enhancer should affect the expression of endogenous Bdnf.For this purpose, we used lentivirus-encoded CRISPR interference and activator (CRISPRi/a) systems where a catalytically inactive Cas9 protein (dCas9) is fused with Krüppel associated box (KRAB) domain (dCas9-KRAB) or 8 copies of VP16 domain (VP64-dCas9-VP64), respectively.For the enhancerspecific repression or activation we used lentiviruses expressing five different gRNAs targeting the À840 kb region.As a control, we used lentiviruses expressing a gRNA that does not target any sequence in the rat genome.Cortical astrocytes were co-transduced at 7 DIV with lentiviruses expressing dCas9-KRAB or VP64-dCas9-VP64 and enhancer-specific or negative control gRNAs (Figure 3a,g).At 15 DIV, rat cortical astrocytes were treated for 3 h with dopamine or DMSO as a vehicle control (Figure 3a-f).We also investigated the effect of À840 kb enhancer after norepinephrine treatment (Figure 3g-l), another catecholamine known to regulate Bdnf expression in astrocytes (Koppel et al., 2018).The À840 kb region eRNAs, mRNA levels of total Bdnf and different Bdnf transcripts were measured using RT-qPCR.
To test the specificity of our system, we first measured the expression of FosB, an unrelated stimulus-dependent gene.Neither repression nor activation of À840 kb enhancer region affected the expression of FosB (Figure 3b,h).The expression levels of À840 kb eRNAs were notably induced in cortical astrocytes-up to $9-fold upon dopamine (Figure 3c) and $25-fold after norepinephrine treatment (Figure 3i).Targeting dCas9-KRAB complex to the À840 kb region remarkably decreased the levels of À840 kb eRNAs $4and $12-fold after dopamine and norepinephrine treatments, respectively, but did not affect the basal eRNA levels (Figure 3c,i).
Activation of the À840 kb enhancer by VP64-dCas9-VP64 increased the À840 kb basal eRNA levels, but not stimulus-dependent levels, up to $9-fold (Figure 3c,i).As the basal expression levels of the À840 kb eRNAs are very low, our data suggests that the À840 kb enhancer is not active in unstimulated rat astrocytes, but is activated upon dopamine and norepinephrine treatments.However, nascent bidirectional transcription from À840 kb region is evident in previously conducted GRO-seq data obtained from human astrocytes (Figure S4) (Bouvy-Liivrand et al., 2017), indicating that this region is transcriptionally active in unstimulated human glial cells.Further mechanistic experiments in human astrocytes would be necessary to unequivocally show that the À840 kb enhancer is active at basal levels and determine whether its activity differs in a species-specific manner.
Repression of the À840 kb region with dCas9-KRAB had no effect on the basal Bdnf mRNA expression level (Figure 3d-f,j-l), confirming the inactive/poised state of the enhancer at basal levels in rat cortical astrocytes.In contrast, dopamine-and norepinephrine-induced levels of Bdnf exon IV-(Figure 3e,k) and VI-containing transcripts (Figure 3f,l) were downregulated $2and $6-7-fold, respectively, after targeting dCas9-KRAB to À840 kb enhancer region.Downregulation of different Bdnf transcripts was reflected in total Bdnf mRNA expression levels, which decreased $2-fold in dopamine (Figure 3d) and $6-fold in norepinephrine-treated cells (Figure 3j).Activation of the À840 kb region increased the basal levels of Bdnf $ 4-25-fold and norepinephrine and dopamine-dependent expression levels $2-5-fold (Figure 3d-f,j-l).Collectively, these results confirm the À840 kb region as a functional Bdnf gene enhancer in the endogenous context regulating catecholamine-dependent but not basal expression of Bdnf in cortical astrocytes.

| Deletion of the À840 kb enhancer decreases the catecholamine-dependent expression of Bdnf in mouse embryonic stem cell-derived astrocytes
To confirm the endogenous function of the À840 kb enhancer, we deleted the enhancer region in mouse embryonic stem cells (mESCs).
Based on genotyping results, wild-type mESCs and clones with homozygous À840 kb enhancer deletion (Figure S5) were differentiated into neuronal precursor cells (NPCs) (Figure S6A,B) and astrocytes (Figure S7, Figure 4a).Because differentiation of astrocytes from mouse stem cells lacks a standardized protocol, we tried three different approaches that were selected based on previously published research, and we aimed to obtain astrocytes that show induction of    In NPCs, deletion of the À840 kb region abolished the expression of the eRNAs and slightly decreased the levels of total Bdnf and Bdnf exon VI-containing transcripts (Figure S6C).In induced astrocytes, the deletion abolished the expression of À840 kb eRNAs and decreased the induction of total Bdnf $ 2-fold, Bdnf exon IV transcripts up to $6-fold and Bdnf exon VI-containing transcripts $2-fold in both dopamine and norepinephrine treated astrocytes independent of the differentiation protocol (Figure 4b).
However, the effects of Bdnf mRNA decreases were not statistically significant due to high biological variability between cell clones (Figure 4b).Similar to the endogenous repression of the enhancer in rat cultured astrocytes (Figure 3), deleting the À840 kb enhancer region did not decrease Bdnf expression in unstimulated mESC-derived astrocytes, further corroborating that the À840 kb enhancer only participates in the regulation of Bdnf in stimulated cells.Collectively, homozygous deletion of the À840 kb core enhancer region decreased the catecholamine-dependent expression of Bdnf in mESC-derived astrocytes, providing further evidence for the functional significance of this element in the regulation of Bdnf expression.

| CREB and AP1 family transcription factors regulate the À840 kb enhancer and Bdnf in rat cultured cortical astrocytes
To understand the molecular mechanism of the À840 kb enhancer in cortical astrocytes, we first used luciferase reporter assays to screen transcription factors (TFs) important for the transcriptional activity of the enhancer region.We hypothesized that if we inhibit the binding of TFs that are important for the activity of À840 kb region, then transcription from the enhancer region will decrease.To study this, rat cortical astrocytes were cotransfected with reporter vectors containing the À840 kb enhancer region in either forward (Figure 5a) or reverse orientation (Figure 5d) in front of the luciferase coding sequence together with a panel of dominant-negative TFs, including A-CREB, A-CEBPα, A-CEBPβ, A-FOS, A-USF, and A-ATF2.Of note, all the selected and screened TFs have predicted binding elements in À840 kb region based on JASPAR computational prediction (Castro-Mondragon et al., 2022).
As a control, we cotransfected the EGFP-encoding vector together with the À840 kb enhancer reporter construct.Transfected astrocytes were treated for 8 h with dopamine or DMSO as a vehicle control.(a-h) At 13 DIV, rat cortical astrocytes were cotransfected with different dominant-negative transcription factors (termed as A in the name), constitutively active form of CREB (VP16-CREB), FOS (VP16-FOS), JUN (VP16-JUN) or EGFP and luciferase reporter vectors as indicated under the graphs: À840 kb enhancer region in forward (a-c) or reverse (d-f) orientation before luciferase coding sequence or rat BDNF promoter IV (g) or VI (h) before luciferase coding sequence and À840 kb enhancer region in forward and reverse orientation after the polyadenylation signal (PolyA).Transfected astrocytes were treated for 8 h with 0.15% DMSO as control or with 150 μM dopamine (DA).The results are shown relative to the luciferase activity measured from DMSO-treated astrocytes that were cotransfected with EGFP and respective reporter vector as shown below each panel.(i, k) At 7 DIV, rat cortical astrocytes were transduced with lentiviruses expressing either EGFP, dominant-negative for the CREB family (A-CREB, i), or dominant-negative for the AP1 family (A-FOS, k).At 15 DIV, astrocytes were treated with 0.15% DMSO as a control or 150 μM dopamine (DA) for 3 h.The RNA expression of À840 kb enhancer RNAs, total Bdnf (Bdnf coding exon), Bdnf exon IV-and exon VI-containing transcripts was measured using RT-qPCR.The expression level of the respective transcript in cells treated with DMSO and transduced with EGFP-encoding lentiviruses was set as 1 (i, k).Error bars indicate SEM (n = 3 independent experiments).(j, l) Endogenous binding of CREB, pCREB and FOS was analyzed with ChIP-qPCR.At 15 DIV, astrocytes were treated for 3 h with 0.15% DMSO as a control or 150 μM DA.A region $5 kb downstream of the À840 kb enhancer was used as a negative control region.Data is depicted as enrichment relative to the control region (dashed line equals to 1).Error bars indicate SEM (n = 4 independent experiments).(a-h) Statistical significance was calculated using two-tailed paired t-test relative to the luciferase signal measured from respectively treated cells cotransfected with the respective reporter vector and EGFP plasmid (a-h), relative to the mRNA levels of respective transcript in respectively treated cells infected with lentiviruses encoding EGFP (i, k), or relative to the enrichment of DNA at the negative control region at respective treatment or the respective region in DMSO-treated astrocytes (j, l).NS, not significant, *p < .05;**p < .01;***p < .001;p-values were corrected for multiple comparisons with Holm-Šidak method.To investigate whether the TFs controlling the activity of the À840 kb region in the reporter assay also participate in the regulation of endogenous Bdnf gene expression, rat cortical astrocytes were transduced at 7 DIV with lentiviruses encoding either EGFP, A-CREB (Figure 5i) or A-FOS (Figure 5k).The overexpression of dominant-negative CREB did not affect the basal expression levels of À840 kb eRNAs, but decreased the basal expression levels of exon IV-containing Bdnf levels $3-fold (Figure 5i).Conversely, A-CREB inhibited the expression of À840 kb eRNAs, total Bdnf and exon IVand VI-containing transcripts $2-10-fold in dopamine-treated cells (Figure 5i).Using ChIP-qPCR, we confirmed direct binding of CREB to À840 kb enhancer region in DMSO-and DA-treated astrocytes ($3-5-fold higher enrichment relative to a control region) and binding of the active form of CREB, phospho-CREB (pCREB), which increased $3-fold following a dopamine treatment (Figure 5j).Overexpression of A-FOS decreased the expression of À840 kb region eRNAs $1.7-fold in dopamine-treated cells (Figure 5k).While overexpression of A-FOS did not affect the basal levels of Bdnf, A-FOS decreased total Bdnf and exon VI-containing Bdnf mRNA (Figure 5k) levels $2-fold in dopamine-treated astrocytes.Interestingly, Bdnf exon IV mRNA levels were upregulated $2.3-fold in dopamine-treated A-FOS overexpressing cells (Figure 5k).We also detected endogenous binding of FOS-family TFs to the À840 kb enhancer region with ChIP-qPCR (Figure 5l), where FOS binding increased $5-fold in dopamine-treated cells.Although CREB and AP1 family TFs also showed some interaction with Bdnf promoters IV and VI, our results collectively point to the importance of the recruitment of these TFs to the À840 kb enhancer region for stimulating Bdnf gene expression in cortical astrocytes.
3.7 | CREB and AP1 cis-elements within the À840 kb enhancer are required for the enhancer activity Finally, we examined the role of CREB and AP1 transcription factor binding sites in the transcriptional regulation of À840 kb enhancer.
First, we used JASPAR cis-element prediction of CREB and AP1 binding sites and found two CRE (cAMP responsive element, CRE-1 and CRE-2) and two AP1 sites (AP1-1 and AP1-2) in rat À840 kb enhancer (Figure 6a).Comparison of the sequences among rat, mouse and human revealed high evolutionary conservation of the identified binding sites, with the exception of AP1-2, which differs by a single nucleotide in human compared to rat and mouse (Figure 6a).Next, we mutated the CREB and AP1 binding sites in the rat À840 kb enhancer reporter constructs and used luciferase reporter assay to study the importance of these sites in the bidirectional transcription of the enhancer.
In DMSO-treated astrocytes, mutation only in the AP1-1 site showed $3-4-fold downregulation of the À840 kb enhancer activity.
Next, we analyzed whether the identified cis-elements in the À840 kb enhancer affect the transcription from Bdnf promoters IV (Figure 6c) and VI (Figure 6d).Out of four tested cis-elements, mutations in the CRE-1, CRE-2 and AP1-1, but not in AP1-2, decreased the dopamine-dependent transcription from Bdnf promoter IV and VI.The strongest effect was seen with the mutations in the CRE-1 site which decreased the dopamine-dependent induction of Bdnf promoter IV $6-9-fold (Figure 6c) and Bdnf promoter VI $4-5-fold (Figure 6d).Taken together, these results show that all four identified transcription factor binding sites are responsible for the stimulus-dependent activity of the rat À840 kb enhancer region, and specifically CRE-1, CRE-2 and AP1-1 sites are necessary to increase the Bdnf promoter IV and VI activity after dopamine treatment in cortical astrocytes.
The multi-promoter composition of Bdnf gene allows generating alternative Bdnf transcripts, which result in identical BDNF protein.This kind of gene regulation ensures complex spatiotemporal expression of BDNF (Esvald et al., 2023;West et al., 2014).For instance, the first cluster of Bdnf transcripts is expressed in neurons, but not in nonneuronal cells, and, in contrast, the second cluster of Bdnf transcripts is expressed ubiquitously in both neuronal and non-neuronal tissues (Aid et al., 2007;Esvald et al., 2023;Pruunsild et al., 2007;Timmusk et al., 1993).Likely, this spatiotemporal regulation is maintained by both cell type-specific activity of proximal promoters (Palm et al., 1999;Zuccato et al., 2003) and recruitment of specific enhancer regions (Brookes et al., 2023;Flavell et al., 2008;Lyons et al., 2012;Tuvikene et al., 2021).In this study we aimed to identify distal regulatory regions that participate in the catecholamine-dependent expression of Bdnf in rodent astrocytes.Of the four candidate enhancer regions that we initially investigated, only one region mapping 840 kb upstream of Bdnf gene showed properties of a stimulus-dependent enhancer in astrocytes, and interaction with Bdnf promoter VI.It is possible that the rest of the bioinformatically predicted enhancer regions (À450, À40 and +37 kb) would function upon different stimuli or in another cell type than studied here, highlighting the uniqueness of each enhancer.
The astrocytic BDNF seems to have similar roles as neuronderived BDNF, for example, it modulates neuronal morphology, plasticity, and survival (De Pins et al., 2019;Giralt et al., 2010Giralt et al., , 2011;;Vignoli et al., 2016).However, in general the kinetics of exocytosis in astrocytes is slower compared to that in neurons (Kreft et al., 2004;Verkhratsky et al., 2016).Furthermore, due to different dynamics of BDNF secretion from neurons and astrocytes, the neuron-and astrocyte-derived BDNF distinctively modulates the hippocampal long-term potentiaton and learning (Liu et al., 2022).The physiological function of dopamine-and norepinephrine-induced BDNF production in astrocytes remains to be investigated.In addition to neuron-derived BDNF, it is plausible that astrocytes within the dopaminergic and  (Joo et al., 2015).However, we acknowledge that our CAPTURE-3Csequencing might lack the resolution and statistical power to detect the possible short-range interactions within the Bdnf locus.We have previously shown that Bdnf is regulated as two clusters of promoters (Aid et al., 2007;Esvald et al., 2023;Pruunsild et al., 2007;Timmusk et al., 1993), and that the intronic enhancer located downstream of exon III is responsible for neuron-specific expression of the first cluster of Bdnf exons (Tuvikene et al., 2021).It is plausible that the enhancers of Bdnf loop together with the anchor point located nearby the first, second, or both of the clusters depending on the cell type and stimulus.Further work is needed to determine the role and/or interplay of the potential enhancer regions to pinpoint the exact mechanism of the catecholamine-dependent Bdnf expression.
Although the transcriptional regulation of Bdnf in astrocytes is poorly investigated, catecholamines have been shown to induce Bdnf gene expression in astrocytes via CREB-dependent signaling (Jurič et al., 2008;Koppel et al., 2018).Our work provides further mechanistic insight by showing that AP1 family transcription factors act as regulators of astrocytic Bdnf expression.Interestingly, Koppel et al. (2018) showed that the well-studied CRE element in Bdnf promoter IV is not involved in catecholamine-dependent Bdnf regulation in astrocytes, whereas this site is critical for stimulusdependent expression of BDNF in neurons (Esvald et al., 2020;Hong et al., 2008;Pruunsild et al., 2011;Shieh et al., 1998;Tao et al., 1998).Similarly, AP1 family members are known to regulate the expression of the neuronal first cluster of Bdnf transcripts (Tuvikene et al., 2016) that are not expressed in astrocytes.These In neurons, the AP1 family plays a crucial role in the regulation of activity-dependent enhancers.The AP1 motifs are one of the most prevalent cis-elements in enhancers, and the regulation of neuronal enhancers relies on the rapid induction and binding of AP1 family transcription factors (Malik et al., 2014).The AP1 family acts as an epigenetic regulator, modifying chromatin accessibility around their binding sites throughout the genome (Malik et al., 2014;Su et al., 2017).To achieve this, the AP1 factors recruit ATP-dependent BRG1/BRM associated factor (BAF) chromatin remodeling complexes, which, in turn, facilitate nucleosome remodeling and generate an accessible chromatin state (Vierbuchen et al., 2017).In our study, we identified two AP1 cis-elements within the À840 kb enhancer region, both of which are significant for enhancer activity.However, only one of these AP1 cis-elements (AP1-1), located in the center of the À840 kb enhancer, is involved in enhancing the activity of Bdnf promoters.Interestingly, the other AP1 cis-element (AP1-2) appears to be non-functional for Bdnf induction in astrocytes, but may play a crucial role in regulating Bdnf activity in other cell types.Currently, little is known about the general mechanism of enhancers in astrocytes, but our results studying this stimulus-specific enhancer in depth suggests a similar central role for AP1 factors in astrocytes as observed in neurons.A natural continuation of this line of work would be to investigate the role of AP1 family transcription factors in astrocyte-specific enhancers using genome-wide approaches.
Notably, the À840 kb enhancer is relevant in vivo.First, in 2006, a case-report of an 8-year-old girl suffering from hyperphagia, severe obesity, impaired memory and hyperactivity was published.Furthermore, it was determined that the BDNF protein levels were decreased in the patient's serum despite the fact that BDNF gene itself was not mutated (Gray et al., 2006).Genetic analysis revealed that the patient had a chromosomal inversion with a breakpoint located within a 9 kb region at the À840 kb enhancer locus.Second, a mouse line with transgene insertion in a conserved region located approximately 840 kb upstream of the Bdnf gene (Sha et al., 2007) was reported to have a similar phenotype as Bdnf knockout mice (Lyons et al., 1999;Rauskolb et al., 2010) and patient described in Gray et al. (2006).
Furthermore, the transgene insertion strikingly reduced both Bdnf mRNA and protein levels in the hypothalamus and in the hippocampus
1. p-values were corrected for multiple comparisons using Benjamini & Hochberg false discovery rate method.For graphical visualization, the number of anchor-interacting reads in each bin was divided with the number of n anchor reads in the respective sample and multiplied by the mean of n anchor reads in both DMSO-and DA-treated samples, to account for slight variations in sequencing depth.Consecutive and overlapping bins with adjusted p-value <.05 were flattened into one interacting region, and maximum read count was used for graphical depiction of the data.All regions analyzed in Capture-3C-sequencing are listed in the Supplementary file 1.

2. 12 |
figure legends, and p-values were corrected for multiple comparisons with Holm-Šidak method using Prism 9.5.1 (GraphPad Software).For graphical representation, data was backtransformed and error bars indicate upper and lower limits of backtransformed means ± SEM.
GeneHancer identifier is a unique tag for the enhancer region in the GeneHancer database.Enhancer score describes the likelihood of the region being an enhancer based on integrated data from different databases, BDNF gene-enhancer score reflects the association between BDNF gene expression and the activity of an enhancer, and total score is the multiplication of gene-enhancer and enhancer scores.Enhancer sources depict datasets that support the region as an enhancer.TSS distance (kb) shows the distance from the human BDNF exon I TSS to the midpoint of enhancer, positive values stand for regions downstream of the BDNF exon I TSS and negative values show upstream regions.Gene targets for enhancer shows the number of possible target genes.The table also provides the information on the size of the enhancer regions, numbers of transcription factors (TFs) that have been shown to bind the enhancer in ChIP-seq experiments, and the coordinates of the regions in mouse, rat and human genome are also listed.Our annotation takes into account the distance variations between rat and human genome and is based on the rat genome.U R E 1 The À840 kb region shows orientation-independent transcriptional activity and potentiates transcription from Bdnf promoters in a dopamine-dependent manner in cultured rat cortical astrocytes.(a) Schematic representation of rodent Bdnf gene structure together with putative enhancer regions screened in luciferase assay.(b-d) Rat cortical astrocytes were transfected at 13 DIV with luciferase reporter vectors indicated on the graphs: rat À840, À450, À40 or +37 kb putative enhancer region in either forward or reverse orientation (b), the conserved core region of either rat, mouse or human À840 kb enhancer in forward or reverse orientation in front of the luciferase coding sequence (Luc2P) followed by polyadenylation signal (PolyA) (c) or luciferase reporter constructs containing rat Bdnf promoter IV or VI upstream of the Luc2P coding sequence and the À840 kb region in either forward or reverse orientation downstream of the PolyA (d).

F
I G U R E 2 Bdnf promoter VI region interacts with various distal genomic regions in cultured cortical astrocytes in a stimulus-dependent manner.(a) Schematic depiction (created with BioRender.com) of CAPTURE-3C-sequencing protocol that was used to determine long-range chromatin interactions with the Bdnf promoter VI region.(b) The statistically significant long-range interactions of Bdnf promoter VI are shown with arcs, and the arc height is proportional to the number of sequencing reads for that interaction.x-axis shows the genomic coordinates in the rat Rn6 genome.Blue-DMSO-treated cells, red-DA-treated cells.The À840 kb enhancer region is shown with an arrow.Only interactions with adjusted p-value <.05 are shown.The number of reads and whole statistical analysis of ±3 Mb region of Bdnf pVI is shown in Supplementary file 1.

Bdnf
expression in response to dopamine and norepinephrine treatments.The differentiation strategies used either serum, serum with B27, or serum with B27 and LIF.These conditions led to the maturation of induced astrocytes in time (FigureS7A) and resulted in varied astrocyte morphology and marker gene expression (FigureS7B,C), possibly representing different subtypes and/or maturation stages of astrocytes.The cells were differentiated for 21 days to obtain astrocytes responding to dopamine and norepinephrine treatment.On the 21st day of differentiation the induced astrocytes were treated for 3 h with dopamine or norepinephrine and the expression levels of different Bdnf transcripts and À840 kb eRNA levels were measured using RT-qPCR.
Dopamine-and norepinephrine-induced Bdnf expression in rat cortical astrocytes is regulated by the À840 kb enhancer.(a, g) At 7 DIV, cultured rat cortical astrocytes were transduced with lentiviruses expressing dCas9-KRAB or VP64-dCas9-VP64 and 5 gRNAs specific for the À840 kb enhancer region (À840 kb gRNA) or negative control gRNA (neg gRNA) as shown in the schematic.(b-f, h-l) At 15 DIV, astrocytes were treated for 3 h with 0.15% DMSO as a vehicle control or 150 μM dopamine (DA, b-f) or 25 μM norepinephrine (NE, h-l).The expression of FosB (b, h), À840 kb enhancer RNA (c, i), Bdnf coding exon (d, j) and Bdnf exon IV-(e, k) and exon VI-containing (f, l) transcripts were measured using RT-qPCR.The expression level of respective transcript in cells transduced with dCas9-KRAB or VP64-dCas9-VP64 and negative gRNAencoding lentiviruses and treated with DMSO was set as 1.Error bars indicate SEM (n = 4 [b-f, n = 3 [h-l] independent experiments).Statistical significance was calculated with two-tailed paired t-test relative to the levels of respective transcript in cells infected with lentiviruses encoding negative gRNA and either dCas9-KRAB or VP64-dCas9-VP64 upon respective treatment.NS, not significant, *p < .05;**p < .01;***p < .001,p-values were corrected for multiple comparisons with Holm-Šidak method.
Deletion of the À840 kb enhancer region decreases the expression of Bdnf in mouse embryonic stem cell-derived astrocytes.(a) CRISPR/Cas9 system was used to delete the $500 bp conserved region of the À840 kb enhancer.Mouse embryonic stem cells with an intact À840 kb region (WT, n = 4 clones) or with a homozygous deletion of the À840 kb enhancer conserved core region (À840 kb KO, n = 3 clones) were differentiated to NPCs and then astrocytes using FBS alone or along with B27 (containing retinoic acid) or together with B27 and LIF as shown in the legend (illustration on panel a was created with BioRender.com).On the 21st day the induced astrocytes were treated for 3 h with 150 μM dopamine (DA) or 25 μM norepinephrine (NE) or with DMSO as a control.(b) The expression levels of À840 kb eRNAs, total Bdnf (Bdnf coding exon) or different Bdnf transcripts were measured using RT-qPCR.The average mRNA expression level relative to the housekeeper Cnot4 expression level is shown with error bars indicating SEM.Statistical significance was calculated with two-tailed unpaired unequal variance t-test relative to the expression level of the respective transcript in wildtype cells at the respective treatment.NS, not significant, #p < .1;*p < .05;**p < .01;***p < .001,p-values were corrected for multiple comparisons with Holm-Šidak method.ND, not detected.
U R E 5 CREB and AP1 family transcription factors regulate the expression of Bdnf and À840 kb eRNA in rat cultured cortical astrocytes.
Basal transcriptional activity of the À840 kb region was decreased $2-3-fold in cells overexpressing A-FOS, a dominantnegative protein for AP1 family (Figure5a,d).After dopamine treatment A-FOS inhibited the activity of the À840 kb enhancer region $2-fold (Figure 5a,d).In addition to AP1 family, the dominantnegative form of the CREB family, A-CREB, almost abolished the induction of À840 kb transcription in cortical astrocytes upon dopamine treatment (Figure 5a,d).Other tested dominant negative proteins did not have a notable effect on the À840 kb region enhancer activity (Figure 5a,d).In addition, we overexpressed CREB or FOS and JUN fused to the VP16 activator domain together with À840 kb enhancer region to elucidate whether the CREB and AP1 family TFs can increase the transcription from the enhancer.The overexpression of constitutively active form of CREB (VP16-CREB) robustly increased the transcriptional activity of the À840 kb enhancer both in DMSO-and dopamine-treated cortical astrocytes (Figure 5b,e), while overexpression of VP16-FOS and VP16-JUN enhanced the transcriptional activity by $2-fold only in DMSO-treated cortical astrocytes (Figure 5c,f), possibly due to saturation of À840 kb enhancer by endogenous AP1 family members after dopamine treatment.Next, we studied whether the two dominant-negative TFs (A-CREB and A-FOS) regulate the À840 kb enhancer regionpotentiated activation of Bdnf promoter IV and VI.For that, cells were cotransfected with A-CREB and A-FOS (Figure 5g,h) together with reporter constructs where Bdnf promoter IV (Figure 5g) or VI (Figure 5h) was in front of the luciferase coding sequence and the À840 kb enhancer region in forward or reverse orientation after the polyadenylation signal (PolyA).Overexpression of A-CREB markedly decreased the À840 kb region-enhanced activity of Bdnf promoter IV (Figure 5g) and VI (Figure 5h) in dopamine-treated cortical astrocytes, whereas A-FOS slightly decreased ($1.3-1.8-fold) the activity of Bdnf promoter IV and VI (Figure 5g,h).These results indicate that CREB and AP1 family members mediate the activation of Bdnf promoters by the À840 kb enhancer region in a heterologous context in cortical astrocytes.
Mutations in the CRE and AP1 sites in the À840 kb region decrease the activity of the enhancer and Bdnf promoters in cultured rat cortical astrocytes.(a) CRE and AP1 sites in the rat, mouse and human À840 kb enhancer.Nucleotides shown in bold indicate the CREB or AP1 family transcription factor core binding sites, numbers indicate the sequence location counted from the 5 0 of the cloned full-length À840 kb enhancer region.Mutated nucleotides are shown below the wild-type binding sites.(b-d) Rat cortical astrocytes were transfected at 13 DIV with luciferase reporter vectors indicated on the graphs: wild-type (WT) or mutated rat À840 kb enhancer region in either forward or reverse orientation (b) or luciferase reporter constructs containing rat Bdnf promoter IV or VI upstream of the Luc2P coding sequence and the WT or mutated À840 kb region in either forward or reverse orientation downstream of the PolyA (c, d).Luciferase vector with WT enhancer region was used as a control.At 15 DIV, transfected astrocytes were treated for 8 h with 0.15% DMSO as a control or 150 μM dopamine (DA).The results are depicted relative to the luciferase signal measured from DMSO-treated astrocytes transfected with WT enhancer construct (b-d).Error bars indicate SEM (n = 3 independent experiments).Statistical significance was calculated with two-tailed paired t-test relative to the luciferase level measured from DMSO-treated cells transfected with luciferase vector with WT enhancer (b), or relative to the luciferase level measured from cells transfected with respective promoter-containing vector with WT enhancer (c, d).NS, not significant, *p < .05;**p < .01;***p < .001,p-values were corrected for multiple comparisons with Holm-Šidak method.noradrenergictripartite synapses express BDNF to augment BDNF signaling, thereby contributing to several neuronal functions.What is the exact role of catecholamine signaling-derived astrocytic BDNF in the cerebral cortex and whether the À840 kb enhancer facilitates the expression of astrocytic BDNF similarly in different brain regions, are important questions to address in the future.Koppel et al. (2018) showed for the first time that the catecholamine-dependent induction of Bdnf in astrocytes is not conferred only by the Bdnf proximal promoter regions.By using BDNFbacterial artificial chromosome (BAC)-based reporters encompassing the whole rat or human BDNF gene locus, they found that at least some distal regulatory regions that facilitate the catecholamineinduced levels of Bdnf in cortical astrocytes are located within or near the Bdnf gene.Notably, the À840 kb region described in the present study is outside of the BDNF-BAC constructs used byKoppel et al. (2018).Together, these results show the existence of at least two enhancers, one 840 kb upstream of Bdnf described in the present study, and another within or nearby of Bdnf gene locus described byKoppel et al. (2018), that are responsible for catecholamine-induced regulation of Bdnf.It is also plausible that Bdnf gene regulation in astrocytes is even more complex, considering the numerous longrange chromatin interactions detected by our CAPTURE-3Csequencing experiment.Namely, we determined that Bdnf pVI interacts with À840 kb region upon dopamine treatment, but we also showed dopamine-dependent interactions with À990 kb and À2.65 Mb regions.Moreover, we found interaction with À790 kb region only in DMSO-treated cells and a stimulus-independent interaction with a region located at À1.82 Mb.This indicates that the transcription of Bdnf gene is subject to specific combinatorial regulation by different enhancer regions, as it has been demonstrated for c-Fos results argue for substantial differences in Bdnf gene regulation between different cell types even when the same transcription factor families are involved, implying involvement of other regulatory regions.Here, we demonstrate that the CREB and AP1 familydependent regulation of Bdnf in astrocytes occurs via the À840 kb enhancer region.We show that mutations in any of the CRE sites and one AP1 site within the À840 kb enhancer decrease the activity of Bdnf promoters IV and VI.Collectively, our study is the first report showing an enhancer coordinating the transcription of Bdnf gene in non-neuronal cells and demonstrating the fine-tuning of gene expression based on cellular context and signaling.