BDNF-DT and BDNF-AS-DT: Novel Genes in the BDNF locus

BDNF-DT and BDNF-AS-DT: Novel Genes in the BDNF locus | medRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-P4HH5NV'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search BDNF-DT and BDNF-AS-DT : Novel Genes in the BDNF locus Svitlana V. Bach , Giovanna Punzi , Nuri E. Smith , Sreya Mukherjee , Joo Heon Shin , Qiang Chen , Geo Pertea , Leonardo Collado-Torres , Kristen R. Maynard , Stephanie C. Page , Joel E. Kleinman , Thomas M. Hyde , Daniel R. Weinberger , View ORCID Profile Keri Martinowich , View ORCID Profile Gianluca Ursini doi: https://doi.org/10.1101/2025.07.10.25331311 Svitlana V. Bach 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Giovanna Punzi 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA 2 Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Nuri E. Smith 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Sreya Mukherjee 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Joo Heon Shin 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Qiang Chen 3 Stem Cell Translation Laboratory (SCTL), Division of Preclinical Innovation, National Center for Advancing Translational Sciences (NCATS), National Institutes of Health (NIH) , Rockville (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Geo Pertea 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Leonardo Collado-Torres 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA 4 Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health , Baltimore, MD, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Kristen R. Maynard 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA 2 Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine , Baltimore (MD), USA 5 The Solomon H. Snyder Department of Neuroscience, Johns Hopkins School of Medicine , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Stephanie C. Page 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Joel E. Kleinman 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA 2 Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Thomas M. Hyde 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA 2 Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine , Baltimore (MD), USA 6 Department of Neurology, Johns Hopkins School of Medicine , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Daniel R. Weinberger 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA 2 Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine , Baltimore (MD), USA 5 The Solomon H. Snyder Department of Neuroscience, Johns Hopkins School of Medicine , Baltimore (MD), USA 6 Department of Neurology, Johns Hopkins School of Medicine , Baltimore (MD), USA 7 McKusick-Nathans Institute, Department of Genetic Medicine, Johns Hopkins School of Medicine , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Keri Martinowich 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA 2 Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine , Baltimore (MD), USA 5 The Solomon H. Snyder Department of Neuroscience, Johns Hopkins School of Medicine , Baltimore (MD), USA 8 Johns Hopkins Kavli Neuroscience Discovery Institute , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Keri Martinowich For correspondence: keri.martinowich{at}libd.org gianluca.ursini{at}libd.org Gianluca Ursini 1 Lieber Institute for Brain Development, Johns Hopkins University Medical Campus , Baltimore (MD), USA 2 Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine , Baltimore (MD), USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Gianluca Ursini For correspondence: keri.martinowich{at}libd.org gianluca.ursini{at}libd.org Abstract Full Text Info/History Metrics Data/Code Preview PDF Abstract Divergent transcription from bidirectional promoters is frequently observed in eukaryotic genomes, but the biological relevance of divergent RNA transcripts (DT) is unknown. We identified and characterized BDNF-DT , a novel DT gene, and BDNF-AS-DT , a novel readthrough gene, in the locus containing BDNF , a gene with key roles in neuronal development, differentiation, and synaptic plasticity. BDNF-DT is independent from the known BDNF antisense ( BDNF-AS ), and its expression is developmentally regulated and positively correlated with BDNF in human postmortem dorsolateral prefrontal cortex (DLPFC). BDNF-DT and BDNF-AS-DT expression increase after induced depolarization, but the temporal dynamics follow expression of BDNF, suggesting a regulatory role. Moreover, CRISPR-mediated upregulation of BDNF in human neural progenitor cells drives BDNF-DT expression. Finally, BDNF-DT shows higher expression in DLPFC from patients diagnosed with schizophrenia compared to neurotypical controls, and genetically predicted lower expression of the BDNF-AS-DT readthrough transcript is associated with schizophrenia and with the schizophrenia-associated C allele of the rs6265 single-nucleotide polymorphism. These data suggest that BDNF-DT and BDNF-AS-DT contribute to BDNF regulation and schizophrenia risk. Introduction Divergent transcription – the phenomenon of RNA polymerase II transcribing in opposing directions from active promoters - has been described in many eukaryotic organisms ( 1 ). Protein-coding/long non-coding RNA (lncRNA) pairs resulting from divergent transcription of bidirectional promoters are often observed among transcriptionally regulated protein-coding transcripts ( 2 , 3 ). Such recurring phenomena raise the possibility that divergent transcription fine-tunes regulation of protein-coding RNAs. Divergent RNAs (divRNAs) contribute to many aspects of gene expression, from transcription to translation to RNA degradation, and are involved in both activating and repressing cis (local) and trans (distant) gene expression ( 1 , 3 – 13 ). In addition, divRNAs that partially overlap protein-coding genes may share features of antisense transcripts ( 14 ) and mediate effects on translation of their coupled protein-coding genes. DivRNAs contribute to intergenic transcription, together with readthrough transcription, a phenomenon that occurs when RNA polymerase II fails to recognize termination signals and continues transcribing beyond the gene’s end ( 15 – 17 ). Cells of neural origin are highly transcriptionally active ( 18 ). DivRNAs are poised to contribute to molecular complexity in the brain because they represent a substantial amount of the lncRNAs that are expressed in transcriptionally active cells ( 19 ), and because they mediate diverse regulatory and functional roles in transcriptional, post-transcriptional, epigenetic, and nuclear processes ( 1 , 3 – 14 ). However, while regulation by divRNAs represents a potential mechanism for tuning expression and splicing of protein-coding RNAs that are important for brain development and function, functional characterization of brain-expressed divRNAs remains limited. Brain-derived neurotrophic factor ( BDNF ) encodes a secretory protein that plays key roles in neuronal development, differentiation and synaptic plasticity ( 20 ), and is implicated in multiple neuropsychiatric disorders ( 21 ). Conflicting results regarding BDNF associations with disease ( 22 ) are complicated by difficulty in disentangling the complex regulation of BDNF transcription. BDNF contains at least 9 unique promoters, which are spliced to a common 3’ coding exon, generating at least 22 transcripts encoding an identical BDNF protein. The locus also comprises several differentially spliced lncRNAs of the antisense BDNF-AS gene ( 23 ) ( Fig.1 ). While BDNF is of significant interest as a therapeutic target, there are caveats associated with globally targeting BDNF signaling. For example, alterations in BDNF signaling map to discrete brain regions for different diseases ( 21 ). In addition, BDNF operates within a narrow window of optimal concentration to support trophic signaling in neurons. Indeed, its levels must be tightly regulated to prevent excessive neural activity, and overexpression can result in hyper-excitability and neuronal death ( 24 – 28 ). Download figure Open in new tab Fig. 1. Structure, location, and expression of BDNF-DT and BDNF-AS-DT transcripts relative to BDNF and BDNF-AS in the human genome. ( Top ) BDNF mRNA structure with corresponding exons ( I - IXa ; red) and the coding exon IX (wide red box). BDNF-AS structure with corresponding exons ( 1-9 ; green). BDNF-AS-DT readthrough transcript structure with corresponding exons ( 1 , 4 , 6 , 8 , 9 , β , γ , and θ ; blue). BDNF-DT1 , 2 , and 3 transcript structure with corresponding exons ( α - θ ; black). ( Bottom ) RNAseq read coverage track from bulk DLPFC postmortem human brain tissue in neurotypical controls (NC; lavender) and patients with schizophrenia (SCZ; navy), illustrating expression of BDNF-DT exons (pink horizontal lines) vs. BDNF coding exon IX (light blue horizontal line). Scale bar on the lower right is 3 kb. We identified evidence for divergent transcription in the BDNF locus by analyzing a large RNA sequencing (RNAseq) dataset from postmortem human brain tissue ( 29 ) ( Fig.1 ). We went on to rigorously characterize BDNF divergent transcripts ( BDNF-DT ) and establish activity-dependent kinetics of BDNF and its divergent transcripts in both human and mouse cells. We further mechanistically determined whether expression of BDNF from endogenous promoters alters divergent transcription. Our findings provide evidence of the existence of novel genes in the BDNF locus: BDNF-DT and the BDNF-AS-DT readthrough gene that is associated with genetic risk for schizophrenia. Our data suggest that BDNF-DT and BDNF-AS-DT provide an additional layer of regulation over BDNF signaling during development and in response to neuronal activity, which are relevant in the etiopathogenesis of neuropsychiatric disorders. Results Characterization of novel divergent antisense transcription in the human BDNF locus Novel transcription and splicing events can be pinpointed in RNAseq data by identifying reads that do not align to known transcript sequences. We re-analyzed a large RNAseq dataset from postmortem dorsolateral prefrontal cortex (DLPFC) tissue from neurotypical control donors and individuals diagnosed with schizophrenia ( 29 ). RNAseq libraries in this dataset were generated using Poly-A + selection, so stable transcripts should be more readily detected compared with unstable transcripts (e.g. short bidirectional transcripts, upstream antisense RNAs or promoter upstream transcripts) ( 30 ). To comprehensively characterize the complex gene regulation of BDNF , we used the Integrative Genomics Viewer software (IGV) ( 31 , 32 ) to identify novel transcription and splicing events. Analyzing this dataset ( 29 ) (see Table S1 for sample characteristics), we noted reads in the 5’ upstream region of BDNF exon I ( Fig.1 ) that are not annotated in the current reference transcriptome (Ensembl build GRCh38/hg38). These reads aligned to the positive DNA strand, antisense to BDNF , which is transcribed from the negative strand ( Fig.1 ). Some reads partially overlapped BDNF exons I and II ; however, most transcripts were detected in the region 5’ to the BDNF exon I transcription start site (TSS). We validated existence of these reads with PCR on cDNA from independent DLPFC postmortem human brain samples, followed by cloning and sequencing ( Fig.S1 ). We named this new gene BDNF divergent transcript (symbol: BDNF-DT ), consistent with the recommendations of the HUGO human gene nomenclature committee ( 33 ). We used 5’ and 3’ RACE-PCR with gene-specific primers on the novel exons to detect the 5’ start and the 3’ end of BDNF-DT ( Fig.S1 ). We found two alternative 3’ ends, located respectively 3,720 and 52,445 kb upstream of BDNF exon I and three alternative 5’ starts, which overlap, respectively, BDNF exon I , BDNF exons II and the exon I of BDNF-AS , the previously annotated antisense in the BDNF locus. After defining the 5’ starts and the 3’ ends, we performed end-to-end PCR to characterize the full-length structure of the divRNAs: we detected three main transcripts of BDNF-DT (here referred to as BDNF-DT1 , 2 , and 3 ) and one BDNF-AS-DT readthrough transcript ( Fig.1 and Fig.S2 ; see Table S2 for transcript sequences), that is, a RNA molecule formed via the splicing of exons from BDNF-AS and BDNF-DT . Of note, BDNF-AS-DT does not overlap the protein-coding exon of BDNF , which differs from BDNF-AS . We also re-analyzed RNAseq data from an independent postmortem DLPFC dataset ( 34 ), as well as datasets from postmortem human hippocampus ( 34 , 35 ) and anterior cingulate cortex ( 36 ) ( Fig.S3, see Table S1 for sample characteristics). Because RNAseq libraries in these datasets were generated using ribosomal RNA depletion rather than Poly-A+ selection, unstable transcripts may have been detected, likely increasing transcriptional noise ( 30 , 37 , 38 ). However, coverage of RNAseq reads in these data is consistent with the existence of BDNF divergent and readthrough transcription not only in DLPFC ( Fig.S3a,b ), but also in hippocampus ( Fig.S3c,d ), including within the dentate gyrus ( Fig.S3e,f ), as well as within the subgenual and dorsal subdivisions of the anterior cingulate cortex ( Fig.S3g,h ). Further, by exploring a comprehensive resource of RNAseq data from 54 non-diseases tissues ( 39 ), we detected RNAseq reads supporting divergent transcription in the BDNF locus in many brain areas, including amygdala, anterior cingulate cortex, cerebellum, frontal cortex, hippocampus, and hypothalamus. Among the non-brain tissues, divergent transcription was blunted in most tissues, but particularly high in the pituitary gland ( Fig.S3i ). Relationship between BDNF-DT, BDNF-AS-DT and BDNF transcripts in human cortex across development To investigate how BDNF-DT expression changes in DLPFC across development we further examined the postmortem human Poly-A + RNAseq dataset described above ( 29 ) ( Table S1 ). Analyzing data from just neurotypical brain donors, BDNF-DT expression is associated with age (N=296, F=11.03, p=2.47e-06); it increases during prenatal and neonatal life, until approximately 1 year of age, slightly decreases during childhood, until 12 years old, rises during adolescence and young adulthood, until 25 years old, and then drops during adulthood and senescence ( Fig.2a ). These changes mirror the developmental trajectory of BDNF in early life and adulthood, where expression increases until early postnatal life, followed by a decrease at the childhood-to-adolescence transition, and a further decrease during adulthood and senescence ( Fig.2b ). The BDNF-AS-DT transcript is also developmentally regulated (N=296, F=7.582, p=8.08e-06) and its association with age is similar to that of BDNF-DT and BDNF ( Fig.2a,b ). Specifically, it increases in prenatal and neonatal life, decreases in the childhood-adolescence transition, and decreases in adulthood ( Fig.2c ). However, the developmental trajectory of BDNF-DT , BDNF-AS-DT and BDNF are different compared with the previously characterized BDNF-AS ( Fig.2d ); BDNF-AS increases in adulthood and senescence, consistent with established patterns of other antisense RNAs ( 40 , 41 ). Download figure Open in new tab Fig. 2. Developmental regulation of BDNF-DT and the known sense and antisense transcripts of the human BDNF locus. Adjusted gene expression (y-axis) of BDNF-DT ( A ), BDNF ( B ), BDNF-AS-DT ( C ), and BDNF-AS ( D ) is shown in relationship with age (orange line, x-axis, logarithmic scale). Age has a strong non-linear association with expression of BDNF-DT (N=296, F=11.03, p=2.47e-06), BDNF (N=296, F=41.91, p<2.2e-16), BDNF-AS-DT (N=296, F=7.58, p=8.08e-06), and BDNF-AS (N=296, F=132.92, p13 years; Table S1 ). The expression of BDNF-DT is positively correlated with BDNF (N=219, r=0.168, t=2.768, p=0.006) although BDNF is highly expressed in some individuals with low BDNF-DT expression ( Fig.S4a ). BDNF-DT is not correlated with BDNF-AS (N=219, r=0.05, t=1.119, p=0.265; Fig.S4b ), while it is weakly correlated with the readthrough transcript BDNF-AS-DT (N=219, r=0.10, t=2.038, p=0.0429; Fig.S4c ). Of note, BDNF-AS-DT has a negative association with BDNF (N=219, r= −0.17, t= −2.109, p=0.036; Fig.S4d ), and BDNF-AS (N=219, r= −0.178, t=-2.655, p=0.008; Fig.S4e ). Finally, BDNF and BDNF-AS are not correlated (N=219, r= −0.017, t= - 0.056, p=0.956; Fig.S4f ). These associations suggest that BDNF-DT , BDNF-AS , and BDNF-AS-DT , are independently regulated lncRNAs with differing relationships to BDNF . While we did not detect an association between BDNF-AS and BDNF , we detected a positive correlation of BDNF-DT with BDNF , and a negative correlation between the BDNF-AS-DT readthrough transcript and BDNF . The similarity of the developmental trajectory of BDNF-DT , BDNF-AS-DT and BDNF supports the possibility that BDNF-DT and BDNF-AS-DT at least in part regulate BDNF, independently of BDNF-AS . Conservation of novel divergent antisense transcription in the mouse Bdnf locus The BDNF-DT promoter shows high homology with the corresponding genomic sequence in the mouse genome. We extracted RNA from mouse frontal cortex and performed RT-PCR, to verify the existence of divergent transcription in the mouse Bdnf locus. We confirmed existence of an antisense amplicon connecting an exon partially overlapping exon I of Bdnf with an exon located in the region upstream the TSS of Bdnf exon I . We performed 5’- and 3’-RACE PCRs, and end-to-end PCR, confirming existence of the mouse Bdnf-DT gene. We detected four different transcripts, with high homology with the human BDNF-DT1 , 2 , 3 ( Fig. S5, Table S3 ). We did not find evidence for expression of a transcript homologous to the human BDNF-AS or the readthrough BDNF-AS-DT in the mouse genome. However, a mouse Bdnf-AS was recently described ( 42 ), which partially overlaps the Bdnf-DT gene. Based on our RACE results, we propose that the transcripts of the previously detected mouse Bdnf-AS are part of the Bdnf-DT gene, which has a further upstream 5’ exon, and a more complex structure ( Fig. S5 ). Consistent with the recommendations of the HUGO human gene nomenclature committee ( 33 ), we propose that the mouse Bdnf-AS should be renamed Bdnf-DT for the following reasons: i) it is not antisense to the genomic span of the protein-coding portion of Bdnf (differently from the human BDNF-AS ), ii) it does not show any homology with the human BDNF-AS , and iii) it is divergent to Bdnf , with which it shares at least one bidirectional promoter. Temporal dynamics of BDNF-DT and BDNF-AS-DT in response to neuronal depolarization The positive correlation between BDNF-DT and BDNF expression across the lifespan ( Figs.2 , S4 ) supports the possibility of a regulatory role played by BDNF-DT with respect to BDNF . Because BDNF is strongly regulated by neuronal activity ( 43 – 45 ), we investigated whether BDNF-DT was also induced by neuronal depolarization. We first probed the relationship between KCl-induced depolarization and Bdnf-DT expression in mouse primary cortical cultures ( Fig.3 ) . We observed a robust upregulation in Bdnf I-IX and Bdnf II-IX expression at 3 and 6 hours after stimulation with 100 mM KCl ( Fig.3a,b ). The highest level of expression, however, was observed after 6 hours of KCl stimulation followed by the wash out (WO) of KCl-containing media. Bdnf I-IX and Bdnf II-IX expression peaked ∼300 and ∼50 fold, respectively, at 3 hours post WO and gradually decreased to baseline over the next 24 hours ( Fig.3a,b ). Bdnf IV-IX displayed a different time-course of upregulation with a ∼ 30-fold induction at 3 hours, and a peak expression of nearly 60-fold at 6 hours after KCl stimulation with a gradual return to baseline after WO over the next 24 hours ( Fig.3c ). Interestingly, Bdnf-DT expression showed a temporal delay with respect to BDNF, in that it remained at baseline during the 6 hours of KCl stimulation, peaked 20-fold at 6 hours following WO, and returned to baseline over the next 24 hours ( Fig.3d ). Download figure Open in new tab Fig. 3. Temporal dynamics of Bdnf and BDNF-DT expression after depolarization in primary mouse cortical neurons. Expression of Bdnf I-IX transcript ( A , BDNF-I , RefSeq: NM_007540.4), Bdnf II-IX transcript ( B , BDNF-II , RefSeq: NM_001048139.1), Bdnf IV-IX transcript ( C , BDNF-IV , RefSeq: NM_001048141.1), and Bdnf-DT ( D ) transcripts is shown at different time points after KCl-induced depolarization and wash-out (WO), in mouse cortical neurons. RNA expression, measured with qPCR, is represented as fold-change versus vehicle (baseline). KCl-induced neuronal depolarization causes an increase of the expression of all the Bdnf transcripts, and of Bdnf-DT , but with different temporal dynamics. RNA expression, measured with qPCR, is represented as fold change vs. Baseline. All data are expressed as mean□±□standard error of the mean (SEM). We then examined BDNF , BDNF-DT and BDNF-AS-DT expression in response to induced depolarization in an immortalized human neural progenitor cell (NPC) line, which is derived from the ventral mesencephalon and can be differentiated into neurons or glia, ReNcell VM ( 46 ). Here, we quantified expression of BDNF IX (pan BDNF encompassing all coding transcript variants), BDNF IV-IX , BDNF-DT , and BDNF-AS-DT at different time points: 6 hours after depolarization alone or 6 hours after depolarization followed by 3, 6, 12, and 24 hours after WO ( Fig.4 ). BDNF - IX and BDNF IV-IX peaked after 6 hours of stimulation, followed by 3 hours after WO, 3-fold and 8-fold, respectively, and returned to baseline over the next 24 hours ( Fig.4a,b ). We observed a strong effect of KCl-induced depolarization on BDNF-DT , an 11-fold increase after 6 hours of stimulation followed by 3 hours after WO, with a return to baseline over the next 24 hours ( Fig.4c ). The BDNF-AS-DT readthrough transcript was also elevated after depolarization, although less strongly compared with BDNF-DT ( Fig.4d ). These data support the hypothesis that BDNF-DT and BDNF-AS-DT are co-regulated or play a regulatory role with respect to BDNF , as their expression peaks simultaneously or follows BDNF ’s upregulation both in mouse and human KCl-depolarized cell cultures. Download figure Open in new tab Fig. 4. Temporal dynamics of BDNF , BDNF-DT , BDNF-AS-DT transcript expression after depolarization in the human ReNcell VM neural progenitor cells. Expression of total BDNF mRNA ( A , BDNF-IX ), the BDNF exon IV containing transcript ( B , BDNF IV-IX, RefSeq: NM_001048141.1), BDNF-DT ( C ), and BDNF-AS-DT readthrough transcript ( D ) is shown at different time-points after KCl-induced depolarization, in human ReNcell VM neural progenitor cells. RNA expression, measured with qPCR, is represented as fold-change vs. vehicle. KCl-induced neuronal depolarization causes an increase of the expression of BDNF , BDNF-DT , and BDNF-AS-DT . RNA expression, measured with qPCR, is represented as fold-change vs. Vehicle. All data are expressed as mean□±□standard error of the mean (SEM). Individual comparisons; *p□<□0.05, **p□<□0.01. CRISPR-mediated regulation of BDNF induces BDNF-DT in human neural progenitor cells The ReNcell VM line has previously been used for CRISPR gene editing and BDNF gene expression studies ( 47 ). To functionally test whether targeted upregulation of BDNF leads to an increase in BDNF-DT expression, we engineered a novel ReNcell VM NPC line to constitutively express a CRISPR activation (CRISPRa) system for gene upregulation ( Fig.S6a ). ReNcell VM cultures were transduced with high titer lentivirus carrying deactivated Cas9 (dCas9) endonuclease fused to a strong transcriptional activator, VP64-p65-Rta, (VPR) ( 45 ). Cells expressing VPR were propagated to create a stably transduced VPR-ReNcell VM line, which, like the traditional ReNcell VM line can be differentiated into neurons ( Fig.S6b ). Stable transduction was confirmed with single molecule in situ hybridization (smFISH) for Cas9 ( Fig.S6c ) and neural activity was verified with electrophysiological recordings ( Fig.S6d-e ). To manipulate the expression of individual BDNF transcript variants, VPR-ReNcell VM cultures were transduced with lentiviruses carrying single guide RNAs (sgRNAs) targeting individual BDNF promoters at coding exons I or IV ( Table S4 ). Previous work in cultured rat neurons and rat hippocampus validated the efficiency and selectivity of this CRISPRa system ( 45 ). Transduction of VPR-ReNcell VM line with sgBDNF I resulted in a 30-fold increase of BDNF I-IX ( Fig.5a ), but not BDNF IV-IX expression ( Fig.5b ). Transduction with sgBDNF IV resulted in an almost 800-fold increase of BDNF IV-IX ( Fig.5b ), but not BDNF I-IX expression ( Fig.5c ). CRISPRa-mediated upregulation of BDNF I-IX and BDNF IV-IX corresponded to >2-fold and to >3-fold significant increase of BDNF-DT ( Fig.5c ). These results indicate that the induction of the expression of BDNF-protein coding transcripts also causes an increase of BDNF-DT . BDNF-AS and BDNF-AS-DT were not detectable in this in vitro system. Download figure Open in new tab Fig. 5: Upregulation of BDNF-DT using CRISPRa in VPR-ReNcell VM culture. Expression of BDNF I-IX ( A ), BDNF IV-IX ( B ) and BDNF-DT ( C ) transcripts is shown after CRISPRa of BDNF I promoter (sg Bdnf I ) and of BDNF IV promoter (sg Bdnf IV ). Activation of BDNF I promoter induces a 30-fold increase of BDNF I-IX transcript ( A ), a 2-fold increase of BDNF-DT ( C ), and has no off-target effect on BDNF IV-IX transcript ( B ). Activation of BDNF IV promoter induces a 800-fold increase of the BDNF IV-IX transcript ( B ), a 3-fold increase of BDNF-DT ( C ), and has no off-target effect on the BDNF I-IX transcript ( A ). RNA expression, measured with qPCR, is represented as fold-change vs. non-targeting control sgRNA, lacZ . All data are expressed as mean□±□standard error of the mean (SEM). Individual comparisons; *p□<□0.05, ****p□<□0.0001. Link between BDNF-DT, BDNF-AS-DT and schizophrenia Our data demonstrate that BDNF-DT and BDNF-AS-DT are developmentally regulated, sensitive to neuronal activity, and can potentially regulate BDNF . We next analyzed a relationship between these novel genes and schizophrenia, a neurodevelopmental disorder linked to alterations in neuronal activation ( 48 ) and with a GWAS significant association in the BDNF locus ( 49 ). Leveraging the postmortem human RNAseq dataset described above ( 29 ), we investigated expression of BDNF-DT and the BDNF-AS-DT readthrough transcript in DLPFC tissue from control brain donors versus schizophrenia. While expression of BDNF-DT is significantly higher in donors with schizophrenia than in neurotypical controls (N=378, t=3.805, p= 0.000167; Fig.6a ), BDNF-AS-DT (N=378, t=-0.875, p= 0.3824), BDNF (N=378, t=0.572, p= 0.5675), and BDNF-AS (N=378, t=0.10, p= 0.920720) remain unchanged. Download figure Open in new tab Fig. 6: Divergent transcription, schizophrenia, and rs6265. A) Expression of BDNF-DT (adjusted values, y-axis) is higher in patients affected with schizophrenia (orange, right box on the x-axis) compared with in neurotypical controls (light yellow, right box on the x-axis); p-value is shown at the top. B) Schizophrenia GWAS plot visualization of the BDNF locus. The SNP rs49234957, indicated with two inverted triangles, is the variant with the most significant association with schizophrenia in the BDNF locus, and it reaches GWAS significance in the combined replication and discovery sample (upper triangle, p=3.89e-08). Rs49234957 is in high linkage disequilibrium with rs6265. C) Expression of the BDNF-AS-DT readthrough transcript (y-axis) is associated with the rs6265 genotype so that it increases with the number of T/Met alleles: specifically, it is higher in the TT/MetMet genotype (red box) and in the CT/ValMet (orange box) compared with the CC/ValVal (light yellow box), in both neurotypical controls (left, x-axis) and patients affected with schizophrenia (right, x-axis); p-value is shown at the top. D) Expression of BDNF-AS (y-axis) is associated with the rs6265 genotype so that it decreases with the number of T/Met alleles: specifically, it is lower in the TT/MetMet genotype (red box) and in the CT/ValMet (orange box) compared with the CC/ValVal (light yellow box), in both neurotypical controls (left, x-axis) and patients affected with schizophrenia (right, x-axis); p-value is shown at the top. Because previous work suggests that differentially expressed genes in postmortem human brain tissue studies of schizophrenia are more likely reflecting consequences of illness rather than underlying genetic causes( 34 , 50 ), we leveraged Summary-based Mendelian Randomization (SMR) ( 51 ) to investigate whether genetic variation in the BDNF locus might be mediating schizophrenia risk through regulating expression of these transcripts. Of note, the most recent GWAS( 49 ) identified the rs4923457 A/T single nucleotide polymorphism (SNP) as the genetic variant with the most significant association with schizophrenia in the BDNF locus. This SNP reaches GWAS-significance in the combined discovery and replication samples (p discovery- sample =9.52e-08, p combined-sample =3.89e-08, OR for A allele=1.05; Fig.6b ) but, because its alleles are complementary base pairs, it is not suitable as an instrumental variable for the SMR analysis, given the risk of potential inconsistent strand alignment across different DNA genotyping batches. Of note, rs4923457 is in high LD (D’=0.955, R 2 = 0.798) with rs6265 (also known as BDNF Val66Met), a highly studied functional variant in psychiatric genetics( 52 ), which lies in the BDNF pro-region domain and produces a non-synonymous substitution at codon 66: C → T/Valine → Methionine (Val66Met). This SNP is directly linked to subcellular trafficking of BDNF, and is associated with behavior in both human and mice: in particular, the Met allele has been shown to disrupt the activity-dependent secretion of BDNF by inhibiting sorting of BDNF into dense-core vesicles ( 52 ). However, GWAS data indicate that the T allele (Met-BDNF), correlated with the rs4923457 T allele, is associated with decreased risk for schizophrenia [odds ratio (OR) = 0.96 and 0.95, respectively for rs6265 T allele and rs4923457 T allele]. We therefore performed SMR with rs6265 as the instrumental variable, and we found that the readthrough transcript BDNF-AS-DT has the most significant potential causal association with schizophrenia, so that lower BDNF-AS-DT genetically predicted expression is associated with the disease (p=9.89e-06 in the whole sample and p=7.64e-05 in the European ancestry sample; Table S5 ). SMR also detected an association with opposite sign between BDNF-AS and schizophrenia, albeit less significantly compared with BDNF-AS-DT ( Table S5 ). Results were consistent when running SMR without pre-selecting any candidate SNP ( Table S6 ). Indeed, the eQTL analysis shows that rs6265 is significantly associated with the expression of BDNF-AS-DT (N=378, t=10.365, p=3.63e-22; Fig.6c ). The relationship of rs6265 with BDNF-AS-DT is stronger compared with rs4923457 (N=378, t=6.829, p=3.72e-11), the top schizophrenia GWAS-index SNP in this locus. This association of rs6265 with BDNF-AS-DT is present both in neurotypical controls (N=219, t=7.267, p=8.38e-12) and in patients with schizophrenia (N=159, t=7.073, p=7.11e-11; Fig.6c ). We also analyzed the relationship of rs6265 with expression of the other genes in the same locus. While we did not detect any significant association of rs6265 with BDNF-DT (N=378, t=0.608, p=0.54) and BDNF (N=378, t=0.802, p=0.42), we found a significant association between rs6265 and BDNF-AS , with opposite directionality compared with the BDNF-AS-DT (N=378, t= −5.328, p=1.77e-07; Fig.6d ). In other words, the T allele (Met-BDNF) is associated with higher expression of BDNF-AS-DT and lower expression of BDNF-AS . Because the BDNF-AS-DT transcripts differs from BDNF-AS not only for the link with BDNF-DT, but also for the skipping of an exon common to all the BDNF-AS transcripts (hg38: chr11:27,658,240-27,658,462), and containing rs6265 (hg38: chr11: 27,658,369), these findings suggest that rs6265 genotype may also be associated with an effect on transcriptional splicing. In particular, the rs6265 T allele appears to be associated with the skipping of this exon common to all the BDNF-AS transcripts, leading to opposite effect on expression of BDNF-AS-DT (increase) and BDNF-AS (decrease). Intragenic DNA methylation has been associated with exon inclusion during transcription ( 53 , 54 ), thus we explored the possible role of DNA methylation in the relationship between rs6265 and BDNF-AS-DT expression. Because the rs6265 C>T substitution abolishes a CpG site in the position chr11:27658369, we analyzed the relationship between methylation of this CpG site and BDNF-AS-DT expression, leveraging DNA methylation bisulfite sequencing DLPFC data available in a sample subset (N=141, Table S1 ). We detected a negative relationship between DNA methylation at this site and BDNF-AS-DT expression (N= 141, Z= −3.86, p= 0.0001). By extending the analyses to other CpGs in the same locus, we note that this CpG in position chr11:27658369 is the top methylation site associated with expression of BDNF-AS-DT ( Table S7 ). As predicted, the rs6265 C(G)/T(G) SNP is associated with methylation of the CpG at position chr11:27658369 (N= 141, t= −19.31, p=4.84e-39; Fig.S7 ), consistent with previous work ( 55 ). Co-expression gene networks associated with BDNF, BDNF-DT and BDNF-AS-DT in the human cortex To explore convergence and divergence in expression patterns linked with BDNF-DT , BDNF-AS-DT , BDNF , and BDNF-AS , we performed brain coexpression network analyses using each of the above as seed genes, in the subset of neurotypical controls (N=318) from the above referenced DLPFC RNAseq dataset ( 29 ) ( Fig. 7a-d , Fig.S8 , and Tables S8-11 ). Genes significantly co-expressed with BDNF-DT were associated with neurogenesis and synaptic function, and with serotonin, dopamine, and opioid signaling ( Fig. 7d , Fig.S8a , and Table S8a ); indeed, these genes were associated with activation of BDNF and inhibition of REST, the neuron-restrictive silencing factor (also known as NRSF), a fundamental transcription repressor element involved in suppressing expression of neural genes ( 56 – 58 ) ( Table S8b ). Genes significantly co-expressed with BDNF were also associated with neurogenesis and synaptic function, and inhibition of REST ( Fig.S8b , and Table S9a,b ). On the other hand, genes significantly co-expressed with BDNF-AS were associated with inhibition of synaptogenesis and dopamine pathways, and REST activation ( Fig.S8c , and Table S10a,b ), while genes co-expressed with BDNF-AS-DT were linked with protein ubiquitination, and SOX2A activation ( Fig.S8d , and Table S11a,b ). Of note, the ubiquitin proteasome pathway is linked to BDNF expression ( 45 ), and can impact transcription termination by RNA polymerase II ( 59 ), which is directly related to readthrough transcription under stress conditions ( 15 – 17 ). Download figure Open in new tab Fig. 7: Coexpression gene networks associated with BDNF-DT , BDNF-AS-DT , BDNF-AS , and BDNF , in human DLPFC. A-C) RRHO2 heatmaps showing concordant (bottom-left and top-right quadrant in each panel, highlighted in a red frame) and discordant (top-left and bottom-right quadrant in each panel, highlighted in a green frame) association of BDNF-DT coexpression network (x-axis, A-C ) with coexpression network of BDNF (y-axis, A ), BDNF-AS (y-axis, B ), BDNF-AS-DT (y-axis, C ). Color bars represent negative logarithm of p value of the overlap from the hypergeometric test. BDNF-DT show a concordant pattern of association with gene expression compared with BDNF ( A ), and discordant pattern of association compared with BDNF-AS ( B ), and BDNF-AS-DT ( C ). D ) Ingenuity Pathway Analysis summary graph with subcellular localization of the activators (orange) and inhibitors (blue) upstream regulators of the BDNF-DT gene coexpression network, and visualization of their effect on pathways predicted to be activated (orange) and inhibited (blue) in presence of the gene expression changes associated with the BDNF-DT gene coexpression network. Legend in the red box. We also used a threshold-free algorithm ( 60 ) to detect concordant and discordant patterns of associations with potential biological relevance of BDNF-DT with BDNF , BDNF-AS , BDNF-AS-DT . Briefly, we analyzed whether the genes co-expressed with BDNF-DT were co-expressed with BDNF , BDNF-AS , BDNF-AS-DT with the same directionality (positive vs. positive, and negative vs. negative: concordant pattern) or with opposite directionality (positive vs. negative, or negative vs. positive: discordant pattern). We detected a highly concordant pattern of associations of gene expression between BDNF-DT and BDNF ( Fig.7a ), while we detected discordant pattern of association of gene co-expression of BDNF-DT with BDNF-AS ( Fig.7b ), and less strongly with BDNF-AS-DT ( Fig.7c ). These results provide additional evidence that BDNF and BDNF-DT share similar expression patterns, while their relationship with co-expressed genes contrasts with that of BDNF-AS and the BDNF-AS-DT readthrough transcript. Discussion RNAseq in human brain has revealed many novel transcripts and isoforms of known genes. In addition, it is now understood that transcription occurs from non-coding regions of the genome. Divergent transcription from bidirectional promoters substantially contributes to overall levels of transcription, and it has been proposed that the majority of mammalian lncRNAs originate from divergent transcription of gene pairs ( 19 ). Next generation sequencing also indicates that readthrough transcription is a common phenomenon ( 17 ). However, it is still not known whether divergent and readthrough transcription are functionally relevant. We performed a genomic characterization of a divRNA in the human and mouse BDNF locus, which we named BDNF-DT . Consistent with other divRNAs, our data suggest that BDNF-DT is evolutionary conserved ( 61 ). BDNF-DT transcription start sites overlap known BDNF exons, they are transcribed antisense to BDNF , and they have a 3’ end located 5’ upstream of BDNF exon I ( Fig.1 ). CpG islands overlap the transcription start site of BDNF-DT ( Fig.S2 , S5 ), and the 5’ region of BDNF-DT overlaps areas of BDNF exon II that bind the RE1-Silencing Transcription factor (REST) ( 56 – 58 , 62 ). The REST binding site (neuron-restrictive silencer element, NRSE) is located on the opposite strand of the BDNF gene, downstream of the 5’ region of BDNF-DT. This is consistent with the possibility that REST regulates directly BDNF-DT expression, which may subsequently regulate BDNF . Indeed, our co-expression network analyses indicate that genes co-expressed with BDNF and BDNF-DT are predicted to be regulated by REST ( Fig.7d , S8a,b ). In addition, we detected BDNF-DT isoforms spliced with the known BDNF-AS , which constitute novel BDNF-AS-DT readthrough transcripts. BDNF-DT and its readthrough transcript are developmentally regulated - their expression increases starting from prenatal and neonatal life, decreases in the transition from childhood to adolescence, peaks again around 25 years of age and then decreases ( Fig.2 ). These changes in expression across development mirror those of BDNF ; while on the other hand, they differ from that of known BDNF-AS , whose expression increases during adulthood. While BDNF-DT and BDNF expression patterns are positively correlated, we detected a negative correlation between BDNF-AS-DT and BDNF , and no correlation between BDNF-AS and BDNF ( Fig.S4 ). These findings suggest that BDNF-DT and BDNF-AS-DT are expressed in similar developmental contexts as BDNF , but they may regulate BDNF in contrast to BDNF-AS . Indeed, gene coexpression networks revealed highly concordant patterns of associations between BDNF-DT and BDNF ( Fig.7a ), and discordant patterns between BDNF-DT and BDNF-AS ( Fig.7b ) or BDNF-DT and BDNF-AS-DT ( Fig.7c ). BDNF , BDNF-DT and BDNF-AS-DT expression increases in response to KCl-induced depolarization ( Fig.3 , 4 ). In cultured mouse neurons we observed differing temporal dynamics in the expression patterns between activity-induced Bdnf and Bdnf-DT . Specifically, the increase of Bdnf transcripts precedes the increase of Bdnf-DT , which occurs simultaneously with a decrease of the Bdnf coding transcript. This evidence raises the possibility that BDNF-DT plays a role in regulating BDNF levels, via an autoregulatory negative feedback loop. Supporting this idea, CRISPR-mediated induction of BDNF exon I and IV transcription in human NPCs drives subsequent BDNF-DT expression ( Fig.5c ). These data raise the possibility that the divRNAs may regulate the optimal ‘balance’ of BDNF . While BDNF expression is essential for synaptic plasticity and trophic support, its levels must be tightly controlled to prevent hyper-excitability( 24 – 28 ). We therefore propose that BDNF-DT and BDNF-AS-DT are part of a regulatory mechanism, where expression of BDNF induces expression of BDNF-DT , which is required for the splicing of BDNF-AS with BDNF-DT , thus originating BDNF-AS-DT . BDNF-AS-DT may therefore represent a key regulator of BDNF. Additional mechanistic experiments will help to definitively test this hypothesis, which is consistent with our data showing an effect of CRISPR mediated BDNF transcription on BDNF-DT expression ( Fig.5c ), and an effect of KCl-induced depolarization on the expression, first of BDNF , and then of BDNF-DT and BDNF-AS-DT ( Fig.4 ). Further, this hypothesis is consistent with our observations of a positive correlation of BDNF-DT with BDNF and BDNF-AS-DT , which is negatively correlated with BDNF ( Fig.S4 ). Schizophrenia is a neurodevelopmental disorder that is linked with altered BDNF signaling ( 20 , 21 , 63 ) and neuronal activity ( 48 ). Hence, we reasoned that the developmental regulation of BDNF-DT and BDNF-AS-DT , and their potential ability to regulate BDNF in response to neuronal activity, may contribute to established links between schizophrenia and BDNF signaling. To probe this hypothesis, we investigated whether BDNF-DT and BDNF-AS-DT were differentially expressed in patients affected with schizophrenia compared with neurotypical controls, and whether their expression was associated with genetic risk for schizophrenia in the BDNF locus. We observed higher expression of BDNF-DT in the DLPFC of patients with schizophrenia, compared with controls, and no significant differences in BDNF-AS-DT expression ( Fig.6a ). These findings should be interpreted with caution since differences in gene expression between patients and controls tend to be confounded by medication, comorbid health issues and socioeconomic factors. Hence, even if observed differences are robust, they may reflect consequences of the disorder, rather than risk or cause. Genetic association, in contrast, is independent of illness state and confounders. In this context, SMR indicates a significant association of genetically predicted expression of BDNF-AS-DT with schizophrenia. BDNF-AS-DT is indeed associated with rs6265, a SNP in the BDNF locus that shows association with schizophrenia in the current GWAS. Rs6265 is in strong linkage disequilibrium with the SNP showing the most significant association with schizophrenia in the BDNF locus ( 49 ). The association is highly significant in the whole sample and, separately, both in patients with schizophrenia and controls. In particular, the C/Val-coding and risk-associated allele is associated with lower expression of BDNF-AS-DT ( Fig. 6c ) and higher expression of the canonical BDNF-AS gene ( Fig. 6d ). Current data about rs6265 may appear potentially paradoxical, given previous allele-specific findings showing deleterious effects of the T/met-coding allele on diverse brain phenotypes, and the association of the C/Val-coding allele with higher risk for schizophrenia. Our results suggest that the effects of the rs6265 val66met variant on BDNF activity and on the expression of the readthrough transcript are two distinct mechanisms, that may impact BDNF biology in a context specific manner. The effects of the C/Val-coding allele on activity-dependent secretion of BDNF and on lower expression of BDNF-AS-DT may converge, in specific developmental stages, on increasing the activity and the expression of BDNF, creating an excess of BDNF which may disrupt brain development, increasing risk for schizophrenia. At the cellular level, the Met allele has been shown to disrupt activity-dependent secretion of BDNF by inhibiting sorting of BDNF into dense core vesicles ( 52 ). Experimental studies have shown that the Met-BDNF results in an approximate 18% decrease in activity-dependent secretion in transfected cells carrying one Met allele and a 29% decrease in those transfected with two Met alleles ( 64 ). However, what is often neglected is that rs6265 lies not only in the BDNF pro-region domain but also in an exon common to all the BDNF-AS isoforms (hg38: chr11:27,658,240-27,658,462) which our findings show is not part of the BDNF-AS-DT readthrough transcript. An effect of rs6265 on the splicing of such exon is plausible, consistent with the fact that splice regulatory elements can lie within alternative exons and can be disrupted by sequence variations ( 65 ); in this regard, rs6265 may affect splicing by influencing the binding of transcription factors on the positive strand of the DNA, from which BDNF-AS and BDNF-AS-DT are transcribed. DNA methylation may contribute to mediate these effects, consistently with the relationship between rs6265 and intragenic methylation ( Fig. S7 ), and the relationship between intragenic DNA methylation and exon inclusion during transcription ( 53 , 54 ). Indeed, we show that the C/Val-coding allele is associated with higher methylation, lower expression of BDNF-AS-DT and higher expression of the canonical BDNF-AS gene, which incorporates the exon missing in the readthrough transcript. The C/Val-coding allele may therefore be associated not only with higher activity-dependent secretion of the BDNF protein but also with blunted expression of BDNF-AS-DT , with a possible impairment in the regulation of BDNF and a susceptibility to a detrimental excess of BDNF which may lead to hyperexcitability during early development, and higher risk for schizophrenia. This aspect of rs6265 biology is likely overlooked by most mechanistical studies on the SNP, in cellular and animal models, which are often based on the ectopic expression of constructs expressing the Val- or the Met-BDNF in a genomic context that does not include the BDNF-AS , BDNF-AS-DT , and BDNF-DT . Even if we cannot exclude the possibility that the relationship between rs6265 and BDNF-AS-DT expression is driven by other genetic variants in high LD with rs6265, the presence of such an association is still relevant to understand the clinical correlates of rs6265. In conclusion, our study sheds a new light and poses new questions on the complex biology of BDNF , revealing that, in addition to the canonical BDNF-AS isoforms, divergent RNAs and readthrough transcripts linking BDNF-AS and the divRNAs may contribute to BDNF regulation. Moreover, our study provides novel insight on the mechanisms mediating or moderating the effect of the functional rs6265 polymorphism, one of the most studied SNPs in psychiatric genetics, raising the possibility that readthrough transcription contributes, directly or indirectly, to the effect of this functional polymorphism. Our study does not clarify whether divergent transcription and readthrough transcription represent just noise or phenomena with biological relevance. However, the evolutionary conservation, the developmental trajectory and the temporal dynamic after neuronal depolarization of BDNF-DT and BDNF-AS-DT , support a regulatory role on BDNF , potentially relevant for brain development and activity, and risk for schizophrenia. Materials and Methods Postmortem brain samples Postmortem human brain tissue was obtained by autopsy primarily from the Offices of the Chief Medical Examiner of the District of Columbia and of the Commonwealth of Virginia, Northern District, all with informed consent from the legal next of kin (protocol 90-M-0142 approved by the NIMH/NIH Institutional Review Board). The National Institute of Child Health and Human Development Brain and Tissue Bank for Developmental Disorders ( http://www.BTBank.org ) provided additional postmortem prenatal, infant, child and adolescent brain tissue samples, by under contracts NO1-HD-4-3368 and NO1-HD-4-3383. The Institutional Review Board of the University of Maryland at Baltimore and the State of Maryland approved the protocol, and the tissue was donated to the Lieber Institute for Brain Development under the terms of a Material Transfer Agreement. Tissue acquisition, handling, processing, dissection, clinical characterization, diagnoses, neuropathological examinations, and quality control measures were performed following established protocols at the LIBD as previously described ( 29 , 66 ). RNA extraction and sequencing Postmortem tissue homogenates of dorsolateral prefrontal cortex gray matter (DLPFC) approximating Brodmann area 46/9 in postnatal samples and the corresponding region of PFC in prenatal samples were obtained from all subjects. Total RNA was extracted from ∼100 mg of tissue using the RNeasy kit (Qiagen) according to the manufacturer’s protocol. The poly(A)- containing RNA molecules were purified from 1 μg DNase-treated total RNA, and sequencing libraries were constructed using the Illumina TruSeq RNA Sample Preparation v2 kit. These products were then purified and enriched with PCR to create the final cDNA library for high-throughput sequencing using an Illumina HiSeq 2000 with paired-end 2 × 100 bp reads, as previously described ( 29 ). Poly(A)+ libraries include divRNAs, since these transcripts do have a Poly(A) tail, together with other features of mRNAs, such as the 7-methylguanosine cap, and being often subject to splicing ( 19 ). RNAseq data processing has been extensively previously described ( 29 ). Rs62625 and other genotypes were extracted from genotyping data obtained with HumanHap650Y_V3, Human 1M-Duo_V3, and Omni5 BeadChips (Illumina, San Diego, CA), as described in previous reference ( 29 ). Sample characteristics and demographics are summarized in Table S1 . Pipeline for detection and validation of divRNAs A similar procedure was performed to detect and validate divRNAs in the human BDNF and mouse Bdnf locus: Detection of divergent transcription Most divRNAs are not annotated in the reference transcriptome (GRCh38/hg38). Moreover, they are easily degraded, and they can have very low expression, so that transcripts assembly procedures are not completely reliable in detecting their expression and their structure. In addition, they can overlap exons of protein-coding genes, so that precise information about the expression of divRNAs cannot be obtained with measures that do not consider the directionality of transcription, such as exons and ‘expressed genomic regions ( 67 )’, without defining their exact structure. For these reasons, we investigated the occurrence of divergent transcription at the promoter of protein-coding genes by visualizing RNAseq data with Integrative Genomics Viewer (IGV), a high-performance software for visualization and exploration of genomic data ( 31 ). After detecting sequencing reads in the promoter region of protein-coding genes, we confirmed their antisense direction by analyzing the consistence with the intron rules ( 68 ) of the genomic sequences overlapping the reads covering exon-exon splice junctions of the potential divRNAs. PCR validation We validated the novel sequences potentially belonging to the divRNAs with PCR. RNAs isolated from postmortem DLPFC were reverse transcribed using SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen) to synthesize cDNA. Based on the RNAseq results, we designed primer pairs to amplify portions of the divRNAs, thus validating the novel exons detected with the RNAseq analysis (primer sequences available in Table S4 ). RACE 5′ RACE and 3′ RACE were performed using SMARTer® RACE 5′/3′ Kit (Clontech) to determine the 5’ start and the 3’ end of divRNAs. RNAs isolated from postmortem human DLPFC were reverse transcribed to full-length cDNA using the reverse transcriptase and the oligo(dT)-anchor primer available in the kit, according to the manufacturer’s protocol (Clontech). The obtained cDNA was further amplified by a second PCR, using divRNA-specific primer, the PCR-anchor primer, and the Advantage Polymerase Mix, according to the manufacturer’s instruction (Clontech). Nested PCR was then performed on the RACE-amplicons, to confirm the presence of the divRNAs. PacBio sequencing PacBio sequencing was performed on the RACE amplicons where the nested PCR confirmed the presence of divRNAs. Three micrograms of purified PCR products were sequenced on an RSII instrument, according to the manufacturer’s protocols (Pacific Biosciences), by the Johns Hopkins Transcriptomics and Deep Sequencing Core Facility. Briefly, PCR products were purified twice by AMPure magnetic beads (Beckman-Coulter) to yield SMRTbell libraries. We size-selected PCR products to produce two distinct SMRTbell libraries: one containing fragments less than 6000 bp and another with fragments greater than 6000 bp. In the final step, the sequencing primer was annealed to the adapters and sequencing polymerase bound to SMRTbell libraries. Each sample was sequenced on a single SMRT cell. The subreads generated by the SMRT analysis tools v2.3.0 were aligned back to the reference genome, using the BLAT search genome browser. End-to-end PCR Based on the PacBio sequencing results, we designed primers overlapping the 5’ starts and the 3’ ends of the divRNAs, and we performed end-to-end PCR, using Platinum® Taq DNA Polymerase High Fidelity (Thermo Fisher Scientific). Nested PCR was then performed on the PCR-amplicons, to confirm the presence of the divRNAs. We then performed PacBio sequencing, as described above, on the End-to-end PCR amplicons where the nested PCR confirmed the presence of divRNAs, thus defining the full-length sequences of the divRNAs. Quantification of divRNAs, BDNF, and BDNF-AS expression in postmortem RNAseq data The quantification of divRNAs in postmortem data poses technical challenges, because their expression, although variable across individuals, is low in basal conditions (mean expression of divRNA in the BDNF locus: 0.03348 RPKM, SD: 0.0557). For this reason, we estimated expression of divRNAs by counting and normalizing the reads that are aligned to the most expressed common exon of the divRNAs of the BDNF locus (hg38: chr11:27722351-27722486). Such exon does not overlap any other gene in this locus, as shown by our RNAseq data and RACE-PCR. Consistently, we estimated the expression of total BDNF by calculating and normalizing the reads overlapping the coding exon of BDNF (chr11:27654893-27658585), without overlapping the BDNF-AS exon in this region (chr11:27658241-27658462). We measured expression of the total BDNF-AS gene by calculating and normalizing the reads overlapping the exon-exon junction common to all the BDNF-AS transcripts (chr11:27658463-27659170), without being part of the readthrough transcript that connects the divergent RNA and BDNF-AS . Finally, we assessed the expression of the BDNF-AS-DT readthrough transcript by counting and normalizing the reads that overlap the most expressed unique junction that characterizes it (hg38: chr11: 27640006-27659170). Mouse primary cortical neuronal cell cultures E16.5 embryos were isolated from CD timed-pregnant females, and cortices were dissected from individual embryos in base medium. Dissociation of cortices was performed as previously described ( 69 ). Briefly, cortices were washed with sterile PBS and incubated for 40 min at 37°C in dissociation medium containing 10 U/ml of papain (Worthington). After quenching the enzymatic reaction with trypsin inhibitor (Sigma) and BSA, cortices were gently triturated in serum-free medium containing glutamate (Invitrogen). Neurons were plated on dishes coated with poly-L-lysine (Sigma) and laminin (Invitrogen) and cultured in Neurobasal medium containing B27 supplement (Invitrogen) and antibiotics. ReNCell VM cultures The ReNcell VM neural progenitor cell (NPC) line (Millipore Sigma), derived from the ventral mesencephalon by overexpression of the myc family transcription factors in human cells, has been used for stem cell therapy and disease modeling ( 47 , 70 ). The ReNcell VM line was maintained in ReNCell NSC Maintenance medium (Millipore Sigma), supplemented with Antibiotic-Antimycotic (Gibco), 20 ng/ml EGF (Peprotech) and 20 ng/ml bFGF (Peprotech). Cells were cultivated on 6, 12, or 24-well plates, or T75 culture flasks (Fisher Scientific), or CytoView MEA plates (Axion) pre-coated with 20 µg/ml laminin (Sigma-Aldrich). Maintenance medium was exchanged every other day until 80% cell confluence was reached, and standard differentiation was performed by removing EGF and bFGF from culture medium. After differentiation, half-medium changes were performed every 2-3 days for two weeks with maintenance medium lacking growth factors. KCl time courses were performed after complete differentiation at Day In Vitro (DIV) 15. Transduction with CRISPR constructs was performed at the same time as differentiation. Immunocytochemistry For staining cultured ReNcell VM NPC line, samples were fixed with ice-cold 4% paraformaldehyde for 10 min. Permeabilization and blocking were performed simultaneously with 5% normal goat serum. Primary antibodies directed against FLAG (dCas-VPR), mCherry (sgRNA), and MAP2 (Millipore Sigma) were prepared in blocking buffer and incubated with the sample at 4 °C overnight. Fluorescently conjugated secondary antibodies were prepared in blocking buffer and incubated with the sample at room temperature for 90 min. Samples were counterstained with DAPI and stored in PBS until imaging. Neuronal depolarization experiment To analyze whether the expression of the divRNAs in the BDNF locus is affected by neuronal activity in vitro, we performed a KCl-induced depolarization experiment using, in human, immortalized human neural progenitor cell line from the ventral mesencephalon (ReNcell VM NPC ( 46 ) and, in mouse, cortical neurons obtained from E14 mouse embryo cortices. In human ReNcell VM NPC ( 46 )), we induced neuronal depolarization by incubation in 100 mM KCl for 0 hours (h), 6 h, 6h + 3h washout (WO), 6h + 6h washout (WO), 6h + 12h washout (WO), 6h + 24h washout (WO). At each time point, we obtained cells for gene expression analysis. We then performed RNA extraction with RNeasy Mini Kit (Qiagen), reverse transcription to cDNA, using SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen), and quantitative PCR (qPCR), using primers for BDNF-DT , the BDNF-AS-DT readthrough transcript, BDNF coding transcripts, and GAPDH as housekeeping gene. In mouse, E14 embryo cortices were dissected and then dissociated in HBSS (Lifetech), papain and L-cysteine, using a papain dissociation system (Worthington), according to the manufacturer’s instructions. Neurons were seeded at an appropriate density for further experiments on poly-ornithine/fibronectin coated plates (VWR), in Neurobasal media (LifeTech) containing 5% Fetal Bovine Serum (LifeTech), 1% Glutamax and 1X B27 supplements (LifeTech), 1X Penicillin-Streptomycin (LifeTech). Neurons were then grown in vitro for 7 more days in the same media but without Fetal Bovine Serum. We then induced neuron depolarization, by incubation in 100 mM KCl for 0h, 3h, 6h, 6h + 3h washout (WO), 6h + 6h WO, 6h + 12h WO, 6h + 24h WO, 6h + 48h WO and, at each time point, we obtained cells for gene expression analysis. We then performed RNA extraction with RNeasy Mini Kit (Qiagen), reverse transcription to cDNA, using SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen), and quantitative PCR (qPCR), using primers for Bdnf-DT locus, Bdnf protein-coding transcripts, and GAPDH as housekeeping gene. CRISPR/dCas9 and sgRNA construct design A lentivirus compatible backbone (a gift from Feng Zhang, Addgene #52961) was used to insert a dCas9-VPR (VP64-p65-Rta) cassette driven by the human synapsin 1 promoter (SYN1). SP-dCas9-VPR was a gift from George Church (Addgene #63798). A guide RNA scaffold (a gift from Charles Gersbach, Addgene #47108) was inserted into a lentivirus compatible backbone containing EF1α-mCherry. A BbsI cut site within the mCherry construct was mutated with a site-directed mutagenesis kit (NEB). Bdnf-specific sgRNA targets were reused from our previous study ( 45 ) or were designed using CHOPCHOP ( http://chopchop.cbu.uib.no/ ). All CRISPR RNA (crRNA) sequences were analyzed with National Center for Biotechnology Information’s (NCBI) Basic Local Alignment Search Tool (BLAST) to ensure specificity. An extensive validation of minimal sgRNA off-target effects was conducted in our previous study ( 45 ). The bacterial lacZ gene was used as a non-targeting sgRNA control. CRISPR activation of BDNF transcripts in ReNcell VM culture Here we describe a novel iteration of the ReNcell VM line with constitutively expressed CRISPR activation (CRISPRa) system for convenient upregulation of desired target genes. ReNcell VM cultures were transduced with high titer lentivirus (multiplicity of infection (MOI) over 10,000) carrying deactivated Cas9 (dCas9) endonuclease fused to a strong transcriptional activator, VPR (VP64-p65-Rta). Successfully transduced cells were sub-cultured and passaged to create a stably expressing VPR-ReNcell VM line ( Fig. S6 ). As the traditional ReNcell VM cultures, VPR-ReNcell VM cultures were differentiated into neurons by the removal of growth factors (EGF and bFGF) from the culture media for 15 DIV. Stably-transduced VPR-ReNCell VM cultures were propagated and transduced again with single guide RNA (sgRNA) lentivirus (MOI over 5,000) and differentiated into neurons. Following CRISPRa of BDNF I and IV promoters, we performed RNA extraction with RNeasy Mini Kit (Qiagen), reverse transcription to cDNA, using SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen), and quantitative PCR (qPCR), using primers for BDNF-DT , the BDNF-AS-DT readthrough transcript, BDNF coding transcripts, and GAPDH as housekeeping gene ( Fig. S4 ). Statistical analyses All the statistical analyses were performed in the R computing environment ( 71 ), version 4.3.3. - Relationship between age and BDNF-DT , BDNF-AS , BDNF-AS-DT , BDNF : because gene expression does not change linearly with age, we created a spline function to modeling the relationship between expression and age, establishing knots at ages associated with evident gene expression changes, that is, 1 year, 12 years, 25 years. We assessed the significance of the relationship with age by calculating an F-statistics and a p-value from the comparison of a full spline model with age and covariates as predictors, and a covariates-only model. Covariates included in this analysis were: ancestry genetic Principal Components (PCs), and sex, as in previous work ( 29 ). - Relationship between BDNF-DT , BDNF-AS , BDNF-AS-DT , BDNF expression: we used a linear model to analyze how the expression of these genes were each other related, adjusting for sex, age, ancestry PCs, and surrogate variables (SVs) accounting for RNA quality ( 29 , 34 , 72 ), in the sample of adult neurotypical subjects (age>13). - Differential expression analysis in patients with schizophrenia and neurotypical controls was performed, in the sample with age >13, using a linear model with expression as dependent variable, and case-control status as predictors, with age, sex, ancestry PCs and SVs accounting for RNA quality ( 29 , 34 , 72 ) as covariates. - Summary-based Mendelian Randomization (SMR): we performed SMR analysis( 51 ) focusing on the gene expression features in the BDNF locus (chr11: 27,403,033-27,759,030). First, we conducted SMR using rs6265 as an instrumental variable. Next, we performed SMR without pre-selecting a candidate SNP. For each feature, we clumped cis-eQTL for each feature (including exons and splice junctions) using PLINK (500 kb window, r 2 > 0.1). We selected independent eQTL clusters including the top eQTL (nominal p<0.05) and the top PGC3 GWAS SNPs (p< 0.05) for SMR analysis. For each cluster, we implemented SMR and HEIDI methods, using the SMR software tool (version 1.3.1) to examine pleiotropic associations of genetic variation with feature expression in the DLPFC RNAseq dataset and case–control status in the PGC3 GWAS schizophrenia data ( 49 ). We performed the analyses using: i) GWAS summary statistics from the primary meta-analysis for autosomal SNPs and eQTL in the full DLPFC sample; ii) GWAS summary statistics from the meta-analysis of European ancestry cohorts for autosomal SNPs and eQTL in the DLPFC sample of European ancestry. - eQTL analysis: the relationship between expression and genotypes was analyzed using a linear model with expression as dependent variable, and genotype, case-control status, age, sex, genetic PCs and surrogate variable (SVs) ( 29 , 34 , 72 ) accounting for RNA quality as covariates; we repeated the same analysis using gene expression PCs instead of SVs with analogous results. In addition, we performed this analysis stratifying by diagnosis. The R package ‘MatrixeQTL’( 73 ) version 2.3 was used for this analysis. - DNA methylation analyses: we analyzed whether expression of BDNF-AS-DT is associated with methylation of the CpG in the position chr11:27658369. For this analysis, we extracted % CpG methylation levels from a published DLPFC whole-genome bisulfite sequencing dataset ( 74 ) partially overlapping the DLPFC RNAseq dataset of this study (N=141, Table S1 ). The relationship between expression and methylation was analyzed using a linear model with expression as dependent variable, and methylation, case-control status, age, sex, genetic PCs, methylation PCs, and surrogate variable (SVs) accounting for RNA quality as covariates. The same model was applied to other CpGs in the BDNF locus. The relationship between methylation at chr11:27658369 and rs6265 genotype was analyzed using a linear model with methylation as dependent variable, and genotype, case-control status, age, sex, genetic PCs, and methylation PCs as covariates. - Co-expression gene networks associated with BDNF-DT , with BDNF , BDNF-AS , and BDNF-AS-DT , and detection of concordant and discordant pattern of associations: we first performed brain coexpression network analyses using each of these genes as seed gene, in RNAseq data from DLPFC of neurotypical controls (N=318), to detect genes coexpressed with BDNF-DT , BDNF, BDNF-AS , and BDNF-AS-DT . Specifically, after removing low-expressed genes (mean reads per kilobase million [RPKM] <0.17) using the expression_cutoff function in the R package ‘jaffelab’ (version 0.99.34), we performed voom normalization in the Limma package ( 75 ) (release 3.19), and we then used the lmTest and ebayes functions in the package to fit the statistical models to estimate log2 fold changes, moderated t-statistics, and corresponding p values. Multiple-testing correction via the false discovery rate (FDR) was applied using the set of expressed genes (23,346 genes), to define the four subsets of genes significantly co-expressed (p FDR-corrected <0.05) respectively with BDNF-DT , BDNF, BDNF-AS , and BDNF-AS-DT. We investigated whether these four sets of genes are enriched for particular biological features, using QIAGEN’s Ingenuity Pathway Analysis (IPA, QIAGEN, Redwood City, CA, USA; http://www.qiagen.com/ingenuity ). The software determines the pathways, biological processes, and upstream regulators, enriched for a given set of genes by considering the number of focus genes that participate in each process and the total number of genes that are known to be associated with that process in the selected reference set. We performed the Ingenuity Pathway Analysis core analysis, using default parameters (reference set: Ingenuity Knowledge Base; relationships: direct and indirect; node types: all; data sources: all; confidence: experimentally observed and high; species: human, mouse, and rat; tissues and cell lines: all; mutations: all). We chose a P value calculation based on the Benjamini– Hochberg method of accounting for multiple testing in the canonical pathway and functional analyses. We finally applied, on the moderated t-statistics from the linear models described above, a threshold-free algorithm, RRHO2 ( 60 ), based on a rank-rank hypergeometric overlap approach, to detect pattern of concordant and discordant association of genes coexpressed with BDNF-DT with genes coexpressed with BDNF , BDNF-AS , and BDNF-AS-DT . Funding This work was supported by funding from the Lieber Institute for Brain Development. The Genotype-Tissue Expression (GTEx) Project was supported by the Common Fund of the Office of the Director of the National Institutes of Health, and by NCI, NHGRI, NHLBI, NIDA, NIMH, and NINDS. GPe and LCT were supported by R01MH123567. Author contributions Conceptualization: SVB, GPu, DRW, KM, GU Data generation: SVB, NES, JHS, GPe, LCT, KRM, SCP, JEK, TMH, KM, GU Data analysis: SVB, GPu, SM, QC, KRM, GU Supervision: GU Writing—original draft: SVB, GPu, KM, GU Writing—review & editing: all authors Competing interests Authors declare that they have no competing interests Data and materials availability All the data analyzed in this manuscript are publicly available and can be obtained via authorized access from the dbGaP database under accession code phs000979.v3.p2. The GTEX data used to generate data of Supplementary Figure S3i were obtained from the GTEx Portal on 04/03/2025. Acknowledgments We are grateful to the Lieber and Maltz families for their visionary support that funded the analytic work of this project. We are grateful to Shizhong Han, Alexander Straub, Andrew Jaffe, Yankai Jia, Bin Xie, Yuan Gao, for their contribution to data generation. We thank the Psychiatric Genomics Consortium for providing the GWAS summary statistics, and the contributors of the Genotype-Tissue Expression (GTEx) Project for generating part of the transcriptomic data that were analyzed in this work. References 1. ↵ W. Wei , V. Pelechano , A. I. Jarvelin , L. M. Steinmetz , Functional consequences of bidirectional promoters . Trends in genetics : TIG 27 , 267 – 276 ( 2011 ). OpenUrl PubMed 2. ↵ M. Uesaka et al. , Bidirectional promoters are the major source of gene activation-associated non-coding RNAs in mammals . BMC genomics 15 , 35 ( 2014 ). OpenUrl CrossRef PubMed 3. ↵ X. Wu , P. A. Sharp , Divergent transcription: a driving force for new gene origination? Cell 155 , 990 – 996 ( 2013 ). OpenUrl CrossRef PubMed Web of Science 4. R. J. Taft et al. , Tiny RNAs associated with transcription start sites in animals . Nature genetics 41 , 572 – 578 ( 2009 ). OpenUrl CrossRef PubMed Web of Science 5. P. Preker et al. , RNA exosome depletion reveals transcription upstream of active human promoters . Science 322 , 1851 – 1854 ( 2008 ). OpenUrl Abstract / FREE Full Text 6. H. Neil et al. , Widespread bidirectional promoters are the major source of cryptic transcripts in yeast . Nature 457 , 1038 – 1042 ( 2009 ). OpenUrl CrossRef PubMed Web of Science 7. A. Kanhere et al. , Short RNAs are transcribed from repressed polycomb target genes and interact with polycomb repressive complex-2 . Molecular cell 38 , 675 – 688 ( 2010 ). OpenUrl CrossRef PubMed Web of Science 8. A. C. Seila et al. , Divergent transcription from active promoters . Science 322 , 1849 – 1851 ( 2008 ). OpenUrl Abstract / FREE Full Text 9. A. C. Seila , L. J. Core , J. T. Lis , P. A. Sharp , Divergent transcription: a new feature of active promoters . Cell Cycle 8 , 2557 – 2564 ( 2009 ). OpenUrl CrossRef PubMed Web of Science 10. Z. Xu et al. , Antisense expression increases gene expression variability and locus interdependency . Molecular systems biology 7 , 468 ( 2011 ). OpenUrl Abstract / FREE Full Text 11. P. G. Engstrom et al. , Complex Loci in human and mouse genomes . PLoS genetics 2 , e47 ( 2006 ). OpenUrl 12. T. K. Kim et al. , Widespread transcription at neuronal activity-regulated enhancers . Nature 465 , 182 – 187 ( 2010 ). OpenUrl CrossRef PubMed Web of Science 13. ↵ U. A. Orom et al. , Long noncoding RNAs with enhancer-like function in human cells . Cell 143 , 46 – 58 ( 2010 ). OpenUrl CrossRef PubMed Web of Science 14. ↵ V. Pelechano , L. M. Steinmetz , Gene regulation by antisense transcription . Nature reviews. Genetics 14 , 880 – 893 ( 2013 ). OpenUrl CrossRef PubMed 15. ↵ A. Vilborg et al. , Comparative analysis reveals genomic features of stress-induced transcriptional readthrough . Proc Natl Acad Sci U S A 114 , E8362 – E8371 ( 2017 ). OpenUrl Abstract / FREE Full Text 16. A. Vilborg , M. C. Passarelli , T. A. Yario , K. T. Tycowski , J. A. Steitz , Widespread Inducible Transcription Downstream of Human Genes . Mol Cell 59 , 449 – 461 ( 2015 ). OpenUrl CrossRef PubMed 17. ↵ P. Caldas et al. , Transcription readthrough is prevalent in healthy human tissues and associated with inherent genomic features . Commun Biol 7 , 100 ( 2024 ). OpenUrl PubMed 18. ↵ A. E. West , E. C. Griffith , M. E. Greenberg , Regulation of transcription factors by neuronal activity . Nat Rev Neurosci 3 , 921 – 931 ( 2002 ). OpenUrl CrossRef PubMed Web of Science 19. ↵ A. A. Sigova et al. , Divergent transcription of long noncoding RNA/mRNA gene pairs in embryonic stem cells . Proc Natl Acad Sci U S A 110 , 2876 – 2881 ( 2013 ). OpenUrl Abstract / FREE Full Text 20. ↵ B. Lu , K. Martinowich , Cell biology of BDNF and its relevance to schizophrenia . Novartis Found Symp 289 , 119-129 ; discussion 129-135, 193 – 115 ( 2008 ). OpenUrl 21. ↵ A. E. Autry , L. M. Monteggia , Brain-derived neurotrophic factor and neuropsychiatric disorders . Pharmacol Rev 64 , 238 – 258 ( 2012 ). OpenUrl Abstract / FREE Full Text 22. ↵ P. Di Carlo , G. Punzi , G. Ursini , Brain-derived neurotrophic factor and schizophrenia . Psychiatr Genet 29 , 200 – 210 ( 2019 ). OpenUrl CrossRef PubMed 23. ↵ P. Pruunsild , A. Kazantseva , T. Aid , K. Palm , T. Timmusk , Dissecting the human BDNF locus: bidirectional transcription, complex splicing, and multiple promoters . Genomics 90 , 397 – 406 ( 2007 ). OpenUrl CrossRef PubMed Web of Science 24. ↵ L. Feng , P. Molnar , J. V. Nadler , Short-term frequency-dependent plasticity at recurrent mossy fiber synapses of the epileptic brain . The Journal of neuroscience : the official journal of the Society for Neuroscience 23 , 5381 – 5390 ( 2003 ). OpenUrl Abstract / FREE Full Text 25. S. Lahteinen et al. , Decreased BDNF signalling in transgenic mice reduces epileptogenesis . The European journal of neuroscience 15 , 721 – 734 ( 2002 ). OpenUrl CrossRef PubMed Web of Science 26. P. S. Murray , P. V. Holmes , An overview of brain-derived neurotrophic factor and implications for excitotoxic vulnerability in the hippocampus . International journal of peptides 2011 , 654085 ( 2011 ). OpenUrl PubMed 27. R. Koyama et al. , Brain-derived neurotrophic factor induces hyperexcitable reentrant circuits in the dentate gyrus . The Journal of neuroscience : the official journal of the Society for Neuroscience 24 , 7215 – 7224 ( 2004 ). OpenUrl Abstract / FREE Full Text 28. ↵ H. E. Scharfman , J. H. Goodman , A. L. Sollas , Actions of brain-derived neurotrophic factor in slices from rats with spontaneous seizures and mossy fiber sprouting in the dentate gyrus . The Journal of neuroscience : the official journal of the Society for Neuroscience 19 , 5619 – 5631 ( 1999 ). OpenUrl Abstract / FREE Full Text 29. ↵ A. E. Jaffe et al. , Developmental and genetic regulation of the human cortex transcriptome illuminate schizophrenia pathogenesis . Nat Neurosci 21 , 1117 – 1125 ( 2018 ). OpenUrl CrossRef PubMed 30. ↵ R. Elkon , A. P. Ugalde , R. Agami , Alternative cleavage and polyadenylation: extent, regulation and function . Nat Rev Genet 14 , 496 – 506 ( 2013 ). OpenUrl CrossRef PubMed 31. ↵ H. Thorvaldsdottir , J. T. Robinson , J. P. Mesirov , Integrative Genomics Viewer (IGV): high-performance genomics data visualization and exploration . Brief Bioinform 14 , 178 – 192 ( 2013 ). OpenUrl CrossRef PubMed 32. ↵ J. T. Robinson et al. , Integrative genomics viewer . Nat Biotechnol 29 , 24 – 26 ( 2011 ). OpenUrl CrossRef PubMed Web of Science 33. ↵ R. L. Seal et al. , A guide to naming human non-coding RNA genes . EMBO J 39 , e103777 ( 2020 ). OpenUrl CrossRef PubMed 34. ↵ L. Collado-Torres et al. , Regional Heterogeneity in Gene Expression, Regulation, and Coherence in the Frontal Cortex and Hippocampus across Development and Schizophrenia . Neuron 103 , 203 – 216 e208 ( 2019 ). OpenUrl CrossRef PubMed 35. ↵ A. E. Jaffe et al. , Profiling gene expression in the human dentate gyrus granule cell layer reveals insights into schizophrenia and its genetic risk . Nat Neurosci 23 , 510 – 519 ( 2020 ). OpenUrl CrossRef PubMed 36. ↵ P. P. Zandi et al. , Amygdala and anterior cingulate transcriptomes from individuals with bipolar disorder reveal downregulated neuroimmune and synaptic pathways . Nat Neurosci 25 , 381 – 389 ( 2022 ). OpenUrl CrossRef PubMed 37. ↵ W. Zhao et al. , Comparison of RNA-Seq by poly (A) capture, ribosomal RNA depletion, and DNA microarray for expression profiling . BMC Genomics 15 , 419 ( 2014 ). OpenUrl CrossRef PubMed 38. ↵ X. Adiconis et al. , Comparative analysis of RNA sequencing methods for degraded or low-input samples . Nat Methods 10 , 623 – 629 ( 2013 ). OpenUrl CrossRef PubMed Web of Science 39. ↵ G. T. Consortium , The Genotype-Tissue Expression (GTEx) project . Nat Genet 45 , 580 – 585 ( 2013 ). OpenUrl CrossRef PubMed 40. ↵ L. Muniz et al. , Control of Gene Expression in Senescence through Transcriptional Read-Through of Convergent Protein-Coding Genes . Cell Rep 21 , 2433 – 2446 ( 2017 ). OpenUrl CrossRef PubMed 41. ↵ N. Mullani et al. , Reduced RNA turnover as a driver of cellular senescence . Life Sci Alliance 4 , ( 2021 ). 42. ↵ F. Modarresi et al. , Inhibition of natural antisense transcripts in vivo results in gene-specific transcriptional upregulation . Nat Biotechnol 30 , 453 – 459 ( 2012 ). OpenUrl CrossRef PubMed 43. ↵ K. Martinowich et al. , DNA methylation-related chromatin remodeling in activity-dependent BDNF gene regulation . Science 302 , 890 – 893 ( 2003 ). OpenUrl Abstract / FREE Full Text 44. B. Lu , BDNF and activity-dependent synaptic modulation . Learn Mem 10 , 86 – 98 ( 2003 ). OpenUrl Abstract / FREE Full Text 45. ↵ S. V. Bach et al. , Distinct roles of Bdnf I and Bdnf IV transcript variant expression in hippocampal neurons . Hippocampus 34 , 218 – 229 ( 2024 ). OpenUrl PubMed 46. ↵ R. Hoffrogge et al. , 2-DE proteome analysis of a proliferating and differentiating human neuronal stem cell line (ReNcell VM) . Proteomics 6 , 1833 – 1847 ( 2006 ). OpenUrl CrossRef PubMed Web of Science 47. ↵ A. Roussel-Gervais et al. , Genetic knockout of NTRK2 by CRISPR/Cas9 decreases neurogenesis and favors glial progenitors during differentiation of neural progenitor stem cells . Front Cell Neurosci 17 , 1289966 ( 2023 ). OpenUrl CrossRef PubMed 48. ↵ M. J. Owen , A. Sawa , P. B. Mortensen , Schizophrenia . Lancet 388 , 86 – 97 ( 2016 ). OpenUrl CrossRef PubMed 49. ↵ V. Trubetskoy et al. , Mapping genomic loci implicates genes and synaptic biology in schizophrenia . Nature 604 , 502 – 508 ( 2022 ). OpenUrl CrossRef PubMed 50. ↵ K. J. M. Benjamin et al. , Analysis of the caudate nucleus transcriptome in individuals with schizophrenia highlights effects of antipsychotics and new risk genes . Nat Neurosci 25 , 1559 – 1568 ( 2022 ). OpenUrl CrossRef PubMed 51. ↵ Z. Zhu et al. , Integration of summary data from GWAS and eQTL studies predicts complex trait gene targets . Nat Genet 48 , 481 – 487 ( 2016 ). OpenUrl CrossRef PubMed 52. ↵ M. F. Egan et al. , The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function . Cell 112 , 257 – 269 ( 2003 ). OpenUrl CrossRef PubMed Web of Science 53. ↵ S. Li , J. Zhang , S. Huang , X. He , Genome-wide analysis reveals that exon methylation facilitates its selective usage in the human transcriptome . Brief Bioinform 19 , 754 – 764 ( 2018 ). OpenUrl CrossRef PubMed 54. ↵ K. Song , L. Li , G. Zhang , The association between DNA methylation and exon expression in the Pacific oyster Crassostrea gigas . PLoS One 12 , e0185224 ( 2017 ). OpenUrl CrossRef PubMed 55. ↵ G. Ursini et al. , BDNF rs6265 methylation and genotype interact on risk for schizophrenia . Epigenetics 11 , 11 – 23 ( 2016 ). OpenUrl PubMed 56. ↵ P. Pruunsild , M. Sepp , E. Orav , I. Koppel , T. Timmusk , Identification of cis-elements and transcription factors regulating neuronal activity-dependent transcription of human BDNF gene . The Journal of neuroscience : the official journal of the Society for Neuroscience 31 , 3295 – 3308 ( 2011 ). OpenUrl Abstract / FREE Full Text 57. C. Zuccato et al. , Huntingtin interacts with REST/NRSF to modulate the transcription of NRSE-controlled neuronal genes . Nature genetics 35 , 76 – 83 ( 2003 ). OpenUrl CrossRef PubMed Web of Science 58. ↵ T. Timmusk , K. Palm , U. Lendahl , M. Metsis , Brain-derived neurotrophic factor expression in vivo is under the control of neuron-restrictive silencer element . The Journal of biological chemistry 274 , 1078 – 1084 ( 1999 ). OpenUrl Abstract / FREE Full Text 59. ↵ F. Geng , S. Wenzel , W. P. Tansey , Ubiquitin and proteasomes in transcription . Annu Rev Biochem 81 , 177 – 201 ( 2012 ). OpenUrl CrossRef PubMed Web of Science 60. ↵ K. M. Cahill , Z. Huo , G. C. Tseng , R. W. Logan , M. L. Seney , Improved identification of concordant and discordant gene expression signatures using an updated rank-rank hypergeometric overlap approach . Sci Rep 8 , 9588 ( 2018 ). OpenUrl CrossRef PubMed 61. ↵ S. Luo et al. , Divergent lncRNAs Regulate Gene Expression and Lineage Differentiation in Pluripotent Cells . Cell Stem Cell 18 , 637 – 652 ( 2016 ). OpenUrl CrossRef PubMed 62. ↵ E. P. Consortium , An integrated encyclopedia of DNA elements in the human genome . Nature 489 , 57 – 74 ( 2012 ). OpenUrl CrossRef PubMed Web of Science 63. ↵ R. Nieto , M. Kukuljan , H. Silva , BDNF and schizophrenia: from neurodevelopment to neuronal plasticity, learning, and memory . Frontiers in psychiatry 4 , 45 ( 2013 ). OpenUrl PubMed 64. ↵ Z. Y. Chen et al. , Genetic variant BDNF (Val66Met) polymorphism alters anxiety-related behavior . Science 314 , 140 – 143 ( 2006 ). OpenUrl Abstract / FREE Full Text 65. ↵ A. B. Rosenberg , R. P. Patwardhan , J. Shendure , G. Seelig , Learning the sequence determinants of alternative splicing from millions of random sequences . Cell 163 , 698 – 711 ( 2015 ). OpenUrl CrossRef PubMed 66. ↵ B. K. Lipska et al. , Critical factors in gene expression in postmortem human brain: Focus on studies in schizophrenia . Biol Psychiatry 60 , 650 – 658 ( 2006 ). OpenUrl CrossRef PubMed Web of Science 67. ↵ L. Collado-Torres et al. , Flexible expressed region analysis for RNA-seq with derfinder . Nucleic Acids Res 45 , e9 ( 2017 ). OpenUrl CrossRef PubMed 68. ↵ J. E. Krebs , B. Lewin , S. T. Kilpatrick , E. S. Goldstein , Lewin’s genes XI . ( Jones & Bartlett Learning , Burlington, Mass ., ed. 11th , 2014 ), pp. xxvii , 940 p. OpenUrl 69. ↵ K. R. Maynard , E. Stein , DSCAM contributes to dendrite arborization and spine formation in the developing cerebral cortex . J Neurosci 32 , 16637 – 16650 ( 2012 ). OpenUrl Abstract / FREE Full Text 70. ↵ R. Donato et al. , Differential development of neuronal physiological responsiveness in two human neural stem cell lines . BMC Neurosci 8 , 36 ( 2007 ). OpenUrl CrossRef PubMed 71. ↵ R. C. Team , in R Foundation for Statistical Computing . ( Vienna, Austria , 2019 ). 72. ↵ A. E. Jaffe et al. , qSVA framework for RNA quality correction in differential expression analysis . Proc Natl Acad Sci U S A 114 , 7130 – 7135 ( 2017 ). OpenUrl Abstract / FREE Full Text 73. ↵ A. A. Shabalin , Matrix eQTL: ultra fast eQTL analysis via large matrix operations . Bioinformatics 28 , 1353 – 1358 ( 2012 ). OpenUrl CrossRef PubMed Web of Science 74. ↵ K. A. Perzel Mandell et al. , Genome-wide sequencing-based identification of methylation quantitative trait loci and their role in schizophrenia risk . Nat Commun 12 , 5251 ( 2021 ). OpenUrl CrossRef PubMed 75. ↵ M. E. Ritchie et al. , limma powers differential expression analyses for RNA-sequencing and microarray studies . Nucleic Acids Res 43 , e47 ( 2015 ). OpenUrl CrossRef PubMed View the discussion thread. Back to top Previous Next Posted July 14, 2025. Download PDF Data/Code Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following BDNF-DT and BDNF-AS-DT: Novel Genes in the BDNF locus Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share BDNF-DT and BDNF-AS-DT : Novel Genes in the BDNF locus Svitlana V. Bach , Giovanna Punzi , Nuri E. Smith , Sreya Mukherjee , Joo Heon Shin , Qiang Chen , Geo Pertea , Leonardo Collado-Torres , Kristen R. Maynard , Stephanie C. Page , Joel E. Kleinman , Thomas M. Hyde , Daniel R. Weinberger , Keri Martinowich , Gianluca Ursini medRxiv 2025.07.10.25331311; doi: https://doi.org/10.1101/2025.07.10.25331311 Share This Article: Copy Citation Tools BDNF-DT and BDNF-AS-DT : Novel Genes in the BDNF locus Svitlana V. Bach , Giovanna Punzi , Nuri E. Smith , Sreya Mukherjee , Joo Heon Shin , Qiang Chen , Geo Pertea , Leonardo Collado-Torres , Kristen R. Maynard , Stephanie C. Page , Joel E. Kleinman , Thomas M. Hyde , Daniel R. Weinberger , Keri Martinowich , Gianluca Ursini medRxiv 2025.07.10.25331311; doi: https://doi.org/10.1101/2025.07.10.25331311 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Psychiatry and Clinical Psychology Subject Areas All Articles Addiction Medicine (568) Allergy and Immunology (863) Anesthesia (299) Cardiovascular Medicine (4426) Dentistry and Oral Medicine (443) Dermatology (382) Emergency Medicine (607) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1507) Epidemiology (15222) Forensic Medicine (30) Gastroenterology (1123) Genetic and Genomic Medicine (6589) Geriatric Medicine (667) Health Economics (997) Health Informatics (4525) Health Policy (1368) Health Systems and Quality Improvement (1612) Hematology (540) HIV/AIDS (1264) Infectious Diseases (except HIV/AIDS) (15910) Intensive Care and Critical Care Medicine (1103) Medical Education (623) Medical Ethics (145) Nephrology (667) Neurology (6588) Nursing (346) Nutrition (998) Obstetrics and Gynecology (1143) Occupational and Environmental Health (956) Oncology (3331) Ophthalmology (971) Orthopedics (369) Otolaryngology (420) Pain Medicine (435) Palliative Medicine (129) Pathology (663) Pediatrics (1690) Pharmacology and Therapeutics (691) Primary Care Research (710) Psychiatry and Clinical Psychology (5440) Public and Global Health (9221) Radiology and Imaging (2195) Rehabilitation Medicine and Physical Therapy (1369) Respiratory Medicine (1196) Rheumatology (593) Sexual and Reproductive Health (710) Sports Medicine (529) Surgery (711) Toxicology (99) Transplantation (289) Urology (265) (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9ffebe842a743fe2',t:'MTc3OTQ4MzU2MQ=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();