Molecular Psychiatry (2006) 11, 633–648 & 2006 Nature Publishing Group All rights reserved 1359-4184/06 $30.00 www.nature.com/mp ORIGINAL ARTICLE Specificity and timing of neocortical transcriptome changes in response to BDNF gene ablation during embryogenesis or adulthood C Glorioso1, M Sabatini1,2, T Unger1, T Hashimoto1, LM Monteggia4, DA Lewis1,3 and K Mirnics1,2 1Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA; 2Department of Neurobiology, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA; 3Department of Neuroscience, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA and 4Department of Psychiatry, University of Texas Southwestern Medical Center, Dallas, TX, USA Brain-derived neurotrophic factor (BDNF) has been reported to be critical for the development of cortical inhibitory neurons. However, the effect of BDNF on the expression of transcripts whose protein products are involved in gamma amino butric acid (GABA) neurotransmission has not been assessed. In this study, gene expression profiling using oligonucleotide microarrays was performed in prefrontal cortical tissue from mice with inducible deletions of BDNF. Both embryonic and adulthood ablation of BDNF gave rise to many shared transcriptome changes. BDNF appeared to be required to maintain gene expression in the SST-NPY-TAC1 subclass of GABA neurons, although the absence of BDNF did not alter their general phenotype as inhibitory neurons. Furthermore, we observed expression alterations in genes encoding early-immediate genes (ARC, EGR1, EGR2, FOS, DUSP1, DUSP6) and critical cellular signaling systems (CDKN1c, CCND2, CAMK1g, RGS4). These BDNF-dependent gene expression changes may illuminate the biological basis for transcriptome changes observed in certain human brain disorders. Molecular Psychiatry (2006) 11, 633–648. doi:10.1038/sj.mp.4001835; published online 9 May 2006 Keywords: DNA microarray; gene expression; RGS4; somatostatin; neuropeptide Y; tachykinin 1; RT qPCR; BDNF; in situ hybridization Introduction their laminar distribution, connectivity and physio- logical properties. However, the dependence of Brain-derived neurotrophic factor (BDNF) is a critical different subclasses of GABA neurons on BDNF regulator of neural development, promoting the signaling remains to be determined. For example, survival of a variety of neurons in the central nervous the inducible deletion of BDNF11 does not result in system (CNS) (for a review, see Angelucci et al.1). For altered mRNA levels of the 67 kDa isoform of glutamic example, the development of certain phenotypic acid decarboxylase (GAD1) or parvalbumin (PARV) in features of cortical GABA neurons requires BDNF.2–6 adult mice.12 In addition, BDNF is produced and released in an Furthermore, the dependence of any given neuro- activity-dependent fashion by pyramidal neurons, a nal type on BDNF may differ as a function of major target of GABA neurons. Together, these data developmental stage.13 For example, BDNF is a suggest that BDNF is a target-derived trophic factor critical mediator of the maturation of glutamatergic for GABA interneurons. In turn, GABA neurons might synapses in developing mouse somatosensory cor- regulate BDNF synthesis and release from pyramidal tex.14 Similarly, developing dorsal thalamus neurons cells in an activity-dependent manner.7,8 need BDNF for survival from the cerebral cortex, and Cortical GABA neurons are heterogeneous and can dorsal thalamic cell death can be regulated by be subdivided into a variety of subclasses based on increasing or blocking cortical levels of BDNF.15 In their expression of different calcium binding proteins addition, behavioral deficits are more pronounced in and neuropeptides (for a review, see Markram et al.9 mice with embryonic than adult ablation of BDNF.11 McBain and Fisahn10). These subclasses also differ in These data suggest that BDNF requirement of neurons may change from development to adulthood, and Correspondence: Dr K Mirnics, Department of Psychiatry, Uni- mice with embryonic and adult deletion of BDNF may versity of Pittsburgh School of Medicine, E1655 Biomed Sci exhibit markedly different neocortical transcriptome Tower, Pittsburgh, PA 15261, USA. profiles. E-mail: karoly þ @pitt.edu Received 31 January 2006; revised 13 March 2006; accepted 24 To obtain comprehensive insight into the BDNF- March 2006; published online 9 May 2006 dependent transcriptome changes and determine BDNF cKO transcriptome C Glorioso et al 634 what aspects of interneuronal phenotypes are affected tional three animals per group were assessed (12 by reduced BDNF levels, we performed high-density animals in total). oligonucleotide microarray profiling of the prefrontal cortex of mice with inducible deletion of BDNF. The Microarray experiments goals of the current study were to: (1) establish Frontal cortices were rapidly dissected, frozen on dry transcriptome changes which are a result of embryo- ice, and stored at À801C until RNA extraction. Total nic deletion of BDNF, (2) identify transcriptome RNA was isolated using the Trizol reagent. RNA changes that are a result of adult deletion of BDNF, quality was assessed using the Agilent Bioanalyzer. (3) describe the genes that are most influenced by the Reverse transcription, in vitro transcription and absence of BDNF across both conditions, (4) define fragmentation were performed according to manufac- the transcriptome differences between the embryonic turer’s recommendation (Affymetrix, Sunnyvale, CA, and adult BDNF-deficient mice, and (5) identify USA). Samples were hybridized onto MOE430A neuronal cortical phenotypes that may be preferen- mouse Affymetrix GeneChips which contained tially affected by the absence of BDNF. > 22 000 probesets using the Affymetrix hybridization station. To avoid microarray batch variation only microarrays from a single lot were used. Microarrays Materials and methods were considered for use only if the average 30:50 ratio Brain-derived neurotrophic factor -deficient mice for GAPDH and actin did not exceed 1:1.2. Segmenta- All mice used in this experimental series were tion of scanned microarray images was performed by described previously.11,12 To generate the animals for Microarray Analysis Suite 5.0 (MAS5). Determination the current experimental series, two genetically- of expression levels and scaling were performed using altered, independently-derived mouse strains were Robust Multi-array Average (RMA).20,21 The resulting used. Reduction of BDNF expression was achieved data set was filtered for genes that reported < 30% using an inducible knockout (KO) of the BDNF gene present calls based on MAS5 analysis. For scale where two transgenes, the tetracycline transactivator linearity, the data were log2 transformed, and differ- (tTA) gene driven by neuron-specific enolase (NSE) ential expression was established using average log2 promoter (nse-tTA)16 and the Cre recombinase gene ratio (ALR) between the studied cohorts (|ALR| = 1 under the control of tTA-responsive tetO promoter corresponds to a twofold increase or decrease, (tetO-cre)17,18 regulate the deletion of floxed exon V of |ALR| = 0.585 represents a 50% change, while BNDF19 in a tetracycline-dependent manner. The two |ALR| = 0.263 depicts a 20% expression alteration). mouse lines with NSE-tTA and tetO-cre transgenes were both maintained as homozygotes. Crossing of Data analysis these lines resulted in bigenic mice.19 For the embryonic deletion of BDNF, mice were bred in the Identification of differentially expressed genes across absence of the tetracycline analogue doxycycline; this mice with adult and embryonic BDNF-deletion: We resulted in late embryonic ablation of BDNF.16 For the identified genes as differentially expressed between adulthood deletion of BDNF, mice were bred in the all wild-type and all BDNF KO samples if they presence of doxycycline (1 mg/ml) in drinking water fulfilled the following four criteria: (1) reported and maintained on this until 3 months of age. At that > 0.263 |average log2 ratio| (|ALR|) in embryonic time doxycycline was removed from the drinking BDNF-deleted mice versus their matched wild-type water, thus inducing recombination. Maximum re- controls, (2) a group-wise Student’s t-test reported combination was achieved 4 weeks after doxycycline P < 0.05 in the embryonic BDNF-deleted mice versus removal11 resulting in a > 70% reduction in BDNF their matched wild-type controls, (3) reported > 0.263 transcript and protein levels.11,12 Littermates with |ALR| in adult BDNF-deleted mice versus their NSE-tTA and tetO-cre transgenes were used as matched wild-type controls and (4) a group-wise controls. These mice expressed similar levels of Student’s t-test reported P < 0.05 in the adult BDNF- BDNF as wild type mice.11 deleted mice versus their matched wild-type controls. Mice with BDNF deletion induced from embryo- These combined criteria were implemented to reduce genesis or in adulthood, together with the corre- false positive findings and eliminate significant, but sponding control animals, were sacrificed between 5 very small, expression changes that may have a and 6 months of age. The experimental series marginal biological effect.22 A gene was ‘definitively consisted of frontal cortices from a total of 38 animals changed’ if it met all four aforementioned criteria, with different animals used for each method of gene while a gene was considered ‘putatively changed’ if it expression analysis. For the microarray experiments, met three