BDNF deficiency and enriched environment treatment affect neurotransmitter gene expression differently across ages

Abstract

Deficiency of activity-induced expression of brain-derived neurotrophic factor (BDNF) disturbs neurotransmitter gene expression. Enriched environment treatment (EET) ameliorates the defects. However, how BDNF deficiency and EET affect neurotransmitter gene expression differently across ages remains unclear. We addressed this question by determining neurotransmitter gene expression across three life stages in wild-type and activity-dependent BDNF deficient (KIV) mice. Mice received 2-months of standard control treatment (SCT) or EET at early-life development (ED: 0–2 months), young adulthood (2–4 months), and old adulthood (12–14 months)(N=16/group). Half of these mice received additional one-month SCT to examine persisting EET effects. High-throughput qRT-PCR measured expression of 81 genes for dopamine, adrenaline, serotonin, GABA, glutamate, acetylcholine, and BDNF systems in the frontal cortex (FC) and hippocampus. Results revealed that BDNF deficiency mostly reduced neurotransmitter gene expression, greatest at ED in the FC. EET increased expression of a larger number of genes at ED than adulthood, particularly in the KIV FC. Many genes downregulated in KIV mice were upregulated by EET, which persisted when EET was provided at ED (e.g., 5HTT, ADRA1D, GRIA3, GABRA5, GABBR2). In both regions, BDNF deficiency decreased the density of gene co-expression network specifically at ED, while EET increased the density and hub genes (e.g., GAT1, GABRG3, GRIN1, CHRNA7). These results suggest that BDNF deficiency, which occurs under chronic stress, causes neurotransmitter dysregulations prominently at ED, particularly in the FC. EET at ED may be most effective to normalize the dysregulations, providing persisting effects later in life.

Graphical Abstract

We report the age-dependency of how BDNF deficiency (with promoter IV defect, KIV) and enriched environment treatment (EET) affect expression of 81 neurotransmitter-related genes. BDNF deficiency reduced expression of the largest number of neurotransmitter genes at early-life development (ED), particularly in the frontal cortex (FC). The EET effects were more specific to BDNF deficiency, largest at ED in the FC but more universal across ages in the hippocampus. These effects were reflected by the density of gene co-expression network. Our results highlight the importance of early-life EET for BDNF deficiency, which occurs under chronic stress.

Introduction

Neuronal activity induces gene expression of brain-derived neurotrophic factor (BDNF) (Zafra et al. 1991; Timmusk et al. 1993). BDNF is the major neuronal growth factor in the brain which promotes neuronal maturation and synaptic function (Thoenen 1995; Poo 2001; Lu 2003). While BDNF expression is controlled by nine promoters (Liu et al. 2005; Aid et al. 2007), promoter IV (previously classified as promoter III) is the predominant activity-dependent promoter, inducing BDNF transcription upon neuronal depolarization (Ca²⁺ influx) (Tao et al. 1998; Shieh et al. 1998). However, promoter IV becomes inactivated upon exposure to negative environments, such as chronic stress, toxins, and early-life maltreatment via epigenetic processes (e.g., DNA methylation and chromatin modification), particularly in the hippocampus and prefrontal cortex (Tsankova et al. 2006; Fuchikami et al. 2009; Roth et al. 2009; Onishchenko et al. 2008; Marwarha et al. 2017). Inactivation of promoter IV abolishes activity-induced BDNF expression in the hippocampus and prefrontal cortex (Sakata et al. 2009; Sakata et al. 2013a). This causes abnormal behavior in reward intake, stress avoidance, and flexible learning, which manifests at young adulthood (Sakata et al. 2010; Sakata et al. 2013a). Defective promoter IV-driven BDNF transcription can even transmit across generations via abnormal maternal behavior (Roth et al. 2009). Deficiency of BDNF, particularly with inactive promoter IV, is widely observed in human mental disorders, such as schizophrenia (Weickert et al. 2003; Hashimoto et al. 2005; Wong et al. 2010), depression (Dwivedi et al. 2003; Keller et al. 2010; Hing et al. 2012), and Alzheimer’s disease (Phillips et al. 1991; Murray et al. 1994; Garzon et al. 2002).

To compensate for deficiency of promoter IV-driven BDNF expression (pIV-BDNF deficiency), we found that enriched environmental treatment (EET) is more effective than chronic drug treatment with four classes of antidepressants (SSRI: selective serotonin transporter reuptake inhibitor, SNRI: serotonin-norepinephrine reuptake inhibitor, TCA: tricyclic antidepressant, and MAOI: monoamine oxidase inhibitor) (Jha et al. 2011; Sakata et al. 2013b). EET, which combines physical exercise, mental stimulation, and social interaction (van Praag et al. 2000), activates almost all other BDNF promoters (Jha et al. 2011; Dong et al. 2018; Russo-Neustadt et al. 2000; Zajac et al. 2010; Adlard et al. 2004), whereas the drug treatments activate only a few BDNF promoters (Sakata et al. 2013b; Sakata 2011). EET ameliorate defects caused by BDNF deficiency (Chourbaji et al. 2008; Rossi et al. 2006; Zhu et al. 2009). Interestingly, we recently found that the EET effects are age-dependent; BDNF induction by EET is greater when EET is given during early-life development than at young adulthood or old adulthood in mice (Jha et al. 2016; Dong et al. 2018). The BDNF induction reflects changes in motivational behaviors (e.g., increased exploratory activity and stress-induced struggles) (Jha et al. 2011; Jha et al. 2016). This BDNF induction and some behavioral changes persist after discontinuation of EET, particularly in pIV-BDNF deficiency, when EET is provided during early-life development (Jha et al. 2016; Dong et al. 2018).

One potential mechanistic link between BDNF and behavioral changes is neurotransmitter gene regulation. BDNF is critical for maturation and function of most types of neurons, including glutamatergic, GABAergic, serotonergic, dopaminergic, adrenergic and cholinergic neurons (Ghosh et al. 1994; Gottmann et al. 2009; Marty et al. 1996; Lyons et al. 1999; Mamounas et al. 2000; Hyman et al. 1991; Baker et al. 2005; Guo et al. 2005; Alderson et al. 1990; Ward & Hagg 2000). BDNF deficiency causes aberrant expression of genes related to these neurotransmitter systems (neurotransmitter genes) (Sakata & Duke 2014; Sakata & Overacre 2017; Chourbaji et al. 2004; Rios et al. 2006; Djalali et al. 2005; Tripp et al. 2012). On the other hand, EET can normalize the expression changes (Sakata & Overacre 2017; Rampon et al. 2000; Lee et al. 2013; Bredy et al. 2004; Komitova et al. 2013; Zhu et al. 2014; Mehta-Raghavan et al. 2016). However, these studies have been done mostly in young adult animals and the age dependency of the effects remain largely unknown. How do pIV-BDNF deficiency and EET affect expression of neurotransmitter genes differently across ages? Here, we addressed this question by determining the effects of pIV-BDNF deficiency and EET across three life stages: early-life development, young adulthood, and old adulthood. Specifically, we asked 1) which life stage pIV-BDNF deficiency largely affected neurotransmitter gene expression, 2) which life stage EET most effectively altered gene expression in both normal and BDNF deficient conditions, and 3) whether the EET effects persisted after discontinuation of EET. We also examined how pIV-BDNF deficiency and EET affect gene co-expression network across ages. We tested our hypothesis that EET during early-life development may produce maximum effects in neurotransmitter gene expression particularly in pIV-BDNF deficiency.

Materials and Methods

Animals

Wild-type (WT) and KIV mice [generated as previously described (Sakata et al. 2009), source: NIH] were used to determine effects of pIV-BDNF deficiency. Briefly, KIV mice contain a green-fluorescent-protein gene with a stop codon in exon IV, which disrupts promoter IV-driven expression of BDNF protein. This largely abolishes activity-induced BDNF expression in the prefrontal cortex and hippocampus (Sakata et al. 2009; Sakata et al. 2013a). KIV mice were generated from 129/sv embryonic stem cells with C57BL/6J blastocytes and then crossed to C57BL/6J females (The Jackson Laboratory, Bar Harbor, ME) for >12 generations. Heterozygous mice were bred to produce WT and KIV littermates of the same genetic background. WT and KIV offspring from these littermates were used. A total of 96 WT and 96 KIV mice were used to collect data from 8 mice/group (4 males and 4 females) for 3 age groups with 4 treatments (see below). The sample size was decided based on previous experiments and power analyses (Dong et al. 2018). All animals were group-housed in a climate-controlled vivarium in a normal 12:12h-dark-light cycle with food and water ad libitum. All animal experiments were approved by the University of Tennessee Laboratory Animal Care and Use Committee (#17–090.0) and were conducted in accordance with NIH guidelines. The study was not pre-registered (exploratory). The custom-made materials/animals will be shared upon reasonable request.

Treatments

WT and KIV mice were arbitrarily assigned to EET or standard control treatment (SCT) to determine EET effects in normal and pIV-BDNF-deficient conditions. Briefly, mice were listed in a Microsoft Excel (RRID:SCR_016137) and sorted by the age and sex for each genotype, and then the mice were assigned for treatment by the order (1 SCT and 2 EET). The EET method is described previously (Jha et al. 2011; Jha et al. 2016; Dong et al. 2018). Briefly, SCT mice were kept in a standard cage (2–5 mice in a small cage). EET mice were provided running wheels to promote voluntary exercise, novel objects (replaced weekly) to promote cognitive/sensory stimulation, 5–10 companion mice to increase social interaction, and bacon-flavored Rodent Foraging Crumbles (#5783: Bio-Serv, Frenchtown, NJ) in bedding to encourage exploration in the enriched environment (5–10 mice in a large cage). The companion mice were siblings (from the same mother) or cage mates (litters from another mother but raised together). In the case of insufficient number of siblings/cage mates, female companion mice were used. This is because introducing unfamiliar adult male mice to an adult male mouse may cause fighting.

EET was provided for 2 months during early-life development (ED: 0–2 months old), young adulthood (YA: 2–4 months old), and old adulthood (OA: 12–14 months old) to examine EET effects at three different life stages (Fig. 1a). These periods were selected to cover from birth to reproductive maturity (ED), after reproductive maturity (YA), and end of reproductive ability (OA) in mice (Flurkely et al. 2007; Jax). After EET, half of these mice (N=8/group) received SCT for 1 month (EET-SCT or SCT-SCT) to examine any persisting EET effects. No animals were excluded.

Fig. 1.

a. Research design. Wild-type (WT) and BDNF deficient (KIV) mice received 2 months of enriched environment treatment (EET; red arrows) or standard control treatment (SCT: blue arrows) from birth (early-life development: ED), 2 months of age (young adult: YA), or 12 months of age (old adult: OA), then received 1 month of SCT. Tissues were collected after the treatment (T1, N=8 mice/group) and after the consequent 1 month of SCT (T2, N=8 mice/group). Modified and reused from Jha et al. Transl. Psychiatry 2016 with a permission. b. Number of genes showing effects of BDNF deficiency and EET.

Tissue collection, RNA/cDNA preparation, and high-throughput qRT-PCR (BioMark)

The frontal cortex (FC) and hippocampus (HIP) were selected because these regions are where KIV mice largely lack activity-induced BDNF protein expression (Sakata et al. 2009) and where BDNF is reduced by stress (Tsankova et al. 2006; Roth et al. 2009) and in many psychiatric conditions (Dwivedi et al. 2003; Weickert et al. 2003; Phillips et al. 1991). The same tissues used in our previous study (Dong et al. 2018) were used in this study. Briefly, 384 tissues (96 WT and 96 KIV mice x 2 brain regions) were collected at the end of 4 treatments (SCT/EET/SCT-SCT/EET-SCT, see Fig. 1a). Mice were given isoflurane (1–3 ml/L vapor, for ~1 min) for rapid anesthesia and sacrificed by decapitation. Tissues were collected within 3 min and immediately frozen on dry ice. Tissues were collected between 14:00–17:00 h to avoid any expression differences due to circadian rhythm. RNA was extracted from each tissue and converted into cDNA, which was subjected to high-throughput qRT-PCR (BioMark), as described previously (Sakata & Duke 2014; Dong et al. 2018). The experimenters knew which tissue samples were from which treatments/ages/genotypes. This is because RNA/cDNA measurements and PCR were performed by machines and therefore did not include experimenters’ subjective measures. Additionally, Cycle threshold (Ct) measures were normalized to Ct values of references in the same samples.

Four 96.96 BioMark chips (#BMK-M-96.96, Fluidigm, South San Francisco, CA) were used to measure expression levels of 96 genes for 384 samples (96 male FC, 96 female FC, 96 male HIP, and 96 female HIP). Ct values were obtained by BioMark software (RRID: SCR_015686) after automatic inspection for quality. CyclophilinD and HGPRT were used as reference genes; they showed the least Ct standard deviation among 6 house-keeping genes (CyclophilinD, HGPRT, TBP, β-actin, β-tubulin, and S19) in HIP and FC samples. We used the average Ct values of CyclophilinD and HGPRT because CyclophilinD increased and HGPRT decreased in expression by age. This combination cancelled the age effects, in order to provide constant reference across ages. Relative gene expression was calculated by the 2^–ΔΔCt method (Livak & Schmittgen 2001). Expression (fold) changes by EET were calculated to the respective control (SCT/SCT-SCT). In a pre-test, we prescreened genes and omitted the genes of which expression levels were too low and unreliable (Ct values >27 after 14 cycle preamplification). As for acetylcholine genes, we selected 6 genes showing BDNF/EET effects in our previous study (Sakata & Overacre 2017). Expression of total 81 genes of dopamine, serotonin, noradrenaline, glutamate, GABA, and acetylcholine and BDNF-related systems were analyzed. The primers, probes, and the full name of genes are listed in Supplementary Table 1.

Gene co-expression network analyses were carried out using an R program. The Pearson correlation coefficient threshold was set at 0.8. To obtain sufficient sample numbers (>13) for reliable correlation analyses, SCT and SCT-SCT groups were combined as SCT group, and EET and EET-SCT groups were combined as EET group. Hub genes were defined as genes showing high node degree (>10).

Statistical Analysis

R program and XLSTAT (RRID:SCR_016299) were used for statistical analyses. Two-way analyses of variance (ANOVA) were used to determine age dependency of genotype effects or treatment effects. Post-hoc Ryan-Einot-Gabriel-Welsch q (REGWQ) test was used to compare a large number of data while controlling family-wise error (Type I errors during multiple comparisons) (Ryan 1960). We removed zero values and outliers that were more than two standard deviations away from the mean, as they likely resulted from technical errors. Otherwise, no exclusion criteria were pre-determined. The Jarque-Bera normality tests were performed to analyze the data distribution for each group. All the data were normally distributed (P>0.05) except for several genes, for which Kruskal-Wallis nonparametric tests were used (Supplementary Tables 2–4). Statistics are presented in Supplementary Table 2–4. Statistical significance was set at p<0.05*.

Results

1. Does pIV-BDNF deficiency affect neurotransmitter gene expression differently across ages?

First, we examined age-dependent effects of pIV-BDNF deficiency by comparing WT and KIV mice with SCT across the three life stages. BDNF deficiency affected a larger number of genes in the FC than in the HIP; two way-ANOVA revealed significant genotype effects for 49 genes in the FC and 20 genes in the HIP, out of 81 genes tested (WT vs. KIV: p<0.05, F values for each gene in Supplementary Table 2). Significant age effects were observed in 22 genes in the FC and 27 genes in the HIP, with significant age-genotype interactions in 16 genes in the FC and 6 genes in the HIP (Supplementary Table 2). Post-hoc tests revealed significant genotype effects at ED, YA, and OA in 23, 13, and 14 genes in the FC and 7, 3, and 7 genes in the HIP, respectively (Fig. 1b). The largest number of affected genes at ED in the FC indicated that deficiency of activity-dependent BDNF expression disturbed expression of neurotransmitter genes more severely at ED than at adulthood particularly in the FC. These affected genes mostly showed downregulation in KIV mice compared to WT mice (FC: 27 down vs. 5 up; HIP: 9 down vs. 3 up, Table 1a). Many of these genes were for the serotonin, glutamate, and GABA systems in the FC (Table 1a). Seven genes (5HTR5b, 5HTT, AdRa1d, CHRNA5, Tac1, Gabra2, PENK) were affected in both FC and HIP with the same (up or down) expression regulation (Table 1a).

Table 1. BDNF deficiency effects on neurotransmitter gene expression across ages.

a. Genes showing significant genotype effects (WT vs. KIV) with SCT at specific ages (in parentheses). Note that many genes showed downregulation, rather than upregulation, in KIV mice compared to WT mice in both brain regions. b. Genes showing significant age effects [early-life development (ED) vs. young adulthood (YA) vs. old adulthood (OA)]. The genes showing both genotype and age effects are listed in blue. Significance by two-way ANOVA with REGWQ test. N=14–16 mice/group.

	a. Genotype Effects			b. Age Effects
	Up (WT<KIV)	Down (WT>KIV)		WT	KIV
Frontal Cortex	TPH(YA)	5HTR1b(ED, YA, OA)	GRIN2B(ED)	5HTR1b(ED>YA/OA)	DRD4(ED/YA<OA)
	DRD4 (OA)	5HTR2a(ED)	Grm5(ED, YA)	DRD5(ED<YA/OA)	DRD5(ED/YA<OA)
	CHRM2(YA, OA)	5HTR2c(ED)	Gabra5(ED, YA, OA)	TH(ED>YA/OA)	CHRM2(ED<YA/OA)
	Gabra2(YA)	5HTR5b(ED, YA, OA)	Gabrr1(ED, OA)	GRIN1(ED>YA/OA)	Grm3(ED/YA>OA)
	PENK(YA, OA)	5HTT(ED)	Gabbr2(ED)	GRIN2A(ED>YA/OA)	Gabra2(ED/YA>OA)
		DRD5 (YA)	GAD2(OA)	GRIN2B(ED>YA/OA)	PALV(ED/YA<OA)
		TH (ED)	CALB(YA)	Grm3(ED/YA>OA)	CALB(ED>YA/OA)
		AdRa1d(ED)	SMST(ED, YA, OA)	Grm5(ED>OA)	GFAP(ED<OA)
		CHRNA5(ED, YA, OA)	Tac1(ED, YA, OA)	Gabra2(ED>YA/OA)
		GRIA1(ED, OA)	p75(OA)	Gabra4(YA>OA)
		GRIA2(ED, OA)	DCX(ED)	Gabrg2(ED>OA)
		GRIA3(ED)	HDAC5(ED)	CALB(ED/YA>OA)
		GRIN1(ED)	CRF(ED)	p75(ED/YA<OA)
		GRIN2A(ED)		DCX(ED>YA/OA)
				CRF(ED>YA/OA)
Hippocampus	DRD1 (OA)	5HTR1a(OA)		DRD3 (ED>YA/OA)	5HTR1a (ED<YA, YA>OA)
	Gabra2(YA, OA)	5HTR5b(ED, YA, OA)		CHRM5 (ED>YA/OA)	TPH (ED<OA)
	PENK(OA)	5HTT(ED)		CHRNA5 (YA>OA)	DRD3 (ED>YA>OA)
		TPH(ED)		Gabra1 (ED<YA, YA>OA)	Grm1 (ED<YA, YA>OA)
		AdRa1d(ED)		Gabra2 (ED>YA/OA)	Grm2 (ED/YA>OA)
		CHRM5(ED)		DCX (ED>YA/OA)	Gabra1 (YA>OA)
		CHRNA5(ED)			Gabrb2 (ED/YA<OA)
		Gabrb2(OA)			Gabrr2 (ED<YA)
		Tac1(ED, YA, OA)			PENK (ED<OA)
					DCX (ED>YA/OA)
					GFAP (ED<YA/OA)

Post-hoc tests revealed significant age effects for 19 genes in the FC and 14 genes in the HIP (Table 1b). Interestingly, about one-half of these age-affected genes were the genes showing significant genotype effects (shown in blue in Table 1b). The genotype x age effects and interactions for each gene are presented in Supplementary Fig. 1 and 2.

2. At which life stage does EET most effectively alter neurotransmitter gene expression?

Next, we examined age-dependency of EET effects for each genotype. Two way-ANOVA revealed significant EET effects (vs. SCT) in 7, 30, 15, and 24 genes in the WT FC, KIV FC, WT HIP, and KIV HIP, respectively, and significant age-treatment interactions in 0, 12, 2, and 9 genes in the WT FC, KIV FC, WT HIP, and KIV HIP, respectively (p<0.05, Supplementary Table 3). Post-hoc tests revealed significant EET effects at ED, YA, OA in 0, 0, and 0 genes in the WT FC; 21, 0, and 1 genes in the KIV FC; 2, 0, and 2 genes in the WT HIP; and 10, 1, and 8 genes in the KIV HIP, respectively (Table 2a, Fig. 1b, Supplementary Fig. 3 and 4). Thus, EET affected a larger number of neurotransmitter genes in KIV mice than in WT mice in both FC and HIP, while affecting the largest number of genes at ED in KIV mice.

Table 2. EET effects on neurotransmitter gene expression across ages.

a. Genes showing significant EET effects (vs. SCT) at specific ages (in parenthesis). Arrows indicate upregulation (↑) or downregulation (↓) by EET. EET increased a larger number of genes in KIV than in WT for both regions. b. Genes showing significant age effects (ED vs. YA vs. OA) on EET effects. Genes showing both age and EET effects are listed in blue. Many genes showing EET effects also showed age effects. c. Genes showing significant genotype differences in EET effects (WT-EET vs. KIV-EET). Almost all genes showed a larger expression changes by EET in KIV mice than in WT mice. Genes showing downregulation in KIV mice (vs. WT) without EET (see Table 1) are listed in bold font. Significance by REGWQ test. N=6–8 mice/group.

	a. EET effects			b. Age effects			c. Genotype effects
	WT	KIV		WT	KIV		KIV >WT		KIV < WT
Frontal Cortex		5HTR5b(ED)↑	Gabra5(ED)↑	Gabre (ED>YA, YA<OA)	5HTR5b(ED>YA/OA)	Gabra3(ED>YA)	5HTR5b(ED)	Gabra5(ED)
		5HTT(ED) ↑	Gabrb1(ED) ↑		5HTT(ED>OA)	Gabra5(ED>YA/OA)	5HTT(ED, YA)	Gabrg2(OA)
		TH (ED) ↑	Gabrr1(OA) ↑		TH(ED>YA/OA)	Gabrb1(ED>YA/OA)	TH(ED)	Gabrd(ED)
		AdRa1a(ED) ↑	Gabbr2(ED) ↑		AdRa1a(ED>YA/OA)	Gabrd(ED>YA/OA)	AdRa1a(ED)	Gabrr1(ED, YA, OA)
		AdRa1d(ED)↑	GAD1(ED)↑		AdRa1d(ED>YA/OA)	CALB(ED>OA)	AdRa1d(ED)
		CHRM2(ED) ↑	Tac1(ED) ↑		AdRb3(YA<OA)	Tac1(ED>OA)	AdRa2a(OA)	GAD1(ED, OA)
		CHRNA5(ED)↑	TrkB(ED) ↑		CHRM2(ED>YA/OA)	p75(ED>YA)	AdRb3(OA)
		GRIA3(ED) ↑	p75(ED) ↑		CHRNA5(ED>OA)	DCX(ED>YA)	CHRM2(ED)	CALB(OA)
		GRIN1(ED) ↑	GFAP(ED) ↑		GRIA3(ED>YA/OA)	GFAP(ED>YA/OA)	GRIA2(ED)	CR(YA)
		GRIN2A(ED) ↑	HDAC5(ED) ↑		GRIN1(ED>YA/OA)	HDAC5(ED>YA/OA)	GRIK1(OA)	p75(ED)
		GRIN2B(ED) ↑	CRF(ED) ↑		GRIN2A(ED>OA)	GR1(ED>OA)	GRIN1(ED)	DCX(ED)
					GRIN2B(ED>YA/OA)	CRF(ED>YA/OA)	GRIN2A(ED)	GFAP(ED)
							GRIN2B(ED)	HDAC5(ED)
								CRF(ED)
Hippocampus	TH (ED)	5HTR2a(YA, OA)↑	Gabrr2(ED, OA)↑	TH (ED>OA)	5HTR2a(ED<OA)	Gabrg2(ED<OA)	5HTR5b(ED)	Gabra5(YA)	Gabra2(OA)
		5HTR5b(ED) ↑		GRIN2A(ED<OA)	5HTR4(ED<YA/OA)	Gabre(ED>YA)	DRD2 (OA)	Gabrg2(OA)
		DRD1 (OA) ↑	Tac1(ED) ↑		5HTR5b(ED>YA/OA)	Gabrr2(ED>YA,YA<OA)	DRD3 (OA)	Gabbre(ED)
	AdRa2a(OA) ↓	DRD3 (OA) ↑	TrkB(ED) ↑		DRD3(ED/YA<OA)		DRD4 (OA)	Gabrr1(OA)
		DRD4 (OA) ↑	p75(ED, OA) ↑	GRIN2B(ED<OA)	DRD4(YA<OA)	Tac1(ED<YA/OA)	AdRa2a(ED)	Gabrr2(ED, OA)
		TH (ED) ↑	CRF(ED) ↑		AdRb3(ED>YA/OA)	p75(ED>YA, YA<OA)	AdRb3(ED)
	PENK(ED, OA)	AdRb3(ED) ↑			CHRM2(ED>YA)	CRF(ED>YA/OA)	GRIK1(OA)	p75(OA)
		GRIN1(ED, OA) ↑		TrkB(ED>OA)	Grm7(ED/YA<OA)		GRIN1(ED, OA)	DCX(OA)
		Grm7(OA) ↑					Grm7(OA)	HDAC5(ED)
		Gabre(ED) ↑					Grm8(ED, OA)	CRF(ED)

EET increased expression of all genes in both regions for both genotypes, except AdRa2a (Table 2a). Interestingly, many of these EET-induced genes were the genes downregulated in KIV mice without EET in the FC (shown in bold in Table 2a, see examples in Fig. 2). The 15 (68%) of 22 genes that EET induced were the same as those downregulated in KIV mice [see the overlap in Fig. 1b, 14 (67%) of 21 genes at ED]. This overlap was much larger than the total number of genes (i.e., 7) that EET induced but were not affected by KIV (Fig. 1b, Table 2a). By contrast, in the HIP, such overlap genes were only two (13 %, 5HTR5b/Tac1) out of 15 genes detected [(2 (20%) of 10 genes at ED, Fig. 1b]. In the HIP, EET induced many other genes not downregulated by BDNF deficiency (13 genes, Table 2a). EET affected genes broadly across the neurotransmitter systems. Seven genes (5HTR5b, TH, GRIN1, Tac1, TrkB, p75, CRF) were commonly affected in both HIP and FC (Table 2a).

Fig. 2. Examples of gene expression.

showing: i) reductions by BDNF deficiency in KIV (vs. WT) in all ages but significant induction by EET (vs. SCT) only at ED; ii) significant reductions by BDNF deficiency only at ED and induction by EET only at ED; iii) no changes by BDNF deficiency but inductions by EET at ED (and at OA for p75); iv) a reduction or an induction by BDNF deficiency across ages but no changes by EET; and v) compensatory induction by EET in BDNF deficiency across ages. The graphs shows data in the frontal cortex unless noted with “hippocampus (HIP).” For each category, top graphs show gene expression levels in SCT (N=14–16 mice/group) and bottom graphs show EET effects by fold changes (EET/SCT, N=6–8 mice/group). All graphs for 81 genes are shown in Supplementary Fig. 1–6. The box and whisker plot shows the minimum, first quartile, median, third quartile, and maximum of a set of data. The cross shows the mean.

Among the genes affected in KIV mice (vs. WT), EET affected the largest number of genes at ED in the FC: EET affected 14, 0, and 1 genes (61%, 0%, and 7%) of 23, 13, and 14 genes in the FC and 2, 0, and 0 genes (29%, 0%, and 0%) of 7, 3, and 7 genes in the HIP at ED, YA, and OA, respectively (Fig. 1b). Thus, EET did not rescue most of the genes at YA (100 % and 100%) or OA (93% and 100%) in either FC or HIP, respectively (Fig. 1b).

Many genes showing EET effects in KIV mice also displayed significant age effects (FC: 18 of 22 genes detected; HIP: 11 of 14 genes detected, shown in blue, Table 2b). EET induction levels were significantly higher at ED than at YA or OA for most genes in the FC (Supplementary Fig. 3). In the HIP, EET induction levels were larger at OA as well as at ED, depending on the gene (Supplementary Fig. 4).

Significant genotype differences in the EET effects (WT-EET vs. KIV-EET) were observed at ED, YA, and OA in 19, 3, and 7 genes in the FC and 9, 1, and 13 genes in the HIP, respectively (Table 2c). The expression increases by EET were larger in KIV mice than in WT mice in both FC and HIP for all genes, except Gabra2 (Table 2c, Supplementary Fig. 3 and 4). Thus, EET neurotransmitter gene effects in both regions depended on BDNF deficiency, being larger in BDNF deficiency, for both the number of genes and the degree of expression changes.

3. Do the EET effects persist after discontinuation of EET?

Further, we examined whether EET gene effects persisted. Two-way ANOVA revealed significant EET effects one month after discontinuation of EET (EET-SCT vs. SCT-SCT: p<0.05) in 1, 9, 5, and 12 genes in the WT FC, KIV FC, WT HIP, and KIV HIP, respectively (Supplementary Table 4). Post-hoc tests revealed significant EET effects at ED, YA, and OA in 1, 0, and 0 genes in the WT FC; 4, 1, and 2, genes in the KIV FC; 0, 1, and 1 genes in the WT HIP; and 4, 2, and 5 genes in the KIV HIP, respectively (Table 3a). AdRa1d, GRIA3, Gabra5, and Gabbr2 in the FC and 5HTT and CHRM5 in the HIP were the genes downregulated in KIV mice without EET compared to WT (Table 3a). All of these genes showed persisting upregulation by EET, only when EET was provided at ED (Table 3a). EET also affected other genes not downregulated in KIV mice, for which the effects were observed across ages including at YA and OA (Table 3a).

Table 3. Persisting EET effects.

a. Genes showing significant upregulation (↑) or down regulation (↓) by EET one month after EET discontinuation (EET-SCT vs. SCT-SCT). b. Genes showing significant age effects in persisting EET effects. Genes showing both age and EET-SCT effects are listed in blue. c. Genes showing significant genotype differences in persisting EET effects (WT EET-SCT vs. KIV EET-SCT). All genes, except DRD3, showed a larger persisting EET effects in KIV mice than in WT mice. Genes downregulated in KIV mice without EET (vs. WT, see Table 1) is shown in bold font. Significance by REGWQ test. N=6–8 mice/group.

	a. EET effects		b. Age effects		c. Genotype effects
	WT	KIV	WT	KIV	KIV >WT	KIV < WT
Frontal Cortex	Gabrb3(ED) ↓	AdRa1d(ED) ↑	CALB(YA<OA)	AdRa1d (ED>YA/OA)	5HTR1b(ED)
		GRIA3(ED) ↑		CHRM5(YA<OA)	5HTR4(ED)
		Gabra5(ED) ↑		Gabra5(ED>OA)	DRD1 (ED)
		Gabre(OA) ↑		Gabre(ED/YA<OA)	CHRM5 (OA)
		Gabrr1 (YA) ↑		Gabrr2(ED/YA<OA)	CHRNA5(OA)
		Gabrr2(OA) ↑		Gabbr2(ED>YA/OA)	GRIA2(ED)
		Gabbr2(ED) ↑		SMST (ED<YA/OA)	GRIA3(ED)
				POMC(YA>OA)	Gabra5(ED)
				GFAP(ED>OA)	Gabre(OA)
					Gabbr2(ED)
					CALB(YA)
					HDAC5(ED, OA)
Hippocampus	DRD3 (YA) ↑	5HTT(ED) ↑	DRD3 (ED<YA>OA)	5HTT(ED>OA)	5HTT(ED)	DRD3 (YA)
	AdRb1(OA) ↑	DRD5 (OA) ↑	AdRb1 (ED/YA<OA)	DRD5(ED/YA<OA)	DRD5 (OA)
		AdRa2a(OA) ↑	Gabre (ED<YA>OA)	AdRa2a (ED/YA<OA)	DAT (ED)
		CHRM5(ED) ↑		CHRM5 (ED>YA)	AdRa2a (OA)
		Gabra2(YA) ↑		Gabra2 (ED<YA/OA)	CHRM5(ED)
		Gabrr1 (YA) ↑		Gabrb2 (ED<OA)	Gabra2(YA, OA)
		GAD1(OA) ↑		Gabrr1 (ED<YA>OA)	GAD1(OA)
		CALB(OA) ↑		GAD1 (ED/YA<OA)	CALB(OA)
		PDYN(ED) ↓		CALB(ED/YA<OA)	POMC(YA)
		TrkB(OA) ↑		POMC (ED<YA>OA)	CRF(ED)
		CRF(ED) ↑		TrkB(YA<OA)
				CRF(ED>YA/OA)

Many of the genes showing persisting EET effects also displayed significant age effects in KIV mice (FC: 5 of 8 genes; HIP: 8 of 12 genes, shown in blue in Table 3b), indicating that persisting EET effects were age-dependent. Effects are larger at ED or OA, depending on the gene (Table 3b, Supplementary Fig. 5 and 6).

Significant genotype differences in persisting EET effects (KIV EET-SCT vs. WT EET-SCT) were observed across ages: at ED, YA, and OA in 8, 1, and 4 genes in the FC and 4, 4, and 4 genes in the HIP, respectively (Table 3c). All of these genes, except DRD3 in the HIP, showed larger induction by EET in KIV mice than in WT mice (Table 3c, Supplementary Fig. 5 and 6). Many of these genes in the FC (8 of 12 genes) but a few genes in the HIP (2 of 12 genes) were the genes downregulated in KIV mice without EET (vs. WT mice, shown in bold, Table 3c).

4. Gene co-expression network

We analyzed how these neurotransmitter genes showed expression correlations. Gene co-expression network analyses revealed higher network density in the FC than the HIP, regardless of genotype and treatment (Fig. 3). EET generally increased the network density except in the KIV FC (Fig. 3), indicating that expression of many neurotransmitter genes became co-regulated under EET. Age effects were further analyzed by separating the data for each age group. KIV mice, compared to WT mice, showed reduced network density specifically at ED in both FC and HIP (Fig. 4). The gene network density increased with age (at YA and OA) for both regions in KIV mice (Fig. 4). The results indicated that deficiency of activity-dependent BDNF expression disturbed co-expression regulation of neurotransmitter genes specifically at ED. Interestingly, the network density was increased by EET in both regions in KIV mice at ED (Fig. 4). Larger network figures showing details of gene names and connections are presented in Supplementary Fig. 7.

Fig. 3. Gene co-expression network with all age groups.

Note that the FC generally showed higher network density than the HIP and that EET generally increased the network density except in the KIV FC. N=13–16 mice/group.

Fig. 4. Gene co-expression network across ages.

Note that KIV mice, compared to WT mice, showed reduced network density(*) at ED in both FC and HIP (decreased frequency of higher (>5) node degrees, indicated by arrows). See larger network figures showing details of gene names and connections in Supplementary Fig. 8. N=6–8 mice/group.

Genes having a high node degree were identified as hub genes. GAT1, Gabrg3, CHRNA7 and GRIN1 genes were identified in all groups, while other genes (e.g., AdRa1d, Gabrd, and Gabbr2) were identified depending on genotype, age, and treatment (Supplementary Table 5). Interestingly, in both genotypes, EET increased the node degree of many hub genes in both regions for all ages, except in the HIP at YA (Supplementary Table 5).

Discussion

Results of the present study led to six major findings. First, pIV-BDNF deficiency reduced expression of a larger number of neurotransmitter genes in the FC than in the HIP (27 vs. 9 genes) and at ED than at YA or OA (23 vs. 13 or 14 genes) in the FC. Second, EET affected expression of a larger number of neurotransmitter genes in pIV-BDNF deficiency than in a normal condition for both regions (KIV vs. WT, FC: 22 vs. 0; HIP: 15 vs. 3 genes). Third, the number of these EET-induced genes in BDNF deficiency was larger at early-life development than at adulthood (ED: 21, YA: 0, OA: 1 genes) in the FC, but at both ED and OA (ED: 10, YA: 1, OA: 8 genes) in the HIP. Fourth, many genes in the FC (15 of 22) but a few genes in the HIP (2 of 15) that EET induced were the genes downregulated in KIV mice without EET (vs. WT mice). Fifth, EET effects were observed one month after EET discontinuation in several genes, particularly in KIV mice (KIV vs. WT, FC: 7 vs. 1 genes; HIP: 11 vs. 2 genes). Sixth, pIV-BDNF deficiency reduced the density of the neurotransmitter gene co-expression network, particularly at ED in both FC and HIP, while EET increased the network density.

To our knowledge, this is the first systematic study showing the age-dependent effects of pIV-BDNF deficiency and EET on neurotransmitter gene expression. Previous studies, including ours, have shown the effects in one life stage, mostly at YA (Sakata & Duke 2014; Sakata & Overacre 2017; Chourbaji et al. 2004; Rios et al. 2006; Djalali et al. 2005; Tripp et al. 2012; Rampon et al. 2000; Lee et al. 2013; Bredy et al. 2004; Komitova et al. 2013; Zhu et al. 2014). Our most striking findings are that ED is the period when pIV-BDNF deficiency affects, mostly reduces, expression of neurotransmitter genes particularly in the FC and when EET most effectively induces expression of these genes. It is noteworthy that the EET gene effects are much larger in BDNF deficiency than in a normal condition, likely compensating for the expression defects, rather than enhancing the normal gene expression. These results support our hypothesis and suggest that pIV-BDNF deficiency, which occurs under chronic stress, disturbs neurotransmitter systems prominently during ED in the FC, and that EET during this stage is critical for normalizing the defects.

1. Do expression levels of neurotransmitter genes correspond to BDNF levels?

We have previously shown that pIV-BDNF deficiency reduces total BDNF (mRNA and protein) levels regardless of age in both FC and HIP (Dong et al. 2018; Jha et al. 2016). Thus, the prominent disturbance in neurotransmitter gene expression in the FC, particularly at ED, in BDNF deficiency likely involves other age-specific mechanisms, rather than just correlation with BDNF levels. For example, it is possible that BDNF is a predominant factor that regulates neurotransmitter gene expression at ED, but other neurotrophic factors may also contribute to the gene regulation later in adulthood. Even though the expression levels are relatively normal at adulthood in BDNF deficiency, the largely downregulated neurotransmitter expression in the FC at ED may disrupt normal development of functional neuronal network, and this may affect behavior later in life.

On the other hand, EET effects on neurotransmitter gene expression likely correspond with EET-induced BDNF levels. This study showed that EET neurotransmitter effects were largest at ED specifically in the FC of KIV mice, but were more universal across ages in the HIP of both genotypes. These results are in accordance with our previous finding that EET effects on BDNF mRNA induction were largest and limited at ED in the FC of KIV mice, but more universal across ages in the HIP for both WT and KIV mice (Dong et al. 2018). Our results suggest that the critical time window for transcriptional regulation may differ depending on the brain regions; it is limited at ED in the FC but more permissive across ages in the HIP. Importantly, many of the genes induced by EET in the KIV FC were the genes downregulated in KIV mice without EET (vs. WT mice). Thus, the neurotransmitter gene regulation by EET in the FC is likely under influence of BDNF and limited at ED. In the HIP, EET also induced many other genes (e.g., TH, Fig. 2) that were not downregulated in KIV mice with SCT (vs. WT mice). EET may upregulate other factors [e.g., VGF (Hunsberger et al. 2007)], as well as BDNF to affect expression of neurotransmitter genes in the HIP.

The persisting EET effects on neurotransmitter gene expression were also larger in KIV mice than in WT mice and larger at ED than at YA or OA for both regions. These results also match the persisting BDNF mRNA induction by EET only in KIV mice when EET is given at ED (Dong et al. 2018). The prolonged inductions of AdR1d, GRIA3, Gabra5 and Gabbr2 in the FC and 5HTT and CHRM5 in the HIP likely depend on age and BDNF, since they were downregulated in KIV mice (vs. WT mice) in the standard condition, but showed increased expression even after EET discontinuation, only when EET was given at ED in KIV mice.

2. How do changes in neurotransmitter gene expression relate to behavior?

We have previously shown that pIV-BDNF deficiency commonly causes depression-like behavior, regardless of age (Sakata et al. 2010; Jha et al. 2011; Jha et al. 2016; Dong et al. 2018) but EET can normalize the behavior with persisting effects, particularly when EET is provided at ED (Jha et al. 2016; Dong et al. 2018). This study revealed that pIV-BDNF deficiency disturbed neurotransmitter gene expression prominently at ED in both FC and HIP, while EET largely changed gene expression at ED, specifically in BDNF deficiency. The results suggest that the disturbed neurotransmitter gene expression at ED may cause abnormal behavior across ages, while EET at ED may most effectively reverse the behavioral abnormality by regulating neurotransmitter gene expression. The largest EET effects on gene expression at ED may be because ED is a period that allows more flexible and dynamic epigenetic changes (Lister et al. 2013).

This study revealed specific genes affected by BDNF deficiency and EET across ages. The causation of whether the expression changes indeed lead to behavioral changes remain to be elucidated in the future. Once verified, these specific genes at the affected life stage will be the targets for developing drug treatment for abnormal behaviors caused by BDNF deficiency. Such treatments may help augment EET effects and help people for whom EET is not applicable because their environmental resources and physical ability are limited.

BDNF deficiency and disturbed neurotransmitter genes have been observed in the postmortem prefrontal cortex and HIP of psychiatric patients (Weickert et al. 2003; Hashimoto et al. 2005; Wong et al. 2010; Dwivedi et al. 2003; Keller et al. 2010; Hing et al. 2012; Tripp et al. 2012; Kim et al. 2012; Phillips et al. 1991; Murray et al. 1994; Garzon et al. 2002). Currently, drug treatments for mental disorders empirically target neurotransmitter systems (e.g., antidepressants targeting 5HTT, antipsychotics targeting dopamine receptors, anxiolytics targeting GABA receptors, ketamine/memantine targeting glutamate receptors (Stahl 2013; Sakata 2014; Sakata 2011; Johnson et al. 2015). Here we showed that BDNF deficiency affected expression of neurotransmitter genes age-dependently. Our results suggest that this age-dependency should be considered when using drug treatments targeting the neurotransmitter systems; while pIV-BDNF deficiency occurs under chronic stress (Tsankova et al. 2006; Roth et al. 2009), the drug response likely differs across ages in pIV-BDNF deficiency.

3. Expression of specific neurotransmitter genes across ages

Some genes showing BDNF/EET effects in this study were the same as previously reported. For example, this study reproduced expression changes of 5HTR1b, 5HTR2a, 5HTR5b, TPH, DRD4, AdRa1d, Tac1, SMST, CHRNA5, and CHRM2 in BDNF deficiency reported during adulthood (Sakata & Duke 2014; Sakata & Overacre 2017; Tripp et al. 2012). This study also reproduced increased or normalized expression of GRIN1 and GRIN2A/B by EET (Guilarte et al. 2003; Bredy et al. 2004). Our results confirmed BDNF/EET effects on specific neurotransmitter genes and expanded the information of when during lifetime the changes occur. Some genes showed age-dependent reductions by pIV-BDNF deficiency and age-dependent induction by EET. For example, 5HTT, TH, and AdR1d, GRIN1 and GRIN2A/2B in the FC showed reductions by pIV-BDNF deficiency and induction by EET only at ED. This age-dependency may explain some inconsistency in previous studies that did or did not detect the gene expression changes by BDNF deficiency and EET.

Some genes were newly identified in this study, particularly at ED for effects of BDNF deficiency and EET. Conversely, some genes showing the BDNF/EET effects in previous studies were not detected in this study (e.g., DRD2/AdRa1a reductions in the KIV FC, ChaT/CHT inductions by EET in the KIV FC) (Sakata & Duke 2014; Sakata & Overacre 2017). One reason for this discrepancy might be due to the differences in statistical methods used. This study used REGWQ test, which includes family wise corrections and thus is more conservative when used with a number of comparison groups (15 comparisons for 2 treatments across 3 ages). Indeed, significant effects were detected for more genes when simple two-group comparisons were done using Student t-tests (Supplementary Table 6, e.g., significant genotype effects with SCT for 52 (FC) and 25 (HIP) genes detected by Student-t test, vs. 32 (FC) and 12 (HIP) genes detected by REGWQ test). Although the gene numbers differed by the statistical methods, the main findings (e.g., larger genotype and EET effects at ED in the FC) were mostly the same.

4. Gene co-expression network

Our results showed reduced density of gene co-expression network in pIV-BDNF deficiency, particularly at ED in both FC and HIP. The reduced gene-co-expression may be due to reductions in neuronal connectivity. For example, neuron A (e.g., glutamate neuron) expresses gene “a” (e.g., glutamate receptors) through activity-induced BDNF. This maturates neuron A, which connects to neuron B (e.g., GABA neuron). Neuron A stimulates neuron B to induce gene “b” (e.g., GAT1) in neuron B. This maturates neuron B and increases its connections to neuron C, which induces gene “c” (e.g., GABA receptors, CHRANA7) in neuron C. The expression correlation among genes ‘a’, ‘b’, and ‘c’ increases when the connection among neurons A, B, and C becomes stronger. Without activity-induced pIV-BDNF, this connection may decrease, thus reducing the expression correlation in the gene network. EET may restore BDNF and induce other neurotrophic factors [e.g., VGF (Hunsberger et al. 2007)] to increase neuronal connection and thus the density of gene co-expression network. The increased gene expression may in turn maturate neurons to further increase neuronal functional connectivity.

Our results showing that the EET increased the gene network density agree with a similar finding reported by Mehta-Raghavan et al. for rats (Mehta-Raghavan et al. 2016). Our results expanded the EET effects across ages for both normal and pIV-BDNF deficient conditions. Our important finding is that ED was the critical period in which BDNF deficiency and EET significantly affected gene co-expression network. The hub genes are likely the key players regulating the gene-expression network, and thus contributing to neuronal network functions and behavior.

5. Limitations and future directions

This study has several limitations, which encourages future investigations. First, this exploratory study determined only mRNA levels of the neurotransmitter genes. Future studies can examine protein levels since mRNA levels do not always reflect protein levels. Once confirmed, studies can further examine the age-dependent functional changes caused by BDNF deficiency and EET (e.g., by using agonists/antagonists of the identified receptors and by measuring the amount of release and re-uptake of the neurotransmitters related to the enzymes and transporters). Second, this study focused on neurotransmitter genes, considering neuronal network effects, but future studies can also examine other genes (e.g., for immune/metabolic systems) affected by BDNF/EET. The BioMark system was useful for measuring gene expression for many (384) samples covering different genotype/age/treatment groups, but its limitation is the number of measurable genes (<96). RNA sequencing/microarray can measure more genes, if resources allow for processing many samples. Third, qPCR was conducted on tissue homogenates and thus did not provide information at specific cell levels. Single cell qPCR may further clarify the age-dependent BDNF/EET effects at cellular levels (e.g., neuron-to-neuron or neuro-to-glial interactions). Fourth, we focused on the FC and HIP related to pIV-BDNF, but other brain regions (amygdala, raphe, ventral tegmental area, etc.) that connect to the FC and HIP can be explored for any age-dependent effects. The extended gene network studies may better explain functional and behavioral changes caused by BDNF deficiency and the amelioration by EET. Fifth, this study used both males and females, but the number used for each group (N=4) was too small to reliably detect sex effects and interactions with age, BDNF and/or EET. Future studies using increased sample numbers may clarify the sex effects and interactions; when all groups were combined (N=96), significant sex effects in expression changes were detected in several genes (e.g., Gabra5 in the FC; GRIN2A/B in the HIP, see Supplementary Fig. 8). Sixth, EET, particularly at ED, activates almost all BDNF promoters to increase total BDNF levels (Dong et al. 2018). This can rescue the perturbed neurotransmitter gene expression due to promoter IV defect. EET can also directly induce neurotransmitter genes without the BDNF pathway by increasing their promoter activity. It is intriguing to study the related molecular event (e.g., individual promoter elements and common/unique transcriptional factors affected by EET). Finally, gene regulation by BDNF/EET across ages differed between the FC and HIP, where the FC showed a critical period at ED. The FC is the last region (among the cortex) to fully differentiate and mature (e.g., matures at ~25 years old in humans) (Benes et al. 1994; Sowell et al. 1999). The hippocampus is the region that retains the stem-cell like feature (neurogenesis) and neuronal plasticity that persist even longer through the adulthood, and these features are enhanced by EET (van Praag et al. 2000). The involved region-specific mechanisms that explain the age-dependent neurotransmitter gene regulation (e.g., transcriptional (co-)factors and epigenetic regulation) remain to be elucidated.

6. Conclusion

The present study demonstrated the age-dependency of BDNF deficiency and EET affecting expression of neurotransmitter genes, which differed by brain region. BDNF deficiency and EET largely affected neurotransmitter gene expression at ED, particularly in the FC, which highlights the importance of early-life treatment for BDNF deficiency. The EET effects were more specific to BDNF deficiency, largest at ED in the FC but more universal across ages in the HIP. The affected neurotransmitter genes are targets for clarifying functional/behavioral mechanisms underlying BDNF deficiency and EET. While pharmacological treatment targets the affected neurotransmitter molecules, the efficacy may differ across ages in pIV-BDNF deficiency under chronic stress.

Abbreviations

5HTT

5-hydroxytryptamine (serotonin) transporter

5HTR

5-hydroxytryptamine (serotonin) receptor

ADRA

adrenergic receptor alpha

ANOVA

analyses o variance

BDNF

brain-derived neurotrophic factor

ChAT

choline acetyltransferase

CHT

choline transporter

CHRNA

nicotinic acetylcholine receptor alpha

CHRM

muscarinic acetylcholine receptor

CRF

corticotropin-releasing factor

Ct

cycle threshold

ED

early-life development

EET

enriched environment treatment

FC

frontal cortex

GABA

gamma aminobutyric acid

GABBR

GABA B receptor

GABRA

GABA A receptor alpha subunit

GABRG

GABA A receptor gamma subunit

GAT

GABA transporter

GRIA

glutamate receptor

GRIN

glutamate ionotropic receptor NMDA type

HGPRT

hypoxanthine-guanine phosphoribosyl-transferase

HIP

hippocampus

KIV

knockin BDNF promoter IV

NIH

National Institute of Health

OA

old adulthood

p75

p75 neurotrophin receptor

pIV-BDNF

promoter IV-driven BDNF expression

PENK

proenkephalin

qRT-PCR

quantitative reverse transcription polymerase chain reaction

RRID

Research Resource Identifier (see scicrunch.org)

SCT

standard condition treatment

TBP

TATA-box binding protein

Tac1

tachykinin precursor 1

TH

tyrosine hydroxylase

TrkB

tyrosine receptor kinase B

VGF

VGF(non-acronymic)- nerve growth factor inducible

WT

wild-type

YA

young adulthood