1521-0103/360/1/84–94$25.00 http://dx.doi.org/10.1124/jpet.116.237305 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS J Pharmacol Exp Ther 360:84–94, January 2017 Copyright ª 2016 by The American Society for Pharmacology and Experimental Therapeutics

Ontogeny of Norepinephrine Transporter Expression and Antidepressant-Like Response to Desipramine in Wild-Type and Serotonin Transporter Mutant Mice

Nathan C. Mitchell, Melodi A. Bowman, Georgianna G. Gould, Wouter Koek, and Lynette C. Daws Departments of Cellular and Integrative Physiology (N.C.M., M.A.B., G.G.G., L.C.D.), Psychiatry (W.K.), and Pharmacology (W.K., L.C.D.), University of Texas Health Science Center, San Antonio, Texas Received August 16, 2016; accepted November 8, 2016 Downloaded from

ABSTRACT Depression is a major public health concern with symptoms [P21]), adolescent (P28), and adult (P90) wild-type (SERT1/1) that are often poorly controlled by treatment with common mice. To model carriers of low-expressing SERT gene vari- antidepressants. This problem is compounded in juveniles and ants, we used mice with reduced SERT expression (SERT1/2) adolescents, because therapeutic options are limited to selec- or lacking SERT (SERT2/2). The potency and maximal jpet.aspetjournals.org tive serotonin reuptake inhibitors (SSRIs). Moreover, therapeutic antidepressant-like effect of desipramine was greater in P21 benefits of SSRIs are often especially limited in certain subpop- mice than in P90 mice and was SERT genotype independent. ulations of depressed patients, including children and carriers NET expression decreased with age in the locus coeruleus and of low-expressing serotonin transporter (SERT) gene variants. increased with age in several terminal regions (e.g., the cornu Tricyclic antidepressants (TCAs) offer an alternative to SSRIs; ammonis CA1 and CA3 regions of the hippocampus). Binding however, how age and SERT expression influence antidepres- affinity of [3H]nisoxetine did not vary as a function of age or SERT sant response to TCAs is not understood. We investigated the genotype. These data show age-dependent shifts for desipra- relation between antidepressant-like response to the TCA de- mine to produce antidepressant-like effects that correlate with at ASPET Journals on September 26, 2021 sipramine using the tail suspension test and saturation binding NET expression in the locus coeruleus and suggest that drugs of [3H]nisoxetine to the norepinephrine transporter (NET), the with NET-blocking activity may be an effective alternative to primary target of desipramine, in juvenile (21 days postnatal SSRIs in juveniles.

Introduction SERT expression (Kessler et al., 2001; Serretti et al., 2007; Bujoreanu et al., 2011). Tricyclic antidepressants (TCAs) are Depression is a major public health concern with symptoms an alternative to SSRIs. However, in juveniles and adoles- that are often poorly controlled with commonly prescribed cents, TCAs are not approved by the Food and Drug Admin- antidepressants. This problem is compounded in juveniles and istration for the treatment of depression, and they are adolescents, because there are fewer pharmacological treat- prescribed infrequently because of serious side effects. Ami- ment options for these patients compared with adults (Bylund triptyline poisoning, a condition that is more dangerous in and Reed, 2007). The U.S. Food and Drug Administration children than in adults, is one example (Paksu et al., 2014). approved two antidepressant drugs, fluoxetine and escitalo- Given that studies on the therapeutic benefit experienced by pram, for use in treating pediatric depression. Both drugs are juveniles and adolescents and by carriers of low-expressing selective serotonin reuptake inhibitors (SSRIs), which prevent SERT gene variants from TCA treatment have reported mixed serotonin (5-HT) uptake via the serotonin transporter (SERT). results (Hazell et al., 1995; Rajewska-Rager et al., 2008; Perlis SSRIs can be effective treatments for adult patients suffering et al., 2010; Hazell and Mirzaie, 2013), there is a need to better from depression, but these drugs often fail to relieve all understand the age and SERT gene variant dependence of depressive symptoms (Kirsch et al., 2008). The therapeutic TCA efficacy as a first step toward developing improved benefit of SSRIs can be especially limited in children and in antidepressants for these populations. carriers of a common SERT gene variant that yields lower TCAs primarily act to block the norepinephrine transporter (NET), but some also block SERT. The ensuing increase in extracellular norepinephrine (NE) and 5-HT is thought This work was supported by the National Institutes of Health National to trigger therapeutic downstream events (Frazer, 1997). Institute of Mental Health [Grants MH106978, MH093320, and MH086708] and Congressionally Directed Medical Research Programs [Award AR110109]. It has been suggested that a developmental delay in the dx.doi.org/10.1124/jpet.116.237305. noradrenergic central nervous system may limit the

ABBREVIATIONS: ANOVA, analysis of variance; CA, cornu ammonis; DMI, desipramine; 5-HT, serotonin; FST, forced swim test; NE, norepinephrine; NET, norepinephrine transporter; P, postnatal day; SERT, serotonin transporter; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressant; TST, tail suspension test.

84 Ontogeny of NET Expression and Antidepressant-Like Response 85 antidepressant potential of NET-targeting TCAs in juveniles previously described (Bengel et al., 1998). Mice were aged P21 and adolescents compared with adults (Bylund and Reed, (juvenile), P28 (adolescent), and P90–P100 (adult) (Spear, 2000) for 2007). However, other reports go against this idea. For all experiments. Animals were housed in a temperature-controlled example, in rats, innervation of noradrenergic neurons into (24°C) vivarium maintained on a 12-hour/12-hour light/dark cycle   higher brain regions, such as the cortex, reach adult-like (lights on at 7:00 AM) in plastic cages (29 cm 18 cm 13 cm) containing rodent bedding (Sani-chips; Harlan Teklad, Madison, WI) morphology by postnatal day (P) 7 (Loizou, 1972; Lauder and with free access to food (irradiated rodent sterilizable diet; Harlan Bloom, 1974; Coyle, 1977; Thomas et al., 1995). Noradrenergic Teklad) and water. Weaning occurred at P28, after which mice were receptors reach adult levels by P14–P21, and norepinephrine housed with five of their same-sex peers. To avoid possible confounds content reaches adult levels by P14–P42 (Loizou and Salt, of treatment effects with litter effects, no more than one mouse from a 1970; Konkol et al., 1978; Morris et al., 1980; for review, see given litter was assigned to a particular treatment condition. All Murrin et al., 2007). In juvenile and adolescent rats, NET procedures were conducted in accordance with the National Institutes expression is reported to be greater or equivalent to that in of Health Guide for the Care and Use of Laboratory Animals (Institute adults in numerous brain regions (Moll et al., 2000; Sanders of Laboratory Animal Resources, Commission on Life Sciences, et al., 2005). These findings suggest that NET-selective TCAs, National Research Council, https://grants.nih.gov/grants/olaw/ such as desipramine (DMI), should produce antidepressant- Guide-for-the-Care-and-use-of-laboratory-animals.pdf) and were approved by the Institutional Animal Care and Use Committee of like effects in juvenile (P21) and adolescent (P28) rodents, the University of Texas Health Science Center at San Antonio. because the noradrenergic system is relatively established at these ages. Downloaded from TST However, there is a paucity of research investigating antidepressant-like effects of TCAs in juvenile and adolescent TST experiments were conducted as originally described by Steru rodents. In rats, the TCAs imipramine and DMI have been et al. (1985) with minor modifications (described in Mitchell et al., reported to be less effective in producing antidepressant-like 2015). In brief, mice were moved from a housing facility to a testing room and then given a 1- to 2-hour acclimation period prior to TST. effects in the forced swim test (FST) in juveniles than in adults Experiments were conducted in the afternoon, between 12:00 PM and jpet.aspetjournals.org (Reed et al., 2008). In contrast, we found that 32 mg/kg DMI 5:00 PM. Mice received an intraperitoneal injection of saline 1 hour produced equivalent antidepressant-like effects in mice aged prior to testing, followed by an intraperitoneal injection of either DMI P21, P28, and P64–P90 (young adult) in the tail suspension (3.2, 10, or 32 mg/kg) or saline vehicle (control condition) 30 minutes test (TST) (Mitchell et al., 2013). These behavioral results prior to testing. This two-injection protocol was used to be consistent were paralleled by [3H]nisoxetine binding in whole hippocam- with previous studies from this laboratory (Baganz et al., 2008; Horton pal homogenates, which revealed no difference in NET et al., 2013) and with future studies that will examine the effects of expression among these ages. Although these studies are drug combinations on immobility time in the TST. Immediately before informative, they provide no information on how age may testing, the distal portion of the tail was secured to a flat aluminum at ASPET Journals on September 26, 2021   affect the potency or maximal effect of DMI to produce bar (2 0.3 10 cm) using adhesive tape. A hole on the opposite end of the bar was secured to a hook in a visually isolated white box (40  antidepressant-like effects in the TST. [3H]Nisoxetine binding 40  40 cm). Each mouse was suspended for 6 minutes while a digital assays in hippocampal homogenates also lack the ability to video camera recorded its ventral surface. Immobility was defined as discriminate potential differences in NET expression among the absence of active movements and included passive swaying. hippocampal subregions as a function of age and do not Immobility time was scored in seconds by observers blind to the provide information about other brain regions that may be randomly assigned treatment conditions. Mice were tested only once. important in mediating antidepressant-like effects. Determining the potency and maximal effect of DMI to Drugs produce antidepressant-like effects in the TST, as well as DMI (desmethylimipramine) hydrochloride (Sigma-Aldrich, St. establishing the ontogeny of NET expression throughout early Louis, MO) was dissolved in physiologic saline and injected intraperi- postnatal development, could help to identify age-dependent toneally at doses expressed as salt weight per kilogram body weight. The mechanisms that may limit the therapeutic benefit of antide- injection volume was 10 ml/kg. pressants. Here, we examined the dose-response relationship for DMI to produce antidepressant-like effects in P21, P28, Autoradiography and P90 SERT-deficient mice, and we quantified NET expres- NET density in the mouse brain was assessed by quantitative sion in a number of brain regions using autoradiography with 3 3 autoradiography using the NET-selective ligand [ H]nisoxetine and [ H]nisoxetine, a NET-selective ligand. SERT-deficient mice methods adapted from Tejani-Butt (1992). Mice were euthanized by were included in this study to assess the effects of constitutive decapitation and brains were flash frozen on powdered dry ice before reduction in SERT expression, which occurs in humans carrying being stored at 280°C. Before sectioning, brains were brought to 220°C low-expressing gene variants of SERT, on antidepressant-like in a cryostat (Leica CM 1850; Meyer Instruments, Houston, TX). effects of DMI and on NET expression and affinity. Coronal sections (20 mm) were collected at the level of plate 12 (pre- frontal cortex), plate 47 (hippocampal subregions including the cornu ammonis regions CA1, CA2, and CA3 and the dentate gyrus), the Materials and Methods amygdala, plate 64 (raphe nuclei), and plate 76 (locus coeruleus) according to the Paxinos and Franklin (1997) mouse brain atlas. Animals Sections were thaw mounted onto gelatin-coated microscope slides. Naïve male and female SERT wild-type (SERT1/1), heterozygote Tissue-mounted slides were vacuum desiccated for 18–24 hours at 4°C (SERT1/2), or homozygote knockout (SERT2/2) mice (backcrossed to before the slides were stored at 280°C. Brain tissue was stored at 280°C for C57BL/6J for .10 generations) were used for all experiments. Dr. 2–4 weeks, and every experiment contained brain tissue from all groups, Dennis Murphy (National Institute of Mental Health, Bethesda, MD) including P21, P28, and P90 SERT1/1, SERT1/2,andSERT2/2 male provided mice to found the colony. Animals were bred by crossing male and female mice. Before incubation, sections on slides were thawed and female SERT1/2 mice, and SERT genotypes were identified as for 1 hour in a vacuum desiccator at 4°C to remove excess moisture 86 Mitchell et al. and maximize brain tissue adherence to the slides. The slides were Quantitative Autoradiography. [3H]Nisoxetine binding densi- then preincubated for 1 hour in a wash buffer solution (43 mM Tris- ties were measured from autoradiograms and analyzed using methods HCl, 124 mM NaCl, and 4.3 mM KCl, pH 7.4) at room temperature described in Mitchell et al. (2016). In brief, nonspecific binding was (approximately 24°C) to remove endogenous ligands bound to NET. fit to an unweighted linear regression and subtracted from total Incubation was carried out in slide mailers (VWR International, [3H]nisoxetine to give specific binding. Unweighted nonlinear re- Radnor, PA) filled with 10 ml ice-cold reaction buffer (50 mM Tris- gression was used to analyze [3H]nisoxetine-specific binding data. HCl, 300 mM NaCl, and 5 mM KCl, pH 7.4) containing [3H]nisoxetine Saturation binding isotherms were fitted according to the following at concentrations of 0.3, 1, 3, 10, or 30 nM for 4 hours. Each one-site model to calculate maximal specific [3H]nisoxetine binding 3 concentration of [ H]nisoxetine was incubated with three brain (Bmax) and affinity (Kd) values: Y 5 Bmax  X/(Kd 1 X) (Figs. 2 and 3). sections per animal per brain region. Nonspecific binding was Bmax and Kd values were analyzed with a two-factor (age, genotype) defined by 2.5 mM mazindol (Pfizer, Groton, CT) and was approxi- ANOVA (Fig. 4; Table 2). There were no statistically significant mately 9%–50% total binding in low binding regions (i.e., CA1) and effects of sex (P $ 0.09) or sex and age interactions (P $ 0.17) for 3 approximately 9%–22% total binding in high binding regions (i.e., [ H]nisoxetine Bmax or Kd values within genotype- and brain locus coeruleus). The incubation was terminated by three 5-minute region–matched groups; thus, data from both sexes were combined. washes in wash buffer solution at 4°C, followed by a 5-second dip in Two-factor ANOVA (age, SERT genotype) (Fig. 4) followed by the deionized water at 4°C. Slides were dried on a slide warmer for Dunnett post hoc test for multiple comparisons was used to analyze 3 20 minutes. [ H]Nisoxetine-labeled sections were exposed to Care- mean Bmax and Kd values. All data are expressed as means 6 S.E.M. stream Biomax MR film for 6 weeks, along with tritium standards P , 0.05 was considered statistically significant.

(American Radiolabeled Chemicals, St. Louis, MO). Films were Correlations. As shown in Fig. 5, Pearson correlation was used to Downloaded from developed in a film processor (AFP Imaging, Elmsford, NY). A digital examine the relation between the Emax for DMI to reduce immobility imaging system that included a 12-bit digital camera (CFW-1612M; time in the TST and maximal specific [3H]nisoxetine binding, as a Scion Corp., Frederick, MD), Nikon Lens, Northern Lights Illumi- function of age. Within each brain region, SERT genotype did not nator, and Kaiser RS1 copy stand (all from InterFocus Imaging Ltd., significantly influence the relation between the Emax for DMI and Linton, UK) was used to capture autoradiogram images. Autoradio- maximal specific [3H]nisoxetine binding (P $ 0.15; data not shown) so grams were calibrated and measured using ImageJ public access data for SERT1/1, SERT1/2, and SERT2/2 were best fit with a

shareware (National Institutes of Health, Bethesda, MD; https:// single line (Fig. 5). All data are expressed as means 6 S.E.M. jpet.aspetjournals.org imagej.nih.gov/ij/download.html) on a MacBook computer (OS 10; Apple, Cupertino, CA). Additional brain sections were stained with thionine (FD NeuroTechologies, Inc., Columbia, MD) to verify tissue Results integrity and neuroanatomical brain regions quantified (Figs. 2B The Potency of DMI to Reduce Immobility Time in the TST and 3B). Depends on Age and SERT Genotype, Whereas Its Maximal Effect Depends on Age Only

Data Analysis Clinical studies have shown that therapeutic potential of at ASPET Journals on September 26, 2021 Statistical analyses were performed using GraphPad Prism (ver- SSRIs is limited in children (Kirsch et al., 2008). Similarly, we sion 6.0; GraphPad Inc., La Jolla, CA) and NCSS 2007 (NCSS, previously found that the antidepressant-like response to Kaysville, UT) software. SSRIs in juvenile mice is limited compared with adult mice TST. Dose-response curves were generated by administering 3.2, (Mitchell et al., 2013, 2016). As an alternative to SSRIs, we 10, or 32 mg/kg DMI or saline vehicle in juvenile, adolescent, or adult began our study by evaluating the antidepressant-like re- mice. Under vehicle conditions, we previously found that time spent sponse to DMI, a potent NET blocker, in SERT1/1 P21, P28, immobile varies by age and SERT genotype (Mitchell et al., 2013, 2015, 2016). Because such differences were also apparent in our study and P90 mice. Using the TST, we found that DMI reduced 5 , (see vehicle data in Fig. 1, A–C), we expressed data as total immobility immobility time in all age groups [F(3,276) 35.06, P 0.01], in seconds (Fig. 1, A–C) and then replotted these data as the and younger mice spent overall less time immobile than adult percentage from the saline vehicle control (Fig. 1, D–F). Within mice [F(2,276) 5 19.73, P , 0.01] (Fig. 1A). An interaction genotype, TST data were analyzed by a two-factor analysis of variance between age and DMI dose showed that younger mice are (ANOVA) (age, drug dose) followed by Dunnett and Tukey multiple more sensitive to the anti-immobility effects of DMI than comparisons tests. Sample sizes for TST data included the following adults [F(6,276) 5 2.34, P 5 0.03]. Post hoc analysis revealed – 1 1 – 1 1 per data point: 9 14 SERT / males and 10 20 SERT / females, that immobility time was significantly lower in P21 mice than – 1 2 – 1 2 – 8 10 SERT / males and 16 20 SERT / females, and 8 10 in P28 and P90 mice after administration of 10 mg/kg DMI SERT2/2 males and 9–14 SERT2/2 females. A multifactor ANOVA (P , 0.01), and immobility time was lower in P21 mice than in showed neither a main effect of sex nor any interaction of sex with , other factors (P 5 0.51 and P $ 0.10, respectively), with the exception P90 mice after administration of 3.2 and 32 mg/kg DMI (P 5 of a sex  genotype interaction (P 5 0.008); however, multiple 0.01 and P 0.012, respectively). Likewise, immobility time at comparisons failed to show significant sex differences for each of the 3.2 mg/kg DMI was significantly lower in P28 mice than in P90 genotypes. Thus, data for males and females were combined. Maximal mice (P , 0.01). Furthermore, 3.2 mg/kg DMI was the lowest effects (Emax) and half-maximally effective dose (ED50) values were effective dose in P21 and P28 mice, and 10 mg/kg was the – derived from data in Fig. 1, D F, and are summarized in Table 1. Emax lowest effective dose in P90 mice (P , 0.01). Taken together, was defined as the greatest observed percent change of immobility the anti-immobility effects of DMI appear to be greater in from the saline control. Emax values were analyzed by two-factor younger mice than adults. ANOVA (age, SERT genotype) followed by the Tukey multiple Clinical studies have also found that the antidepressant comparisons test (Table 1). ED values were calculated using 50 effects of SSRIs are limited in individuals with a SERT gene methods detailed in Koek et al. (2009). Briefly, the linear portion of the dose-response curves was analyzed by log-linear regression polymorphism that yields a reduction in SERT expression/ 1 2 [effect 5 slope  log(dose) 1 intercept] of data from individual function (Serretti et al., 2007). We used SERT / mice, which 1 1 subjects. All data are expressed as means 6 S.E.M. except for ED50 express 50% less SERT than SERT / mice, to evaluate the values, which are expressed as the mean. P , 0.05 was considered antidepressant-like potential of NET blockade in individu- statistically significant. als with reduced SERT expression/function. DMI reduced Ontogeny of NET Expression and Antidepressant-Like Response 87 Downloaded from jpet.aspetjournals.org at ASPET Journals on September 26, 2021

Fig. 1. Influence of age and SERT genotype on the antidepressant-like effect of DMI. (A–C) Dose-dependent reductions in immobility time in the TST in P21, P28, and P90 SERT+/+ mice (A), SERT+/2 mice (B), and SERT2/2 mice (C). (D) Data from (A) expressed as a percentage of vehicle control. (E) Data from (B) expressed as a percentage of vehicle control. (F) Data from (C) expressed as a percentage of vehicle control. Data obtained in males and females are pooled, because a multifactor ANOVA (sex, genotype, DMI) showed no main effect or interaction of sex with other factors (P . 0.05), with one exception between sex and genotype (P , 0.01); however, multiple comparisons failed to show significant sex differences for each genotype (P . 0.05). Data are means 6 S.E.M. Filled symbols represent data points that are significantly different from the SERT genotype– and age-matched vehicle control as determined by the Dunnett post hoc multiple comparisons test after a two-factor ANOVA. *P , 0.05 (significant difference from SERT genotype–matched P90); **P , 0.01 (significant difference from SERT genotype–matched P90); #P , 0.05 (significant difference from SERT genotype–matched P28 with the Tukey post hoc multiple comparisons test after a two-factor ANOVA); ##P , 0.01 (significant difference from SERT genotype–matched P28 with the Tukey post hoc multiple comparisons test after a two-factor ANOVA). Sample sizes per data point were as follows: SERT+/+, n =21–31 (9–14 males and 10–20 females, pooled); SERT+/2, n =16–20 (8–10 males and 8–12 females, pooled); and SERT2/2, n =18–23 (8–10 males and 9–14 females, pooled).

immobility time in the TST [F(3,212) 5 16.58, P , 0.01] (Fig. as seen with P28 SERT1/2 mice may limit the interpretation 1B), and younger mice spent overall less time immobile in the of our findings. TST than adult mice [F(2,212) 5 9.34, P , 0.01]. In contrast Although DMI is a potent NET blocker, it does have affinity with SERT1/1 mice, no interaction was found between age for SERT. The antidepressant-like effect of DMI in SERT2/2 and DMI drug dose in SERT1/2 mice [F(6,212) 5 0.81, P 5 mice was evaluated to investigate the proportion of the 0.57]. Post hoc analysis showed that immobility time in the antidepressant-like response that was SERT dependent in TST was lower in vehicle treated P28 mice than vehicle P21, P28, and P90 mice. Immobility time in the TST was treated P90 mice (P , 0.01). The lowest effective dose for reduced after DMI administration [F(3,217) 5 9.00, P , 0.01], P21 and P90 mice was 10 mg/kg DMI (P , 0.01). Although no although no dose of DMI significantly lowered immobility time dose of DMI significantly reduced immobility time in P28 in P28 mice (P 5 0.13). A main effect of age was also found mice, immobility times after 32 mg/kg DMI trended to be lower [F(2,217) 5 6.14, P , 0.01] (Fig. 1C). No interaction between than immobility times after vehicle treatment (P 5 0.087). age and drug dose was found [F(6,217) 5 0.29, P 5 0.94]. These data suggest that the antidepressant-like effects of DMI Immobility time was lower in P21 and P90 mice than in P28 are not greater in younger SERT1/2 mice than in adult mice; mice after administration of 10 mg/kg DMI (P 5 0.015, P , however, variation in basal immobility (saline vehicle control) 0.01). The lowest effective dose for P21 and P90 mice was 88 Mitchell et al.

TABLE 1 possible exception of P28 SERT2/2 mice, in which (as for Influence of age and SERT genotype on the potency (ED50) and maximal SERT1/2 mice) no dose of DMI significantly reduced immo- effect (Emax) of DMI to reduce immobility time in the TST bility time (Fig. 1C). Data are expressed as means 6 S.E.M. Values were calculated from data shown in Fig. 1, D–F, and Table 1. It should be noted that basal immobility time in the TST was lower in SERT2/2 mice (Fig. 1C) compared with SERT1/1 Age SERT ED E 50 max mice (Fig. 1A), and P28 SERT1/2 mice showed less basal mg/kg %D Control immobility compared with P90 SERT1/2 mice (Fig. 1B). To , P21 +/+ 7.5* ** 70.1 6 5.4* account for age- and SERT genotype–related variations in +/2 17.4** 68.5 6 7.0 2/2 17.1** 62.4 6 8.9** basal immobility time, data were normalized to a percentage P28 +/+ 23.7* 57.5 6 10.6 of vehicle control (Fig. 1, D–F). The analysis of the normalized +/2 36.3 41.3 6 8.6 SERT1/1 data was similar to the analysis of raw SERT1/1 2 2 6 / 45.2*** 32.3 9.2 data (in seconds) with minor exceptions. DMI reduced immo- P90 +/+ 37.3 47.0 6 6.1 +/2 29.1 47.3 6 7.7 bility time in all SERT1/1 age groups [F(3,276) 5 36.23, P , 2/2 30.5 51.9 6 9.2 0.01], and a main effect of mouse age was found [F(2,276) 5 , Sample sizes per data point were as follows: SERT+/+, n =21–31 (9–14 males and 10.36, P 0.01] (Fig. 1D). Younger mice showed greater 10–20 females, pooled); SERT+/2, n =16–20 (8–10 males and 8–12 females, pooled); sensitivity to the anti-immobility effects of DMI in the TST 2 2 – – – and SERT / , n =18 23 (8 10 males and 9 14 females, pooled). For ED50 values, than adult mice [F(6,276) 5 2.58, P 5 0.02]. Immobility time comparisons were made by the F ratio test to compare intercepts. For Emax values, Downloaded from comparisons made by the Tukey post hoc multiple comparisons test after a two-factor was significantly lower in P21 SERT1/1 mice than in P90 ANOVA. Because there were no statistically significant sex differences, data for male mice after 3.2 mg/kg (P , 0.01) and 10 mg/kg DMI (P 5 0.04). and female mice were pooled. *P , 0.05 (difference from SERT genotype–matched P90); **P , 0.05 (difference Immobility time was lower in SERT1/1 P28 mice than in P90 from SERT genotype–matched P28); ***P , 0.05 (difference from age-matched SERT+/+ mice after 3.2 mg/kg DMI (P , 0.01). The lowest effective dose group). for DMI was 3.2 mg/kg in SERT1/1 mice aged P21 and P28 (P , 0.01), and the lowest effective dose for P90 mice was 10 mg/kg DMI (P , 0.01 and P 5 0.04, respectively). These 10 mg/kg (P , 0.01). jpet.aspetjournals.org data suggest that NET blockade is primarily responsible for Analysis of normalized SERT1/2 data showed that DMI the antidepressant-like effects of DMI in the TST, with the reduced immobility time in the TST [F(3,212) 5 15.7, P , at ASPET Journals on September 26, 2021

Fig. 2. Specific [3H]nisoxetine binding to NET in hippocampal regions as a function of age and SERT genotype. Brain sections from P21, P28, and P90 SERT- deficient mice incubated with the NET-specific ligand [3H]nisoxetine. Nonspecific binding was defined by mazindol (2.5 mM). (A) Representative coronal sections at the level of plate 47 (Paxinos and Franklin, 1997) in SERT+/+, SERT+/2, and SERT2/2 mice aged P21, P28, or P90. The boxed area in (A) is enlarged in (B), which shows representative thionine-stained brain sections labeled with hippocampal regions quantified, which include the CA1, CA2, and CA3 regions and the dentate gyrus (DG). (C) Example of saturation binding isotherms used to calculate Bmax and Kd values. Curves include specific [3H]nisoxetine binding values for the CA3 of P21, P28, and P90 SERT+/+ mice. There was no main effect of sex on Bmax or Kd values, so male and female data are pooled (P . 0.05). Bmax values are summarized in Fig. 4. There were no significant differences in Kd among ages or between SERT+/+ and SERT+/2 mice. Sample sizes of mice per group were as follows: SERT+/+, n =5–9(4malesand3–5females,pooled);SERT+/2, n =6–10 (3–5 males and 2–4 females, pooled); and SERT2/2, n =4–7(2–4 males and 2–4 females, pooled). See Table 2 and Fig. 4 for a summary of data. Ontogeny of NET Expression and Antidepressant-Like Response 89

Fig. 3. Specific [3H]nisoxetine binding to NET in the locus coeruleus as a function of age and SERT genotype. Brain sections from P21, P28, and P90 mice were incubated with increasing concentrations of [3H]nisoxetine. Nonspecific binding was defined by mazindol (2.5mM). (A) Representative coronal sec- tions at the level of plate 76 (Paxinos and Franklin, 1997) in SERT+/+, SERT+/2, and SERT2/2 mice aged P21, P28, or P90. The boxed area in (A) is enlarged in (B), which shows a representative thionine-stained brain sections including the locus

coeruleus (LC). (C) Example of saturation binding Downloaded from isotherms used to calculate Bmax and Kd values. Curves include specific [3H]nisoxetine binding val- ues for the LC of P21, P28, and P90 SERT+/+ mice. Bmax values for this region are summarized in Fig. 4. There was no main effect of sex on Bmax or Kd values, so male and female data are pooled (P . 0.05). Sample sizes of mice per group were as follows: – – SERT+/+, n =5 9 (4 males and 3 5 females, pooled); jpet.aspetjournals.org SERT+/2, n =6–10 (3–5 males and 2–4 females, pooled); and SERT2/2, n =4–7(2–4 males and 2–4 females, pooled). See Table 2 and Fig. 4 for a summary of data. at ASPET Journals on September 26, 2021

0.01], with no main effect of age [F(2,212) 5 2.091, P 5 0.13] adults. ED50 values for DMI varied as a function of age in (Fig. 1E). DMI reduced immobility time in all age groups, and SERT1/1, SERT1/2, and SERT2/2 mice [F(2,205) 5 12.57, it did so equally across all age groups [F(6,212) 5 0.46, P 5 P , 0.01; F(2,163) 5 3.22, P 5 0.04; and F(2,149) 5 3.38, P 5 0.84]. No differences among age groups were found. The lowest 0.04, respectively]. In SERT1/1 mice, the ED50 value was effective dose for P21 and P90 SERT1/2 mice was 3.2 mg/kg lower in P21 mice than in P28 mice and P90 mice [F(1,125) 5 DMI (P , 0.01 and P 5 0.03, respectively). The lowest effective 6.75, P 5 0.02; and F(1,143) 5 25.0, P , 0.01], and P28 mice dose for P28 SERT1/2 mice was 32 mg/kg DMI (P 5 0.02). By had a lower ED50 value than P90 mice [F(1,143) 5 5.2, P 5 normalizing SERT1/2 data, we found that DMI reduced 0.02]. In SERT1/2 mice, ED50 values were lower in P21 mice immobility in all age groups, including P28 SERT1/2 mice. than in P28 mice [F(1,105) 5 5.7, P 5 0.02]. In SERT2/2 mice, Normalized data for SERT2/2 mice showed a main effect of ED50 values were lower in P21 mice than in P28 mice [F(1,96) 5 DMI [F(3,217) 5 9.01, P , 0.01] and age [F(2,217) 5 4.20, P 5 9.33, P , 0.01]. Among SERT genotype comparisons showed 0.02] (Fig. 1F). Like SERT1/2 mice, there was no interaction an effect of SERT genotype in P28 mice [F(2,167) 5 5.50, P , between age and drug dose [F(6,217) 5 0.27, P 5 0.95]. After 0.01]. P28 SERT1/1 mice showed a lower ED50 than P28 10 mg/kg DMI, immobility time was significantly lower in P21 SERT2/2 mice [F(1,116) 5 12.25, P , 0.01]. SERT genotype and P90 SERT2/2 mice than in P28 mice (P 5 0.02 and P 5 had no significant main effect on ED50 values in P21 or P90 0.03, respectively). The lowest effective dose for P21 and P90 mice [F(2,177) 5 1.29, P 5 0.28; and F(2,193) 5 0.82, P 5 0.44]. SERT2/2 mice was 10 mg/kg (P 5 0.02); in addition, no dose In P28 mice, an effect of SERT genotype was found [F(2,167) 5 significantly reduced immobility time in P28 mice, including 5.50, P , 0.01]. P28 SERT1/1 mice had a lower ED50 value 32 mg/kg DMI (P 5 0.17). than P28 SERT2/2 mice [F(1,116) 5 12.25, P , 0.01]. To investigate a possible interaction between postnatal age The Emax for DMI to reduce immobility time in the TST was and SERT genotype on the antidepressant-like response to significantly greater in young mice than in adults [F(2,178) 5 DMI, we compared the potency (ED50) and maximal effect 7.06, P , 0.01]. P21 SERT1/1 mice showed a significantly (Emax) of DMI at different ages and in different genotypes greater Emax than P90 SERT1/1 mice (P 5 0.046). In (Table 1). With the exception of P28 SERT1/2 and SERT2/2 addition, the Emax for DMI in P21 SERT2/2 mice was greater mice, ED50 values for DMI were lower in younger mice than in than P28 SERT2/2 mice (P 5 0.02). SERT genotype did not 90 Mitchell et al.

TABLE 2 3 Summary of Bmax values for specific [ H]nisoxetine binding in SERT+/+, SERT+/2, and SERT2/2 mice Data are expressed as means 6 S.E.M. in femtomoles per milligram protein.

Genotype P21 P28 P90 Prefrontal cortex SERT+/+ 228 6 23* 275 6 26 314 6 36 SERT+/2 229 6 23 251 6 21 290 6 30 SERT2/2 206 6 19 271 6 35 225 6 26 Amygdala SERT+/+ 193 6 27 137 6 14* 233 6 29 SERT+/2 150 6 29 165 6 14 191 6 24 SERT2/2 157 6 24 146 6 20 155 6 21 Dorsal raphe SERT+/+ 619 6 40 551 6 35 539 6 35 SERT+/2 560 6 57 562 6 31 578 6 27 SERT2/2 565 6 20 543 6 24 536 6 36

Sample sizes of mice per group were as follows: SERT+/+, n =5–9 mice per group (4 males and 3–5 females, pooled); SERT+/2, n =6–10 (3–5 males and 2–4 females, pooled); and SERT2/2, n =4–7(2–4 males and 2–4 females, pooled). Downloaded from *P , 0.05 (different from SERT genotype–matched P90 group; Dunnett multiple comparisons test after two-factor ANOVA).

of Figs. 2A and 3A. Figures 2C and 3C show summary data for specific binding in SERT1/1 mice in the CA3 region of the hippocampus and locus coeruleus, respectively. Bmax and Kd values were analyzed with a two-factor (age, genotype) jpet.aspetjournals.org

3 ANOVA (Fig. 4; Table 2) to determine how age and SERT Fig. 4. Bmax for [ H]nisoxetine binding to SERT in SERT+/+ and SERT+/2 mice aged P21, P28, and P90. Bmax values from P21, P28, and P90 SERT+/+, genotype influence NET binding densities in brain regions of SERT+/2, and SERT2/2 mice in the CA1 region (A), CA2 region (B), CA3 potential importance for antidepressant-like response. These region (C), dentate gyrus (D), and locus coeruleus (E) were determined from results are described below. one-site curve fits as described in the Materials and Methods.Dataare 3 B K means 6 S.E.M. *P , 0.05 (significant difference from SERT genoty- [ H]Nisoxetine max and d Values in Terminal 3 pe–matched P90, Dunnett post hoc multiple comparisons test after a two- Regions. In the hippocampus, specific [ H]nisoxetine Bmax factor ANOVA for age and SERT genotype). Data are means 6 S.E.M. values significantly increased with age in the CA1 and CA3 at ASPET Journals on September 26, 2021 pooled from male and female mice. Sample sizes of mice per group were as 5 5 5 , follows: SERT+/+, n =5–9(4malesand3–5 females, pooled); SERT+/2, n = regions [F(2,52) 4.53, P 0.02; and F(2,52) 5.40, P 0.01] 6–10 (3–5malesand2–4 females, pooled); and SERT2/2, n =4–7(2–4 but not in the CA2 region of the hippocampus or dentate gyrus males and 2–4 females, pooled). [F(2,52) 5 2.71, P 5 0.08; and F(2,52) 5 0.42, P 5 0.66] (Fig. 4, A–D). Bmax values in P90 SERT1/1 mice were greater than in P21 SERT1/1 mice in the CA1 and CA3 regions (P 5 0.02 and affect the E for DMI to reduce immobility time [F(2,178) 5 max P , 0.01, respectively) (Fig. 4, A and C). B values did not vary 1.043, P 5 0.35], and SERT genotype did not interact with the max as a function of SERT genotype in the CA1, CA2, and CA3 regions effects of age [F(4,178) 5 0.94, P 5 0.44]. of the hippocampus or dentate gyrus [F(4,52) 5 1.60, P 5 0.19; Collectively, the ED values for DMI to decrease immobil- 50 F(2,52) 5 0.21, P 5 0.80; F(2,52) 5 0.78, P 5 0.46; and F(2,52) 5 ity time in the TST were lowest in P21 mice, regardless of 0.87, P 5 0.42, respectively]. In addition, no interactions between SERT genotype. E values for DMI were greater in P21 mice max age and SERT genotype on B values were found in these re- than in P28 and P90 mice. The role of SERT genotype in max gions [F(2,52) 5 1.76, P 5 0.18; F(4,52) 5 0.57, P 5 0.69; F(4,52) 5 antidepressant-like response to DMI was most apparent in 0.68, P 5 0.61; and F(2,52) 5 0.42, P 5 0.66, respectively]. P28 mice, in which drug potency decreased with the reduction Although they were not statistically significant, B values and loss of SERT, although a similar trend was observed with max tended to increase with age in the prefrontal cortex and P21 mice as well (Table 1). amygdala [F(2,52) 5 3.04, P 5 0.056; and F(2,52) 5 2.32, P 5 0.11] (Table 2). Between age group comparisons showed NET Expression Increased with Age in Some Noradrenergic greater Bmax values for P90 SERT1/1 mice than P21 SERT1/1 Terminal Regions and Decreased with Age in the Locus mice in the prefrontal cortex (P 5 0.05), and Bmax values were Coeruleus greater in P90 SERT1/1 mice than P28 SERT1/1 mice in the In an effort to explain the age-dependent, antidepressant- amygdala (P , 0.01). As for the hippocampus, Bmax values in like response to DMI, NET binding densities were quantified the prefrontal cortex and amygdala did not vary as a function in limbic (CA1, CA2, CA3, dentate gyrus, prefrontal cortex, of SERT genotype [F(2,52) 5 1.30, P 5 0.28; and F(2,52) 5 and amygdala) and cell body (dorsal raphe and locus coeru- 1.81, P 5 0.18, respectively]. No interaction between age and leus) regions using quantitative autoradiography to measure SERT genotype on Bmax values was found in these regions specific [3H]nisoxetine binding. Figures 2A and 3A show [frontal cortex: F(4,52) 5 0.79, P 5 0.54; amygdala: F(4,52) 5 representative autoradiograms. Figures 2B and 3B are 1.30, P 5 0.28]. Within the dorsal raphe, Bmax values were not thionine-stained tissue sections to confirm tissue integrity dependent on either age or SERT genotype [F(2,52) 5 0.78, and to highlight brain regions quantified where significant P 5 0.46; and F(2,52) 5 0.36, P 5 0.70] (Table 2). No differences were found. These sections are a representative interaction between age and SERT genotype on Bmax values enlargement of the boxed areas shown in the top right panels was found [F(4,52) 5 0.58, P 5 0.68]. Ontogeny of NET Expression and Antidepressant-Like Response 91

Fig. 5. Relationship between Emax values for DMI to produce antidepressant-like effects in the TST and Bmax values for specific [3H]nisoxetine binding in the locus coeruleus as a function of age and SERT genotype. The CA1 region (A), CA2 region (B), CA3 region (C), dentate gyrus (D), and locus coeruleus (E) are shown. Relation- ship between Emax and Bmax did not vary by SERT genotype; thus, one line was used to fit data regardless Downloaded from of genotype. Data are taken from Fig. 1, D–F (per data point: SERT+/+, n =21–31; SERT+/2, n =16–20; and SERT2/2, n =18–23), and Fig. 4 (per age group: SERT+/+, n =5–9; SERT+/2, n =6–10; and SERT2/2, n =4–7). Data are means 6 S.E.M. (male and female data are pooled). jpet.aspetjournals.org at ASPET Journals on September 26, 2021

Kd values in all terminal regions ranged from 1.0 to 5.5 nM fitted with a single line (Pearson correlation: r 520.66, 20.74, and were not dependent on age (P $ 0.44) or SERT genotype 20.65, and 20.26 for the CA1, CA2, and CA3 regions and the (P $ 0.5). dentate gyrus, respectively). 3 [ H]Nisoxetine Bmax and Kd Values in Cell Body Cell Body Regions. Figure 5E shows a positive relation 3 Regions. Specific [ H]nisoxetine Bmax values decreased with between Emax values for DMI’s anti-immobility effects in the 3 age in the locus coeruleus [F(2,52) 5 5.11, P , 0.05] (Fig. 4E). TST and Bmax values for specific [ H]nisoxetine binding in the Bmax values were greater in P21 SERT2/2 mice than in P90 locus coeruleus. As for the hippocampus, the relation between SERT2/2 mice (P 5 0.03). Bmax values did not appear to Emax and Bmax in the locus coeruleus was not affected by SERT depend on SERT genotype [F(2,52) 5 2.53, P 5 0.09] and no genotype (P 5 0.88); data from SERT1/1, SERT1/2 and interaction between age and SERT genotype was found SERT2/2 mice were therefore pooled and fitted with a single [F(4,52) 5 1.00, P 5 0.41]. line (Pearson correlation: r 5 0.41). In the locus coeruleus, Kd values ranged from 0.52 to 0.75 nM and did not depend on age (P 5 0.67) or SERT genotype (P 5 0.25). Discussion

Here, we found the potency and Emax of DMI to be greater Relationship between Maximal Antidepressant-Like Effects 1 1 3 in juvenile (P21) and adolescent (P28) SERT / mice than of DMI in the TST and Maximal Binding Values for [ H] in adult (P90) SERT1/1 mice (Table 1). Regardless of Nisoxetine in the Hippocampus and Locus Coeruleus SERT genotype, NET expression, quantified by specific Terminal Regions. Figure 5, A–D, shows a negative [3H]nisoxetine binding using autoradiography, increased with relation between Emax values for DMI’s anti-immobility effects age in the hippocampus (Fig. 4, A–D) and decreased with age 3 in the TST and Bmax values for specific [ H]nisoxetine binding in the locus coeruleus (Fig. 4E). Age-related changes in Emax in the CA1, CA2, and CA3 regions and dentate gyrus of the for DMI to reduce immobility were negatively related with 3 hippocampus of P21, P28, and P90 SERT1/1, SERT1/2, and Bmax values for specific [ H]nisoxetine binding in the hippo- SERT2/2 mice. In all regions, the relation between Emax and campus and were positively related with Bmax values in the Bmax was not significantly influenced by SERT genotype (P $ locus coeruleus (Fig. 5). Thus, there is an inverse relation 0.15); thus, the data were pooled across genotype and were between the maximal antidepressant-like response after DMI 92 Mitchell et al. and NET expression in the hippocampus, in which the pharmacokinetic parameters, and expression of NE-synthesizing antidepressant-like response to DMI decreases with age and enzymes) (Murrin et al., 2007). Future studies are needed to NET expression increases. In contrast, NET expression in the better understand the ontogeny of the noradrenergic system and locus coeruleus and antidepressant-like response to DMI how changes in these parameters may influence the response to decreases with increasing age. These data reveal a complex psychoactive therapeutics. relationship between the antidepressant-like response to DMI The therapeutic benefit of a drug may be affected by its and NET expression, suggesting that the antidepressant-like tolerability. Pediatric patients treated with TCAs consistently effect of DMI may be reliant on brain regions such as the locus showed side effects such as vertigo, tremors, low blood coeruleus. pressure, and dry mouth (Hazell and Mirzaie, 2013). Side The TCA imipramine, a NET and SERT blocker, is known to effects do not necessarily limit antidepressant-like drug reduce immobility time of adolescent mice (P28 and P35) in effects in the TST. Thus, our findings suggest that TCAs the FST, an assay of antidepressant-like response (Bourin may be efficacious antidepressants in juveniles (lower ED50 et al., 1998; David et al., 2001; Mason et al., 2009). Adolescent and greater Emax), but low tolerability of TCAs may limit the (P28 and P35) mice also respond to the anti-immobility effects therapeutic benefit of TCAs in pediatric patients. Therefore, of imipramine and DMI in the TST (Mason et al., 2009; our results encourage further investigation of the utility of Mitchell et al., 2013), and DMI reduces immobility time in the NET, and/or dual NET and SERT, blockers in the treatment FST in adolescent (P28 and P30) rats (Pechnick et al., 2008; of pediatric depression and the development of new drugs that Reed et al., 2008). Our SERT1/1 mouse data agree with target transporters for NE and 5-HT but have fewer side Downloaded from literature showing that adolescents (P28) are sensitive to the effects. antidepressant-like effects of DMI (Mitchell et al., 2013). We To elucidate the mechanism underlying age-related found that juvenile (P21) SERT1/1 mice are more sensitive to changes in the antidepressant-like response to DMI, we used the antidepressant-like effects of DMI in the TST than P28 quantitative autoradiography to assess [3H]nisoxetine bind- mice. These data deviate from experiments using juvenile ing to NET (Figs. 2 and 3). Specific [3H]nisoxetine binding (P21) rats showing that DMI does not significantly reduce FST densities generally increased with age in noradrenergic jpet.aspetjournals.org immobility (Reed et al., 2008). Possible reasons for this terminal regions. Bmax values were greater in adult CA1 discrepancy include differences in the species (mice versus and CA3 hippocampal regions than in younger mice (Fig. 4, rats), behavioral assay (TST versus FST) and injection AandC).Bmax valuestrendedtoincreasewithageinthe schedule. The FST and TST differ in their sensitivity to detect CA2 region, dentate gyrus, prefrontal cortex, and amygdala antidepressant-like activity, which can vary with rodent (Fig. 4, B and D; Table 1). These findings partially diverge strain; thus, discrepancies such as this are expected (Cryan from experiments using [3H]nisoxetine saturation binding et al., 2005). with whole hippocampus mouse homogenate, in which no at ASPET Journals on September 26, 2021 The SERT-deficient mouse provides a model of SERT gene age-related differences in Bmax values among P21, P28, and variants that confer a reduction in SERT expression and/or P90 mice were found (Mitchell et al., 2013). It is conceivable function and a reduction in SSRI efficacy (Fox et al., 2007; that hippocampal subregion differences in NET expression, Serretti et al., 2007; Homberg and Lesch, 2011). Our findings which can be revealed by quantitative autoradiography, are consistent with reports that DMI and imipramine reduce limited the ability to detect age-related changes in Bmax immobility in adult SERT1/1, SERT1/2, and SERT2/2 mice values for [3H]nisoxetine binding when using hippocam- in the TST (Holmes et al., 2002; Fig. 1). Studies investigating pal homogenate preparations. Notably, the Kd values for 3 treatment outcomes of patients with low-expressing SERT [ H]nisoxetine here (0.5–5.5 nM) are consistent with Kd gene variants treated with TCAs are limited. Reports using values from hippocampal homogenate preparations duloxetine (a SERT and NET blocker) and nortriptyline (a (2.5–7.4 nM) (Mitchell et al., 2013). NET-selective blocker) have failed to show any association In the rat hippocampus, triphasic [3H]nisoxetine binding with treatment outcome and low-expressing SERT gene patterns from birth through adulthood have been observed variants (Rajewska-Rager et al., 2008; Perlis et al., 2010). (Sanders et al., 2005). Binding densities increased postnatally, Our results from P28 SERT1/2 mice suggest that DMI may peaked between P15 and P25, and decreased into adulthood. 3 have especially limited therapeutic efficacy in adolescents In contrast, we found that Bmax values for [ H]nisoxetine harboring low-expressing SERT gene variants, although the binding increased with age from P21 to P90 (CA1 and CA3) or exact mechanism for this remains unknown. remained unchanged (CA2 and dentate gyrus) (Fig. 4, A–D). In 3 ED50 values increased with age from P21, confirming a the locus coeruleus of rats, [ H]nisoxetine binding densities similar effect of age in adult (P90) to middle-aged (P300) mice were found to peak at P10 and then decrease into adulthood (Mitchell et al., 2015; Table 1). Like DMI potency, Emax values (Sanders et al., 2005). Our data in the mice locus coeruleus are were greatest in SERT1/1 P21 and P28 mice compared with consistent with this finding (Fig. 4E). adults (Table 1). These findings are consistent with a study It is important to evaluate NET expression in SERT- showing a greater antidepressant-like Emax of imipramine deficient mice because non-SERT transporters of 5-HT, such in adolescent mice (P28) than in adults using the FST (David as NET, could upregulate to compensate for the loss of SERT et al., 2001). Emax values were not dependent on SERT (Daws, 2009; Daws et al., 1998, 2005; Baganz et al., 2008). We 3 genotype, which is consistent with studies showing no found that Bmax for [ H]nisoxetine binding was similar among difference in Emax values between P90 SERT1/1 and P90 SERT genotypes (Fig. 4), which is consistent with previous SERT1/2 or P90 SERT2/2 mice (Mitchell et al., 2015). findings showing that NET expression remains unchanged in It should be noted that we cannot rule out other age-dependent the CA3 region of the hippocampus (Montañez et al., 2003). factors as modulators of antidepressant-like response NET protein levels appear to be uninfluenced by a constitutive (i.e., expression and function of noradrenergic receptors, reduction or loss of SERT expression. Ontogeny of NET Expression and Antidepressant-Like Response 93

Increased NET expression in the hippocampus as a function Bylund DB and Reed AL (2007) Childhood and adolescent depression: why do chil- dren and adults respond differently to antidepressant drugs? Neurochem Int 51: of increasing age correlated with a reduction in the maximal 246–253. antidepressant-like effect of DMI (Fig. 5, A–D). This relation- Coyle JT (1977) Biochemical aspects of neurotransmission in the developing brain. Int Rev Neurobiol 20:65–103. ship was reversed in the locus coeruleus, such that decreasing Cryan JF, Mombereau C, and Vassout A (2005) The tail suspension test as a model NET expression as a function of age correlated with a for assessing antidepressant activity: review of pharmacological and genetic reduction in the maximal antidepressant-like effect for DMI studies in mice. Neurosci Biobehav Rev 29:571–625. David DJ, Bourin M, Hascoët M, Colombel MC, Baker GB, and Jolliet P (2001) (Fig. 5E). The antidepressant action of TCAs is hypothesized Comparison of antidepressant activity in 4- and 40-week-old male mice in the to occur by increasing extracellular levels of NE, primarily via forced swimming test: involvement of 5-HT1A and 5-HT1B receptors in old mice. Psychopharmacology (Berl) 153:443–449. NET blockade in noradrenergic terminal regions such as the Daws LC (2009) Unfaithful neurotransmitter transporters: focus on serotonin uptake hippocampus (Przegalinski et al., 1997; Herr et al., 2012). and implications for antidepressant efficacy. Pharmacol Ther 121:89–99. Daws LC, Montañez S, Owens WA, Gould GG, Frazer A, Toney GM, and Gerhardt Given that we found an inverse relationship between NET GA (2005) Transport mechanisms governing serotonin clearance in vivo revealed expression in the hippocampus and antidepressant response by high-speed chronoamperometry. J Neurosci Methods 143:49–62. Daws LC, Toney GM, Gerhardt GA, and Frazer A (1998) In vivo chronoamperometric (increasing and decreasing, respectively) as a function of age, measures of extracellular serotonin clearance in rat dorsal hippocampus: contribution of our data suggest that brain regions other than hippocampus, serotonin and norepinephrine transporters. J Pharmacol Exp Ther 286:967–976. Fox MA, Andrews AM, Wendland JR, Lesch KP, Holmes A, and Murphy DL (2007) A such as the locus coeruleus, are likely prominent contributors pharmacological analysis of mice with a targeted disruption of the serotonin to the antidepressant-like effects of DMI in the TST and transporter. Psychopharmacology (Berl) 195:147–166. Frazer A (1997) Pharmacology of antidepressants. J Clin Psychopharmacol 17 (Suppl 1): warrant further study. –

2S 18S. Downloaded from This study evaluated the dose dependence of the Hazell P and Mirzaie M (2013) Tricyclic drugs for depression in children and ado- lescents. Cochrane Database Syst Rev (6):CD002317. antidepressant-like effects of DMI and provided a survey of ’ 3 Hazell P, O Connell D, Heathcote D, Robertson J, and Henry D (1995) Efficacy of [ H]nisoxetine binding to brain NET in juvenile, adolescent, tricyclic drugs in treating child and adolescent depression: a meta-analysis. BMJ and adult SERT wild-type and SERT-deficient mice. Regard- 310:897–901. Herr NR, Park J, McElligott ZA, Belle AM, Carelli RM, and Wightman RM (2012) In less of SERT genotype, DMI potency and Emax were generally vivo voltammetry monitoring of electrically evoked extracellular norepinephrine greater in juvenile mice than in adults. Of note, P28 SERT1/2 in subregions of the bed nucleus of the stria terminalis. J Neurophysiol 107: 1731–1737. jpet.aspetjournals.org and SERT2/2 mice were relatively insensitive to the Holmes A, Yang RJ, Murphy DL, and Crawley JN (2002) Evaluation of antidepressant-like effects of DMI, suggesting that NET- antidepressant-related behavioral responses in mice lacking the serotonin trans- porter. Neuropsychopharmacology 27:914–923. blocking antidepressants may be especially ineffective in Homberg JR and Lesch KP (2011) Looking on the bright side of serotonin transporter adolescents harboring low-expressing/function variants of gene variation. Biol Psychiatry 69:513–519. Horton RE, Apple DM, Owens WA, Baganz NL, Cano S, Mitchell NC, Vitela M, Gould SERT, whereas the opposite may be true for juveniles (of GG, Koek W, and Daws LC (2013) Decynium-22 enhances SSRI-induced any SERT genotype). Regardless of SERT genotype, NET antidepressant-like effects in mice: uncovering novel targets to treat depression. – binding increased with age in several noradrenergic terminal J Neurosci 33:10534 10543. Kessler RC, Avenevoli S, and Ries Merikangas K (2001) Mood disorders in children regions and decreased with age in the locus coeruleus. These and adolescents: an epidemiologic perspective. Biol Psychiatry 49:1002–1014. at ASPET Journals on September 26, 2021 findings help lay the groundwork for future studies investi- Kirsch I, Deacon BJ, Huedo-Medina TB, Scoboria A, Moore TJ, and Johnson BT (2008) Initial severity and antidepressant benefits: a meta-analysis of data sub- gating the age and SERT genotype dependence of NET-acting mitted to the Food and Drug Administration. PLoS Med 5:e45. drugs, such as reboxetine. Studies to examine the mechanism Koek W, Mercer SL, Coop A, and France CP (2009) Behavioral effects of gamma- hydroxybutyrate, its precursor gamma-butyrolactone, and GABA(B) receptor ago- by which age-dependent variation in NET expression contrib- nists: time course and differential antagonism by the GABA(B) receptor antagonist ute to antidepressant response of NET blockers remain an 3-aminopropyl(diethoxymethyl)phosphinic acid (CGP35348). J Pharmacol Exp Ther 330:876–883. important avenue for future inquiries. The ultimate goal is to Konkol RJ, Bendeich EG, and Breese GR (1978) A biochemical and morphological develop effective antidepressants for patients whose symp- study of the altered growth pattern of central catecholamine neurons following 6-hydroxydopamine. Brain Res 140:125–135. toms of depression are resistant to treatment with SSRIs. Lauder JM and Bloom FE (1974) Ontogeny of monoamine neurons in the locus coeruleus, raphe nuclei and substantia nigra of the rat. I. Cell differentiation. J Comp Neurol 155:469–481. Acknowledgments Loizou LA (1972) The postnatal ontogeny of monoamine-containing neurones in the central nervous system of the albino rat. Brain Res 40:395–418. The authors thank Melissa Vitela and Myrna Herrera-Rosales for Loizou LA and Salt P (1970) Regional changes in monoamines of the rat brain during technical assistance. postnatal development. Brain Res 20:467–470. Mason SS, Baker KB, Davis KW, Pogorelov VM, Malbari MM, Ritter R, Wray SP, Gerhardt B, Lanthorn TH, and Savelieva KV (2009) Differential sensitivity to SSRI Authorship Contributions and tricyclic antidepressants in juvenile and adult mice of three strains. Eur J Pharmacol 602:306–315. Participated in research design: Mitchell, Gould, Koek, Daws. Mitchell NC, Gould GG, Koek W, and Daws LC (2016) Ontogeny of SERT expression Conducted experiments: Mitchell, Bowman, Gould. and antidepressant-like response to escitalopram in wild-type and SERT mutant Performed data analysis: Mitchell. mice. J Pharmacol Exp Ther 358:271–281. Mitchell NC, Gould GG, Smolik CM, Koek W, and Daws LC (2013) Antidepressant- Wrote or contributed to the writing of the manuscript: Mitchell, like drug effects in juvenile and adolescent mice in the tail suspension test: re- Bowman, Gould, Koek, Daws. lationship with hippocampal serotonin and norepinephrine transporter expression and function. Front Pharmacol 4:131. Mitchell NC, Koek W, and Daws LC (2015) Antidepressant-like effects and basal References immobility depend on age and serotonin transporter genotype. Genes Brain Behav BaganzNL,HortonRE,CalderonAS,OwensWA,MunnJL,WattsLT,Koldzic-Zivanovic 14:543–549. N, Jeske NA, Koek W, Toney GM, et al. (2008) Organic cation transporter 3: keeping the Moll GH, Mehnert C, Wicker M, Bock N, Rothenberger A, Rüther E, and Huether G brake on extracellular serotonin in serotonin-transporter-deficient mice. Proc Natl Acad (2000) Age-associated changes in the densities of presynaptic monoamine trans- Sci USA 105:18976–18981. porters in different regions of the rat brain from early juvenile life to late adult- Bengel D, Murphy DL, Andrews AM, Wichems CH, Feltner D, Heils A, Mössner R, hood. Brain Res Dev Brain Res 119:251–257. Westphal H, and Lesch KP (1998) Altered brain serotonin homeostasis and Montañez S, Owens WA, Gould GG, Murphy DL, and Daws LC (2003) Exaggerated locomotor insensitivity to 3, 4-methylenedioxymethamphetamine (“Ecstasy”)in effect of fluvoxamine in heterozygote serotonin transporter knockout mice. J serotonin transporter-deficient mice. Mol Pharmacol 53:649–655. Neurochem 86:210–219. Bourin M, Colombel MC, Redrobe JP, Nizard J, Hascoët M, and Baker GB (1998) Morris MJ, Dausse JP, Devynck MA, and Meyer P (1980) Ontogeny of alpha 1 and Evaluation of efficacies of different classes of antidepressants in the forced swim- alpha 2-adrenoceptors in rat brain. Brain Res 190:268–271. ming test in mice at different ages. Prog Neuropsychopharmacol Biol Psychiatry Murrin LC, Sanders JD, and Bylund DB (2007) Comparison of the maturation of the 22:343–351. adrenergic and serotonergic neurotransmitter systems in the brain: implications Bujoreanu S, Benhayon D, and Szigethy E (2011) Treatment of depression in children for differential drug effects on juveniles and adults. Biochem Pharmacol 73: and adolescents. Pediatr Ann 40:548–555. 1225–1236. 94 Mitchell et al.

Paksu S, Duran L, Altuntas M, Zengin H, Salis O, Ozsevik SN, Albayrak H, Murat N, Sanders JD, Happe HK, Bylund DB, and Murrin LC (2005) Development of the Guzel A, and Paksu MS (2014) Amitriptyline overdose in emergency department of norepinephrine transporter in the rat CNS. Neuroscience 130:107–117. university hospital: evaluation of 250 patients. Hum Exp Toxicol 33:980–990. Serretti A, Kato M, De Ronchi D, and Kinoshita T (2007) Meta-analysis of serotonin Paxinos G and Franklin KBJ (1997) The Mouse Brain in Stereotaxic Coordinates, transporter gene promoter polymorphism (5-HTTLPR) association with selective se- Academic Press, San Diego, CA. rotonin reuptake inhibitor efficacy in depressed patients. Mol Psychiatry 12:247–257. Pechnick RN, Bresee CJ, Manalo CM, and Poland RE (2008) Comparison of the Spear LP (2000) The adolescent brain and age-related behavioral manifestations. effects of desmethylimipramine on behavior in the forced swim test in peripubertal Neurosci Biobehav Rev 24:417–463. and adult rats. Behav Pharmacol 19:81–84. Steru L, Chermat R, Thierry B, and Simon P (1985) The tail suspension test: a new Perlis RH, Fijal B, Dharia S, Heinloth AN, and Houston JP (2010) Failure to replicate method for screening antidepressants in mice. Psychopharmacology (Berl) 85: genetic associations with antidepressant treatment response in duloxetine-treated 367–370. patients. Biol Psychiatry 67:1110–1113. Tejani-Butt SM (1992) [3H]nisoxetine: a radioligand for quantitation of norepineph- Przegalinski E, Tatarczynska E, Deren-Weso łek A, and Chojnacka-Wojcik E (1997) rine uptake sites by autoradiography or by homogenate binding. J Pharmacol Exp Antidepressant-like effects of a partial agonist at strychnine-insensitive glycine Ther 260:427–436. receptors and a competitive NMDA receptor antagonist. Neuropharmacology 36: Thomas SA, Matsumoto AM, and Palmiter RD (1995) Noradrenaline is essential for 31–37. mouse fetal development. Nature 374:643–646. Rajewska-Rager A, Dmitrzak-Weglarz M, Kapelski P, Skibinska M, Kaczmarkiewicz- Fass M, and Hauser J (2008) [Association between polymorphisms of ins/del in the 5-HTT gene and T102C in the 5HTR2A gene and the drug response for escitalo- Address correspondence to: Dr. Lynette C. Daws, Department of Cellular pram and nortriptyline in depressed patients]. Psychiatr Pol 42:903–914. and Integrative Physiology, University of Texas Health Science Center, Reed AL, Happe HK, Petty F, and Bylund DB (2008) Juvenile rats in the forced-swim 7703 Floyd Curl Drive, San Antonio, TX 78229-3900. E-mail: daws@uthscsa. test model the human response to antidepressant treatment for pediatric de- edu pression. Psychopharmacology (Berl) 197:433–441. Downloaded from jpet.aspetjournals.org at ASPET Journals on September 26, 2021