Abstract
Individual variation in the addiction liability of amphetamines has a heritable genetic component. We previously identified Hnrnph1 (heterogeneous nuclear ribonucleoprotein H1) as a quantitative trait gene underlying decreased methamphetamine-induced locomotor activity in mice. Here, we showed that mice (both females and males) with a heterozygous mutation in the first coding exon of Hnrnph1 (H1+/−) showed reduced methamphetamine reinforcement and intake and dose-dependent changes in methamphetamine reward as measured via conditioned place preference. Furthermore, H1+/− mice showed a robust decrease in methamphetamine-induced dopamine release in the NAc with no change in baseline extracellular dopamine, striatal whole-tissue dopamine, dopamine transporter protein, dopamine uptake, or striatal methamphetamine and amphetamine metabolite levels. Immunohistochemical and immunoblot staining of midbrain dopaminergic neurons and their forebrain projections for TH did not reveal any major changes in staining intensity, cell number, or forebrain puncta counts. Surprisingly, there was a twofold increase in hnRNP H protein in the striatal synaptosome of H1+/− mice with no change in whole-tissue levels. To gain insight into the mechanisms linking increased synaptic hnRNP H with decreased methamphetamine-induced dopamine release and behaviors, synaptosomal proteomic analysis identified an increased baseline abundance of several mitochondrial complex I and V proteins that rapidly decreased at 30 min after methamphetamine administration in H1+/− mice. In contrast, the much lower level of basal synaptosomal mitochondrial proteins in WT mice showed a rapid increase. We conclude that H1+/− decreases methamphetamine-induced dopamine release, reward, and reinforcement and induces dynamic changes in basal and methamphetamine-induced synaptic mitochondrial function.
SIGNIFICANCE STATEMENT Methamphetamine dependence is a significant public health concern with no FDA-approved treatment. We discovered a role for the RNA binding protein hnRNP H in methamphetamine reward and reinforcement. Hnrnph1 mutation also blunted methamphetamine-induced dopamine release in the NAc, a key neurochemical event contributing to methamphetamine addiction liability. Finally, Hnrnph1 mutants showed a marked increase in basal level of synaptosomal hnRNP H and mitochondrial proteins that decreased in response to methamphetamine, whereas WT mice showed a methamphetamine-induced increase in synaptosomal mitochondrial proteins. Thus, we identified a potential role for hnRNP H in basal and dynamic mitochondrial function that informs methamphetamine-induced cellular adaptations associated with reduced addiction liability.
Introduction
Addiction to psychostimulants, including methamphetamine (MA), is a major public health concern in the United States, with an estimated 535,000 individuals currently meeting the criteria for MA dependence (Lipari et al., 2016). Despite the prevalence of MA addiction, there is currently no FDA-approved treatment, in part because the neurobiological mechanisms underlying MA addiction are still not clear. Variation in sensitivity to the locomotor stimulant response to psychostimulants is a heritable trait and can sometimes predict differences in drug self-administration in rodents (Hooks et al., 1991; Yamamoto et al., 2013) because shared neurocircuits and neurochemical mechanisms underlie these behaviors. We recently used quantitative trait locus mapping, positional cloning, and gene editing via transcription activator-like effector nucleases to identify Hnrnph1 as a quantitative trait gene for MA sensitivity in mice (Yazdani et al., 2015). Hnrnph1 (heterogenous nuclear ribonucleoprotein H1) encodes an RNA-binding protein (RBP) that is expressed throughout the brain, and is a part of a subfamily of hnRNPs that includes hnRNP H1, hnRNP H2, and hnRNP F, which possess structurally unique quasi-RNA recognition motifs (Honoré et al., 1995). hnRNP H1 regulates all aspects of RNA metabolism, including pre-mRNA splicing through binding at specific intronic sites, mRNA stability and translational regulation via binding to the 5′UTR and 3′UTR, and poly-adenylation control (Chou et al., 1999; Arhin et al., 2002; Katz et al., 2010; Witten and Ule, 2011; Wang et al., 2012; Song et al., 2017).
We previously demonstrated that Hnrnph1 polymorphisms and heterozygous deletion in the first coding exon of Hnrnph1 affect behavioral sensitivity to acute MA-induced locomotor stimulation; however, the effects on MA reward and reinforcement are unknown. Additionally, the neurobiological mechanism(s) underlying the mutational effects of Hnrnph1 on MA-induced behaviors remain to be established. Hnrnph1 mRNA is ubiquitously expressed throughout the adult mouse brain (Lein et al., 2007). While the protein expression of hnRNP H1 appears to be nuclear-restricted, studies assessing hnRNP H1 protein in the brain are limited (Kamma et al., 1995; Honoré et al., 1999; Van Dusen et al., 2010). With regard to CNS function, hnRNP H family proteins are described as master regulators of neuron and oligodendrocyte differentiation via alternative splicing control (Wang et al., 2007; Grammatikakis et al., 2016). Whole-exome sequencing identified coding variants in human HNRNPH1 and HNRNPH2 (located on the X chromosome) associated with severe neurodevelopmental disorders (Bain et al., 2016; Pilch et al., 2018), implicating a crucial role of the hnRNP H protein family in neurodevelopment.
The purpose of the present study was threefold. First, to expand beyond MA locomotor stimulant sensitivity, we examined the effect of the Hnrnph1 mutation on oral MA reinforcement and intake via operant-conditioning and MA reward via conditioned place preference (CPP). This mutation comprises a small, frameshift deletion in the first coding exon of Hnrnph1 (H1+/−) that causes reduced MA-induced locomotor activity (Yazdani et al., 2015). To gain insight into the neurobiological mechanisms underlying behavioral differences in MA sensitivity, we examined drug-induced dopamine (DA) release via in vivo microdialysis, DA content of striatal tissue, and DA clearance from striatal tissue. Second, because we previously implicated Hnrnph1 polymorphisms in dopaminergic neuron development, we assessed the effect of H1+/− on TH levels in cell bodies and processes of the mesolimbic pathway via immunoblotting and immunohistochemistry (IHC). Finally, to gain insight into neural dysfunction in H1+/− mice at the protein level that could underlie behavioral and neurochemical deficits, we examined the synaptosomal proteome of the striatum between the H1+/− and WT mice at baseline and in response to MA.
Materials and Methods
Mice
Experimental mice were generated by mating H1+/− males with C57BL/6J females purchased from The Jackson Laboratory (for studies at Boston University) or in Sacramento, California (for studies at the University of California at Santa Barbara). Offspring were genotyped as previously described (Yazdani et al., 2015). Unless otherwise indicated, both female and male littermates (56- to 100-d-old at the start of the experiment) were used in the studies. Mice were housed in same-sex groups of 3–5 in standard mouse cages and housed within ventilated racks under standard housing conditions. All procedures conducted in mice were approved by the Boston University, University of California at Santa Barbara, and Virginia Commonwealth University Animal Care and Use Committees. All experiments were conducted in strict accordance with the Guide for the care and use of laboratory animals. Colony rooms were maintained on a 12:12 h light-dark cycle.
Genotyping of H1+/− and H1−/− mice
The genotyping protocol for the H1+/− mice was published by Yazdani et al. (2015). For genotyping of H1−/− embryo tissue, an Hnrnph1 forward primer (GATGATGCTGGGAGCAGAAG) and reverse primer (GGTCCAGAATGCACAGATTG) were designed to target upstream and downstream of the deleted region in exon 4 of Hnrnph1. Genomic DNA was used to amplify a 204 bp PCR product using DreamTaq Green PCR Mastermix (Thermo Fisher Scientific) followed by overnight restriction enzyme digest with BstNI (New England Biolabs).
Hnrnph1 and Hnrnph2 qRT-PCR for mouse embryo tissue
Oligo-dT primers were used to synthesize cDNA from total RNA to examine mRNA expression and qPCR for evaluating gene expression performed using TaqMan SYBR Green (Thermo Fisher Scientific, catalog #4309155). The primer sequences used for evaluating expression of Hnrnph1 (targeting exons 6 and 7) were ACGGCTTAGAGGACTCCCTTT and CGTACTCCTCCCCTGGAAGT. The primer sequences used for quantifying expression of Hnrnph2 (targeting exons 1 and 2) were TAGCCGTTTGAGGGAAGAAG and CCCTGTTAGAGTTTCTTCCAGGTA. The housekeeping gene used was Hprt (targeting exons 7 and 9) with following primer sequence: GCTGGTGAAAAGGACCTCT and CACAGGACTAGAACACCTGC.
Oral MA self-administration
The procedures for MA operant conditioning were similar to those recently described (Lominac et al., 2016). Testing was conducted in operant chambers equipped with 2 nose-poke holes, 2 cue lights, a tone generator, and a liquid receptacle for fluid reinforcement (Med Associates). Mice were not water-restricted at any point during oral MA self-administration procedures. The vehicle for MA was filtered tap water. Under a fixed-ratio 1 schedule of reinforcement, mice were trained daily to self-administer MA during 1 h sessions where a single active nose-poke response resulted in delivery of 20 μl of liquid MA into the receptacle, with a 20 s illumination of the cue light, and the sounding of the tone. During the 20 s period, further responding resulted in no programmed consequences. Inactive hole responses were recorded but had no consequences, serving to gauge the selectivity of responding in the MA-reinforced hole. Mice were initially trained to nose-poke 80 mg/L MA, with the concentration of MA progressively increased over weeks (120, 160, 200, 300, and 400 mg/L MA; 5 d per dose). Upon the completion of each daily session, the volume of MA remaining in the receptacle was determined by pipetting and was subtracted from the volume delivered to calculate MA intake (Lominac et al., 2016).
CPP
Mice were trained for 1 h each day in Plexiglas activity boxes within sound-attenuating chambers (40 cm length × 20 cm width × 45 cm tall; divided into two sides with different plastic floor textures for CPP). Mice were recorded from above using infrared cameras (Swan) and tracked with ANY-maze (RRID:SCR_014289). The CPP paradigm was described by Kirkpatrick and Bryant (2015).
On training days 2–5, mice were injected with either saline (days 2 and 4) or MA (days 3 and 5; saline, 0.5 or 2 mg/kg, i.p.) and confined to either the saline- or MA-paired side for 1 h.
Stereotaxic surgery
The procedures to implant indwelling microdialysis guide cannulae bilaterally over the NAc of mice were described previously (Lominac et al., 2014, 2016). Mice were anesthetized <1.5%–2% isoflurane with 4% oxygen as a carrier gas, mounted in a Kopf stereotaxic device with tooth and ear bars adapted for mice. The mouse skull was exposed, leveled, and holes were drilled based on coordinates from bregma for the NAc (AP: 1.3 mm, ML: ±1 mm, DV: −2.3 mm), according to the mouse brain atlas of Paxinos and Franklin (2001). The guide cannulae were lowered bilaterally such that the tips of the cannulae were 2 mm above the NAc. The skull was then prepared for polymer resin application, and the guide cannulae were secured to the skull with dental resin. After surgery, mice were injected subcutaneously with warm saline and 250 μl of 2.5 mg/ml Banamine (Henry Schein Animal Health) and allowed to recover on a heating pad. Postoperative care was provided for 4 d, during which mice were injected with 250 μl of 2.5 mg/ml Banamine subcutaneously daily for the first 2 d. Mice were allowed a minimum 1 week recovery before in vivo microdialysis assessments.
In vivo microdialysis and high-performance liquid chromatography (HPLC) analysis
Conventional microdialysis was conducted using a within-subjects design to examine saline and acute MA-induced DA release (0.5 or 2 mg/kg, i.p.), using procedures similar to those described previously (Lominac et al., 2014, 2016). Microdialysis probes were inserted unilaterally and perfused with aCSF for 3 h (2 μl/min), allowing for neurotransmitter equilibration. For DA no net flux analysis, DA was infused at 0, 2.5, 5, and 10 nm, and dialysate was collected in 20 min intervals for 1 h/concentration. On a subsequent day, mice were probed on the contralateral side; and following the 3 h equilibration period and 1 h of baseline dialysate collection, mice were injected intraperitoneally with either 0.5 or 2.0 mg/kg MA, and dialysate was collected in 20 min intervals for 3 h after injection. HPLC analysis of DA was conducted as described previously (Lominac et al., 2014). Cannulae placement was determined on Nissl-stained coronal sections, and only mice exhibiting correct placement within the NAc were included in analyses.
Behavioral test battery
Prepulse inhibition of acoustic startle.
This test was used to assess sensorimotor gating. The apparatus and procedures were identical to those previously described by Szumlinski et al. (2005). Six trial types were conducted: startle pulse (st110, 110 dB/40 ms), low prepulse stimulus alone (st74, 74 dB/20 ms), high prepulse stimulus alone (st90, 90 dB/20 ms), low or high prepulse stimulus given 100 ms before the onset of the startle pulse (pp74 and pp90, respectively), and no acoustic stimulus (st0; only background noise). All trials were presented in a randomized order; st0, st110, pp74, and pp90 trials were given 10 times, whereas st74 and st90 were presented five times. Background noise in each chamber was 70 dB, and the average intertrial interval lasted 15 s.
Novel object test.
To assess anxiety-like behavior, mice were placed into a rectangular box (9.25 × 17.75 × 8 inches high) containing one small, inedible object for 2 min. During that time, animals were allowed to explore and interact with the object. The number of contacts and time in contact (seconds) with the novel object were video-recorded and tracked with ANY-maze tracking software (RRID:SCR_014289). The apparatus and procedures used were identical to those previously described by Szumlinski et al. (2005).
Marble burying.
The marble burying test was used to measure anxiety-like defensive burying (Njung'e and Handley, 1991). In our paradigm, 12 square glass pieces (2.5 cm2 × 1.25 cm high) were placed in the animals' home cage, 6 at each end. The latency to start burying the marbles was determined by a blind observer using a stopwatch, and the total number of marbles buried following a 20 min trial was recorded.
Light/dark shuttle box.
The light/dark shuttle box test was used to assess exploratory and anxiety-like behaviors. Mice were placed into a polycarbonate box (46 cm long × 24 cm high × 22 cm wide) wide containing distinct open (light) and closed (dark) environments for a 15 min trial. These two environments were separated by a central divider with an opening. Mice were first placed on the dark side, and the latency to enter the light side, number of light-side entries, and total time spent in the light-side of the shuttle box were recorded using ANY-maze tracking software (RRID:SCR_014289). An increase in latency to enter the light, uncovered, side was interpreted as an index of anxiety-like behavior.
Porsolt swim test.
To assess depressive-like behavior (Porsolt et al., 1977), mice were placed into a pool (30 cm in diameter; 45 cm high) filled with room-temperature water up to 35 cm and allowed to swim for a total of 6 min. Time immobile (seconds), immobile episodes, and immobile latency (seconds) were video-recorded and tracked by ANY-maze tracking software (RRID:SCR_014289).
Accelerating rotarod.
To assess motor coordination, mice were trained on the rotarod (IITC Life Science, ROTO-ROD series) for a total of 10 trials over 3 d: 3 trials the first 2 d and 4 trials on the final day. The rotarod started at 4 RPM and accelerated to 40 RPM in 60 s. The time (seconds) it took a mouse to fall (physically falling or hanging off rotarod) was manually scored. Time on the rotarod was averaged across the total 10 trials for each mouse.
Bitter/quinine taste sensitivity
H1+/− and WT mice were allowed continuous access in the home cage to 4 sipper tubes containing 0 (filtered tap water), 0.1, 0.3, and 0.6 mg/ml quinine (Sigma-Aldrich). The quinine concentrations selected for study were based on Eastwood and Phillips (2014). The mice and bottles were weighed before initial presentation, and the bottle was weighed every 24 h thereafter. The difference in bottle weight was used to determine the volume consumed from each solution over each 24 h period, the average intake from each solution, and the average total volume consumed.
Quantification of baseline monoamine neurotransmitters from whole striatal tissue
Drug-naive striatum were harvested from H1+/− and WT littermates and flash-frozen on dry ice. The dissected tissue was sent to Vanderbilt University Neurotransmitter Core for the quantification of monoamine neurotransmitters using HPLC wth electrochemical detection.
DA transporter (DAT)-mediated DA uptake
Saline or MA (2.0 mg/kg) was administered interperitoneally in a volume of 10 ml/kg. After 2 h after administration (2 h was chosen based on the microdialysis results), mice were decapitated, and DAT-specific [3H]DA uptake from synaptosome preparations was conducted as described previously (Kivell et al., 2014). Mice were rapidly decapitated, and striatal regions were dissected from the brain and collected in 10 volumes (w/v) of prechilled 0.32 m sucrose buffer (0.32 m sucrose in 5 mm HEPES, pH 7.5). The striatal tissue was homogenized and centrifuged at 1000 × g for 15 min at 4°C. The supernatant was centrifuged at 12,000 × g for 20 min, and the pellet was suspended in 0.32 m sucrose buffer. Striatal synaptosomes (30 μg) were incubated in a total volume of 0.3 ml of Krebs-Ringer-HEPES buffer consisting of 120 mm NaCl, 4.7 mm KCl, 2.2 mm CaCl2, 10 mm HEPES, 1.2 mm MgSO4, 1.2 mm KH2PO4, 5 mm Tris, 10 mm d-glucose, pH 7.4 containing 0.1 mm ascorbic acid, and 0.1 mm pargyline at 37°C for 10 min with or without DAT-specific blocker GBR12909 (50 nm). Following incubation, 5 nm [3H]DA (63.2 Ci/mmol-dihydroxyphenylethylamine [2,5,6,7,8–3H]; PerkinElmer) and further incubated for additional 5 min. Uptake of DA was terminated with the addition of 500 nm DAT blocker GBR12909. The samples were filtrated over 0.3% polyethylenimine-coated GF-B filters on a Cell Harvester (Brandel) and washed rapidly with 5 ml cold PBS. Radioactivity bound to the filter was counted using a liquid scintillation counter. DAT-mediated [3H]DA uptake was determined by subtracting total accumulation of [3H]DA (absence of GBR12909) and in the presence of GBR12909. Uptake assays were performed in triplicates.
IHC
For IHC, drug-naive H1+/− and WT mice were anesthetized with pentobarbital, and transcardially perfused with PBS followed by 4% PFA in PBS at room temperature. Next, brains were dissected and processed for TH DAB IHC and analysis as previously described (Burke et al., 1990; Hutson et al., 2011), or double immunofluorescent IHC for hnRNP H and TH colocalization. For DAB IHC, coronal slices were blocked with 4% normal goat serum and then incubated for 48 h at 4°C with anti-hnRNP H (1:50,000, Bethyl, catalog #A300-511A, RRID:AB_203269) or TH (1:500, Santa Cruz Biotechnology, catalog #sc-14007, RRID:AB_671397), and processed for DAB staining and analyzed as previously described (Hutson et al., 2011). For costaining studies with hnRNP H and TH, tissues were blocked with superblock (Thermo Fisher Scientific, catalog #37515), and incubated with anti-hnRNP H (Santa Cruz Biotechnology, catalog #sc-10042, RRID:AB_2295514) and TH (1:500, Santa Cruz Biotechnology, catalog #sc-14007, RRID:AB_671397) for 48 h at 4°C. Next, tissues were incubated with donkey anti-rabbit AlexaFluor-488 (1:500, Invitrogen, catalog #A-21206, RRID:AB_141708) and donkey anti-goat AlexaFluor-633 (1:500, Invitrogen, catalog #A-21082, RRID:AB_141493), washed, and then coated with ProLong Diamond Antifade Mountant (Thermo Fisher Scientific, catalog #P36961), mounted onto slides, and imaged on the SPE Confocal microscope (Leica Microsystems).
TH puncta quantification in the striatum
Entire coronal slices of rostral, medial, and caudal striatum were imaged at 40× magnification using a Nikon Eclipse 600 microscope. A 225,000 μm2 grid was overlaid onto these images using ImageJ and every third FOV within the striatum was graded. Number of puncta within a FOV was graded in ImageJ by subtracting out background signal, creating binary images from these files, and then counting puncta meeting roundness and diameter criteria (roundness <0.6, diameter 1–45 μm). Average puncta densities for ventral, dorsal, and total striatum were calculated. Grading of the puncta was performed in ImageJ. The image was duplicated into a 1,000,000 pixel area followed by brightness/contrast adjustment and background subtraction. The threshold was set to 106. A binary image was then generated for puncta measurement. Roundness was set to 0.6–1, and size was set to 5–200 pixels.
Dissection of mouse brain regions: striatum and midbrain
Live, rapid decapitation was used to avoid the effects of anesthesia or CO2 asphyxiation on gene expression. Immediately followed live decapitation using large, sharpened shears with an incision just posterior from the ears, the mouse brain was removed quickly with forceps and transfer to a cold surface. The striatum has a somewhat darker appearance than the surrounding cortex. To dissect the dorsal and ventral striatum, fine-tip forceps were used to separate the midline of the brain, and then the cortex and hippocampus were removed to reveal the striatum. To dissect the midbrain, a razor blade was used to make a rostral cut where the cerebral aqueduct begins and another caudal cut just before the start of the cerebellum.
MA-induced locomotor activity followed by tissue harvesting
On days 1 and 2, all mice received a saline injection (10 ml/kg, i.p.) and were recorded for locomotor activity in Plexiglas chambers (40 cm length × 20 cm width × 45 cm height) for 1 h. On day 3, mice receive either saline or MA (2 mg/kg, i.p.; Sigma-Aldrich) and were recorded for locomotor activity for 30 min, and the whole striatum was harvested as described above at 30 min after injection. Whole striata (left and right sides) were flash frozen in ethanol/dry ice bath and stored at −80°C for long-term storage. Four cohorts of animals were run in this behavioral paradigm for tissue collection: (1) hnRNP H immunoprecipitation followed by mass spectroscopy; (2) synaptosome mass spectroscopy and hnRNP H immunoblot; (3) validation studies for mitochondrial protein immunoblots; and (4) measurement of MA concentration in MA-treated striatal tissues. The mice that were used for hnRNP H immunoprecipitation were all MA-treated on day 3.
Quantification of MA metabolites MA-treated whole striatal tissue
Whole striata (left and right sides) were flash frozen in ethanol/dry ice bath and stored at −80°C. Tissue were then shipped on dry ice to University of Utah Health Science Center for MA and amphetamine metabolite quantification as previously described (Slawson et al., 2002; DeYoung et al., 2016).
Preparation of synaptosomes
Striatal tissue collection from saline- or MA-treated mice was performed as described above. The tissues were subsequently processed to obtain synaptosomes using a Percoll (Sigma-Aldrich, catalog #1644) gradient fractionation method, which was adapted from Dunkley et al. (2008). Whole striata (left and right hemisphere) were placed in 1 ml sucrose homogenization buffer (2 mm HEPES, pH 7.4, 320 mm sucrose, 50 mm EDTA, 20 mm DTT) supplemented with protease and phosphatase inhibitor (Thermo Fisher Scientific, catalog #78440). The brain tissues were lightly homogenized using a hand-held motorized pestle. The homogenate for each sample was then centrifuged for 2 min at 3000 rcf, and the supernatant (S1) was collected. The pellet (P1) was then resuspended in 500 μl of sucrose homogenization buffer and respun for 3 min at 3000 rcf. The supernatant (S1′) was collected and combined with S1 and then centrifuged for 15 min at 9200 rcf. The supernatant S2 was then removed, and the pellet (P2) was resuspended in 500 μl of sucrose homogenization buffer and loaded onto a Percoll density gradient consisting of 23%, 10%, and 3% Percoll (1 ml each) in polycarbonate centrifuge tubes (13 × 51 mm; Beckman Coulter, catalog #349622). The gradients were then centrifuged for 15 min at 18,700 rcf. The distinct band between the 10% and 23% Percoll was collected as the synaptosome. The synaptosome fraction was then washed to 5 ml with 1× HBM buffer, pH 7.4 (140 mm NaCl, 5 mm KCl, 5 mm NaHCO3, 1.2 mm NaH2PO4, 1 mm MgCl2-6H2O, 10 mm glucose, 10 mm HEPES) to dilute out the Percoll by centrifuging for 12 min at 18,700 rcf. The pellet was then resuspended in 100 μl of HBM buffer to yield the final synaptosme fraction, and BCA assay was used to determine protein concentration. A total of 30 μg of synaptosome was loaded per sample for SDS-PAGE and Western blotting as described below.
SDS-PAGE and Western blot
Brain tissues were homogenized using hand-held homogenizer in RIPA buffer with Halt protease and phosphatase inhibitor mixture (Thermo Fisher Scientific, catalog #78840). For each sample, 30 μg of protein was heated in a 70°C water bath for 10 min before loading into into a 4%–15% Criterion TGX precast Midi protein gel (Bio-Rad) for SDS-PAGE followed by wet transfer to nitrocellulose membrane. The membrane was then blocked with 5% milk for 1 h and probed with primary antibodies. For evaluating TH expression in brain tissues, overnight incubation of the membrane at 4°C with anti-TH (1:50,000, Santa Cruz Biotechnology, catalog #sc-14007, RRID:AB_671397) was performed followed by 1 h incubation with donkey anti-rabbit HRP (1:10,00, Jackson ImmunoResearch Laboratories, catalog #711-035-152, RRID:AB_10015282). For evaluating hnRNP H protein expression in mouse embryo tissues and in striatal synaptosome, the following antibodies were used: hnRNP H (C-term: 1:50,000, Bethyl, catalog #A300-511A, RRID:AB_203269; N-term: 1:50,000, Santa Cruz Biotechnology, catalog #sc-10042, RRID:AB_2295514) followed by 1 h incubation with the appropriate secondary antibodies. For validation of mitochondrial protein expression following MA treatment, the following primary antibodies were used: ATP5A1 (1:2000, Abcam, catalog #ab14748, RRID:AB_301447); ATP5F1 (1:5000, Proteintech, catalog #15999-1-AP, RRID:AB_2258817); and NDUFS2 (1:5000; Abcam, ab192022). We used the following loading controls: anti-β-actin (1:20,000, Sigma-Aldrich, catalog #A2228, RRID:AB_476697); anti-GAPDH (1:20,000; Millipore, catalog #MAB374, RRID:AB_2107445); and PSD95 (1:10,000, Cell Signaling Technology, catalog #3450, RRID:AB_2292883).
Mouse brain tissue processing for SDS-PAGE and DAT immunblotting was modified from Staal et al. (2007). Briefly, tissue was triturated using a 20–22 gauge needle in RIPA buffer (10 mm Tris, pH 7.4, 150 mm NaCl, 1 mm EDTA, 0.1% SDS, 1% Triton X-100) supplemented with Halt protease and phosphatase inhibitor mixture (Thermo Fisher Scientific, catalog #78840); 30 μg of protein of each sample was allowed to rotate at room temperature before SDS-PAGE instead of heating the sample at high temperature. For evaluating DAT expression in whole striatal tissue and striatal synaptosome, we conducted overnight incubation anti-DAT (1:2000; Millipore, catalog #MAB369, RRID:AB_2190413) followed by 1 h incubation with goat anti-rat (1:500, Jackson ImmunoResearch Laboratories, catalog #112-035-003, RRID:AB_2338128).
All processed membranes were imaged via enhanced chemiluminescence photo-detection. Densitometry analysis in ImageJ was used for quantification.
hnRNP H immunoprecipitation
Following the 3 d locomotor paradigm assessing acute locomotor sensitivity in H1+/− versus WT mice as described above, striata were dissected from the mice at 30 min after injection of MA or saline. Striatum was dissected from WT or H1+/− mice and frozen on dry ice and stored at −80°C. Striatal tissues were then homogenized using a microcentrifuge pestle in ice-cold RIPA buffer (50 mm Tris-HCl, pH 6.8, 150 mm NaCl, 5 mm EDTA, 1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate) supplemented with protease and phosphatase inhibitor cocktails (Thermo Fisher Scientific, catalog #78840) and incubated overnight at 4°C with gentle agitation. Lysates were centrifuged at 10,000 rpm for 15 min at 4°C, and the supernatant fraction was saved for protein quantification via BCA assay; 1 mg of striatal lysates was precleared for 1 h with 80 μl Protein G Sepharose-coated beads (Thermo Fisher Scientific, catalog #101243) and then centrifuged at 4°C for 5 min at 1000 rpm. The precleared lysates (supernatant) were then incubated overnight with 10 μg of rabbit anti-hnRNP H (Bethyl, catalog #A300-511A, RRID:AB_203269) or control normal rabbit IgG antibody (Millipore, catalog #NI01-100UG, RRID:AB_490574). The next day, 80 μl of Protein G Sepharose-coated beads was added to the antibody-lysate mixture and incubated for an additional 2 h. The beads were then washed 4 times in 1 ml lysis buffer, resuspended, and centrifuged for 1 min at 1000 rpm each time. The beads were eluted by adding 60 μl of nonreducing SDS buffer and heating at 95°C for 10 min. hnRNP H immunoprecipitates were eluted in nonreducing SDS buffer as described above; 50 μl of each sample was separated by SDS-PAGE at 100 V for 30 min (∼2 cm) on a Novex Bolt 4%–12% Bis-Tris gel. The gel was then washed 3 times in deionized H2O and stained with Simply Blue Coomassie SafeStain (Thermo Fisher Scientific, catalog #LC6060). The gel was then cut at ∼160 kDa to exclude the prominent nonreduced IgG band. Individual gel lanes were then separately excised and stored in prewashed microcentrifuge tubes at 4°C before shipping to the University of Massachusetts Worchester Proteomics and Mass Spectrometry facility for analysis by OrbiTrap liquid chromatography tandem mass spectrometry (LC-MS/MS).
TMT labeling, high pH reverse phase HPLC fraction, followed by LC-MS/MS
Following the 3 d locomotor paradigm assessing acute locomotor sensitivity in H1+/− versus WT mice as described above, striata were dissected from the mice at 30 min after injection. Synaptosomes were isolated following the procedure outlined above for proteomic characterization. Samples were resuspended in 8 m urea for 30 min, followed by the addition of 5 mm DTT for 1 h. Lodoacetamide was then add to the samples that were incubated in the dark for 30 min. The urea concentration was diluted <1 m with the addition of 50 mm of ammonium bicarbonate. The samples were then digested with trypsin (50:1, protein to enzyme ratio) overnight at 37°C and terminated with the addition of formic acid to 1%. The samples were desalted with a C18 tip. Peptide was determined by Pierce Quantitative Colorimetric Assay (Thermo Fisher Scientific, catalog #23275); then 100 μg of peptide was resuspended in 0.1 m triethylammonium bicarbonate and incubated with the TMT plex isobaric labeling for 1 h at room temperature. To quench the reaction, 5% hydroxylamine was added to each sample and incubated for 15 min. Each sample was combined at equal amount and cleaned with C18 tips. One milligram of labeled peptides was fractioned using a Waters XBridge BEH C18 (3.5 μm, 4.6 × 250 mm) column on an Agilent Technologies 1100 HPLC system. A total of 48 fractions were collected and combined to 12 fractions and then dried. A C18 Acclaim PepMap 100 precolumn (3 μm, 100 Å, 75 μm × 2 cm) hyphenated to a PepMap RSLC C18 analytical column (2 μm, 100 Å, 75 μm × 50 cm) was used to separate peptide mixture. LC-MS/MS analyses were completed using an EASY nLC 1200 system coupled to a Q Exactive HF-X mass spectrometer. Full MS spectra were collected at a resolution of 120,000 with an AGC of 3 × 106 or maximum injection time of 50 ms and a scan range of 350 to 1650 m/z. MS2 scans were performed at 45,000 resolution and using 32% total normalized collision energy. Source ionization parameters were optimized with the spray voltage at 2.1 kV; dynamic seclusion was set to 45 s.
Proteomics data analysis and pathway enrichment results
The acquired data were searched by MaxQuant against the UniProt mouse proteome database with standard settings (fragment ion mass tolerance of 20 ppm, maximum missed cleavage of 2, oxidation as variable modification, false discovery was 1%, only protein groups identified with at least 2 or more peptides). The intensity data were filtered and normalized using R (RRID:SCR_001905) and the LIMMA package (RRID:SCR_010943) was used for differential analysis with Genotype and Treatment as factors. A ranked list was generated from the analysis, and the fgsea R package was used to perform preranked analysis, with proteins filtered for absolute log2FC > 0.2 and p < 0.05. Enrichment results for the comparisons were visualized using Cytoscape (RRID:SCR_003032) and EnrichmentMap (RRID:SCR_016052), with nodes representing pathways and edges representing overlap genes. Pathways were clustered and annotated with themes using AutoAnnotate (Reimand et al., 2019).
Experimental design and statistical analyses
Characterization of Hnrnph1/2 gene expression and hnRNP H protein expression in H1+/− mice.
For gene expression analysis of Hnrnph1/2 via qPCR, each sample was run in triplicate and averaged. Differential gene expression was reported as the fold change in H1+/− and H1−/− relative to WT littermates using the 2−(ΔΔCT) method (Livak and Schmittgen, 2001). To examine genotypic differences in transcript and protein expression among WT, H1+/−, and H1−/−, one-way ANOVA was conducted. To examine genotypic difference in level of peptides unique to either hnRNP H1 or hnRNP H2, two-tailed unpaired t test was performed for each peptide, with the exception of Peptide 3 for hnRNP H1, where we used a one-tailed unpaired t test based on a priori results published by Yazdani et al. (2015), indicating a decrease in exon 4 usage of Hnrnph1 in the H1+/− versus WT. Thus, we hypothesized that Peptide 3, which is encoded by part of exon 4, would also be decreased in the H1+/− mice.
Oral self-administration.
To determine the effects of the H1+/− mutation upon MA reinforcement and intake, an operant-conditioning procedure was used, and the data for the mean number of active nose pokes, inactive nose pokes, and MA intake compared between WT and H1+/− mice. This comparison was evaluated using mixed-model ANOVA with the between-subjects factors of Genotype and Sex and the within-subjects factor of MA concentration (80, 120, 160, 200, 300, and 400 mg/L). In a follow-up study, WT and H1+/− mice were trained to self-administer the 200 mg/L MA concentration over the course of 14 d. The data for this study were analyzed using mixed-model ANOVA with Genotype and Sex as between-subjects factors and day as the repeated measure (days 1–14). To interrogate potential genotypic differences in bitter tastant sensitivity, the dose-response function for quinine intake was determined in a separate cohort of mice, and the data were analyzed using a mixed-model ANOVA with between-subjects factors of Genotype and Sex and repeated measures on the Quinine Concentration factor (0.003, 0.01, 0.03, 0.1, 0.3, and 0.6 mg/ml).
In vivo microdialysis.
To relate genotypic differences in behavior to changes in extracellular DA and glutamate within the NAc, we conducted a series of in vivo microdialysis experiments. The first series of experiments used no net flux microdialysis procedures, and the ensuing dose-response data were analyzed by linear regression to determine the extraction fraction/clearance of the neurotransmitter (determined by the slope of the regression) and the basal neurotransmitter content (determined by x = 0 from the regression). The extraction fraction and content were analyzed by a univariate ANOVA with the between-subjects factors of Sex and Genotype. The second series of experiments examined the time course of MA-induced changes in extracellular DA, 3,4-dihydroxyphenylacetic acid (DOPAC), and glutamate. The average baseline neurotransmitter content for the 60 min before MA injection was analyzed using a univariate ANOVA along the between-subjects factors of Genotype and Sex. Mice were slated to receive one of two MA doses (0.5 and 2.0 mg/kg), and the data for each neurotransmitter were expressed as a percent change from the average baseline and analyzed using a mixed-model ANOVA with Genotype and Sex as between-subjects factors and repeated measures on the Time factor (12, 20 min, bins), separately for each MA dose.
CPP.
To examine genotypic differences in MA reward, we assessed MA-CPP by quantifying the change in time spent on drug-paired side on day 8 − day 1 or on day 9 − day 1. Preference was analyzed using a three-way ANOVA with Genotype, Sex, and Treatment (saline, 0.5 mg/kg MA, and 2 mg/kg MA) as between-subjects factors. Because this analysis indicated genotype- and dose-dependent change in MA-CPP, we ran two-tailed unpaired t tests at each dose to identify genotypic differences in MA-CPP. To assess potential genotypic differences in MA-induced locomotor activity in the smaller arena containing the textured floor that comprised the MA-CPP side (days 2–5) as well as in the entire arena (days 1, 8, and 9), we analyzed distance traveled in 5 min bin using a mixed-effects ANOVA with Genotype, Sex, and Treatment (saline, 0.5 mg/kg, or 2 mg/kg MA) as between-subjects factors and repeated measures on Time (12, 5 min bins), separately for each day. Two-tailed unpaired t tests were used to determine the sources of main effects of Genotype and Treatment and interactions with Time.
Behavioral test battery.
To determine the effects of Genotype upon sensorimotor function and affect, WT and H1+/− mice were subjected to a behavioral test battery consisting of prepulse inhibition, the novel object test, marble-burying, light-dark shuttle-box, Porsolt swim test, and motor coordination on a rotarod. The data generated from these assays were analyzed using two-tailed unpaired t tests.
Quantification of DA content and MA metabolites and DAT-mediated DA uptake.
To determine whether there was a genotypic difference in DA content at baseline and MA metabolite after MA injection, a two-tailed unpaired t test was performed. In examining whether there is an MA-induced change on DAT-mediated DA uptake between WT and H1+/− mice, a two-way ANOVA with Genotype and Treatment (saline and MA) as between-subjects factors was used.
MA-induced locomotor activity assessment before tissue collection for mass spectrometry.
To replicate the reduction in MA-induced locomotor activity in H1+/− mice that we previously observed under identical conditions (Yazdani et al., 2015), we examined total distance traveled in 5 min bin, separately for days 1–3 via mixed-effects ANOVA with Genotype, Sex, and Treatment (saline and 2 mg/kg MA) as between-subjects factors and Time as the repeated-measures factor. Two-tailed unpaired t tests were used to determine sources of the main effects of Genotype and interactions with Time.
Immunoblotting and IHC.
To examine the effect of Genotype on TH protein expression via immunoblotting and IHC, two-tailed unpaired t test was performed on loading control-normalized data between WT and H1+/−. To determine whether there was an effect of Sex on DAT protein expression, two-way ANOVA with Genotype and Sex as between-subjects factors was conducted. Two-tailed unpaired t tests were used to determine sources of the main effects of Sex and interactions with Genotype.
To examine MA-induced changes in protein expression, including DAT, hnRNP H, and mitochondrial proteins, two-way ANOVA with Genotype and Treatment as between-subjects factors was conducted. For hnRNP H expression in the synaptosome, the significant main effect of Genotype was followed up with two-tailed unpaired t tests to examine differences between WT and H1+/−.
For all behavioral studies, we used an effect size of Cohen's d = 0.9 to guide our sample size based on genotypic differences in MA-induced locomotor activity at 30 min after MA in our previously published data from H1+/− mice (Yazdani et al., 2015). We used G*Power 3 (Faul et al., 2007) to calculate a sample size of n = 16 per genotype required for 80% power (p < 0.05). All behavioral studies used both male and female mice. All data are presented as means of replicates from each experiment ± SEM. In summary, for experiments in which two conditions were compared, a two-tailed unpaired t test was used to analyze the data unless otherwise specified. For experiments in which multiple factors were evaluated, ANOVA was used and significant interaction was deconstructed along the relevant factor to examine main effects and group differences followed up using t tests. Statistical p value threshold for t test and ANOVA was set to 0.05. The data comprising oral self-administration in vivo microdialysis were analyzed in SPSS version 21 (IBM). All other data were analyzed in R (RRID:SCR_001905).
Data and code availability
All related data and materials are available upon request. The mass spectrometry proteomics data are available in MassIVE under proteome exchange accession number PXD014813.
Results
Characterization of Hnrnph1/2 gene expression and hnRNP H protein expression in H1+/− mice
Hnrnph1 mutant mice (H1+/−) were generated using transcription activator-like effector nucleases targeting the first coding exon of Hnrnph1 (exon 4, UCSC Genome Browser). Deletion of a small region in the first coding (exon 4) in Hnrnph1 leads to a premature stop codon and transcription of a truncated mRNA message. The mice were genotyped using a restriction enzyme-based assay (Fig. 1A) (Yazdani et al., 2015). Gene expression via qPCR with primers specific for exons 6 and 7 (not targeting the deleted exon 4) detected a 50% increase of Hnrnph1 transcript in the H1+/− (heterozygous for deletion) relative to WT (Yazdani et al., 2015) and a 1.5-fold increase in the H1−/− (homozygous for deletion) (Fig. 1B; one-way ANOVA, F(2,9) = 2.55, p = 0.133; pairwise comparison between WT vs H1−/−: t(6) = −2.82, p = 0.031; n = 4 per group) with no change in gene expression of Hnrnph2 (Yazdani et al., 2015; Fig. 1C; one-way ANOVA, F(2,9) = 1.28, p = 0.324; n = 4 per group). At the protein level, there was no change in hnRNP H protein expression in H1+/− or H1−/− embryonic stage brain tissue homogenate (Fig. 1D–G; one-way ANOVA, C terminus: F(2,9) = 0.89, p = 0.440; N terminus: F(2,9) = 1.64, p = 0.250; n = 4 per group). Because hnRNP H1 and H2 have 96% amino acid sequence homology, there is no commercially available antibody that can differentiate the two proteins. Quantitative analysis of protein peptide from an hnRNP H immunoprecipitation mass spectrometry study identified no difference in peptide level unique to hnRNP H1 between the H1+/− mice compared with that of WT (Fig. 2A,B; unpaired t test for each individual peptide, Peptide 1: t(4) = −0.45, p = 0.673; Peptide 2: t(4) = −1.39, p = 0.236; Peptide 3: t(4) = −1.94, p = 0.125; Peptide 4: t(4) = −0.85, p = 0.445; Peptide 5: t(4) = −1.06, p = 0.351; Peptide 6: t(4) = 0.46, p = 0.672; Peptide 7: t(4) = 2.21, p = 0.091; Peptide 8: t(4) = −1.07, p = 0.346; n = 3 per group). Because we previously detected a decrease in transcription of exon 4 in the H1+/− mice (Yazdani et al., 2015) (Fig. 1B), we hypothesized that there would be a decrease in the level of Peptide 3 (GLPWSCSADEVQR), which contains amino acid residues encoded by the deleted region in exon 4 of Hnrnph1. Indeed, there was a decrease in the level of Peptide 3 in H1+/− mice, although this result did not reach statistical significance (Fig. 2A,B; one-tailed unpaired t test, t(4) = −1.94, p = 0.062). No genotypic difference was observed in unique peptides associated with hnRNP H2 (Fig. 2C,D; unpaired t test for each individual peptide, Peptide 1: t(4) = −0.83, p = 0.454; Peptide 2: t(4) = 0.86, p = 0.441; Peptide 3: t(4) = 1.56, p = 0.197; Peptide 4: t(4) = 0.95, p = 0.396; Peptide 5: t(4) = −0.516, p = 0.633; Peptide 6: t(4) = −0.07, p = 0.950; n = 3 per group).
Oral MA self-administration
We previously demonstrated a robust reduction in sensitivity to the locomotor stimulant response to MA in H1+/− mice (Yazdani et al., 2015). To more directly test MA-induced motivated behaviors, we investigated MA reinforcement and intake in H1+/− mice using an oral MA self-administration paradigm as conducted previously in C57BL/6J mice (Szumlinski et al., 2017). This study showed that mice allocated the majority of their nose-poking behavior toward the MA-reinforced hole, indicating drug reinforcement. H1+/− mice presented fewer active nose pokes at the 160, 200, 300, and 400 mg/L MA doses compared with WT mice (Fig. 3A; mixed-effects ANOVA, F(1,41)Genotype = 6.33, p = 0.02; F(5,205)Dose = 23.93, p < 0.0001; F(5,205)Genotype×Dose = 1.58, p = 0.166; unpaired t tests for 80 and 120 mg/L, p values >0.65; for 160–400 mg/L, p values <0.03; WT: n = 22 and H1+/−: n = 23), with no difference in nonspecific, inactive nose pokes across a range of MA concentrations (80–400 mg/L) (Fig. 3B; mixed-effects ANOVA, F(1,41)Genotype = 0.76, p = 0.39; F(5,205)Dose = 5.84, p < 0.0001; F(5,205)Genotype×Dose = 1.27, p = 0.28). Consistent with their lower level of MA reinforcement, H1+/− mice also consumed less MA (intake; mg/kg/d) at the 200, 300, and 400 mg/L concentrations (Fig. 3C; mixed-effects ANOVA, F(1,41)Genotype = 6.85, p = 0.01; F(5,205)Dose = 52.11, p < 0.0001; F(5,205)Genotype×Dose = 4.47, p < 0.0001; unpaired t tests for 80–160 mg/L, p values >0.12; for 200–400 mg/L, p values < 0.03). These results indicate that the H1+/− mice are less motivated to work for the MA reinforcer. No interactions between Sex and Genotype were observed for any measure during self-administration testing (mixed-effects ANOVAs for active and inactive hole pokes and for MA intake, F(1,41)Sex×Genotype < 2.83, p values >0.10; F(5,205)Genotype× Sex × Dose < 1.63, p values > 0.15). The genotypic differences in high-dose MA intake did not relate to bitter taste sensitivity as quinine intake in the home cage was equivalent between H1+/− and WT mice (Fig. 4; mixed-effects ANOVA, F(5,100)Quinine = 44.63, p < 0.0001; F(1,20)Genotype = 2.07, p = 0.170; F(5,100)interaction = 0.50, p = 0.780).
The blunted escalation of oral MA intake observed at high MA concentrations (≥200 mg/L) in H1+/− mice initially trained to respond for 80 mg/L MA prompted us to determine whether the Hnrnph1 mutation would blunt the acquisition of oral MA self-administration when 200 mg/L MA served as the reinforcer. However, we were unable to detect any genotypic difference in MA intake and active nose pokes when mice were trained at this higher MA concentration (Fig. 5; mixed-effects ANOVA, F(13,273)day = 8.39, p < 0.0001; F(1,21)Genotype = 0.01, p = 0.91; F(13,273)Genotype×day = 0.39, p = 0.97). These results indicate that the Hnrnph1 mutation interferes with the transition from low- to high-dose MA self-administration, without altering the reinforcing effects of MA during acquisition.
MA-conditioned reward
To further investigate why H1+/− mice showed less oral self-administration of MA, we assessed MA reward via CPP. When mice were tested in an MA-free state (day 8), the genotypic difference in the time spent in the MA-paired side between day 8 and varied with MA doses (Fig. 6A; two-way ANOVA, F(2,121)interaction = 3.92; p = 0.023; F(1,121)Genotype = 0.07, p = 0.788; F(2,121)Dose = 6.15, p = 0.003). H1+/− mice showed lower CPP to 0.5 mg/kg (WT: n = 23, H1+/−: n = 22) MA but higher CPP to 2 mg/kg (WT: n = 19, H1+/−: n = 16) MA compared with the WT mice (Fig. 6A; unpaired t test; 0.5 mg/kg: t(45) = −2.13, p = 0.039; 2 mg/kg: t(45) = 2.18, p = 0.036). Similar nonsignificant effects were observed during drug state-dependent CPP whereby H1+/− mice exhibited lower and higher CPP at 0.5 mg/kg and 2 mg/kg MA doses, respectively (Fig. 6B; two-way ANOVA, F(2,121)interaction = 1.86; p = 0.160; F(1,121)Genotype = 0.23, p = 0.634; F(2,121)Dose = 2.08, p = 0.129). No sex differences or interactions were observed for any measure during MA-CPP testing (three-way ANOVA, day 8 to day 1: F(1,115)Sex = 0.77, p = 0.381, F(2,115)interaction = 0.35, p = 0.707; day 9 to day 1: F(1,115)Sex = 0.09, p = 0.761, F(2,115)interaction = 1.26, p = 0.287). The dose-dependent difference in CPP indicates that H1+/− showed a reduced sensitivity to the rewarding effect of low-dose MA and that a higher MA dose was required to elicit CPP in H1+/− mice.
We also identified significant genotypic differences in MA-induced locomotor activity in the confined one-half of the CPP box during CPP training on days 3 and 5 (Fig. 7; mixed-effects ANOVA, day 3: F(1,109)Genotype = 5.04, p = 0.037, F(1,109)Sex = 0.06, p = 0.800, F(22,1199)Genotype×Treatment×Time = 1.69, p = 0.024; day 5: F(1,109)Genotype = 3.07, p = 0.073, F(1,109)Sex = 2.93, p = 0.090, F(22,1199)Genotype×Treatment×Time = 2.291, p = 0.0006; days 2 and 4: F(1,109) Genotype×Treatment×Time < 1, p > 0.5). H1+/− mice traveled less distance at 60 min after 2 mg/kg MA injection on day 3 and 5 (unpaired t tests, day 3: t(31) = −2.26, p = 0.031; day 5: t(31) = −2.32, p = 0.027). During state-dependent CPP (following the two previous MA exposures) on day 9 when mice had twice as much open access space, we observed a decrease in MA-induced locomotor activity in H1+/− mice (Fig. 8; mixed-effects ANOVA, day 9: F(1,113)Genotype = 3.05, p = 0.083, F(1,113)Sex = 0.16, p = 0.695, F(2,113)Genotype×Treatment = 3.78, p = 0.026; days 1 and 8: F(2,115) Genotype×Treatment = 1.16, 0.39 for days 1 and 8, respectively, p values >0.3). This result was similar to the previously published results in the same-sized open arena minus the CPP divider and floor textures in response to a single, acute MA dose of 2 mg/kg (Yazdani et al., 2015). Furthermore, we later demonstrate replication of reduced MA-induced locomotor activity in H1+/− mice using the same 3 d protocol from the original publication (Yazdani et al., 2015) before tissue harvesting for mass spectrometry analysis (for later description of the results, see Fig. 20). There was no genotypic difference in the striatal level of either MA or its metabolite amphetamine at 30 min (i.p.) after injection of MA (Fig. 9; unpaired t test, t(14) = −0.72, −0.22, p = 0.480, 0.826 for MA and amphetamine, respectively; WT: n = 8 and H1+/−: n = 8). Thus, differences in transport or metabolism of MA are unlikely to explain the decrease in MA-induced behavior in H1+/− mice at 30 min after MA.
MA-induced DA release in H1+/− mice
Because H1+/− mice showed reduced MA self-administration and reward, we next examined basal extracellular DA and MA-elicited DA release using in vivo microdialysis in the NAc, a brain region and neurochemical event that are necessary for MA reinforcement and reward (Prus et al., 2009; Keleta and Martinez, 2012; Bernheim et al., 2016). Linear regression analysis of no net flux in vivo microdialysis in the NAc indicated no significant change in either DA release/reuptake (extraction fraction or slope; Fig. 10A; two-way ANOVA, F(1,31)Genotype = 2.86, p = 0.10; F(1,31)Genotype×Sex = 2.52, p = 0.12) or basal extracellular DA content or (Fig. 10B; two-way ANOVA, F(1,31)Genotype = 0.06, p = 0.81; F(1,31)Genotype×Sex = 0.03, p = 0.86). Following administration of 0.5 mg/kg MA (i.p.), WT mice exhibited an increase in extracellular DA, whereas H1+/− mice showed a decrease in extracellular DA below baseline (Fig. 11A; mixed-effects ANOVA, F(1,25)Genotype = 18.57, p < 0.0001; F(11,275)Time = 1.63, p = 0.09; F(11,275)Genotype×Time = 4,71, p < 0.0001; unpaired t tests for time = −60 to 0 min, t(25) < 0.60, p values > 0.60; for time = 20 to 180 min, t(25) > 2.25, p values < 0.033). Administration of 2 mg/kg MA (i.p.) induced an increase in DA within the NAc of both genotypes, but H1+/− mice again showed markedly lower MA-induced extracellular DA levels, in particular from 100 to 180 min after MA (Fig. 11B; mixed-effects ANOVA, F(1,31)Genotype = 6.41, p = 0.02; F(11,341)Time = 8.29, p < 0.0001; F(11,341)Genotype×Time = 5.58, p < 0.0001; unpaired t tests, for time −60 to 0 and 40 min, t(31) < 1.5, p values > 0.15; for time 20, 60–180 min, t(31) > 2.0, p values < 0.05). Consistent with the genotypic differences in extracellular DA levels, for both the 0.5 and 2 mg/kg MA doses, H1+/− mice also showed a lower level of DOPAC (metabolite of DA) compared with WT mice (Fig. 11C,D; 0.5 mg/kg MA: mixed-effects ANOVA, F(1,28)Genotype = 1.50, p = 0.23; F(11,308)Time = 13.33, p < 0.0001; F(11,308)Genotype×Time = 2.76, p = 0.002; unpaired t tests, for all times but 160 min, t(28) < 1.90, p values > 0.08; for time = 160 min, t(28) = 2.37, p = 0.03; 2.0 mg/kg MA: mixed-effects ANOVA, F(1,25)Genotype = 24.09, p < 0.0001; F(11,275)Time = 29.04, p < 0.0001; F(11,275)Genotype×Time = 10.42, p < 0.0001; unpaired t tests, for time −60 to 0 and 40 min, t(25) < 1.6, p values > 0.10; for time 20, 60–180 min, t(25) > 2.0, p values < 0.05).
In examining glutamate levels, no net flux microdialysis study revealed no difference in basal extracellular level of glutamate (Fig. 12A; for slope, t(14) = 0.77, p = 0.45; for x intercept, t(14) = 0.49, p = 0.64). Acute administration of 0.5 or 2 mg/kg MA did not induce much effect on extracellular glutamate levels, nor were there any genotypic differences in glutamate levels (Fig. 12B,C; mixed-effects ANOVA, for 0.5 mg/kg MA, F(1,28)Genotype = 0.54, p = 0.47; F(11,308)Time = 0.56, p = 0.86; F(11,308)Genotype×Time = 0.59, p = 0.84; for 2.0 mg/kg MA, F(1,30)Genotype = 1.46, p = 0.24; F(11,330)Time = 1.89, p = 0.04; F(11,330)Genotype×Time = 0.59, p = 0.84). To summarize, H1+/− mice showed a significantly blunted MA-induced DA release in the NAc in the absence of any significant difference in baseline DA levels and in the absence of any significant differences in glutamate neurotransmission. No sex differences or interactions were observed. These results point toward an MA-induced deficit in DA release as a plausible functional mechanism underlying the deficit in the MA-induced behavioral responses in H1+/− mice.
To further investigate the possible causes of the decreased capacity of MA to induce DA release in H1+/− mice, we quantified basal striatal tissue levels of biogenic amines in drug-naive H1+/− and WT mice. No genotypic difference was detected in the amount of DA and DOPAC (Fig. 13A; unpaired t test, DA: t(14) = −0.30, p = 0.766; DOPAC: t(14) = −0.93, p = 0.372; n = 8 per group), or in the levels of other biogenic amines, including HVA, 3-MT, 5-HT, etc. (Fig. 13A; unpaired t test, HVA: t(14) = 0.08, p = 0.938; 3-MT: t(14) = −0.53, p = 0.372; 5-HT: t(14) = 1.02, p = 0.323; 5-HIAA: t(14) = 1.68, p = 0.114; noradrenaline: t(14) = 0.722, p = 0.482). We also tested whether H1+/− impacts the function of DAT by examining DAT-mediated DA uptake in synaptosomes prepared from whole striatal tissue. However, no difference in basal DAT function was detected (Fig. 13B; two-way ANOVA, F(1,14)interaction = 0.01, p = 0.911; F(1,14)Genotype = 0.02, p = 0.903; F(1,14)Treatment = 0.26, p = 0.619; n = 4 per group). In addition, there was no change in the protein level of DAT in the total striatal brain lysate in H1+/− compared with WT mice (Fig. 13C,D; unpaired t test, t(23) = 0.32, p = 0.750; WT: n = 14 and H1+/−: n = 11). An increase in DAT protein level was detected in the female mice (Fig. 13C,D; two-way ANOVA, F(1,21)Sex = 5.74, p = 0.026, F(1,21)Genotype = 0.12, p = 0.730; F(1,21)interaction = 0.07, p = 0.800; unpaired t test, females vs males t(23) = 2.53, p = 0.019; females: n = 13; males: n = 12). There was also no effect of MA on DAT protein level in the synaptosome of either H1+/− or WT mice (Fig. 13E,F; two-way ANOVA, F(1,20)Genotype = 0.031, p = 0.86; F(1,20)Treatment = 0.037, p = 0.85; F(1,20)interaction = 1.24, p = 0.28; WT SAL: n = 5, WT MA: n = 6, H1+/− SAL: n = 6, and H1+/− MA: n = 7).
TH expression in the mesolimbic and nigrostriatal brain regions of H1+/− mice
In our previous transcriptome analysis of congenic mice that captured Hnrnph1 polymorphisms and decreased MA behavioral sensitivity, we identified a decrease in expression of Nurr1 (Nr4a2), a transcription factor critical for dopaminergic neuron development (Yazdani et al., 2015). Thus, we hypothesized that dysfunctional mesolimbic and nigrostriatal DA pathways at the neuroanatomical level could underlie decreased MA-induced behaviors and DA release in H1+/− mice. We tested this hypothesis by examining changes in the number of neurons and the number of projections from neurons in the dopaminergic mesolimbic and nigrostriatal pathways, which mediate reward and motor activity, respectively (Adinoff, 2004). Diagrams outlining the different brain regions assessed are presented in Figure 14. We first examined coexpression of hnRNP H (there are currently no commercially available antibodies for differentiating between hnRNP H1 and hnRNP H2) and TH (the rate-limiting enzyme for the synthesis of DA) (Daubner et al., 2011). Results showed that hnRNP H and TH were coexpressed in the same midbrain TH neurons of the VTA and substantia nigra pars compacta (SNc) at a similar level in both genotypes (Fig. 15A). hnRNP H immunostaining was primarily nuclear, whereas TH immunostaining was cytoplasmic. We next examined expression of TH in the dopaminergic neurons of the mesolimbic and nigrostriatal circuit (Figs. 15, 16). TH DAB immunohistostaining in the VTA and SNc of the midbrain (Fig. 15B) indicated no significant genotypic differences in TH optical density (OD) (Fig. 15C; unpaired t test, VTA: t(13) = −1.29, p = 0.220; SNc: t(13) = −0.75, p = 0.466; WT: n = 7 and H1+/−: n = 8). There was also no genotypic difference in the number (Fig. 15D; unpaired t test, t(14) = 0.02, t = 0.985; n = 8 per genotype) or diameter (Fig. 15E; t(14) = 0.88, p = 0.395; n = 8 per genotype) of TH-positive neuron of the VTA and SNc. Immunoblot analysis indicated no significant difference in TH protein level in the midbrain (Fig. 15F,G; unpaired t test, t(18) = 0.67, p = 0.510; n = 10 per genotype).
TH DAB immunostaining in the striatum indicated no change in TH OD in the dorsal striatum between H1+/− and WT mice (Fig. 16A–C; unpaired t test, t(13) = 2.07, p = 0.10; WT: n = 8; H1+/−: n = 7). However, a small, but statistically significant increase in TH OD was observed in the ventral striatum (which includes the NAc) of the H1+/− mice (Fig. 16D–F; unpaired t test, t(13) = 2.30, p = 0.040; WT: n = 8 and H1+/−: n = 7). To measure TH-positive puncta as an indirect estimate of the number of DA terminals in the striatum, we performed stereology under higher magnification (Fig. 17A) and detected no difference in the number of TH-positive puncta between the H1+/− and WT mice in the dorsal striatum or in the ventral striatum (Fig. 17B; dorsal: t(14) = 0.55, p = 0.588; ventral: t(14) = −1.49, p = 0.159, n = 8 per genotype). Immunoblot analysis indicated no significant difference in TH protein level in the whole striatum (Fig. 16G,H; unpaired t test, t(23) = 0.62, p = 0.540; WT: n = 14 and H1+/−: n = 11). Together, analysis of the dopaminergic mesolimbic and nigrostriatal DA circuit did not provide strong evidence for changes in the expression of TH within cell bodies or terminals or changes in the number of TH-positive terminals that could explain behavioral differences in H1+/− mice.
An important message gleaned from the above results is that the behavioral and neurochemical deficits exhibited by H1+/− mice appear to manifest only under the influence of MA. In further support of this notion, a screen of WT (n = 33) and H1+/− (n = 38) mice in a behavioral test battery (Fig. 18A) did not detect any genotypic differences in sensorimotor gating (Fig. 18B,C; prepulse inhibition of acoustic startle: unpaired t test, 75 Hz: t(69) = −1.26, p = 0.213, 95 Hz: t(69) = −0.53, p = 0.598; novel object: unpaired t test, contact number: t(69) = 0.51, t = 0.615, contact time: t(69) = 1.62, p = 0.111), anxiety-like behaviors (Fig. 18D,E; marble burying, unpaired t test, t(69) = −0.44, p = 0.660; light/dark shuttle box: unpaired t test, t(69) = −1.48, p = 0.145), depressive-like behaviors (Fig. 18F; t(69) = 0.24, p = 0.810), or in sensorimotor coordination (Fig. 18G; unpaired t test, t(69) = −0.44, p = 0.663). The null results from this behavioral battery, combined with the lack of genotypic differences in saline-induced locomotion/response to a novel environment, and in various indices of basal DA transmission in the striatum argue for an active, MA-induced cell biological mechanism linking Hnrnph1 dysfunction to MA behavior.
Methamphetamine-induced changes in total and synaptic level of hnRNP H
The above neuroanatomical studies failed to support a neuroanatomical hypothesis of reduced neurodevelopment of DA projection pathways that could underlie reduced MA-induced DA release and behavior in H1+/− mice. Thus, to further explore the possibility of a synaptic, MA-induced cell biological mechanism that could underlie H1+/− behavior, we next examined changes in the synaptic localization of hnRNP H, which is potentially relevant for understanding how H1+/− could alter MA-induced DA release. Surprisingly, we identified a twofold increase in hnRNP H protein level in the striatal synaptosome of H1+/− mice, regardless of treatment (Fig. 19A,B; two-way ANOVA, F(1,27)Genotype = 24.36, p = 3.63E-5; F(1,27)Treatment = 0.23, p = 0.636; F(1,27)interaction = 0.41, p = 0.528; WT [n = 15] vs H1+/− [n = 16]: unpaired t test, t(29) = −5.51, p = 2.176E-5). In contrast, there was no significant genotypic difference in hnRNP H protein level in bulk striatal tissue in response to saline or MA (Fig. 19C,D; two-way ANOVA, F(1,20)Genotype = 0.60, p = 0.450; F(1,20)Treatment = 0.01, p = 0.952; F(1,27)interaction = 0.37, p = 0.548; n = 6 per group), indicating a change in localization rather than total hnRNP H protein in H1+/− mice.
To identify changes in the levels of other synaptic proteins that could mechanistically link the robust increase in synaptic hnRNP H with decreased MA-induced DA release and behavior, we examined genotypic differences in the synaptosomal proteome in the striatum of H1+/− and WT mice treated with MA versus saline using LC-MS/MS. At the behavioral level, H1+/− showed reduced MA-induced locomotor activity only in response to MA on day 3 (Fig. 20C; mixed-effects ANOVA, F(1,12)Genotype = 8.54, p = 0.013; F(1,12)Treatment = 77.42, p = 1.4E-6; F(1,12)Genotype×Treatment = 5.85, p = 0.032; F(5,60)Time = 1.83, p = 0.120; F(5,60)Genotype×Treatment×Time = 1.23, p = 0.305; unpaired t test for WT vs H1+/− after MA at each 5 min bin, t(6) = −1.63, −1.67, −2.44, −2.46, −2.41, −3.01, p = 0.154, 0.147, 0.050, 0.049, 0.052, 0.023 for 5, 10, 15, 20, 25, and 30 min time point, respectively). In contrast, no difference in locomotor activity was detected between WT and H1+/− in response to saline (Fig. 20A,B; mixed-effects ANOVA, day 1: F(1,12)Genotype = 2.40, p = 0.147; F(1,12)Treatment = 0.203, p = 0.660; F(1,12)Genotype×Treatment = 0.17, p = 0.684; F(11,132)Time = 1.89, p = 0.046; F(11,132)Genotype×Treatment×Time = 1.19, p = 0.301; day 2: F(1,12)Genotype = 1.27, p = 0.283; F(1,12)Treatment = 0.466, p = 0.508; F(1,12)Genotype×Treatment = 0.21, p = 0.656; F(11,132)Time = 0.45, p = 0.930; F(11,132)Genotype×Treatment×Time = 1.32, p = 0.219). Overall, proteomic analysis for the main effect of Genotype identified a highly enriched upregulation of mitochondrial proteins in the H1+/− striatal synaptosome regardless of treatment (Fig. 21A). Enrichment analysis for the set of top differentially expressed proteins (absolute log2FC > 0.2; p < 0.05) between H1+/− and WT mice revealed a highly significant enrichment for mitochondrial respiratory chain complex I assembly (Table 1). In examining MA-induced changes in synaptic proteins in H1+/− versus WT mice, we again identified an enrichment for alterations in metabolic processes involving components of mitochondrial complex I and V (Fig. 21B; Fig. 21-1, available at https://doi.org/10.1523/JNEUROSCI.1808-19.2019.f21-1; Table 2). Interestingly, in response to MA, there was a decrease in the level of mitochondrial proteins in the H1+/− mice, but an increase in the WT mice (Fig. 22). The findings were independently validated with a separate cohort of mice, which also pointed to a trending, nonsignificant pattern of an increase in expression of all three mitochondrial proteins in saline-treated H1+/− mice and a trending, nonsignificant decrease in all three mitochondrial proteins in response to MA (Fig. 23; unpaired t test, NDUFS2: t(6) = −1.39 p = 0.107; ATP5A1: t(6) = −1.64, p = 0.077; ATP5F1: t(6) = −1.24, p = 0.131; n = 4 per group).
Discussion
Here, we extend a role for Hnrnph1 in MA reinforcement and reward (Fig. 24). A heterozygous frameshift 16 bp deletion of the first coding exon of Hnrnph1 reduced MA oral operant self-administration and MA-induced CPP (Figs. 3, 6) and induced a robust reduction in MA-induced DA release in the NAc (Fig. 11). This DA anomaly occurred without any differences in (1) basal extracellular DA (Fig. 10A,B); (2) basal DA content within striatal tissue (Fig. 13A); (3) DA uptake (Fig. 13B); (4) the number or staining of midbrain DA neurons within the mesolimbic and nigrostriatal dopaminergic circuits (Fig. 15); or (5) any robust changes in the number of forebrain striatal puncta originating from these neurons (Fig. 16). The combined results suggested an alternate mechanism underlying the decreases in MA-induced DA release and behavior in H1+/− mice. In further support of an MA-induced cell biological mechanism, there was no effect of H1+/− on spontaneous locomotion, anxiety- and depressive-like behaviors, or sensorimotor function (Fig. 18).
H1+/− mice showed less MA-CPP at 0.5 mg/kg but greater CPP at 2 mg/kg than WT mice (Fig. 6). One interpretation is that MA-CPP exhibits an inverted U-shaped dose-response curve (Uhl et al., 2014) in WT and that H1+/− shifts this curve to the right, yielding reduced sensitivity to positive and negative motivational valence of MA at lower versus higher doses. Consistent with this interpretation, H1+/− mice exhibited blunted operant oral MA reinforcement. The combined data are consistent with H1+/− mice self-administering less MA because they are less sensitive to the physiological and interoceptive effects of MA due to a reduced MA-induced DA release.
H1+/− mice exhibited a blunted DA response to 0.5 and 2 mg/kg MA (Fig. 11A,B). This effect on MA-induced DA release could not be explained by alterations in total DA levels (Fig. 13A) or in DAT levels or function at the presynaptic membrane (Fig. 13B–F). An alternative explanation to DAT dysfunction is that the Hnrnph1 mutation somehow decreases MA binding to DAT, limiting its entry into presynaptic dopaminergic neuronal terminals and decreasing DA release. Future studies are necessary to examine MA binding to DAT in H1+/− mice.
Our findings support a dopaminergic mechanism underlying reduced MA reward and reinforcement in H1+/− mice. Nevertheless, we acknowledge the potential involvement of additional neurotransmitter systems and brain regions. We did not identify any difference in basal or MA-induced changes in extracellular glutamate levels in the NAc of H1+/− mice. While MA reward is generally attributed to an increase in DA release in the NAc (Segal and Kuczenski, 1997; Adinoff, 2004), MA also increases release of norepinephrine and serotonin by targeting their respective transporters that could modulate the locomotor stimulant and/or rewarding response to psychostimulants, such as MA (Haughey et al., 2000; Rothman et al., 2001; Zaniewska et al., 2015). Future studies are warranted to address these other neurotransmitters in MA-induced behavioral dysfunction in H1+/− mice as well as the possibility that H1+/− has a pleiotropic influence on behavioral (e.g., cognitive, antidepressant) and neurochemical responses to drugs targeting other membrane transporters, such as NET and SERT. For example, phosphorylation of the RBP hnRNP K increases expression of SERT protein via changes in binding to the distal polyadenylation element of the transporter (Yoon et al., 2013).
The twofold increase in hnRNP H protein in the striatal synaptosome of H1+/− mice with no change in total hnRNP H protein was surprising. This finding was observed in multiple replication studies and suggests a redistribution of hnRNP H protein to the synapse in H1+/− mice. We performed LC-MS/MS analyses on striatal synaptosomes following 2 mg/kg MA (i.p.) to further understand the effect of increased synaptosomal hnRNP H on global changes in protein expression and the underlying cell biological mechanisms. We identified a higher abundance of several mitochondrial proteins, in particular, complex I of the mitochondrial respiratory chain in H1+/− mice. The mammalian complex I consists of 38 nuclear DNA-encoded subunits (Sharma et al., 2009), and our proteomic analysis identified 8 of the 38 subunits that showed higher expression in H1+/− mice (Figs. 21–23). Proteomics has been widely used to study the effects of MA on protein expression in the brain tissues of animals (Liao et al., 2005; Faure et al., 2009; Bosch et al., 2015). Iwazaki et al. (2006) used two-dimensional gel electrophoresis proteomics and found that a single low dose of MA (1 mg/kg) administration in rats induced differential expression of proteins involved in mitochondria/oxidative metabolism. Furthermore, chronic exposure of 1 mg/kg MA induced locomotor sensitization and neurotoxicity along with a downregulation of numerous striatal proteins indicating mitochondrial dysfunction and an oxidative response (Iwazaki et al., 2007; Chin et al., 2008).
Recent studies showed that postsynaptic dendritic mitochondrial fission and fusion processes mediate cellular and behavioral plasticity, spine and synapse formation, and synaptic function (Li et al., 2004; Oettinghaus et al., 2016; Chandra et al., 2017; Divakaruni et al., 2018). Dynamin-related protein (Drp1), a GTPase involved in mitochondria fission, has been shown to regulate addiction-relevant behavior during early cocaine abstinence (Chandra et al., 2017), with inhibition of mitochondrial fission blunting cocaine-seeking and locomotor sensitization. While the results of our synaptosomal proteome dataset did not reveal any mitochondrial fission and fusion mediators, it is still possible that changes in mitochondrial proteins in the postsynaptic dendrites contribute to behavioral differences. Future studies will isolate potential presynaptic versus postsynaptic mechanisms.
Most mitochondrial proteins are nuclear-encoded and must be transported to mitochondria for organelle-coupled translation (Williams et al., 2014). RBPs play a critical role in targeting mRNAs to membrane-bound organelles, such as mitochondria. RBPs interact with mRNAs and chaperone them toward mitochondrial outer membranes where they are translated (Gerber et al., 2004; García-Rodríguez et al., 2007; Eliyahu et al., 2010; Gehrke et al., 2015). RBPs recognize and bind mitochondria-targeting RNA elements to form higher-order units called mRNA ribonucleoprotein (RNP) complexes consisting of mRNAs and associated RBPs (Béthune et al., 2019; Rossoll and Bassell, 2019). The robust increase in hnRNP H protein in the striatal synaptosome accompanied by the increase in several complex I subunits suggests a novel function for hnRNP H in targeting mRNAs encoding for subunits of mitochondrial complex I to the mitochondria, thus regulating local translation (Fig. 24). Once the mRNAs are transported nearby the mitochondria, hnRNP H could coordinate with other RBPs to form a ribonucleoprotein (RNP) complex to stabilize mRNAs to the mitochondrial membrane where translational activators initiate translation. An example of an RBP with such function is Puf3pm, which binds to ATP2 mRNA encoding mitochondrial components of the F1F0 ATPase and localizes to the outer mitochondrial membrane (García-Rodríguez et al., 2007). Future studies involving cross-linking immunoprecipitation combined with RNA-seq (CLIP-seq) will identify target mRNAs bound to hnRNP H and determine the degree of enrichment for mRNAs encoding mitochondrial complex I subunits.
Mitochondria are abundant at the synapse where they generate ATP for Ca2+ buffering, vesicle release, and recycling (Vos et al., 2010; Devine and Kittler, 2018). Mitochondria consist of five oxidative phosphorylation complexes (I-V) (Sharma et al., 2009; Mimaki et al., 2012). Complex I is the first enzyme of the respiratory chain and initiates electron transport continuing to complex II-IV to generate redox energy for Complex V to produce ATP. The increase in Complex I subunits in H1+/− mice could increase Complex I activity and synaptic ATP production to support vesicle fusion, DA release, and DA transport back into the cells via DAT and Na+/K+ ATPase to counteract MA-induced DA release. We found a Genotype × Treatment effect on protein levels of F0F1 ATP synthase subunits (Atp5a1, ATP5f1, and Atp5o) of Complex V (Fig. 22) in the synaptosomal proteome whereby MA decreased these ATP synthase subunits in H1+/− mice, which could decrease ATP production in response to MA in H1+/− mice and affect extracellular DA levels.
In addition to binding and targeting RNAs to the mitochondria for translation, hnRNP H, like other RBPs, could bind and target proteins via its glycine-rich domain in an activity-dependent manner (Tiruchinapalli et al., 2008). In an animal model for frontotemporal dementia, ploy(GR) aggregates (resulting from hexanucleotide repeats in C9ORF72) bind to mitochondrial complex V protein ATP5A1 to increase its ubiquitination and degradation through the proteasome pathway, thus disrupting mitochondrial function (Choi et al., 2019). The higher level of synaptosomal hnRNP H protein in H1+/− mice could increase binding to mitochondrial complex I and V proteins (e.g., via the glycine-rich domain) to prevent degradation, yielding higher protein levels at baseline (Figs. 21, 22). MA administration could then decrease hnRNP H–protein interactions in H1+/− mice, thus decreasing synaptic mitochondrial proteins and synaptic function.
Together, the opposing effects of MA treatment on synaptic abundance of mitochondrial complex I proteins between H1+/− and WT mice could represent a mechanism underlying blunted MA-induced DA release in H1+/− mice. Future studies will focus on the interaction between hnRNP H and mRNA encoding mitochondrial complex I and V subunits at the protein–RNA and protein–protein level (presynaptically and postsynaptically) and determine whether disruption of such interactions can alter ATP production and DA release.
Footnotes
This work was supported by National Institutes of Health/National Institute on Drug Abuse Grants R00DA029635 to C.D.B., R01DA039168 to C.D.B., F31DA40324 to N.Y., and N01DA-14-7788; National Institutes of Health/National Institute of General Medical Sciences Grant T32GM008541 to N.Y., Q.T.R., and J.A.B.; and Burroughs Wellcome Fund Transformative Training Program in Addiction Science Grant 1011479 to N.Y., Q.T.R., and J.A.B. We thank Dr. Haley Melikian for providing guidance on the dopamine transporter immunoblot; and Drs. David Moody, Wenfang B. Fang, Olga Averin, and Jonathan Crites for providing the analytical service for measuring methamphetamine concentration in our biological samples.
The authors declare no competing financial interests.
- Correspondence should be addressed to Camron D. Bryant at camron{at}bu.edu