Histamine‐induced plasticity and gene expression in corticostriatal pathway under hyperammonemia

Abstract Aims Histamine H3 receptor (H3R) antagonists/inverse agonists increase vigilance. We studied brain histaminergic pathways under hyperammonemia and the transcriptome of receptors and their signaling cascades to provide a rationale for wake‐promoting therapies. Methods We analyzed histamine‐induced long‐lasting depression of corticostriatal synaptic transmission (LLDhist). As the expression of dopamine 1 receptors (D1R) is upregulated in LGS‐KO striatum where D1R‐H3R dimers may exist, we investigated actions of H3R and D1R agonists and antagonists. We analyzed transcription of selected genes in cortex and dorsal striatum in a mouse model of inborn hyperammonemia (liver‐specific glutamine synthetase knockout: LGS‐KO) and compared it with human hepatic encephalopathy. Results LGS‐KO mice showed significant reduction of the direct depression (DD) but not the long‐lasting depression (LLD) by histamine. Neither pharmacological activation nor inhibition of D1R significantly affected DDhist and LLDhist in WT striatum, while in LGS‐KO mice D1R activation suppressed LLDhist. Histaminergic signaling was found unchanged at the transcriptional level except for the H2R. A study of cAMP‐regulated genes indicated a significant reduction in the molecular signature of wakefulness in the diseased cortex. Conclusions Our findings provide a rationale for the development of aminergic wake‐promoting therapeutics in hyperammonemic disorders.

narcolepsy 2 which is characterized by orexin-deficiency. Histamine and orexins are responsible for the complementary and synergistic control of wakefulness. 3 Further clinical studies are on their way testing the possible benefit of H3R antagonists in other neurodegenerative diseases characterized by fatigue and day-time sleepiness.
Although not accompanied by neurodegeneration, hyperammonemic conditions are associated with fatigue especially during extreme exercise or liver pathology. 4,5 Thus, patients suffering from hyperammonemia may benefit from symptomatic treatment with H3R antagonists/inverse agonists.
Histamine is an important regulator of a variety of brain functions including attention, motor activity, and learning. 1 Histaminergic neurons are silent during sleep and the histamine synthesizing enzyme, histidine decarboxylase, shows maximal expression during active periods of the day. 6 Data from patients with hepatic encephalopathy (HE) and from animal experiments demonstrate significant alterations in the striatal histamine content, expression of histamine receptors [7][8][9] and in the modulation of corticostriatal neurotransmission by histamine. 10 In the rodent striatum, histamine negatively modulates excitatory glutamatergic transmission from cortical and thalamic afferents acting through presynaptic H3 heteroreceptors. 11,12 Postsynaptic excitatory H2R mediates depolarization of striatal principle cells in the mouse by ca 7 mV. 12 Through cAMP and PKA-signaling, the H2R is known to trigger long-term potentiation of synaptic transmission, neuronal excitability, 1,13 and oscillations 14 in the hippocampus which impacts learning and memory. These actions contribute to the wake-promoting function of histamine. Activation of presynaptic H3Rs not only negatively modulates glutamatergic inputs, but also reduces lateral inhibition between striatal projection neurons, 12 thus improving signal/noise ratio during wakefulness.
Moreover, H3Rs regulate striatal dopamine release 15 and dopamine receptor-mediated signaling. 16 It is suggested that histamine-dopamine interactions are important in several motor disorders, because in both Parkinson's disease and its experimental rat models the H3R expression and radioligand binding are altered in substantia nigra and caudate putamen. [17][18][19] Some rare cases of Parkinsonism-like disorders result from liver failure or liver dysfunction, thus classifying hepatic encephalopathy (HE) as a "basal ganglia disorder". 20 Some HE patients indeed profit from L-Dopa medication. 20 However, the hypodopaminergic phenotype of HE is not a general rule and a recent analysis of clinical studies did not recommend the use of dopaminergic agents in HE. 21 Dopamine-histamine interaction within the basal ganglia under hyperammonemia and in HE remains to be characterized. Our earlier studies showed a reduction of long-lasting depression in response to histamine (LLDhist) in rats with portacaval anastomosis, 10 but no change in LLDhist after the treatment of in vitro slices with ammonium chloride (5 mM) for 4-8 hours. 22 The symptoms of hyperammonemia alone do not reflect the symptoms of HE sufficiently well.
Thus, in addition to hyperammonemia, elevated bile acids and glutamine levels in the blood of HE patients may play a role for the brain pathology. 23,24 In accordance, a more severe phenotype is achieved by combining hyperammonemic diet with bile duct ligation in rats. 25 A recently developed mouse model displays hyperammonemia without changes in further blood components such as amino acids or inflammatory mediators. 26 We reported that these mice with liverspecific deletion of glutamine synthetase ("LGS-KO") show reduced exploratory activity in the open field and delayed habituation to a novel environment which goes along with alterations in glutamatereceptor dependent synaptic plasticity in striatum and hippocampus. 27 In addition, we found upregulated expression and function of the D1 dopamine receptor (D1R), but not the D2 dopamine receptor (D2R) in the striatum of LGS-KO mice. 27 As H3R can associate with D1R or/and D2R in the striatum building heterodimers with altered pharmacology, 28,29 we compare now histamine-mediated plasticity in the corticostriatal pathway of WT and LGS-KO mice and the modulatory action of D1R activation/inhibition. We analyze the transcriptome related to histaminergic neurotransmission in the striatum of WT and LGS-KO mice.
We discuss our results in comparison with previously published 30,31 cortical transcriptome data, obtained from patients with hepatic encephalopathy.

| Animals
Male liver-specific GS knockout (LGS-KO) mice and their wild-type littermates (WT) at the age from 2 to 4 months were used. Gene-targeted mice lacking functional hepatic GS were obtained and genotyped as described previously. 26 In some experiments for the control of action of H3R-ligands, we used histaminergic neurons from Tmt-HDC mice. 32 Slice preparation and cell-attached voltage clamp recordings from histaminergic neurons are described in detail in De Luca et al. 33 Mice were kept on a 12 hours day-12 hours night light schedule with ad libitum access to food and water. All procedures were in compliance with the guidelines for the use of experimental animals, as given by the Directive 2010/63/EU of the European Parliament, the German "Tierschutzgesetz" (animal protection law) and approved by the local authorities (LANUV NRW: Landesamt für Umwelt, Natur und Verbraucherschutz Nordrhein Westfalen, Bezirksregierung Düsseldorf; permission number O58/91). All efforts were made to minimize the number of animals and their suffering.

| Real-time RT-PCR
Striatal tissues were isolated from 1 to 3 horizontal brain slices, and total cellular mRNA was extracted using an mRNA isolation kit (Quickprep Micro mRNA Purification Kit, GE Health care, GB).
Real-time RT-PCR was used to detect genotype-related alterations in gene expression. Detailed description of the applied protocol has been presented previously. 27 Primer sequences used for the amplification of histamine receptors (H1R and H3R) and HNMT (histamine N-methyl transferase) were published previously. 32,34 Primers for the amplification of mouse H2R were as follows: up:5′-GGCCAAGAAGTGAGTGTAGA-3′and lo: 5′-GAAGAGGTTGAGGATGGAAG-3′ (as in 35 ; expected PCR product size: 366 b.p.). Primers for the organic cation transporter 3 (OCT 3) were the same as in our previous study. 33 Primers for mouse Homer1A, 36 ARC, EGR1, and EGR2 37 are listed in Table 1. Several house keeping genes (HKG): GAPDH, RPl13a, beta-actin (actB), and Hsp90 27 were compared in our initial experiments. Using selection criteria of Schmittgen and Livak (highest p value obtained from group comparison, unpaired t test), we selected RPl13a as HKG in experiments with mouse samples. Homer1A, ARC, EGR1, and EGR2 mRNA levels were quantified in total RNA preparations from postmortem human brain tissue from the European cohort by real-time PCR as described recently. 38 RNA quality was verified using the Agilent 2100 Bioanalyzer (Agilent Technologies), SYBR ® Green qPCR was performed on a ViiA7 real-time PCR system (Applied Biosystems).
Primers for the amplification were designed to match, when possible, the position of the mouse primers (see Table 1). Relative mRNA levels of a gene were estimated by the "2 −ΔΔCt " method using human beta-actin as a HKG. 38 All Real-time PCR reactions were validated by gel electrophoresis and by sequencing of selected products. The relative mRNA level encoding each gene in relation to HKG was estimated by the "2 −ΔΔCt " method. Average 2 −ΔΔCt values for the WT genotype were taken as 1.0, and the individual values were expressed relative to this value. The fold change (FC) of mRNA according to the "2 −Ct " method was calculated in each PCR run where cDNA of the same quantity (eg 100 ng in 1 µL) from at least 3 WT and 3 LGS-KO mice was used as a template. Averages of two replicates per mouse were expressed as 2 −Ct and normalized on the average of all 2 −Ct values in the WT group. 39

| Agilent microarray analysis
Expression levels of selected genes were extracted from data sets acquired in two earlier studies by Agilent™ whole human genome microarray analysis. 30,31 We based our search on the gene list from the Synaptic plasticity RT 2 Profiler PCR array (Qiagen, Cat. no.PAMM-126Z). Among 84 genes, we selected 18, which are expressed and characterized in the mouse and are related to cAMP/ PKA-dependent signal cascades. We added to this list histamine and dopamine receptors and their signaling pathways. For microarray analysis, RNA was isolated from human postmortem brain tissue using a commercial kit (RNeasy mini kit, Qiagen) according to the instructions of the manufacturer. Agilent microarray analysis was performed by Miltenyi-Biotech (Bergisch-Gladbach, Germany) as described in. 30,31 Data sets were deposited at the public genomic data repository "Gene Expression Omnibus" (GEO, accession no. GSE41919 and GSE57193) from the National Center for Biotechnology Information (NCBI). The "European cohort" of patients with liver cirrhosis and HE consisted of postmortem human brain tissue taken from the intersection parietal to occipital cortex area. Tissue was provided by the body donor program of the Department of Anatomy of the University of Düsseldorf, Germany. For detailed information on the microarray analysis and the patient characteristics see. 30 Statistical group analysis was performed using Student's t test with equal variances. In the "Australian cohort" postmortem brain tissue was taken from the fusiform gyrus from four control subjects and four patients with liver cirrhosis with HE. 31 Tissue was provided by the Australian Brain Donor Programs NSW Tissue Resource Centre. Statistical analysis was performed using analysis of variance (ANOVA) and Tukey's multiple comparison post hoc test.

| Electrophysiology
Horizontal slices (400 μm thick) were prepared with a vibratome as described previously. 27 After at least two hours preincubation at room temperature, a single slice was transferred to a recording chamber perfused with artificial cerebrospinal fluid (aCSF) at a flow rate of 1.5-2 mL/min at 32°C. Cortical fibers to the striatum were stimulated as previously described. 27 After the initial testing of stimulus-response relationships, the stimulus intensity was

TA B L E 1 Primers used for amplification of human (h) or mouse (m) transcripts
adjusted to induce a field response of approximately 60%-70% of its maximal amplitude and the stimulation frequency was set to 0.033 Hz. Each experiment included a 15-20 minutes period of control recording, application of chemical stimulus, and 60-90 minutes of poststimulus period.
Signals were amplified, digitized at 10 kHz, stored on a hard disk of a PC using Clampex software of pClamp (Axon Instruments), and analyzed off-line, using Clampfit and Excel software. Ten consecutive field responses (5-minutes recordings) were averaged, the amplitude of corticostriatal postsynaptic peak response (N2 peak) was measured as in Chepkova et al 27  Differences were considered statistically significant when P < .05.

| Drugs and chemicals
The following substances were obtained from Tocris: Sigma/RBI. Histamine was applied at 10 µM as previous studies demonstrated that this is an effective concentration to be studied in vitro. 11,12,14,40 To block H3R in slice recordings, previous studies used thioperamide 10 µM. 11,12 As this drug shows several additional activities, 41 we decided to use another H3R antagonist, clobenpropit, which is more potent than thioperamide in vitro but less potent when applied in vivo due to the poor blood-brain permeability. 42 At 20 µM, clobenpropit was able to block maximal R-(alpha)-methylhistamine (RAMH, selective H3R agonist) responses in histaminergic neurons (see Figure 1A). Furthermore, we compared the action of the RAMH (2 µM) with the action of imetit (3 µM) and found no difference in the amplitude of response ( Figure 1B). Concentrations of dopamine receptor modulators were chosen in accordance with previous studies in mouse. 43,44 All substances were diluted and stored as recommended by the provider. Working solutions were freshly prepared immediately before application.

| RE SULTS
Application of 10 μM histamine to slice preparations from WT mice significantly inhibited corticostriatal neurotransmission in the vast major-   2 and 3). Histamine-induced long-lasting depression (LLDhist) was not significantly affected by inhibition of the cAMP-dependent protein kinase (PKA) whose activity is known to be regulated through D1R ( Figure 3C,D). We noticed that PKA inhibition by KT5720 (1 µM) delayed the DDhist maximum. To analyze this effect, we calculated the slope of onset with a linear regression.
All data points were scaled to the control and only experiments with inhibition larger than 20% were considered. Slopes of control responses and responses in the presence of KT5720 differed significantly ( Figure 3E). As direct depolarization of striatal medium spiny neurons by histamine through H2R was previously reported 12 and our experiments with KT5720 indicated a H2R (PKA)-dependent component, we applied the selective H3R ligand imetit (3 µM), which showed the same action as RAMH on histaminergic neurons ( Figure 1). Imetit-induced direct depression of corticostriatal transmission was significantly smaller compared to the response to histamine 10 µM ( Figure 3F) whereas long-lasting depression did not differ. Thus, LLDhist is solely dependent on H3R, whereas DDhist involves H2R and H3R activation.
LGS-KO mice displayed a significantly less pronounced DDhist than that found in the WT striatum ( Figure 4A did not affect DDhist but significantly suppressed the maintenance of LLDhist ( Figure 4C,D). We did not detect differences between LGS-KO and WT mice either in inhibition of firing of histaminergic neurons by RAMH or in effects of clobenpropit 20 µM ( Figure 5).
Transcription of several genes related to the histaminergic pathway was compared between WT and LGS-KO striatum. Using the "2 − ΔΔCt " method, we found no significant difference in mRNA levels for the histamine receptors H1, H2 and H3 (HRH1, HRH2, HRH3 nomenclature is used for the human and H1R, H2R, H3R for the mouse brain), Dunn's multiple comparison test, Figure 6C). The same analysis performed with mouse cortical samples did not reveal any difference in transcripts encoding for histamine receptors ( Figure 6D).
In search for further gene transcripts which might be affected by hyperammonemia and involved in histamine-or dopamine-mediated synaptic plasticity, we screened for relevant genes in the cortical transcriptome of HE patients postmortem. We found similarities between Australian and European cohorts in the reduced expression of ARC (activity-regulated cytoskeleton-association protein), EGR1 (early growth response protein 1), and EGR4 but did not detect changes in HNMT (histamine N-methyltransferase: a correlate of histaminergic activity 46 ). Transcript levels of HRH2, HOMER1, BDNF (brain derived neurotrophic factor), EGR2, and DRD5 (dopamine receptor D5) were significantly reduced in Australian but not in European HE patients ( Figure 7A,B). Our real-time PCR validation of whole transcriptome data 30 showed differences similar to those obtained with microarray analysis ( Figure 8A), except for Homer1A, whose down-regulation detected with our primers (Table 1)

| D ISCUSS I ON
This study shows that mice with inborn hyperammonemia (LGS-KO) display lower sensitivity of corticostriatal neurotransmission  Expression changes of genes related to dopamine and histamine signaling in postmortem brain tissue from the cerebral cortex of patients with liver cirrhosis and HE. Gene expression changes were measured by microarray analysis in two independent patient cohorts: one from Europe (A), gene expression was analyzed with Student t test 30 ; and another from Australia (B), gene expression was analyzed with one-way ANOVA due to the study design. 31 In A and B, gene expression changes are given relative to the respective controls. *P < .05; #P ≤ .01; $P ≤ .001 kinase A inhibitor KT5720 (200nM). 49 We took advantage of the slightly upregulated D1R expression in LGS-KO mice 27  Upregulation of mRNA encoding for the early growth response proteins (EGR1, EGR2) was previously reported in mouse cortex in response to stress 52 or weaning 53 and down-regulation in response to metamphetamine abuse. 37 Lozeva et al 9 reported increased histamine levels in postmortem cortex of HE patients and increased occupancy of H3R. We did not detect in mouse or man brain tissue samples changes in the expression of the histamine degrading enzyme HNMT, whose level correlates with the histaminergic activity in the brain. 46 We found no changes in histamine receptor expression except for the down-regulation of cAMP-stimulating H2R in Australian HE patients and in the mouse striatum (with "2 −Ct " method). Although we obtained only cortical samples from HE patients, mouse brain samples showed region-specific gene alterations under hyperammonemia, which may play a role for the neurological symptoms. Taking into consideration subtle F I G U R E 8 Real-time PCR analysis of "molecular signature of wakefulness" in human cortex postmortem (A) and mouse (B) cortex and striatum of WT and LGS-KO mice. A, Same samples as for microarray experiments shown in Figure 7A were amplified and analyzed with the "2 −ΔΔCt " method, using beta-actin expression as a reference. Note that results similar to those from gene array were obtained. *P = .05, **P = .01 (As values in each group fulfilled criteria of normal distribution (Kolmogorov-Smirnov normality test), the unpaired t test was applied for the group comparison). B, Transcriptional changes in LGS-KO mouse cortex (upper plot) or dorsal striatum (lower plot) in comparison to WT (significant difference is indicated as *P < .05 (MWT) changes in histamine-plasticity and gene expression in the histaminergic system of hyperammonemic mice and in the end-stage of human HE as well as unimpaired H3R signaling, H3R antagonists/ inverse agonists can be suggested as possible vigilance-increasing therapeutics.
Wake-promoting medication with H3R antagonism is indicated to combat day-time sleepiness. We show that in inborn hyperammonemia accompanied by fatigue and motor disturbances, transcription of H3R, and histamine-induced long-lasting depression (LLDhist) are not changed. We demonstrate a misbalance between modulatory functions of histamine (decreases) and dopamine (increases through D1R) in mice with inborn hyperammonemia. These findings provide a rationale for the development of therapies directed toward aminergic neurotransmission.

ACK N OWLED G M ENTS
This study was supported by the German Research Foundation

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest. All applicable international, national, and institutional guidelines for the care and use of animals were followed.