Dopaminergic neurotransmission in the nucleus accumbens is important for various reward‐related cognitive processes including reinforcement learning. Repeated cocaine enhances hippocampal synaptic plasticity, and phasic elevations of accumbal dopamine evoked by unconditioned stimuli are dependent on impulse flow from the ventral hippocampus. Therefore, sensitized hippocampal activity may be one mechanism by which drugs of abuse enhance limbic dopaminergic activity. In this study, in vivo microdialysis in freely moving adult male Sprague–Dawley rats was used to investigate the effect of repeated cocaine on ventral hippocampus‐mediated dopaminergic transmission within the medial shell of the nucleus accumbens. Following seven daily injections of saline or cocaine (20 mg/kg, ip), unilateral infusion of N‐methyl‐d ‐aspartate (NMDA, 0.5 μg) into the ventral hippocampus transiently increased both motoric activity and ipsilateral dopamine efflux in the medial shell of the nucleus accumbens, and this effect was greater in rats that received repeated cocaine compared to controls that received repeated saline. In addition, repeated cocaine altered NMDA receptor subunit expression in the ventral hippocampus, reducing the NR2A : NR2B subunit ratio. Together, these results suggest that repeated exposure to cocaine produces maladaptive ventral hippocampal‐nucleus accumbens communication, in part through changes in glutamate receptor composition.
Cocaine is a recreational drug of abuse that binds to the dopamine transporter, preventing reuptake of dopamine into pre‐synaptic terminals. The increased presence of synaptic dopamine results in stimulation of both pre‐ and post‐synaptic dopamine receptors, considered an important mechanism by which cocaine elicits its reinforcing properties. However, the effects of acute cocaine administration on pre‐synaptic dopamine function remain unclear. Non‐invasive imaging techniques such as positron emission tomography have revealed impaired pre‐synaptic dopamine function in chronic cocaine users. Similar impairments have been seen in animal studies, with microdialysis experiments indicating decreased basal dopamine release. Here we use micro positron emission tomography imaging techniques in mice to measure dopamine synthesis capacity and determine the effect of acute cocaine administration of pre‐synaptic dopamine function. We show that a dose of 20 mg/kg cocaine is sufficient to elicit hyperlocomotor activity, peaking 15–20 min post treatment (p < 0.001). However, dopamine synthesis capacity in the striatum was not significantly altered by acute cocaine treatment (: 0.0097 per min vs. 0.0112 per min in vehicle controls, p > 0.05). Furthermore, expression levels of two key enzymes related to dopamine synthesis, tyrosine hydroxylase and aromatic l ‐amino acid decarboxylase, within the striatum of scanned mice were not significantly affected by acute cocaine pre‐treatment (p > 0.05). Our findings suggest that while the regulation of dopamine synthesis and release in the striatum have been shown to change with chronic cocaine use, leading to a reduced basal tone, these adaptations to pre‐synaptic dopaminergic neurons are not initiated following a single exposure to the drug.
Ghrelin is an appetite‐stimulating peptide. Serine 3 on ghrelin must be acylated by octanoate via the enzyme ghrelin‐O‐acyltransferase (GOAT) for the peptide to bind and activate the cognate receptor, growth hormone secretagogue receptor type 1a (GHSR1a). Interest in GHSR1a increased dramatically when GHSR1a mRNA was demonstrated to be widespread in the brain, including the cortex and hippocampus, indicating that it has multifaceted functions beyond the regulation of metabolism. However, the source of octanoylated ghrelin for GHSR1a in the brain, outside of the hypothalamus, is not well understood. Here, we report the presence of GOAT and its ability to acylate non‐octanoylated ghrelin in the hippocampus. GOAT immunoreactivity is aggregated at the base of the dentate granule cell layer in the rat and wild‐type mouse. This immunoreactivity was not affected by the pharmacological inhibition of GHSR1a or the metabolic state‐dependent fluctuation of systemic ghrelin levels. However, it was absent in the GHSR1a knockout mouse hippocampus, pointing the possibility that the expression of GHSR1a may be a prerequisite for the production of GOAT. Application of fluorescein isothiocyanate (FITC)‐conjugated non‐octanoylated ghrelin in live hippocampal slice culture (but not in fixed culture or in the presence of GOAT inhibitors) mimicked the binding profile of FITC‐conjugated octanoylated ghrelin, suggesting that extracellularly applied non‐octanoylated ghrelin was acylated by endogenous GOAT in the live hippocampus while GOAT being mobilized out of neurons. Our results will advance the understanding for the role of endogenous GOAT in the hippocampus and facilitate the search for the source of ghrelin that is intrinsic to the brain.
Kiss1, a neuropeptide predominantly expressed in the habenula, modulates the serotonin (5‐HT) system to decrease odorant cue [alarm substance (AS)]‐evoked fear behaviour in the zebrafish. The purpose of this study was to assess the interaction of Kiss1 with the 5‐HT system as well as to determine the involvement of the 5‐HT receptor subtypes in AS‐evoked fear. We utilized 0. 28 mg/kg WAY 100635 (WAY), a selective 5‐HT1A receptor antagonist, to observe the effects of Kiss1 administration on AS‐evoked fear. We found WAY significantly inhibited the anxiolytic effects of Kiss1 (p <0.001) with an exception of freezing behaviour. Based on this, we utilized 92.79 mg/kg methysergide, a 5‐HT1 and 5‐HT2 receptor antagonist, and found that methysergide significantly blocked the anxiolytic effects of Kiss1 in the presence of the AS (p <0.001). From this, we conclude that Kiss1 modulates AS‐evoked fear responses mediated by the 5‐HT1A and 5‐HT2 receptors.
Amphetamine is a central nervous system psychostimulant with a high potential for abuse. Recent literature has shown that genetic and drug‐induced elevations in dopamine transporter (DAT) expression augment the neurochemical and behavioral potency of psychostimulant releasers. However, it remains to be determined if the well‐documented differences in DAT levels across striatal regions drive regionally distinct amphetamine effects within individuals. DAT levels and dopamine uptake rates have been shown to follow a gradient in the striatum, with the highest levels in the dorsal regions and lowest levels in the nucleus accumbens shell; thus, we hypothesized that amphetamine potency would follow this gradient. Using fast scan cyclic voltammetry in mouse brain slices, we examined DAT inhibition and changes in exocytotic dopamine release by amphetamine across four striatal regions (dorsal and ventral caudate‐putamen, nucleus accumbens core and shell). Consistent with our hypothesis, amphetamine effects at the DAT and on release decreased across regions from dorsal to ventral, and both measures of potency were highly correlated with dopamine uptake rates. Separate striatal subregions are involved in different aspects of motivated behaviors, such as goal‐directed and habitual behaviors, that become dysregulated by drug abuse, making it critically important to understand regional differences in drug potencies.
The attribution of incentive salience to reward‐predictive stimuli has been shown to be associated with substance abuse‐like behavior such as increased drug taking. Evidence suggests that glutamate neurotransmission and sequential N‐methyl‐D‐aspartate (NMDA) activation are involved in the attribution of incentive salience. Here, we further explore the role of second‐by‐second glutamate neurotransmission in the attribution of incentive salience to reward‐predictive stimuli by measuring sign‐tracking behavior during a Pavlovian conditioned approach procedure using ceramic‐based microelectrode arrays configured for sensitive measures of extracellular glutamate in awake behaving Sprague‐Dawley rats. Specifically, we show that there is an increase in extracellular glutamate levels in the prelimbic cortex (PrL) and the nucleus accumbens core (NAcC) during sign‐tracking behavior to a food‐predictive conditioned stimulus (CS+) compared to the presentation of a non‐predictive conditioned stimulus (CS?). Furthermore, the results indicate greater increases in extracellular glutamate levels in the PrL compared to NAcC in response to the CS+, including differences in glutamate release and signal decay. Taken together, the present research suggests that there is differential glutamate signaling in the NAcC and PrL during sign‐tracking behavior to a food‐predictive CS+.
The psychostimulant amphetamine (AMPH) is frequently used to increase catecholamine levels in attention disorders and positron emission tomography imaging studies. Despite the fact that most radiotracers for positron emission tomography studies are characterized in non‐human primates (NHPs), data on regional differences of the effect of AMPH in NHPs are very limited. This study examined the impact of AMPH on extracellular dopamine (DA) levels in the medial prefrontal cortex and the caudate of NHPs using microdialysis. In addition to differences in magnitude, we observed striking differences in the temporal profile of extracellular DA levels between these regions that can likely be attributed to differences in the regulation of dopamine uptake and biosynthesis. The present data suggest that cortical DA levels may remain elevated longer than in the caudate which may contribute to the clinical profile of the actions of AMPH.
Tolerance to the neurochemical and psychoactive effects of cocaine after repeated use is a hallmark of cocaine addiction in humans. However, comprehensive studies on tolerance to the behavioral, psychoactive, and neurochemical effects of cocaine following contingent administration in rodents are lacking. We outlined the consequences of extended access cocaine self‐administration as it related to tolerance to the psychomotor activating, dopamine (DA) elevating, and DA transporter (DAT) inhibiting effects of cocaine. Cocaine self‐administration (1.5 mg/kg/inj; 40 inj; 5 days), which resulted in escalation of first hour intake, caused reductions in evoked DA release and reduced maximal rates of uptake through the DAT as measured by slice voltammetry in the nucleus accumbens core. Furthermore, we report reductions in cocaine‐induced uptake inhibition and a corresponding increase in the dose of cocaine required for 50% inhibition of DA uptake (Ki) at the DAT. Cocaine tolerance at the DAT translated to reductions in cocaine‐induced DA overflow as measured by microdialysis. In addition, cocaine‐induced elevations in locomotor activity and stereotypy were reduced, while rearing behavior was enhanced in animals with a history of cocaine self‐administration. Here, we demonstrate both neurochemical and behavioral cocaine tolerance in an extended‐access rodent model of cocaine abuse, which allows for a better understanding of the neurochemical and psychomotor tolerance that develops to cocaine in human addicts.
We have previously shown that the selective sigma‐1 receptor (σ1R) antagonist S1RA (E‐52862) inhibits neuropathic pain and activity‐induced spinal sensitization in various pre‐clinical pain models. In this study we characterized both the behavioral and the spinal neurochemical effects of S1RA in the rat formalin test. Systemic administration of S1RA produced a dose‐related attenuation of flinching and lifting/licking behaviors in the formalin test. Neurochemical studies using concentric microdialysis in the ipsilateral dorsal horn of awake, freely moving rats revealed that the systemic S1RA‐induced antinociceptive effect occurs concomitantly with an enhancement of noradrenaline levels and an attenuation of formalin‐evoked glutamate release in the spinal dorsal horn. Intrathecal pre‐treatment with idazoxan prevented the systemic S1RA antinociceptive effect, suggesting that the S1RA antinociception depends on the activation of spinal α2‐adrenoceptors which, in turn, could induce an inhibition of formalin‐evoked glutamate release. When administered locally, intrathecal S1RA inhibited only the flinching behavior, whereas intracerebroventricularly or intraplantarly injected also attenuated the lifting/licking behavior. These results suggest that S1RA supraspinally activates the descending noradrenergic pain inhibitory system, which may explain part of its antinociceptive properties in the formalin test; however, effects at other central and peripheral sites also account for the overall effect.
The functional roles of the orphan nuclear receptor, Nurr1, have been extensively studied and well established in the development and survival of midbrain dopamine neurons. As Nurr1 and other NR4A members are widely expressed in the brain in overlapping and distinct manners, it has been an open question whether Nurr1 has important function(s) in other brain areas. Recent studies suggest that up‐regulation of Nurr1 expression is critical for cognitive functions and/or long‐term memory in forebrain areas including hippocampal formation. Questions remain about the association between Nurr1 expression and Alzheimer's disease (AD) brain pathology. Here, using our newly developed Nurr1‐selective antibody, we report that Nurr1 protein is prominently expressed in brain areas with Aβ accumulation, that is, the subiculum and the frontal cortex, in the 5XFAD mouse and that Nurr1 is highly co‐expressed with Aβ at early stages. Furthermore, the number of Nurr1‐expressing cells significantly declines in the 5XFAD mouse in an age‐dependent manner, accompanied by increased plaque deposition. Thus, our findings suggest that altered expression of Nurr1 is associated with AD progression.
Both dopamine and glutamate are critically involved in cognitive processes such as working memory. Astrocytes, which express dopamine receptors, are essential elements in the termination of glutamatergic signaling: the astrocytic glutamate transporter GLT‐1 is responsible for > 90% of cortical glutamate uptake. The effect of dopamine depletion on glutamate transporters in the prefrontal cortex (PFC) remains unknown. In an effort to determine if astrocytes are a locus of cortical dopamine–glutamate interactions, we examined the effects of chronic dopamine denervation on PFC protein and mRNA levels of glutamate transporters. PFC dopamine denervation elicited a marked increase in GLT‐1 protein levels, but had no effect on levels of other glutamate transporters; high‐affinity glutamate transport was positively correlated with the extent of dopamine depletion. GLT‐1 gene expression was not altered. Our data suggest that dopamine depletion may lead to post‐translational modifications that result in increased expression and activity of GLT‐1 in PFC astrocytes.
Vesicular monoamine transporter‐2 (VMAT2) inhibitors reduce methamphetamine (METH) reward in rats. The current study determined the effects of VMAT2 inhibitors lobeline (LOB; 1 or 3 mg/kg) and N‐(1,2R‐dihydroxylpropyl)‐2,6‐cis‐di(4‐methoxyphenethyl)piperidine hydrochloride (GZ‐793A; 15 or 30 mg/kg) on METH‐induced (0.5 mg/kg, SC) changes in extracellular dopamine (DA) and its metabolite dihydroxyphenylacetic acid (DOPAC) in the reward‐relevant nucleus accumbens (NAc) shell using in vivo microdialysis. The effect of GZ‐793A (15 mg/kg) on DA synthesis in tissue also was investigated in NAc, striatum, medial prefrontal cortex and orbitofrontal cortex. In NAc shell, METH produced a time‐dependent increase in extracellular DA and decrease in DOPAC. Neither LOB nor GZ‐793A alone altered extracellular DA; however, both drugs increased extracellular DOPAC. In combination with METH, LOB did not alter the effects of METH on DA; however, GZ‐793A, which has greater selectivity than LOB for inhibiting VMAT2, reduced the duration of the METH‐induced increase in extracellular DA. Both LOB and GZ‐793A enhanced the duration of the METH‐induced decrease in extracellular DOPAC. METH also increased tissue DA synthesis in NAc and striatum, whereas GZ‐793A decreased synthesis; no effect of METH or GZ‐793A on DA synthesis was found in medial prefrontal cortex or orbitofrontal cortex. These results suggest that selective inhibition of VMAT2 produces a time‐dependent decrease in DA release in NAc shell as a result of alterations in tyrosine hydroxylase activity, which may play a role in the ability of GZ‐793A to decrease METH reward.
Autonomic control of heart rate is mediated by cardioinhibitory parasympathetic cholinergic neurons located in the brainstem and stimulatory sympathetic noradrenergic neurons. During embryonic development the survival and cholinergic phenotype of brainstem autonomic neurons is promoted by brain‐derived neurotrophic factor (BDNF). We now provide evidence that BDNF regulates heart rate by a mechanism involving increased brainstem cardioinhibitory parasympathetic activity. Mice with a BDNF haploinsufficiency exhibit elevated resting heart rate, and infusion of BDNF intracerebroventricularly reduces heart rate in both wild‐type and BDNF+/? mice. The atropine‐induced elevation of heart rate is diminished in BDNF+/? mice and is restored by BDNF infusion, whereas the atenolol‐induced decrease in heart rate is unaffected by BDNF levels, suggesting that BDNF signaling enhances parasympathetic tone which is diminished with BDNF haploinsufficiency. Whole‐cell recordings from pre‐motor cholinergic cardioinhibitory vagal neurons in the nucleus ambiguus indicate that BDNF haploinsufficiency reduces cardioinhibitory vagal neuron activity by increased inhibitory GABAergic and diminished excitatory glutamatergic neurotransmission to these neurons. Our findings reveal a previously unknown role for BDNF in the control of heart rate by a mechanism involving increased activation of brainstem cholinergic parasympathetic neurons
(R)‐3‐[2,6‐cis‐Di(4‐methoxyphenethyl)piperidin‐1‐yl]propane‐1,2‐diol (GZ‐793A) inhibits methamphetamine‐evoked dopamine release from striatal slices and methamphetamine self‐administration in rats. GZ‐793A potently and selectively inhibits dopamine uptake at the vesicular monoamine transporter‐2 (VMAT2). This study determined GZ‐793A's ability to evoke [3H]dopamine release and inhibit methamphetamine‐evoked [3H]dopamine release from isolated striatal synaptic vesicles. Results show GZ‐793A concentration‐dependent [3H]dopamine release; nonlinear regression revealed a two‐site model of interaction with VMAT2 (High‐ and Low‐EC50 = 15.5 nM and 29.3 μM, respectively). Tetrabenazine and reserpine completely inhibited GZ‐793A‐evoked [3H]dopamine release, however, only at the High‐affinity site. Low concentrations of GZ‐793A that interact with the extravesicular dopamine uptake site and the High‐affinity intravesicular DA release site also inhibited methamphetamine‐evoked [3H]dopamine release from synaptic vesicles. A rightward shift in the methamphetamine concentration‐response was evident with increasing concentrations of GZ‐793A, and the Schild regression slope was 0.49 ± 0.08, consistent with surmountable allosteric inhibition. These results support a hypothetical model of GZ‐793A interaction at more than one site on the VMAT2 protein, which explains its potent inhibition of dopamine uptake, dopamine release via a High‐affinity tetrabenazine‐ and reserpine‐sensitive site, dopamine release via a Low‐affinity tetrabenazine‐ and reserpine‐insensitive site, and a low‐affinity interaction with the dihydrotetrabenazine binding site on VMAT2. GZ‐793A inhibition of the effects of methamphetamine supports its potential as a therapeutic agent for the treatment of methamphetamine abuse.
Adenosine modulates dopamine in the brain via A1 and A2A receptors, but that modulation has only been characterized on a slow time scale. Recent studies have characterized a rapid signaling mode of adenosine that suggests a possible rapid modulatory role. Here, fast‐scan cyclic voltammetry was used to characterize the extent to which transient adenosine changes modulate stimulated dopamine release (5 pulses at 60 Hz) in rat caudate–putamen brain slices. Exogenous adenosine was applied and dopamine concentration monitored. Adenosine only modulated dopamine when it was applied 2 or 5 s before stimulation. Longer time intervals and bath application of 5 μM adenosine did not decrease dopamine release. Mechanical stimulation of endogenous adenosine 2 s before dopamine stimulation also decreased stimulated dopamine release by 41 ± 7%, similar to the 54 ± 6% decrease in dopamine after exogenous adenosine application. Dopamine inhibition by transient adenosine was recovered within 10 min. The A1 receptor antagonist 8‐cyclopentyl‐1,3‐dipropylxanthine blocked the dopamine modulation, whereas dopamine modulation was unaffected by the A2A receptor antagonist SCH 442416. Thus, transient adenosine changes can transiently modulate phasic dopamine release via A1 receptors. These data demonstrate that adenosine has a rapid, but transient, modulatory role in the brain.
There are significant differences between acetyl‐CoA and ATP levels, enzymes of acetyl‐CoA metabolism, and toll‐like receptor 4 contents in non‐activated microglial N9 and non‐differentiated cholinergic SN56 neuroblastoma cells. Exposition of N9 cells to lipopolysaccharide caused concentration‐dependent several‐fold increases of nitrogen oxide synthesis, accompanied by inhibition of pyruvate dehydrogenase complex, aconitase, and α‐ketoglutarate dehydrogenase complex activities, and by nearly proportional depletion of acetyl‐CoA, but by relatively smaller losses in ATP content and cell viability (about 5%). On the contrary, SN56 cells appeared to be insensitive to direct exposition to high concentration of lipopolysaccharide. However, exogenous nitric oxide resulted in marked inhibition pyruvate dehydrogenase and aconitase activities, depletion of acetyl‐CoA, along with respective loss of SN56 cells viability. These data indicate that these two common neurodegenerative signals may differentially affect energy‐acetyl‐CoA metabolism in microglial and cholinergic neuronal cell compartments in the brain. Moreover, microglial cells appeared to be more resistant than neuronal cells to acetyl‐CoA and ATP depletion evoked by these neurodegenerative conditions. Together, these data indicate that differential susceptibility of microglia and cholinergic neuronal cells to neurotoxic signals may result from differences in densities of toll‐like receptors and degree of disequilibrium between acetyl‐CoA provision in mitochondria and its utilization for energy production and acetylation reactions in each particular group of cells.
To survive, individuals must learn to associate cues in the environment with emotionally relevant outcomes. This association is partially mediated by the nucleus accumbens (NAc), a key brain region of the reward circuit that is mainly composed by GABAergic medium spiny neurons (MSNs), that express either dopamine receptor D1 or D2. Recent studies showed that both populations can drive reward and aversion, however, the activity of these neurons during appetitive and aversive Pavlovian conditioning remains to be determined. Here, we investigated the relevance of D1- and D2-neurons in associative learning, by measuring calcium transients with fiber photometry during appetitive and aversive Pavlovian tasks in mice. Sucrose was used as a positive valence unconditioned stimulus (US) and foot shock was used as a negative valence US. We show that during appetitive Pavlovian conditioning, D1- and D2-neurons exhibit a general increase in activity in response to the conditioned stimuli (CS). Interestingly, D1- and D2-neurons present distinct changes in activity after sucrose consumption that dynamically evolve throughout learning. During the aversive Pavlovian conditioning, D1- and D2-neurons present an increase in the activity in response to the CS and to the US (shock). Our data support a model in which D1- and D2-neurons are concurrently activated during appetitive and aversive conditioning.
Long‐term nicotine exposure induces alterations in dopamine transmission in nucleus accumbens that sustain the reinforcing effects of smoking. One approach to understand the adaptive changes that arise involves measurement of endogenous dopamine release using voltammetry. We therefore treated rats for 2–3 months with nicotine and examined alterations in nAChR subtype expression and electrically evoked dopamine release in rat nucleus accumbens shell, a region key in addiction. Long‐term nicotine treatment selectively decreased stimulated α6β2* nAChR‐mediated dopamine release compared with vehicle‐treated rats. It also reduced α6β2* nAChRs, suggesting the receptor decline may contribute to the functional loss. This decreased response in release after chronic nicotine treatment was still partially sensitive to the agonist nicotine. Studies with an acetylcholinesterase inhibitor demonstrated that the response was also sensitive to increased endogenous acetylcholine. However, unlike the agonists, nAChR antagonists decreased dopamine release only in vehicle‐ but not nicotine‐treated rats. As antagonists function by blocking the action of acetylcholine, their ineffectiveness suggests that reduced acetylcholine levels partly underlie the dampened α6β2* nAChR‐mediated function in nicotine‐treated rats. As long‐term nicotine modifies dopamine release by decreasing α6β2* nAChRs and their function, these data suggest that interventions that target this subtype may be useful for treating nicotine dependence.
Taste information from type III taste cells to gustatory neurons is thought to be transmitted via synapses. However, the molecular mechanisms underlying taste transduction through this pathway have not been fully elucidated. In this study, to identify molecules that participate in synaptic taste transduction, we investigated whether complexins (Cplxs), which play roles in regulating membrane fusion in synaptic vesicle exocytosis, were expressed in taste bud cells. Among four Cplx isoforms, strong expression of Cplx2 mRNA was detected in type III taste cells. To investigate the function of CPLX2 in taste transduction, we observed taste responses in CPLX2‐knockout mice. When assessed with electrophysiological and behavioral assays, taste responses to some sour stimuli in CPLX2‐knockout mice were significantly lower than those in wild‐type mice. These results suggested that CPLX2 participated in synaptic taste transduction from type III taste cells to gustatory neurons.
Temozolomide (TMZ) has been widely used in the treatment of glioblastoma (GBM), although inherent or acquired resistance restricts the application. This study was aimed to evaluate the efficacy of sulforaphane (SFN) to TMZ‐induced apoptosis in GBM cells and the potential mechanism. Biochemical assays and subcutaneous tumor establishment were used to characterize the function of SFN in TMZ‐induced apoptosis. Our results revealed that β‐catenin and miR‐21 were concordantly expressed in GBM cell lines, and SFN significantly reduced miR‐21 expression through inhibiting the Wnt/β‐catenin/TCF4 pathway. Furthermore, down‐regulation of miR‐21 enhanced the pro‐apoptotic efficacy of TMZ in GBM cells. Finally, we observed that SFN strengthened TMZ‐mediated apoptosis in a miR‐21‐dependent manner. In conclusion, SFN effectively enhances TMZ‐induced apoptosis by inhibiting miR‐21 via Wnt/β‐catenin signaling in GBM cells. These findings support the use of SFN for potential therapeutic approach to overcome TMZ resistance in GBM treatment.
下载免费PDF全文