Dopamine dysregulation in a mouse model of paroxysmal nonkinesigenic dyskinesia

Paroxysmal nonkinesigenic dyskinesia (PNKD) is an autosomal dominant episodic movement disorder. Patients have episodes that last 1 to 4 hours and are precipitated by alcohol, coffee, and stress. Previous research has shown that mutations in an uncharacterized gene on chromosome 2q33–q35 (which is termed PNKD) are responsible for PNKD. Here, we report the generation of antibodies specific for the PNKD protein and show that it is widely expressed in the mouse brain, exclusively in neurons. One PNKD isoform is a membrane-associated protein. Transgenic mice carrying mutations in the mouse Pnkd locus equivalent to those found in patients with PNKD recapitulated the human PNKD phenotype. Staining for c-fos demonstrated that administration of alcohol or caffeine induced neuronal activity in the basal ganglia in these mice. They also showed nigrostriatal neurotransmission deficits that were manifested by reduced extracellular dopamine levels in the striatum and a proportional increase of dopamine release in response to caffeine and ethanol treatment. These findings support the hypothesis that the PNKD protein functions to modulate striatal neuro­transmitter release in response to stress and other precipitating factors.

The paroxysmal dyskinesias consist of clinically and genetically distinct phenotypes, including paroxysmal kinesigenic dyskinesia, paroxysmal exercise-induced dyskinesia, and paroxysmal nonkinesigenic dyskinesia (PNKD) (1, 2). PNKD is a highly penetrant autosomal dominant disorder in which individuals have 1- to 4-hour attacks consisting of dystonia and choreoathetosis (3). These attacks can be induced reliably by administration of caffeine or alcohol and frequently when patients are stressed. The causative gene was mapped to chromosome 2q33–q35 (4, 5), and mutations in the PNKD gene (formerly called MR-1) were subsequently identified in PNKD families (6–10).

The PNKD gene has at least 3 alternate splice forms, which encode proteins of 385, 361, and 142 amino acids. The long isoform of PNKD (PNKD-L) is specifically expressed in CNS, while the medium isoform (PNKD-M) and short isoform (PNKD-S) are ubiquitously expressed (7). Two missense mutations (Ala to Val) located at amino acids 7 or 9 of PNKD-L and PNKD-S were found in most patients, and a third mutation (Ala to Pro) at position 33 was reported in 1 family (11). Both PNKD-L and PNKD-M have a putative catalytic domain that is homologous to hydroxyacylglutathione hydrolase (HAGH), a member of the zinc metallo-hydrolase enzyme family, which contains β-lactamase domains. HAGH functions in a pathway to detoxify methylglyoxal, a by-product of oxidative stress (12). The normal role of PNKD in cells and the contribution of mutations to pathophysiology of PNKD are not known.

Dyskinesia is seen with many genetic and acquired disorders of the brain. Theoretically, such hyperkinetic movements could have their genesis in the basal ganglia, the cerebellum, or even in the cortex. Having cloned the gene and shown by in situ hybridization that it is widely expressed, we were interested in probing the pathophysiology of this fascinating disorder.

In this study, we generated polyclonal antibodies specific for detecting PNKD isoforms. We also generated WT and mutant Pnkd-transgenic and Pnkd-KO mice in order to determine whether they can recapitulate human PNKD phenotypes. Together, these reagents provided a unique opportunity to begin addressing questions regarding the pathophysiology of PNKD. We specifically began by determining the expression pattern of the Pnkd gene and protein. Next, we set out to see whether attacks in mice could be precipitated by the same stimuli that cause attacks in human PNKD patients. Since alcohol and caffeine are known to be “dirty” drugs that act on many receptor systems in the brain, targeted neuropharmacological agents were used to test specific pathways through which they might be acting. Finally, the neurotransmitter systems and receptors involved in transducing the abnormal dyskinetic movements in PNKD and the brain region or regions involved were also investigated. Thus, these studies were aimed at a more systems-level understanding of the pathophysiology as opposed to the molecular or cellular basis of PNKD. Such understanding, along with more work aimed at the molecular basis of PNKD, will be necessary to ultimately develop better therapies for paroxysmal dyskinesias and, potentially, other movement disorders.

Nomenclature. In this study, standard nomenclature for names of genes and proteins was used. In vitro experiments were performed in cells transfected with the human cDNA, and in vivo experiments were done in mice. PNKD represents the human gene name, while PNKD is the human protein name and the acronym for the disorder paroxysmal nonkinesigenic dyskinesia. Pnkd is the mouse gene name, and Pnkd is the name for the mouse protein. “Pnkd mice” is used to denote the animal model we created that harbored the PNKD phenotype (i.e., mice transgenic for a BAC harboring both the A7V- and A9V-encoding Pnkd mutations).

Mapping Pnkd expression. In situ hybridization analysis previously showed that Pnkd-L mRNA is widely expressed in neurons of the CNS, but not in other tissues (7). We developed antibodies induced by oligopeptides from the N terminus (N-terminal antibody, expected to detect PNKD-L and -S) or C terminus (C-terminal antibody, expected to detect PNKD-M and -L) (Figure 1A). The C-terminal antibody detected 2 main bands (PNKD-L, ~47 kDa; PNKD-M, ~40 kDa), and the N-terminal antibody detected 2 bands (PNKD-L and PNKD-S, ~18 kDa) in mouse brain extracts (Figure 1B). We also tested the PNKD antibodies by detecting different isoforms of PNKD-EGFP transfected in human embryonic kidney 293 (HEK293) cells. In this heterologous expression system, the size of PNKD-L–EGFP, PNKD-M–EGFP, and PNKD-S–EGFP are approximately 75 kDa, approximately 70 kDA, and approximately 44 kDa, respectively (Supplemental Figure 1; supplemental material available online with this article; doi: 10.1172/JCI58470DS1).

Figure 1

Pnkd immunochemistry in mouse brain. (A) Oligopeptides used to produce PNKD antibodies. Black rectangles represent exons shared by PNKD-L and PNKD-S; white rectangles represent exons shared by PNKD-L and PNKD-M; red rectangle represents the unique exon 1 of PNKD-M; blue rectangle represents the unique exon 3 of PNKD-S. Cysteines (underlined) were introduced in both oligopeptides to improve the coupling process. (B) Whole brain extracts on Western blots. The C-terminal antibody (1:5000 dilution) detected bands representing Pnkd-L (~47 kDa) and Pnkd-M (~40 kDa). The N-terminal antibody (1:500 dilution) detected bands representing Pnkd-L and Pnkd-S (~18 kDa). Pre-I, blots incubated with preimmune sera; Ag/Ab, PNKD antisera incubated with the oligopeptide against which they were raised; Ab, C- or N-terminal antisera only. (C) Double labeling in mice striatum with PNKD antibodies and either DARPP-32 (medium spiny neurons), ChAT (for cholinergic interneurons), or parvalbumin (Parv) (for GABAergic interneurons). (D) In striatum, Pnkd colocalizes with NeuN, a neuron-specific marker. (E) In striatum, Pnkd did not colocalize with oligodendrocyte-specific marker MBP. (F) Spinal cord cross section. Pnkd did not colocalize with astrocyte-specific marker GFAP. Original magnification, ×400.

Immunohistochemistry using these antibodies confirmed that Pnkd is expressed widely in the CNS, including striatum, substantia nigra, cerebellum, and spinal cord (Figure 1, C–F). Pnkd is expressed both in striatal medium spiny neurons and in interneurons (Figure 1C) and also in substantia nigra pars compacta (SNC), which is constituted by dopaminergic neurons projecting to striatum, and substantia nigra pars reticulata (SNR), which is the output structure of the basal ganglia (Supplemental Figure 2A). In cerebellum, Pnkd is expressed in the granule cell and molecular layers and in Purkinje cells (Supplemental Figure 2B). Double immunochemical staining was performed with these antisera and markers specific for neurons or glia, and the results showed that Pnkd is expressed in neurons but not in oligodendrocytes or astrocytes (Figure 1, D–F). No difference in protein distribution was noted in Pnkd mutant mice versus littermate controls or Pnkd WT transgenic mice (data not shown).

There is one weak transmembrane segment prediction in PNKD-L (7). To determine whether PNKD-L is a transmembrane protein, we performed studies with permeabilized and nonpermeabilized PNKD-L–EGFP–transfected HEK293 cells. PNKD-L can be detected by both N- and C-terminal antibodies at the membrane in permeabilized cells, but not in nonpermeabilized cells (Figure 2A). We also prepared subcellular membrane fractions of HEK293 cells transfected with PNKD-L–EGFP, performed detergent partitioning with Triton X-114 (TX-114) onto subcellular membrane fractions, and showed that after TX-114 treatment, PNKD-L–EGFP can be detected in the hydrophilic layer, but not in the hydrophobic layer, using PNKD antibodies (Figure 2B). Thus, PNKD-L appears to be membrane associated, though not a transmembrane protein integral to the membrane.

Figure 2

PNKD-L is a membrane-associated protein. (A) Immunostaining studies of permeabilized and nonpermeabilized HEK293 cells indicate that PNKD-L is a membrane-associated protein. HEK293 cells are transfected with PNKD-EGFP followed by immunohistochemical staining using N- and C-terminal PNKD antibodies and Cy3-conjugated goat anti-rabbit IgG secondary antibody. Both PNKD N- and C-terminal antibodies detect transfected PNKD-L–EGFP in permeabilized HEK cells, but fail to detect it in nonpermeabilized HEK cells. Scale bars: 10 microns. (B) Detergent phase partitioning of HEK293 membrane fraction transfected with PNKD-L–EGFP. The PNKD-L–EGFP (~75 kDa) can be detected in crude membrane fraction by both PNKD antibodies, but after TX-114 treatment, the PNKD-L–EGFP can be detected in the hydrophilic layer, but not in the hydrophobic layer.

A mouse model harboring the PNKD mutations recapitulates the PNKD phenotype. Three lines of Pnkd mutant BAC transgenic mice (mut-Tg, Pnkd mice) were generated by introducing both A7V- and A9V-encoding mutations into Pnkd (Figure 3A). The resulting BAC construct included the murine promoter and a large flanking region (extending 62 kb upstream) and was expected to contain all of the cis-acting regulatory elements needed to recapitulate the expression pattern of the endogenous loci. In addition, we generated Pnkd WT BAC transgenic mice (WT-Tg). Copy numbers of all lines were assessed by Southern blot; mut-Tg lines contained 1–2 copies, and the WT-Tg lines contained 2–3 copies (Figure 3B). Western blotting of mut-Tg and WT-Tg brain extracts showed that protein levels were consistent with the transgene copy number (Figure 3C). We also generated 2 lines of Pnkd-KO mice (Figure 3D). Both RT-PCR and Western blotting confirmed that they were true Pnkd nulls (Figure 3, E and F).

Figure 3

Generation of Pnkd-transgenic and -KO mice. (A) 2 human mutations (A7V and A9V) were introduced into the BAC containing Pnkd. 62 kb of flanking sequence 5′ of the ATG start site is present in the BAC and is expected to contain all of the cis-acting regulatory elements leading to expression in a pattern faithful to the endogenous alleles. An internal ribosome entry site (IRES) followed by an enhanced red fluorescent protein gene (IRES/DsRed) was introduced into the 3′ UTR of the Pnkd gene. The original BAC without mutations was used for generation of Pnkd WT-Tg mice. (B) Southern blot analyses of mutant and WT-Tg mice. (C) Western blot analyses of brain extracts from mut-Tg and WT-Tg mice with C-terminal antibody. Tubulin was used as loading control. (D) Pnkd-KO mice. A short homologous arm upstream of exon 5 and a long homologous arm downstream of exon 9 were amplified and subcloned into the pMCIDT-A PGKNeo vector to replace 1.5 kb of genomic Pnkd DNA (including exons 5–9) with a neomycin resistance gene. (E) Genotype analyses of Pnkd-KO mice by RT-PCR. (F) Western blot analyses of mut-Tg and KO mice using C-terminal antibody. Tubulin was used as loading control. *Pnkd-L (~47 kDa); **Pnkd-M (~40 kDa).

Multiple lines of mut-Tg, WT-Tg, and KO mice were observed, and all were fertile, with normal growth, and had no gross phenotypic abnormalities. None of the lines showed neuropathological defects by Nissl staining (data not shown). The body weight of PNKD mice trended lower than that of WT littermates, but the difference was not statistically significant. However, in mut-Tg mice, we observed dyskinesia after stress, such as prolonged handling (>15–30 minutes; see Supplemental Video 1). The dyskinesias included oral-facial movements (e.g., tongue protrusions) and stereotypic movements (repetitive sniffing and rearing in one location; see Supplemental Video 1). Dyskinesias were not seen in WT littermates, WT-Tg mice, or Pnkd-KO mice.

Mice were challenged with i.p. injection of caffeine (25 mg/kg) or ethanol (1.5 g/kg, 20% v/v). In mut-Tg mice, caffeine induced dyskinetic attacks approximately 10–15 minutes after treatment that persisted for 2 hours (Figure 4A and Supplemental Video 2). The severity of attacks in mut-Tg mice was significantly increased compared with that in predrug control or postdrug control using vehicle (saline, 10 ml/kg; Figure 4B). Behavioral effects of caffeine injection in WT littermates and WT-Tg mice were moderate hyperlocomotion during the first hour, but no dyskinesias were observed (Figure 4A). Ethanol also induced dyskinetic attacks in the mut-Tg mice (but not in WT littermates) beginning 10 minutes after injection and lasting 2–4 hours, and the attacks were more severe than those induced by caffeine. The mut-Tg mice suffered severe axial stiffness, with some abnormal movements of the limbs (Figure 4C and Supplemental Video 3, A and B). We did not observe any abnormal movements in the WT-Tg or KO mice after injection of caffeine or ethanol (data not shown). We examined multiple lines of mut-Tg, WT-Tg, and KO mice for i.p. injection of caffeine or ethanol, and the results showed no differences among genotypes. Unlike in dt^sz hamsters, there were no obvious age-dependent differences in the expression of dyskinesias of mut-Tg mice.

Figure 4

Pnkd mutant mice (mut-Tg, Pnkd mice) recapitulate the PNKD phenotype and have altered dopamine signaling. (A) Quantifying activity of Pnkd mice and WT littermates after caffeine injection (25 mg/kg, i.p.). Caffeine-induced hyperkinetic movements in mutant mice. (B) Activity ratings of Pnkd mice and WT littermates after saline injection (10 ml/kg, i.p.). A slight increase in activity was observed in Pnkd mice 30–50 minutes after saline injection. (C) The mut-Tg mice manifested profound dyskinesia after injection of ethanol (1.5 g/kg, 20% v/v, i.p.), but the same treatment failed to induce hyperkinetic movements in WT littermates. Activity ratings are expressed as mean ± SEM (n = 8 for each group). (D) There were no obvious differences in the dopamine levels in striatal extracts from mice of various genotypes at rest or after caffeine treatment. (E and F) Both DOPAC levels and DOPAC/dopamine ratios were significantly increased in Pnkd mice (P < 0.01) after caffeine treatment, but decreased in WT and KO mice (P < 0.01). (G) The HVA levels at rest versus after caffeine treatment were significantly increased in Pnkd mice (P < 0.05). All data are presented as mean ± SEM. *P < 0.05; **P < 0.01.

Neuroanatomical characterization of Pnkd mice. Since Pnkd is widely expressed in brain, we performed immunohistochemical studies with anti–c-Fos antibody to identify neuronal populations activated during PNKD attacks. Under basal conditions, almost no c-Fos–reactive cells were found in the brains of mut-Tg mice or WT littermates. Induction of c-Fos was seen in globus pallidus, subthalamic nucleus, and substantia nigra reticulata of mut-Tg mice, but not WT littermates, after caffeine injection (Supplemental Figure 3A). Similar patterns of c-Fos–positive signal were also seen in mutant mice after an attack was provoked with ethanol (data not shown). Minimal c-Fos–positive immunostaining was seen in cortical areas, but was not different in mut-Tg mice versus littermates. There were no obvious changes of c-Fos expression in basal ganglia or other brain regions in either WT-Tg or KO mice (data not shown). Thus, basal ganglia neurons are activated specifically in mut-Tg mice after induction of attacks.

Neurochemical studies show alteration of striatal dopamine signaling system of PNKD mice. Dopamine plays an important role in modulation of cortical and thalamic glutamatergic signal processing in the striatum, thus regulating movement (13–15). Therefore, we measured dopamine and its metabolites in the striatum by HPLC both before and after stimulating attacks. Pnkd and control mice had similar dopamine levels in striatum at rest (Figure 3D). After i.p. injection of caffeine (25 mg/kg), Pnkd mice had significantly higher levels of the dopamine metabolite 3,4-dihydroxyphenylacetic acid (DOPAC) and higher DOPAC/dopamine ratios than untreated animals (P < 0.01, Figure 4, E and F) or WT littermates treated with caffeine (P < 0.05 for DOPAC, Figure 4E; P < 0.01 for DOPAC/dopamine ratio; Figure 4F). The DOPAC levels and the DOPAC/dopamine ratio were significantly decreased in Pnkd-KO mice after injection of caffeine (P < 0.01) or ethanol (data not shown), but did not change in WT-Tg mice (Figure 4, E and F). Homovanillic acid (HVA), the terminal metabolite of dopamine metabolism, was also significantly increased in Pnkd mice after caffeine stimulation (P < 0.05; Figure 4G), while the levels of serotonin (5-HT) and its metabolized product (5-hydroxyindoleacetic acid [5-HIAA]) were not different in Pnkd versus control or before versus after caffeine challenge (Supplemental Figure 3, B and C). No obvious changes of DOPAC, DOPAC/dopamine ratio, and HVA levels were found in the Pnkd WT-Tg mice (Figure 4, D–G). Thus, dysfunction of dopamine signaling is associated with PNKD pathophysiology. Since the striatum is the primary basal ganglia region receiving dopaminergic and glutamatergic inputs from other brain structures (16–18), it is possible that expression of proteins involved in dopamine, glutamate, and/or adenosine signaling is altered in PNKD. To test this, striatal extracts from Pnkd and control mice were run on Western blots and probed with antibodies to proteins important for striatal function.

Because striatal dopamine signaling in Pnkd mice is altered (Figure 4, E–G), we examined the expression levels of proteins involved in dopamine synthesis, signaling, reuptake, and metabolism. Tyrosine hydroxylase (TH), the rate-limiting enzyme in the generation of dopamine, was no different for Pnkd mice versus controls (Supplemental Figure 4A). But D₁AR and D₂R were both significantly upregulated in the striata of Pnkd mice (P < 0.05; Supplemental Figure 4, B and C). We noted a trend (although not statistically significant) of decreased expression of vesicular monoamine transporter 2 (VMAT2) in PNKD striata versus control (Supplemental Figure 4D). Dopamine transporter (DAT) (~55 kDa in size) expression was significantly higher in the Pnkd mice compared with WT littermates (P < 0.05; Supplemental Figure 4E), and Pnkd mice also showed significantly higher expression of another main band of DAT in Western blots (~85 kDa, data not shown). There were no obvious differences in the expression of soluble catechol-methyltransferase (S-COMT) or membrane-bound COMT (MB-COMT) (19) in Pnkd mice versus littermate controls (Supplemental Figure 4F). Immunoblotting did not reveal any changes in expression levels of monoamine oxidase A (MAO-A) among genotypes (Supplemental Figure 4G). However, the level of MAO-B expression was significantly higher in Pnkd mice compared with littermate controls (P < 0.05; Supplemental Figure 4H).

We also examined striatal levels of synaptic vesicle proteins Rab3a, synapsin-1, and synaptophysin. Expression of these proteins was not altered in mutant versus control mice (Supplemental Figure 5, A–C). We also examined striatal proteins involved in GABA synthesis and glutamate reuptake (GAD65 and GAD67, EAAT3, and VGLUT1) by Western blotting. Expression of these proteins was no different in Pnkd versus control mice (Supplemental Figure 5, D–G) suggesting that GABA synthesis in striatum and glutamate reuptake are normal in Pnkd mice. We next examined expression of adenosine A₁ and adenosine A_2A receptors (A₁R and A_2AR), and adenosine kinase (ADK), a key regulator of adenosine metabolism. These were all unchanged in Pnkd mice versus WT littermates (Supplemental Figure 5, H–J). These results indicate that the alteration of striatal dopamine signaling system of Pnkd mice may contribute to dyskinetic attacks of Pnkd mice.

Pnkd mice have lower striatal dopamine release at rest, but an increased percentage of striatal dopamine release in response to challenges. To examine the pattern of nigrostriatal dopamine release, we performed in vivo microdialysis in alert, freely moving mice. Basal extracellular dopamine concentrations were assessed using the no net flux method to provide an unbiased estimate of transmitter concentration. The extracellular concentration of dopamine in Pnkd mice was less than 40% of that of control mice under basal conditions (P < 0.05; Figure 5A). In contrast, the extraction fraction was comparable in control and Pnkd mice, suggesting that the reduction in the basal extracellular dopamine concentrations in mutant mice is attributable to abnormalities in transmitter release (Figure 5B). Next, response to stress and caffeine was assessed by conventional microdialysis. Consistent with the no net flux method, a significant reduction in extracellular dopamine concentrations in mut-Tg mice was seen both before and after challenge with stress and caffeine (Figure 5C). Two-factor ANOVA with repeated measures revealed a significant effect of genotype (F_1,15 = 10.8; P < 0.005), but no effect of treatment (F_2,30 = 0.3; P < 0.5). Interestingly, dopamine release in response to stress and caffeine (described as proportion of baseline release) was higher than in WT littermates (Figure 5D). These results indicate that abnormalities in dopamine release at rest and in response to different challenges are present in the striata of Pnkd mice.

Figure 5

In vivo microdialysis of Pnkd mice. (A) Striatal dopamine concentrations were measured by no net flux microdialysis (mut-Tg, n = 7; WT, n = 5). Decreased levels of extracellular dopamine were observed in striata of Pnkd mice (P < 0.005). ** P < 0.01. (B) No significant difference in extraction fractions was found between genotypes. (C) Dopamine concentrations at basal levels and after stress/caffeine stimulation were reduced in striata of Pnkd mice (P < 0.005, n = 7) versus controls (n = 5). (D) Higher levels of dopamine release in response to stress and caffeine injection were observed in striata of Pnkd versus WT mice (P < 0.05). All data are shown as mean ± SEM.

Pnkd mice have deficits in nigrostriatal neurotransmission. We next performed carbon fiber amperometry studies to assess evoked release of dopamine in striatal slices from mutant and control mice. Electrically evoked release of dopamine was assessed by amperometry in striatal slices. The amperometry traces of striatal slices showed obvious differences between Pnkd mice and WT littermates (Figure 6A). Significantly lower dopamine signals were observed in PNKD mice versus controls (P < 0.01; Figure 6B). Application of the selective DAT inhibitor, nomifensine (3 μM/l for 30 minutes), did not significantly increase the dopamine signals in slices from mutant mice compared with slices from WT littermates (P < 0.01; Figure 6C). Dopamine release in response to electric stimulation was also significantly lower in slices from Pnkd mice versus WT littermates (P < 0.05; Figure 6D). Overall stimulated dopamine release was restored to normal levels in Pnkd slices after application of nomifensine. The spike width was significantly smaller and the half width (t_1/2) was significantly larger in Pnkd than in WT controls (P < 0.05; Figure 6, D and E). The spike width returned to normal in the presence of nomifensine (Figure 6D), but the t_1/2 did not (P < 0.01; Figure 6E). These results suggest that Pnkd mice feature significant deficits in nigrostriatal neurotransmission characterized by low levels of dopamine release and enhanced dopamine reuptake. The latter is compatible with DAT expression being significantly higher in the Pnkd mice compared with WT littermates (Supplemental Figure 4E).

Figure 6

Amperometry studies of Pnkd mice. (A) The amperometry traces of Pnkd mice versus WT littermates. (B) The mean evoked dopamine signal amplitude was significantly decreased in striatal slices from Pnkd mice versus controls with or without nomifensine (P < 0.01). (C) Striatal slices from Pnkd mice showed significantly reduced dopamine release per stimulation (P < 0.05). Application of nomifensine restored the overall stimulated dopamine release. (D and E) Striatal slices of Pnkd mice showed significant reduction of the spike width and elevated half width (t_1/2) of dopamine release (P < 0.05). Only the spike width of Pnkd slices was restored in the presence of nomifensine. In contrast, application of nomifensine did not restore the half width of Pnkd mouse striatal slices (P < 0.01). All data are presented as mean ± SEM. * P < 0.05; ** P < 0.01.

Neuropharmacologic characterization of Pnkd mice. Striatal A₁R and A_2AR are the major modulators of striatal signaling (20, 21). The A₁Rs are enriched in striatonigral-striatoentopenduncular medium spiny neurons that constitute the direct output pathway expressing dopamine D₁ receptors (D₁Rs). A_2ARs and dopamine D₂ receptors (D₂Rs) are both expressed in striatopallidal medium spiny neurons that represent the indirect output pathway (22–24). Since caffeine is a nonselective adenosine receptor antagonist (25–27), we evaluated the contribution of A₁R and A_2AR to attacks in PNKD. Dyskinesias were induced in Pnkd mice approximately 20 minutes after i.p. injection of the selective A_2AR antagonist 8-(3-chlorostryl) caffeine (CSC) (5 mg/kg; Figure 7A) and lasted approximately 2 hours. The A₁R antagonist 8-cyclopentyl-1,3-dipropylxanthine (DPCPX) (3 mg/kg) did not induce dyskinesias in Pnkd mice (Figure 7B). The A₁R agonist N6-cyclopentyladenosine (CPA) (0.1 mg/kg) had no obvious effects on Pnkd mice or controls (Supplemental Figure 6A), and the A_2AR agonist 2p-(2 carboxyethyl)phenethylamine-5′-N-ethylcarboxyamindoadenosine (CGS 21680) (0.5 mg/kg) led to a reduction of activity in both Pnkd mice and controls, with no difference between genotypes (Supplemental Figure 6B).

Figure 7

Pharmacological studies of Pnkd mice. (A) There are significant hyperkinetic movements in Pnkd (mut-Tg) mice (versus littermate controls) after injection of the adenosine A₂ receptor antagonist CSC (5 mg/kg, i.p.). (B) No difference in activity ratings of Pnkd mice and WT littermates was observed after injection of the adenosine A₁ receptor antagonist DPCPX (3 mg/kg, i.p.). (C) There are significant hyperkinetic movements in Pnkd mice after injection of the dopamine D₂ receptor agonist quinpirole (2.5 mg/kg, i.p.). (D) Increased activity was observed in both Pnkd mice and WT controls after injection of SKF 82958 (D₁ receptor agonist, 0.75 mg/kg, i.p.), but no difference was found between genotypes. In all pharmacological studies, behavioral ratings are expressed as mean ± SEM (n = 8 for each group).

We speculated that dyskinesias induced by caffeine are caused by altered signaling between adenosine receptors and dopamine receptors in striatum. D₁R and D₂R are major receptors in the regulation of striatal function, and antagonistic interactions between A₁R-D₁R and A_2AR-D₂R in the basal ganglia are important in motor control (13). Since D₂Rs colocalize with A_2ARs in neurons of the indirect pathway of the striatum, we hypothesized that the selective D₂R agonist quinpirole would induce dyskinetic attacks in Pnkd mice. Mutant mice developed dyskinesias approximately 15 minutes after quinpirole treatment (2.5 mg/kg), but no phenotype was seen in controls (Figure 7C). The D₁R agonist SKF 82958 (0.75 mg/kg) increased the behavioral activity in all mice, without significant differences among genotypes (Figure 7D). Taken together, these results support a model of adenosine-dopamine signaling dysregulation in PNKD.

PNKD is a rare disorder in which dyskinetic attacks can be induced by ingestion of caffeine or alcohol and frequently when patients are stressed (3). A transgenic Pnkd mouse model carrying both human mutations recapitulates the human phenotype. Pnkd-KO mice do not have a phenotype, arguing that the mutation of PNKD is a gain-of-function allele. It is possible that these PNKD mutations cause the disease by altering enzyme activity or specificity. Alternatively, they may not alter its enzymatic properties, but rather, cause altered trafficking and localization of the protein in neurons or through alterations of interactions between PNKD-L and other neuronal proteins. Finally, it is also possible that PNKD has evolved a novel function and is not an enzyme.

We presented in vivo evidence of increased dopamine turnover. Administration of caffeine has been shown to increase dopamine release accompanied by decreased DOPAC levels in rat striatum (28–30), similar to what we saw in WT control and Pnkd-KO mice. It is also known that acute ethanol administration can cause an increase of dopamine turnover as assessed by increased DOPAC in rat striatum, suggesting that both pre- and postsynaptic dopaminergic mechanisms may be involved in the mediation of some of the central effects of ethanol in striatum (31, 32). Increased c-Fos expression in basal ganglia of mutant versus control mice after caffeine administration shows that neurons in this part of brain are activated with the induction of attacks. Furthermore, Pnkd mice have significantly lower evoked dopamine release in real time and higher expression levels of dopamine receptors, DAT, and MAO-B in the striatum compared with controls. Taken together, these results suggest dysfunction of dopamine signaling in basal ganglia of Pnkd mice with induction of dyskinesias.

Neuropharmacological experiments implicate a role for A_2ARs and D₂Rs in pathogenesis, as attacks can be triggered with a selective A_2AR antagonist and a selective D₂R agonist. This, in turn, leads to dysfunction of antagonistic interactions between adenosine and dopamine receptors in modulation of motor outputs from striatum (22, 24). These results indicate strong involvement of the striatal indirect pathway in PNKD pathophysiology, but we cannot rule out the possibility that the direct pathway is also involved. Furthermore, though it is clear that the striatum is important for transducing abnormal dopamine signaling in PNKD, both A_2A and D₂ receptors are also expressed in other CNS regions. Besides, adenosine A1Rs are expressed presynaptically in CNS, including at glutamatergic terminals on medium spiny neurons. Thus, the genesis of PNKD could be outside the striatum (e.g., cortex).

Alterations of dopaminergic function in striatum play an important role in primary dystonias. Dopa-responsive dystonia (DYT5) is caused by mutations of the GTP–cyclohydrolase 1 gene involved in catecholamine and serotonin biosynthesis or by mutations of TH (2, 33). In dt^sz hamsters, an elevation of extracellular striatal dopamine levels has been observed during dystonic episodes (34). Interestingly, unlike dt^sz hamsters, Pnkd mice have reduced extracellular dopamine in striatum in vivo, but, when challenged by stress, caffeine, or alcohol, there is a relative increase in dopamine compared with controls. Pnkd mice have normal dopamine content in striatal dopaminergic terminals and apparently normal dopamine production. But dopamine receptors are upregulated, and when animals are stressed, striatal dopamine release is increased in Pnkd mice and receptor sensitivity is increased due to low basal extracellular dopamine levels. Excessive dopaminergic signaling under these conditions may lead to abnormal neuronal activity in Pnkd basal ganglia accompanied by significantly increased dopamine turnover. Although, typically, less dopamine release might be expected to yield less movement, dopamine loss occurring early in development can result in abnormal and excessive movements. This is true in humans with DYT5, where a developmental loss of dopamine caused by mutations that reduce dopamine synthesis causes dystonia (35, 36). It is also true for rodents, since dopamine depletion by 6-hydroxydopamine (6-OHDA) in adult rats causes akinesia (37–39). However, neonatal 6-OHDA treatment in rats causes hyperactivity (40–42).

PNKD-L is a membrane-associated protein, and Pnkd mice exhibit alterations of exocytosis, suggesting that PNKD may be involved in modulation of neurotransmitter release at nigrostriatal dopaminergic terminals. Alternatively, PNKD may participate in the modulation of striatal glutamatergic inputs projecting from cerebral cortex and thalamus. Adenosine receptors and dopamine receptors not only interact with each other, but also cooperate with other signaling systems, such as metabotropic glutamate receptors and cannabinoid receptors. These interactions between different signaling systems are critical for modulation of normal striatal function and plasticity (14, 20, 43, 44). These findings imply that dysfunction of the striatal dopamine signaling system plays a pivotal role in PNKD pathophysiology. Although expression levels of adenosine receptors and glutamate transporters are not different in striatum of Pnkd mice, we cannot rule out the possibility of altered sensitivity of adenosine receptors and/or glutamate release in glutamatergic terminals. Dopamine is critical for the induction of bidirectional plasticity at glutamatergic synapses on the medium spiny neurons of both direct and indirect pathways, and this balance is interrupted in models of Parkinson disease that cause unidirectional changes in striatal synaptic plasticity (45). Glutamatergic synapses onto the medium spiny neurons of the indirect pathway show higher release probability than glutamatergic synapses onto the medium spiny neurons of the direct pathway, and they selectively express endocannabinoid-mediated long-term depression that is absent in a model of Parkinson disease (46). Since the striatal dopamine signaling system may be upregulated, Pnkd mice may also display alterations of striatal synaptic plasticity under stress or with caffeine/ethanol treatment. In turn, transient alterations of neuronal activity in basal ganglia may occur during PNKD attacks. Further study will provide better understanding of the role of the PNKD protein in cellular and synaptic regulation and contribution of PNKD mutations to pathophysiology. Such work may allow development of better therapies for PNKD and potentially for other episodic disorders. Since PNKD is a novel gene participating in modulating striatal neurotransmitter release, further investigations both in vitro and in vivo will give new insights into understanding normal basal ganglia regulation of motor function and potentially have implications for developing an understanding other dystonias and striatal diseases.

View Supplemental data

View Supplemental video 1

View Supplemental video 2

View Supplemental video 3

View Supplemental video 4

We thank Anatol Kreitzer and Shu-Ting Lin for helpful discussions. This work was supported by grants from the Dystonia Medical Research Foundation, the Bachmann-Strauss Dystonia & Parkinson Foundation, NIH grant U54 RR19481, the Sandler Neurogenetics Fund, and P30 NS047243 (Tufts Center for Neuroscience Research). L.J. Ptácek is an investigator of the Howard Hughes Medical Institute.

Address correspondence to: Louis Ptácek or Ying-Hui Fu, Department of Neurology, UCSF MC 2922, 548F Rock Hall, 1550 Fourth St., San Francisco, California 94158-2324, USA. Phone: 415.502.5614; Fax: 415.502.5641; E-mail: ljp@ucsf.edu (L. Ptácek); ying-hui.fu@ucsf.edu (Y.-H. Fu).

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J Clin Invest. 2012;122(2):507–518. doi:10.1172/JCI58470.

Junko Nakayama’s present address is: Department of Pediatrics, Ibaraki Prefectural University of Health Sciences, Ibaraki, Japan.

Ying Xu’s present address is: Model Animal Resource Center, Nanjing University, Nanjing, China.

1. Demirkiran M, Jankovic J. Paroxysmal dyskinesias: clinical features and classification. Ann Neurol. 1995;38(4):571–579.
   View this article via: PubMed CrossRef Google Scholar
2. Tarsy D, Simon DK. Dystonia. N Engl J Med. 2006;355(8):818–829.
   View this article via: PubMed CrossRef Google Scholar
3. Bruno MK, et al. Genotype-phenotype correlation of paroxysmal nonkinesigenic dyskinesia. Neurology. 2007;68(21):1782–1789.
   View this article via: PubMed CrossRef Google Scholar
4. Fink JK, et al. Paroxysmal dystonic choreoathetosis: tight linkage to chromosome 2q. Am J Hum Genet. 1996;59(1):140–145.
   View this article via: PubMed Google Scholar
5. Fouad GT, Servidei S, Durcan S, Bertini E, Ptacek LJ. A gene for familial paroxysmal dyskinesia (FPD1) maps to chromosome 2q. Am J Hum Genet. 1996;59(1):135–139.
   View this article via: PubMed Google Scholar
6. Chen DH, et al. Presence of alanine-to-valine substitutions in myofibrillogenesis regulator 1 in paroxysmal nonkinesigenic dyskinesia: confirmation in 2 kindreds. Arch Neurol. 2005;62(4):597–600.
   View this article via: PubMed CrossRef Google Scholar
7. Lee HY, et al. The gene for paroxysmal non-kinesigenic dyskinesia encodes an enzyme in a stress response pathway. Hum Mol Genet. 2004;13(24):3161–3170.
   View this article via: PubMed CrossRef Google Scholar
8. Stefanova E, et al. Clinical characteristics of paroxysmal nonkinesigenic dyskinesia in Serbian family with Myofibrillogenesis regulator 1 gene mutation. Mov Disord. 2006;21(11):2010–2015.
   View this article via: PubMed CrossRef Google Scholar
9. Rainier S, et al. Myofibrillogenesis regulator 1 gene mutations cause paroxysmal dystonic choreoathetosis. Arch Neurol. 2004;61(7):1025–1029.
   View this article via: PubMed CrossRef Google Scholar
10. Hempelmann A, Kumar S, Muralitharan S, Sander T. Myofibrillogenesis regulator 1 gene (MR-1) mutation in an Omani family with paroxysmal nonkinesigenic dyskinesia. Neurosci Lett. 2006;402(1–2):118–120.
    View this article via: PubMed CrossRef Google Scholar
11. Ghezzi D, et al. Paroxysmal non-kinesigenic dyskinesia is caused by mutations of the MR-1 mitochondrial targeting sequence. Hum Mol Genet. 2009;18(6):1058–1064.
    View this article via: PubMed Google Scholar
12. Thornalley PJ. Pharmacology of methylglyoxal: formation, modification of proteins and nucleic acids, and enzymatic detoxification--a role in pathogenesis and antiproliferative chemotherapy. Gen Pharmacol. 1996;27(4):565–573.
    View this article via: PubMed Google Scholar
13. Fisone G, Hakansson K, Borgkvist A, Santini E. Signaling in the basal ganglia: postsynaptic and presynaptic mechanisms. Physiol Behav. 2007;92(1–2):8–14.
    View this article via: PubMed CrossRef Google Scholar
14. Surmeier DJ, Ding J, Day M, Wang Z, Shen W. D1 and D2 dopamine-receptor modulation of striatal glutamatergic signaling in striatal medium spiny neurons. Trends Neurosci. 2007;30(5):228–235.
    View this article via: PubMed CrossRef Google Scholar
15. Werkman TR, Glennon JC, Wadman WJ, McCreary AC. Dopamine receptor pharmacology: interactions with serotonin receptors and significance for the aetiology and treatment of schizophrenia. CNS Neurol Disord Drug Targets. 2006;5(1):3–23.
    View this article via: PubMed Google Scholar
16. Bolam JP, Hanley JJ, Booth PA, Bevan MD. Synaptic organisation of the basal ganglia. J Anat. 2000;196(pt 4):527–542.
    View this article via: PubMed CrossRef Google Scholar
17. Mink JW. The basal ganglia: focused selection and inhibition of competing motor programs. Prog Neurobiol. 1996;50(4):381–425.
    View this article via: PubMed CrossRef Google Scholar
18. Smith Y, Bevan MD, Shink E, Bolam JP. Microcircuitry of the direct and indirect pathways of the basal ganglia. NeuroScience. 1998;86(2):353–387.
    View this article via: PubMed CrossRef Google Scholar
19. Lundstrom K, Salminen M, Jalanko A, Savolainen R, Ulmanen I. Cloning and characterization of human placental catechol-O-methyltransferase cDNA. DNA Cell Biol. 1991;10(3):181–189.
    View this article via: PubMed CrossRef Google Scholar
20. Schiffmann SN, Fisone G, Moresco R, Cunha RA, Ferre S. Adenosine A2A receptors and basal ganglia physiology. Prog Neurobiol. 2007;83(5):277–292.
    View this article via: PubMed CrossRef Google Scholar
21. Schwarzschild MA, Agnati L, Fuxe K, Chen JF, Morelli M. Targeting adenosine A2A receptors in Parkinson’s disease. Trends Neurosci. 2006;29(11):647–654.
    View this article via: PubMed Google Scholar
22. Fuxe K, Ferre S, Genedani S, Franco R, Agnati LF. Adenosine receptor-dopamine receptor interactions in the basal ganglia and their relevance for brain function. Physiol Behav. 2007;92(1–2):210–217.
    View this article via: PubMed CrossRef Google Scholar
23. Ferre S, Fredholm BB, Morelli M, Popoli P, Fuxe K. Adenosine-dopamine receptor-receptor interactions as an integrative mechanism in the basal ganglia. Trends Neurosci. 1997;20(10):482–487.
    View this article via: PubMed CrossRef Google Scholar
24. Kim DS, Palmiter RD. Interaction of dopamine and adenosine receptor function in behavior: studies with dopamine-deficient mice. Front Biosci. 2008;13:2311–2318.
    View this article via: PubMed CrossRef Google Scholar
25. Fisone G, Borgkvist A, Usiello A. Caffeine as a psychomotor stimulant: mechanism of action. Cell Mol Life Sci. 2004;61(7–8):857–872.
    View this article via: PubMed CrossRef Google Scholar
26. Ferre S. An update on the mechanisms of the psychostimulant effects of caffeine. J Neurochem. 2008;105(4):1067–1079.
    View this article via: PubMed CrossRef Google Scholar
27. Xie X, Ramkumar V, Toth LA. Adenosine and dopamine receptor interactions in striatum and caffeine-induced behavioral activation. Comp Med. 2007;57(6):538–545.
    View this article via: PubMed Google Scholar
28. Watanabe H, Uramoto H. Caffeine mimics dopamine receptor agonists without stimulation of dopamine receptors. Neuropharmacology. 1986;25(6):577–581.
    View this article via: PubMed Google Scholar
29. Morgan ME, Vestal RE. Methylxanthine effects on caudate dopamine release as measured by in vivo electrochemistry. Life Sci. 1989;45(21):2025–2039.
    View this article via: PubMed CrossRef Google Scholar
30. Kirch DG, Taylor TR, Gerhardt GA, Benowitz NL, Stephen C, Wyatt RJ. Effect of chronic caffeine administration on monoamine and monoamine metabolite concentrations in rat brain. Neuropharmacology. 1990;29(6):599–602.
    View this article via: PubMed CrossRef Google Scholar
31. Reggiani A, Barbaccia ML, Spano PF, Trabucchi M. Dopamine metabolism and receptor function after acute and chronic ethanol. J Neurochem. 1980;35(1):34–37.
    View this article via: PubMed CrossRef Google Scholar
32. Barbaccia ML, Bosio A, Spano PF, Trabucchi M. Ethanol metabolism and striatal dopamine turnover. J Neural Transm. 1982;53(2–3):169–177.
    View this article via: PubMed CrossRef Google Scholar
33. Breakefield XO, Blood AJ, Li Y, Hallett M, Hanson PI, Standaert DG. The pathophysiological basis of dystonias. Nat Rev Neurosci. 2008;9(3):222–234.
    View this article via: PubMed CrossRef Google Scholar
34. Hamann M, Richter A. Striatal increase of extracellular dopamine levels during dystonic episodes in a genetic model of paroxysmal dyskinesia. Neurobiol Dis. 2004;16(1):78–84.
    View this article via: PubMed CrossRef Google Scholar
35. Clot F, et al. Exhaustive analysis of BH4 and dopamine biosynthesis genes in patients with Dopa-responsive dystonia. Brain. 2009;132(pt 7):1753–1763.
    View this article via: PubMed CrossRef Google Scholar
36. Segawa M. Dopa-responsive dystonia. Handb Clin Neurol. 2011;100:539–557.
    View this article via: PubMed Google Scholar
37. Marin C, Aguilar E, Mengod G, Cortes R, Obeso JA. Concomitant short- and long-duration response to levodopa in the 6-OHDA-lesioned rat: a behavioural and molecular study. Eur J Neurosci. 2007;25(1):259–269.
    View this article via: PubMed Google Scholar
38. Paille V, Henry V, Lescaudron L, Brachet P, Damier P. Rat model of Parkinson’s disease with bilateral motor abnormalities, reversible with levodopa, and dyskinesias. Mov Disord. 2007;22(4):533–539.
    View this article via: PubMed CrossRef Google Scholar
39. Zigmond MJ, Berger TW, Grace AA, Stricker EM. Compensatory responses to nigrostriatal bundle injury. Studies with 6-hydroxydopamine in an animal model of parkinsonism. Mol Chem Neuropathol. 1989;10(3):185–200.
    View this article via: PubMed CrossRef Google Scholar
40. Breese GR, Knapp DJ, Criswell HE, Moy SS, Papadeas ST, Blake BL. The neonate-6-hydroxydopamine-lesioned rat: a model for clinical neuroscience and neurobiological principles. Brain Res Brain Res Rev. 2005;48(1):57–73.
    View this article via: PubMed CrossRef Google Scholar
41. Fan X, Bruno KJ, Hess EJ. Rodent models of ADHD [published online ahead of print March 15, 2011]. Curr Top Behav Neurosci . doi:10.1007/7854_2011_121.
42. Masuo Y, Morita M, Oka S, Ishido M. Motor hyperactivity caused by a deficit in dopaminergic neurons and the effects of endocrine disruptors: a study inspired by the physiological roles of PACAP in the brain. Regul Pept. 2004;123(1–3):225–234.
    View this article via: PubMed CrossRef Google Scholar
43. Kreitzer AC, Malenka RC. Striatal plasticity and basal ganglia circuit function. Neuron. 2008;60(4):543–554.
    View this article via: PubMed Google Scholar
44. van der Stelt M, Di Marzo V. The endocannabinoid system in the basal ganglia and in the mesolimbic reward system: implications for neurological and psychiatric disorders. Eur J Pharmacol. 2003;480(1–3):133–150.
    View this article via: PubMed Google Scholar
45. Shen W, Flajolet M, Greengard P, Surmeier DJ. Dichotomous dopaminergic control of striatal synaptic plasticity. Science. 2008;321(5890):848–851.
    View this article via: PubMed CrossRef Google Scholar
46. Kreitzer AC, Malenka RC. Endocannabinoid-mediated rescue of striatal LTD and motor deficits in Parkinson’s disease models. Nature. 2007;445(7128):643–647.
    View this article via: PubMed Google Scholar
47. Maher PA, Singer SJ. Anomalous interaction of the acetylcholine receptor protein with the nonionic detergent Triton X-114. Proc Natl Acad Sci U S A. 1985;82(4):958–962.
    View this article via: PubMed CrossRef Google Scholar
48. Wetterhall M, Shevchenko G, Artemenko K, Sjodin MO, Bergquist J. Analysis of membrane and hydrophilic proteins simultaneously derived from the mouse brain using cloud-point extraction. Anal Bioanal Chem. 2011;400(9):2827–2836.
    View this article via: PubMed CrossRef Google Scholar
49. Page ME, Brown K, Lucki I. Simultaneous analyses of the neurochemical and behavioral effects of the norepinephrine reuptake inhibitor reboxetine in a rat model of antidepressant action. Psychopharmacology (Berl). 2003;165(2):194–201.
50. Sulzer D, Edwards R. Vesicles: equal in neurotransmitter concentration but not in volume. Neuron. 2000;28(1):5–7.
    View this article via: PubMed Google Scholar
51. Goldberg MS, et al. Nigrostriatal dopaminergic deficits and hypokinesia caused by inactivation of the familial Parkinsonism-linked gene DJ-1. Neuron. 2005;45(4):489–496.
    View this article via: PubMed CrossRef Google Scholar
52. Richter A, Loscher W. Pathology of idiopathic dystonia: findings from genetic animal models. Prog Neurobiol. 1998;54(6):633–677.
    View this article via: PubMed CrossRef Google Scholar
53. Wang JQ, McGinty JF. The full D1 dopamine receptor agonist SKF-82958 induces neuropeptide mRNA in the normosensitive striatum of rats: regulation of D1/D2 interactions by muscarinic receptors. J Pharmacol Exp Ther. 1997;281(2):972–982.
    View this article via: PubMed Google Scholar