|Home | About | Journals | Submit | Contact Us | Français|
Ghrelin targets the hypothalamus to regulate food intake and adiposity. Endogenous ghrelin receptors (growth hormone secretagogue receptor, GHSR) are also present in extrahypothalamic sites where they promote circuit activity associated with learning and memory, and reward seeking behavior. Here, we show that the substantia nigra pars compacta (SNpc), a brain region where dopamine (DA) cell degeneration leads to Parkinson’s disease (PD), expresses GHSR. Ghrelin binds to SNpc cells, electrically activates SNpc DA neurons, increases tyrosine hydroxylase mRNA and increases DA concentration in the dorsal striatum. Exogenous ghrelin administration decreased SNpc DA cell loss and restricted striatal dopamine loss after 1-methyl-4-phenyl-1,2,5,6 tetrahydropyridine (MPTP) treatment. Genetic ablation of ghrelin or the ghrelin receptor (GHSR) increased SNpc DA cell loss and lowered striatal dopamine levels after MPTP treatment, an effect that was reversed by selective reactivation of GHSR in catecholaminergic neurons. Ghrelin-induced neuroprotection was dependent on the mitochondrial redox state via uncoupling protein 2 (UCP2)-dependent alterations in mitochondrial respiration, ROS production and biogenesis. Taken together, our data reveals that peripheral ghrelin plays an important role in the maintenance and protection of normal nigrostriatal dopamine function by activating UCP2-dependent mitochondrial mechanisms. These studies support ghrelin as a novel therapeutic strategy to combat neurodegeneration, loss of appetite and body weight associated with PD. Finally, we discuss the potential implications of these studies on the link between obesity and neurodegeneration.
PD is characterized by the progressive degeneration of DA neurons projecting from the SNpc to the dorsal striatum. The resulting loss of dopamine in the striatum leads to debilitating motor dysfunction including rigidity, resting tremor, postural instability and bradykinesia. Familial or genetic causes of PD only account for 10% of all cases whereas 90% are considered sporadic and may manifest as a result of variety of factors.
Obesity is considered one of the most important health concern of our generation and is predicted to reduce life expectancy in the future (Olshansky et al., 2005). While obesity is a known salient risk factor in diabetes, cardiovascular disease and cancer, the risk of neurological disease after prolonged obesity is only starting to emerge. Recent human studies show that body mass index, midlife adiposity and diabetes are associated with the neurodegenerative illness, PD (Abbott et al., 2002; Hu et al., 2006; Hu et al., 2007). Furthermore, obesity is a risk factor for chemically-induced neurodegeneration in mice (Choi et al., 2005). In examining potential mechanisms linking obesity and neurodegeneration, we noted that the gut hormone, ghrelin, is inversely related to obesity, such that levels are higher during negative energy balance or calorie restriction and lower during positive energy balance or obesity (Tschop et al., 2001). Indeed, calorie restriction, in which ghrelin levels are increased, attenuates MPTP-induced neurotoxicity in non-human primates (Maswood et al., 2004) and mice (Duan and Mattson, 1999). Further studies showed that cultured cells treated with serum from calorie-restricted rats displayed mitochondrial biogenesis, enhanced bioenergetic capacity and reduced production of reactive oxygen species (ROS) (Lopez-Lluch et al., 2006) indicating that the effects of calorie restriction may be mediated by a hormonal factor affecting mitochondrial metabolism. Ghrelin also preserves mitochondrial integrity during oxygen–glucose deprivation, (Chung et al., 2007) and mitochondrial dysfunction lies at the heart of PD (Abou-Sleiman et al., 2006) suggesting that the elevated ghrelin levels during calorie restriction may promote neuroprotection against MPTP-intoxication in mice. Recent studies show that ghrelin attenuates MPTP neurotoxicity in dopamine neurons (Jiang et al., 2008) by enhancing mitochondrial function (Dong et al., 2009). Although, the molecular mechanisms regulating mitochondrial function are unknown, our previous studies identified UCP2 as a critical mitochondrial protein regulating nigrostriatal dopamine function by buffering mitochondrial ROS production, enhancing mitochondrial respiration and increasing mitochondrial biogenesis (Andrews et al., 2005a; Andrews et al., 2005b; Conti et al., 2005). We recently demonstrated ghrelin as robust activator of mitochondrial function in the hypothalamus and this action was dependent on UCP2 (Andrews et al., 2008). Therefore, we tested the hypothesis that ghrelin promotes nigrastriatal dopamine function by increasing mitochondrial function in an UCP2-dependent manner.
The Institutional Animal Care and Use Committee of Yale University have approved all procedures described below. UCP2 knockout mice are on a C57/B6 background. Ghrelin knockout mice (on a C57/B6 background) were obtained from Regeneron Pharmaceuticals and bred in our facilities. These genetic mouse lines have been described previously (Wortley et al., 2004). Mice expressing GHSRs only in catecholaminergic cells (on a C57/B6 background) were generated by first crossing mice homozygous for a recombinant “null” GHSR allele (Zigman et al., 2005) with mice expressing Cre recombinase under the control of a tyrosine hydroxylase promoter (TH-Cre) (Savitt et al., 2005).
We used mice with GHSRs only in catecholaminergic cells by first crossing mice homozygous for a recombinant “null” GHSR allele (Zigman et al., 2005) with mice expressing Cre recombinase under the control of a tyrosine hydroxylase promoter (TH-Cre) (Savitt et al., 2005). As described previously, the recombinant null GHSR allele differs from the wild-type GHSR allele in that it has been altered by the insertion of a loxP-flanked transcriptional blocking cassette into an intron located downstream of the transcriptional start site and upstream of the translational start site of the murine GHSR gene. The replacement of two wild-type GHSR alleles with two recombinant GHSR-null alleles results in GHSR-null mice that no longer acutely increase food intake in response to exogenous ghrelin administration and are resistant to the development of diet-induced obesity (Zigman et al., 2005). The offspring of GHSR-null mice X TH-Cre mice were then bred with mice containing either no, one or two copies of the recombinant null GHSR allele, in order to generate the three desired study animal genotypes: wild-type mice (containing two copies of the wild-type GHSR allele and no copies of the TH-Cre transgene), GHSR-null mice (containing two copies of the recombinant null GHSR allele and no copies of the TH-Cre transgene), and “TH-only” mice (containing two copies of the recombinant null GHSR allele and one copy of the TH-Cre transgene). In the presence of Cre recombinase, the transcriptional blocking cassette and one of the loxP sites is removed from the recombinant null GHSR alleles, thus re-activating the ability to synthesize functional GHSR mRNAs. The specificity of Cre recombinase expression afforded by the promoter elements used in the TH-Cre transgene allows for selective re-activation of GHSR expression within tyrosine hydroxylase-expressing cells that have the latent genetic capacity to express GHSR. To validate the re-activation of GHSR expression selectively within TH cells of “TH-only” mice, brains were examined for the presence of GHSR mRNA by in situ hybridization histochemistry (ISHH), using a mouse GHSR-specific riboprobe (Zigman et al., 2006). GHSR transcripts were consistently visualized by ISHH within the substantia nigra and ventral tegmental area (Fig. 3A-C), within certain hypothalamic nuclei (the arcuate nucleus, the anteroventral periventricular nucleus, the dorsomedial nucleus, and the capsule of the ventromedial nucleus), and occasionally within scattered cells of the nucleus of the solitary tract. The pattern of GHSR expression observed within the “TH-only” mice is similar to the GHSR-TH co-expression pattern that we observe in the wild-type mouse brain (not pictured) and involves a few more sites than what has previously been described in the rat brain (Zigman et al., 2006). Similar to what we had previously described for GHSR-null mice (Zigman et al., 2005), strong binding of the GHSR riboprobe was also observed in the Edinger-Westphal nucleus of “TH-only” mice, which is neither an area of TH expression nor TH-Cre activity. In situ hybridization histochemistry for GHSRs was performed as previously described (Zigman et al., 2006).
To test whether ghrelin binds to cells in the SN, 100μm sections containing these regions were processed for binding studies as described previously (Cowley et al., 2003). Briefly, saline-perfused rat brains (n=4) were removed, sectioned, and immediately reacted with biotinylated ghrelin (1M; Phoenix Pharmaceuticals, Belmont, CA) alone or in combination with an equal amount of unlabeled (cold) ghrelin (1M; Phoenix Pharmaceuticals) for 20 min at 4°C. Sections were fixed with 4% paraformaldehyde and reacted with avidin-Texas red and analyzed using a Zeiss microscope equipped with fluorescent filters.
Brain slices (300 μm) containing the SNpc, were cut on a vibratome from 2-3 week old male and female mice (n= 9) and rats (n=5). Briefly, animals were anesthetized with Nembutal (80 mg/kg) and then decapitated. The brains were rapidly removed and immersed in cold (4C) oxygenated bath solution (containing (mM): NaCl 150, KCl 2.5, CaCl2 2, MgCl2 2 Hepes 10, and glucose 10, pH 7.3 with NaOH). After being trimmed to contain only the SNpc, slices were transferred to a recording chamber where they were constantly perfused with bath solution at 2 ml/min. Dopamine neurons in the SNpc were identified by presence of a large Ih current (>100 pA) evoked by hyperpolarizing voltage steps from −50 to −120 mV for 2(Abizaid et al., 2006). This approach identifies dopaminergic cells with >90% accuracy(Abizaid et al., 2006). In brain slices, whole-cell current clamp was used to observe spontaneous action potentials. Slices were maintained at 33°C and perfused continuously with ASCF (bubbled with 5% CO2 and 95% O2) containing (in mM): NaCl, 124; KCl, 3; CaCl2, 2; MgCl2, 2; NaH2PO4, 1.23; NaHCO3, 26; glucose, 10; pH 7.4 with NaOH. Ghrelin was applied to the recording chamber via bath application. The pipette solution contained (mM): Gluconic Acid 140, CaCl2, MgCl2 2, EGTA 1, HEPES 10, Mg-ATP 4, and Na2-GTP 0.5, pH 7.3 with -/-H.
RNA was extracted from the midbrain of 8 ghrelin-treated (10 nmols, ip) mice or 8 saline-treated mice using TRIzol (Invitrogen) according to the manufacturer’s instructions. cDNA was synthesized using the First-Strand cDNA synthesis kit (Amersham Biosciences) using a total of 3 μg of total RNA in 15 μl of total volume. Real-time PCR analysis of TH mRNA (5′ ggaggctttccagcttctg 3′, 5′ gtcagccaacatgggtacg 3′) was performed using iQ SYBR green supermix (Bio-Rad) and 0.5 μm of each primer. Primers for 18S were used as controls. Measurements were performed on an iCycler (Bio-Rad).
Animals were perfused and their brains were processed for TH immunolabeling for electron microscopic examination. Ultrathin sections were cut on a Leica ultra microtome, collected on Formvar-coated single-slot grids and analyzed with a Tecnai 12 Biotwin (FEI Company) electron microscope. Mitochondria were counted blindly from randomly selected sections, and Scion Image was used to normalize cytoplasmic area so that mitochondrial number per cell is expressed in square micrometers.
Mice were injected IP with 40 mg/kg of MPTP in saline as described previously (Andrews et al., 2005b). Control animals were given saline. Animals were sacrificed and perfused 7 days later and processed for immunohistochemistry or HPLC for DA and metabolites. n= 8-10 for ghrelin wt and -/- mice and n = 6 for GHSR wt/wt, homo/wt and homo/TG.
Free floating sections were stained with TH (1:5000, Chemicon International) and visualized with DAB using standard procedures. We then used design-based unbiased stereology methods to quantify TH immunoreactive cells in the SNpc. Cells were visualized by a Zeiss microscope and relayed via a MicroFibre digital camera to a computer where they were counted by a blinded observer using the optical fractionator StereoInvestigator software (MicroBrightField, Williston, VT, USA ). Systematic sampling of every fourth section was collected through the SNpc beginning at approximately bregma −2.06 mm and finishing at approximately −4.16 mm. Sections were cut at 50 μm to allow for a 20 μm optical dissector within each section after dehydration and mounting. The first sample section from the first 1-4 sections collected was chosen at random. TH cells were counted in grids randomly positioned by the software in the outlined counting area through all optical planes, thus creating a 3 dimensional counting area. The counting frame width and height was 55 μm producing an area of 3025 μm. With this counting frame area we discovered that we needed to sample approximately 150 sites throughout the entire SNpc and count approximately 120 neurons throughout the entire SNpc to give an acceptable coefficient of error (using the Gunderson method) of 0.1 using the smoothness factor m=0. The smoothness factor defines the cell distribution of a structure that is sensitive to small changes in sampling. We have applied the more rigorous smoothness factor of m=0 to account for sudden changes in cell distribution across our region of interest. The coefficient of error provides a means to estimate sampling precision, which is independent of natural biological variance. As the value approaches 0, the uncertainty in the estimate precision reduces. A coefficient of error = 0.1 is deemed acceptable. Cells were only counted if they touched the inclusion border or did not touch the exclusion border of the sampling grid.
At the time of sacrifice (7 days after MPTP treatment) both striata were rapidly dissected on a chilled glass plate and frozen at −70°C. The samples were subsequently thawed in 0.4 ml of chilled 0.1 M perchloric acid and sonicated. Aliquots were taken for protein quantification using a spectrophotometric assay. Other aliquots were centrifuged, and DA levels were measured in supernatants by HPLC with electrochemical detection. Concentrations of DA and metabolites were expressed as ng/mg protein (mean ± SEM).
Wild type or ucp2-/- mice were treated with ghrelin or saline 3 hours prior to sacrifice. The SN was rapidly dissected and homogenized in the isolation buffer (215 mM mannitol, 75 mM sucrose, 0.1% fatty acid-free BSA, 20 mM HEPES, 1 mM EGTA, pH adjusted to 7.2 with KOH). The homogenate was spun at 1300 × g for 3 min, the supernatant was removed, and the pellet was resuspended with isolation buffer and spun again at 1300 × g for 3 min. The two sets of supernatants from each sample were topped off with isolation buffer and spun at 13,000 × g for 10 min. The supernatant was discarded, and the step was repeated. After this second spin at 13,000 × g, the supernatant was discarded, and the pellets were resuspended with isolation buffer without EGTA and spun at 10,000 × g for 10 min. The final synaptosomal pellet was resuspended with 50 μl of isolation buffer without EGTA. Protein concentrations were determined with a BCA protein assay kit (Pierce, Rockford, IL). Mitochondrial respirations were assessed using a Clark-type oxygen electrode (Hansatech Instruments, Norfolk, UK) at 37°C with pyruvate and malate (5 and 2.5 mM) as oxidative substrates in respiration buffer (215 mM mannitol, 75 mM sucrose, 0.1% fatty acid-free BSA, 20 mM HEPES, 2 mM MgCl, 2.5 mM KH2PO4, pH adjusted to 7.2 with KOH). After the addition of oligomycin, proton conductance was measured as increased fatty acid-induced respiration(Echtay et al., 2002). Total uncoupled respiration was also measured after the addition of the photonophore FCCP. For analysis of ADP dependent respiration, ADP was added after the addition of oxidative substrates.
Wild type and ucp2-/- mice were injected with ghrelin (10 nmol) and the midbrain was harvested 3 hours later. Mitochondria were isolated as described above. ROS production was quantified using dichlorodihydrofluorescein diacetate (DCF). DCF is a cell permeant indicator for intracellular ROS that is nonfluorescent until the acetate groups are removed by intracellular esterases and oxidation occurs within the cell. Using a fluorescence plate reader (excitation 485 nm, emission 528 nm), hypothalamic ROS in ghrelin-treated or saline-treated samples was estimated in isolated synaptosomes in the presence and absence of oligomycin. Oligomycin induces maximal ROS production as it inhibits the ATP synthase. Data are expressed as arbitrary fluorescence units. In vivo ghrelin-induced ROS production in immunostained TH neurons in wt or ucp2-/- mice was measured by injecting dihydroethidium (DHE) as it is specifically oxidized by superoxide to red fluorescent ethidium. 1 mg/ml was injected into the femoral vein of lightly anathesized mice. Mice were allowed to recover for approximately 90 minutes and were then injected ip with ghrelin (10 nmol) and transcardially perfused 3 hours later. Sections were processed normally for TH immunohistochemistry. Fluorescent mitochondria (red channel) were counted blindly in TH neurons (Andrews et al., 2005b).
All data are presented as mean ± sem. Statistical differences among groups were determined by unpaired 2-tailed student t-tests or two-way ANOVA followed by bonferroni post hoc test as stated.
Biotinylated ghrelin binding was present throughout the SNpc with a similar distribution pattern to GHSR immunolabelled cells (Fig. 1A1-A3). Ghrelin binding was punctate and associated with neuronal perikarya similar to GHSR immunostaining suggesting ghrelin binding to GHSRs in the SNpc. No binding was observed when unlabelled ghrelin was added to the incubation solution.
DA neurons (n=11) in the SNpc were identified based on their characteristic Ih currents (Johnson and North, 1992). Spontaneous action potentials were recorded under current clamp for at least 10 minutes of stable recording. Ghrelin (1-3 μmol) was applied via bath application and significantly increased action potential frequency above baseline (Fig 1B & C), providing direct evidence that ghrelin promotes action potential firing in SNpc DA neurons. 10/11 identified DA neurons responded to ghrelin.
We measured striatal DA concentration using HPLC 3 hours after ip ghrelin injection (10 nmols) in mice. Ghrelin produced a robust and reproducible increase in DA concentration in the dorsal striatum (saline 119.4±9.9 vs ghrelin 153.7±4.7 ng/mg protein, n=7, p<0.05). These results are consistent with our recent study (Abizaid et al., 2006) and reports using in vivo microdialysis, in which ghrelin increases extracellular DA concentration in the ventral striatum (Jerlhag et al., 2006; Jerlhag et al., 2007).
Tyrosine hydroxylase (TH) is the rate-limiting enzyme of dopamine biosynthesis. Three hours after ghrelin injection (10 nmols, ip), TH mRNA expression was significantly increased in the midbrain compared to saline-injected controls (TH mRNA; saline 1.0±0.1 vs ghrelin 2.9±0.4 fold increase, n=10-11, p<0.05).
As ghrelin increased mitochondrial number in NPY neurons in the arcuate nucleus of the hypothalamus (Andrews et al., 2008), we tested whether ghrelin may also increase mitochondrial number in SNpc TH neurons using electron microscopy. Three hours after ghrelin injection, mice exhibited increased mitochondrial number in TH perikarya of the SNpc (saline 0.46±0.03 vs ghrelin 0.62±0.03 mitochondrial number/μm2, n=14-12, p<0.05. In order to validate our electron microscopic analysis of mitochondrial biogenesis in SNpc TH neurons, we examined NRF1 mRNA, an important transcriptional regulator of mitochondrial biogenesis, in the substantia nigra in saline and ghrelin-treated mice (Dhar et al., 2008). After 3 hours, ghrelin stimulated the fold increase of NRF1 mRNA relative to saline (saline 1.042±0.21 vs ghrelin 3.285±0.68, n=4&7, p<0.05) and thus substantiating the claim that ghrelin promotes SNpc dopamine function by increasing mitochondrial biogenesis as seen in NPY neurons of the hypothalamus (Andrews et al., 2008)
The observation that ghrelin promotes SNpc DA neuronal function and mitochondrial proliferation raises the possibility that ghrelin strengthens the nigrostriatal dopaminergic system during increased cellular stress. Seven days after mice were implanted osmotic minipumps delivering ghrelin (10 nmol/day) or saline, we injected one dose of MPTP (40 mg/kg) or saline. Seven days after MPTP injection mice were sacrificed and processed for TH immunohistochemistry or HPLC determination of striatal dopamine, DOPAC levels and dopamine turnover. In this paradigm, there was no protective effect of continuous ghrelin administration on total TH cell number or striatal dopamine levels after MPTP treatment (Supplemental Fig. 1).
We next analyzed the effect of exogenous ghrelin only when ghrelin levels would be elevated in physiological conditions, i.e., in the absence of food, as ghrelin is increased during negative energy balance (Fig. 2A-G). We treated mice with daily IP injections of exogenous ghrelin (10 nmols) or saline for 7 days before and 7 days after MPTP injections. Mice were injected immediately before the dark phase and food was removed overnight. Ghrelin- and saline-treated mice were given 3 grams of standard chow the following morning at 9:00 AM. On the seventh day of ghrelin treatment, MPTP was administered IP (40 mg/kg) to saline and ghrelin-treated mice.
There was no difference in estimated total number of TH cells in the entire SNpc in mice treated with ghrelin and saline (designated ghr/sal) or saline and saline (designated sal/sal) (Fig 2D). Mice treated with saline followed by MPTP (designated sal/MPTP) and mice treated with ghrelin followed by MPTP (designated ghr/MPTP) both displayed significantly TH cell loss in the SNpc. However, ghrelin treatment significantly attenuated TH cell loss in the SNpc in response to MPTP compared to sal/MPTP controls (Fig 2D). Ghrelin treatment restricted dopamine cell loss to 33% of control mice whereas saline treatment resulted in 49% dopamine cell loss compared to controls after MPTP lesion, indicating that ghrelin promoted dopamine SNpc cell survivability after MPTP intoxication. In line with this, ghrelin significantly attenuated the MPTP-induced loss of dopamine (Fig 2E) and DOPAC (Fig. 2F) while striatal DOPAC/dopamine ratios did not show a statistical difference (Fig. 2G). Dopamine in sal/MPTP mice was 65% lower than controls whereas ghr/MPTP showed a 45% loss of striatal dopamine. These experiments demonstrate that ghrelin can restrict dopamine and dopamine cell loss in a mouse model of PD when circulating ghrelin levels are boosted at the time when endogenous ghrelin is naturally induced. It should be noted that these mice normally ate approximately 3.5 grams of food per day, thus our paradigm represents a mild form of food restriction. Despite that mice fed in this manner did not significantly loss body weight data (Supplemental Fig. 2), only ghrelin-treated mice showed neuroprotection in the SNpc compared to saline-treated mice fed in the same restriction paradigm, it is likely that restricted, exogneous ghrelin treatment as we applied enhances cellular mechanisms that exert beneficial effects of calorie restriction.
To further examine the role of ghrelin and GHSR receptor in maintaining the integrity of the SN dopamine system, we analyzed the effect of MPTP in transgenic animals in which either ghrelin or the ghrelin receptor, GHSR, was ablated (Fig. 2H-N). Our results revealed that after saline injection there was no observable difference in TH-immunoreactive cells in the SNpc of ghrelin-/- mice compared to wild type littermates (Fig. 2K). MPTP produced a significant loss of TH neurons in both wt and ghrelin-/- mice compared to saline controls. However, the loss of TH neurons in ghrelin-/- mice was significantly increased, such that ghrelin-/- mice lost almost 50% more TH neurons compared to wt mice (Fig. 2K). To determine whether this loss of TH-immunolabelled neurons correlated with altered dopamine levels, we measured dopamine content in the dorsal striatum. Ghrelin-/- mice challenged with MPTP exhibited a significantly greater reduction in striatal dopamine and DOPAC content compared to wt mice treated with MPTP (Fig. 2L and M), and, DOPAC/dopamine ratios were also significantly increased ghrelin-/- mice treated with MPTP. There was no difference in dopamine content between wt and ghrelin-/- mice injected with saline.
Next we analyzed whether GHSRs mediate this effect of ghrelin. This approach utilized ghsr-/- mice generated by inserting a loxP-flanked transcription-blocking cassette into the endogenous GHSR allele (Zigman et al., 2005). When exposed to promoter specific Cre recombinase, the GHSR expression is established in a neuronal phenotype of choice. We used mice that express GHSR selectively in TH neurons, including those in the SNpc and VTA (designated GHSR homo/TG, Fig. 3A-C) as well as ghsr-/- mice (designated homo/wt, i.e. without cre recombinase reactivation) and controls (designated wt/wt). Mice were injected with MPTP or saline and TH neurons in the SNpc were counted. While MPTP significantly reduced TH neurons in SNpc of wt/wt mice (Fig. 3H), it produced a significantly greater loss of TH neurons in the SNpc in GHSR homo/wt mice (Fig. 3H). However, mice with GHSR reactivation in TH neurons exhibited restricted MPTP TH cell loss similar to that seen in wt/wt MPTP-treated mice (Fig. 3H). Taken together, these data indicate that ghrelin and ghrelin signaling directly in TH neurons through the GHSR are important to mediate ghrelin’s neuroprotective properties in the SNpc.
We showed that ghrelin promotes mitochondrial respiration and proliferation via activation of the AMPK-CPT1-UCP2 pathway (Andrews et al., 2008) that promotes fatty acid beta-oxidation. Interestingly, UCP2 was previously shown to be neuroprotective against MPTP-induced nigral dopamine cell loss (Andrews et al., 2005b; Conti et al., 2005). Hence, we tested whether ghrelin enhances mitochondrial respiration and biogenesis in an UCP2-dependent fashion.
First we revealed that GHSR containing neurons in the SNpc also express UCP2 (Fig. 3I-K). Next, we isolated mitochondria from the midbrain of wt and ucp2-/- mice treated with ghrelin for 3 hours (10nmols, ip) or saline and measured mitochondrial respiration. Ghrelin-treated wt mice exhibited significantly greater basal uncoupled respiration after oligomycin blocked electron transport through the ATP synthase (Fig. 3L), whereas ghrelin did not increase respiration after oligomycin in ucp2-/- mice (Fig. 3L). Activation of UCP2 with the fatty acid palmitate (Echtay et al., 2002) resulted in a significant increase in mitochondrial respiration in ghrelin-treated wt but not in ucp2-/- mice (Fig. 3M). Finally, FCCP driven respiration, which measures the total respiratory capacity, was also increased in ghrelin-treated wt but not in ucp2-/-mice (Fig. 3N). These results demonstrate that ghrelin regulates UCP2-dependent mitochondrial respiration in the SN. As we observed that ghrelin increased mitochondrial number in SNpc TH neurons (see above) and UCPs are important in promoting mitochondrial proliferation in response to increased uncoupled respiration in many tissues including the brain (Diano et al., 2003; Andrews et al., 2008) and muscle (Wu et al., 1999), we tested whether ghrelin produces a UCP2-dependent increase in mitochondrial number in the SNpc DA neurons. Three hours after ghrelin injection, wt mice exhibited increased mitochondrial number (Fig. 3O). However, ghrelin had no effect on mitochondrial number in ucp2-/- mice (Fig. 3O). To further demonstrate the mitochondrial biogenic potential of ghrelin via UCP2, we measured NRF1 mRNA 3 hours after ghrelin treatment relative to saline controls. In wt mice, ghrelin increased NRF1 mRNA relative to saline (wt saline 0.983±0.09 vs wt ghrelin 2.240±0.03, n=6, p<0.05) whereas ghrelin had no effect in ucp2-/- mice (ucp2-/- saline 0.929±0.19 vs ucp2-/- ghrelin 1.120±0.35, n=6, p<0.05). These results confirm that ghrelin-induced mitochondrial respiration and proliferation is dependent on UCP2 in the SNpc.
The effects of ghrelin in mitochondrial responses in the SN are in line with ghrelin-mediated fatty acid beta oxidation recently unmasked in the hypothalamus (Andrews et al., 2008). We propose that under conditions of mild negative energy balance increased levels of ghrelin will activate mitochondrial respiration and contribute to enhance fatty acid oxidation in nigral neurons. To determine whether ghrelin affects the long chain fatty acyl CoA pool in the midbrain, we examined long chain fatty acyl CoA (LCFA) species in the SN using mass spectrometry analysis. Three hours after peripheral ghrelin injection, there was a significant increase in total LCFA in the midbrain compared to saline controls (saline 40.9±0.3 vs ghrelin 52.0±1.3 nmol metabolite/g tissue, n=5, p<0.05). These results are consistent with the idea that as ghrelin promotes mitochondrial respiration, it also supports continued oxidation by supplying the SNpc with LCFA for ongoing fuel utilization (Horvath et al., 2009). The primary LCFAs in the midbrain were palmitic acyl CoA (C 16:0), stearic acyl CoA (C18:0) and oleic acyl CoA (C18:1) (Fig. 4A).
As UCP2 buffers and neutralizes ROS while promoting fatty acid oxidation (Andrews et al., 2008) and neuroprotection (Andrews et al., 2005a), we tested whether ghrelin treatment of wt and ucp2-/- mice also affects ROS production in the midbrain. There was no difference in ROS production after saline or ghrelin injection in wt mice (Fig. 4B). However, in ucp2-/-mice, ghrelin produced a significant increase in ROS production relative to saline controls (Fig. 4C, E and G). In order to examine how ghrelin affects ROS production in identified TH neurons in the SNpc, we injected dihydroethidium as it is oxidized by superoxide to red fluorescent ethidium. Injection of ghrelin lowers ROS production in nigral TH cells in UCP2 wt mice (Fig 4D-H) due to the activation of the UCP2 ROS buffering ability, however ghrelin injection into ucp2-/- mice dramatically increases in vivo TH ROS levels in the SNpc.
The data presented above suggest that UCP2 may play a role in mediating the neuroprotective effects of ghrelin on SNpc dopamine neurons. To test this, we administered ghrelin to wt and ucp2-/- mice using the feeding paradigm described above. Total TH cell number in the SNpc was analyzed after the treatment ended. At baseline, sal/sal wt mice had approximately 10288 ± 574 TH cells and ghrelin treatment had no significant effect as TH cell numbers were not significantly different (9568 ± 168 TH cells). MPTP significantly reduced TH cell numbers in both sal/MPTP and ghr/MPTP wt mice, however, ghrelin restricted this MPTP-induced loss (sal/MPTP 6059 ± 88 vs ghr/MPTP 7202 ± 254, n=6, p<0.05). In sal/sal and ghr/sal ucp2-/- mice there was no difference in total SNpc TH cell number (sal/sal 9082 ± 156 vs ghr/sal 9338 ± 388, n=6). As expected MPTP treatment produced a significant reduction in TH cell number, however unlike in the ghr/MPTP wt mice, ghrelin did not attenuate the loss of TH neurons caused by MPTP in ucp2-/- mice (sal/MPTP 5580 ± 195 vs ghr/MPTP 5734 ± 431) suggesting that ghrelin requires UCP2 to initiate compensatory neuroprotection through enhanced mitochondrial function and biogenesis.
The present study shows that ghrelin promotes the electrical activity and dopamine output of the nigrostriatal DA system and that ghrelin has neuroprotective effects in a mouse model of PD. We suggest that the neuroprotective effects of ghrelin on the nigrostriatal dopamine system stem from two major effects of this metabolic hormone. First, ghrelin has an acute effect on the firing rate of SNpc DA neurons, which enhances dopamine availability during the course of degeneration and lowers the loss of dopamine levels in the dorsal striatum. Second, ghrelin-induced enhancement of UCP2-dependent mitochondrial respiration and proliferation provides a bioenergetic status that makes these neurons more resistant to cellular stress. UCP2 is a mitochondrial protein known to protect SNpc dopamine cells from MPTP intoxication (Andrews et al., 2005b). Ghrelin also regulates UCP2 mRNA in the brain (Andrews et al., 2008) consistent with the role of UCP2 as a component of ghrelin-induced neuroprotection.
Ghrelin-/- and ghsr-/- mice were more susceptible to DA cell loss in the SNpc and DA loss in the striatum after MPTP than their wt littermates. This suggests that any factor leading to reduced ghrelin production or secretion, whether genetic or environmental, may predispose individuals to nigrostriatal dopaminergic dysfunction. Obesity is one such physiological condition that lowers serum ghrelin (Tschop et al., 2001) and this metabolic variation of ghrelin in humans may underlie an inherent predisposition to nigrostriatal dopaminergic dysfunction. This may have implications for PD, as currently 90% of PD is considered sporadic (Dauer and Przedborski, 2003) and epidemiological reports show that obesity, diabetes and elevated body mass index predispose to future risk of PD. In contrast with obesity, ghrelin levels are increased during calorie restriction (Tschop et al., 2000; Ariyasu et al., 2001), the only known experimental manipulation known to extend lifespan (Bordone and Guarente, 2005). Calorie restriction confers neuroprotection against MPTP intoxication (Guan et al., 1997; Maswood et al., 2004) suggesting that at least some of the beneficial effects of calorie restriction on neuroprotection may be mediated by ghrelin.
The effects of ghrelin on neuroprotection were dependent on metabolic state (redox level), as ghrelin had no protective effect when administered via osmotic minipumps during ad libitum feeding. Ghrelin only resulted in neuroprotection when it was administered during its physiological “window period”. This is consistent with the recently uncovered cellular effects of ghrelin, in which activation an AMPK-CPT1-UCP2 pathway promotes fatty acid beta oxidation and neuronal function in the hypothalamus (Andrews et al., 2008; Suppl. Fig. 3). In the SNpc, we suggest that ad libitum feeding, specifically glucose from carbohydrate load, prevents the action of ghrelin on the AMPK-CPT1-UCP2 pathway and therefore does not allow for mitochondrial neuroprotective mechanisms to affect cell survivability. Consistent with this idea, brain glucose regulates malonyl-CoA and inhibits the AMPK-CPT1-UCP2 pathway (Wolfgang et al., 2007). Interestingly, this raises the intriguing and novel hypothesis that SNpc TH neurons are metabolic sensors of energy status. Dysfunction of this energy sensing ability may underlie how obesity affects nigrostriatal neurodegeneration. In our second model, removal of food at the time of ghrelin injection allows for full activation of the AMPK-CPT1-UCP2 pathway allowing increased mitochondrial respiration and biogenesis and subsequent neuroprotection (Suppl. Fig. 3). This model allows for enhanced fatty acid ß-oxidation, which maybe the dominant fuel supply during negative energy balance. Ghrelin mobilizes fatty acid release in to the circulation after 30 minutes (Andrews et al., 2008) and therefore can supply the brain with energy substrates for ß-oxidation. In the present study showing that ghrelin increases LCFA CoAs availability in the midbrain as well is consistent with this hypothesis. We propose that under conditions of negative energy balance, such as calorie restriction, ghrelin increases the neuronal fatty acid fuel supply to midbrain dopamine neurons that results in sustained fatty acid ß-oxidation and maintenance of cellular energy status of dopamine cells. In line with this, LCFAs including oleic acid, are known to promote dopamine production in an immortalized cell line (Heller et al., 2005).
Fatty acid ß-oxidation generates ROS, which is then buffered by activation of UCP2. In support of this, ucp2-/- mice displayed increased midbrain ROS production in response to ghrelin compared with wt mice, as observed in the hypothalamus of ucp2-/- mice (Andrews et al., 2008). Further analysis of ROS production in TH neurons showed that ghrelin significantly lowers ROS in UCP2 wt mice but greatly enhances ROS in ucp2-/- mice. Finally, ghrelin treatment of ucp2-/- mice did not attenuate SNpc TH cell loss after MPTP treatment as observed in wt controls. In this situation, the ghrelin-induced fatty-acid driven mitochondrial respiration and subsequent ROS production cannot be buffered or neutralized by UCP2, thus leading to dangerously high levels of ROS and promoting cell degeneration in response to MPTP. This is consistent with previous reports showing that UCP2 buffers ROS and protects nigral TH cells against MPTP-induced TH cell death (Andrews et al., 2005b; Conti et al., 2005).
Furthermore, ghrelin promotes UCP2 function in both the hypothalamus (Andrews et al., 2008) and the SNpc suggesting that this is a common mechanistic phenomenon for the action of ghrelin. Indeed, UCP2 may play an important role in other regions of the brain where ghrelin promotes neuronal function such as the hippocampus and the VTA (Abizaid et al., 2006; Diano et al., 2006).
Beyond the implications of our current study on neurodegeneration, our observations may have relevance for the regulation of food intake and depression. Ghrelin was originally found to stimulate appetite via the central nervous system (Tschop et al., 2000), with sites of action in various areas, including the hypothalamus (Cowley et al., 2003) and the ventral tegmental dopamine system (Abizaid et al., 2006). Experimental evidence suggests that the nigrostriatal dopamine system is critically involved in the regulation of feeding (Szczypka et al., 2001; Hnasko et al., 2006). Moreover, a recent study showed that ghrelin defends against depression (Lutter et al., 2008), which is mediated by ghrelin’s action on the midbrain dopaminergic system described herein and elsewhere (Abizaid et al., 2006; Krishnan et al., 2007). PD is also associated with depression (Storch et al., 2008) and a decrease in appetite leading to weight loss (Bachmann and Trenkwalder, 2006), events that may be related to the lack of ghrelin’s action in the SN and dorsal striatum PD patients. Thus, ghrelin as a therapeutic treatment for PD may not only prevent TH cell degeneration but also promote food intake, weight gain and anti-depressive symptoms.
This work was supported by a grant for the Michael J Fox Foundation and NIH grant DK-060711 to TLH, a New Zealand Foundation for Research, Science and Technology fellowship, Monash Fellowship and NHMRC grant (NHMRC 546131) to ZBA, and, NIH grant NS-056181 to JDE.