Green Coffee Extract Blockades Haloperidol-Induced Catalepsy and Contralateral Rotations Induced by Apomorphine After Unilateral 6-OHDA Lesion on Striatum

Epidemiological studies have shown an inverse association between coffee consumption and the development of Parkinson’s disease (PD). The present investigation aimed to evaluate the effects of caffeine and green (non-roasted) coffee extract in experimental models of PD. The effects of the oral treatment with green coffee extracts (CE, Coffea arabica 100 or 400 mg/kg) and caffeine (31.2 mg/kg) were evaluated on the catalepsy induced by haloperidol in mice and unilateral 6-OHDA lesion of medial forebrain bundle (MFB) or striatum in rats. Also, the in vitro antioxidant activity and the monoamine levels in the striatum were investigated. CE presented a mild antioxidant activity and administration decreased the catalepsy index. CE at the dose of 400 mg/kg induced ipsilateral rotations 14 days after the lesion. However, the 30-day CE and caffeine treatments did not interfere with the animals’ rotation after apomorphine or methamphetamine challenges in animals with MFB lesion, nor on monoamines levels. Furthermore, CE and caffeine were effective in inhibiting the asymmetry between ipsilateral and contralateral rotations induced by methamphetamine and apomorphine in animals with lesion in the striatum but did not avoid the monoamines depletion. These results suggest a pro-dopaminergic action of CE, although its mechanism remains unclear.


Introduction
Parkinson's disease (PD) is considered the second most common neurodegenerative disorder, with an estimated prevalence of 1 to 3% in the population over the age of 65 years, and with an increasing prevalence among older people 1 . PD is mainly characterized as a progressive degeneration of dopaminergic neurons in substantia nigra pars-compacta, and the presence of Lewy bodies (intraneuronal aggregates of α-synuclein and other misfolded proteins) is one of the most important neuropathological hallmarks 2 . Although the pathophysiology of PD is complex, several evidences have shown that the increased dopamine oxidation associated with decreased endogenous antioxidant defense are key elements in the development of the disease 2 .
The current medicines used in PD are able to improve the quality of life, but cannot avoid the progressive neuron degeneration, making it important to identify potential protective factors such as diet or life style that could diminish the risk of developing PD, and also, to nd new drugs which could avoid the disease progression. Several observational studies and meta-analysis have shown an inverse correlation between coffee consumption and incidence of PD, what leads to the hypothesis that caffeine presents a protective effect on PD [3][4][5][6] .
Coffee is the most consumed beverage worldwide, being usually prepared by the infusion of roasted seeds of Coffea arabica L. or its varieties. The psychological effects of coffee are largely due to the presence of caffeine, which acts mainly through the adenosine receptors antagonism, more speci cally as antagonist of adenosine A 1 and A 2A receptors 7 . The adenosinergic system is co-localized with dopaminergic neurons in several brain regions, in special in the basal ganglia, where it negatively modulates the dopaminergic system 7,8 . The blockage of adenosine A 2A receptors on striatal neurons has been suggested to explain the neuroprotective action of caffeine in PD [8][9][10][11] . Recently, istradefylline, a selective adenosine A 2A receptor antagonist, was approved by FDA as an add-on treatment of PD patients under treatment with levodopa presenting off episodes 12 .
Nevertheless, although the neuroprotective effect of coffee on PD is attributed mainly to caffeine, several authors emphasize the importance of polyphenols and other constituents, suggesting that they can strongly contribute to the bene cial effects of coffee, tea and other caffeinated beverages 13,14 . It has been previously demonstrated that the amount of chlorogenic acids and its derivatives, caffeic acid, ferulic acid, quinic acid and several avonoids, depends on the variety of coffee, roasting process and methods of preparation 15,16 . Depending on the temperature and roasting time, most chlorogenic acids and other constituents are changed or decomposed, while caffeine remains almost unaffected 15,17 .
Chlorogenic acids (quinic esters of hydroxycinnamic acids) is a family of esters formed between quinic acid and three most common trans-cinnamic acids: caffeic acid, ferulic acid and p-coumaric acid. The main chlorogenic acids classes in green coffee are: caffeoylquinic acids, dicaffeoylquinic acids, and feruloylquinic acids 18 .
Currently, there is evidence of the protective role of chlorogenic acids and its derivatives in PD and Alzheimer's disease, based on both epidemiological and experimental studies 13,14,19,20 . Considering this, the present investigation aimed to evaluate the effects of green (non-roasted) coffee extract, rich in chlorogenic acids, in the following experimental models of Parkinson's disease: catalepsy induced by haloperidol, and rotational behavior after unilateral 6-OHDA lesion of medial forebrain bundle or striatum.

Extract
The green coffee extract (CE) was produced by spray dryer technique with dry unroasted seeds of coffee (Coffea arabica L) by Centro ora (Botucatu, Brazil). The extract was evaluated for its physical and chemical characteristics by IT2 100 method (standardized in 21.04% of chlorogenic acid, 27-33% of caffeoylquinic acid and 7.5% to 8% of caffeine), according to the report provided by the manufacturer.

Drugs
Apomorphine, caffeine and 6-hydroxydopamine (6-OHDA) were purchased from Sigma (St. Louis, USA) and haloperidol (Haldol ® ) from Janssen Pharmaceutica N.V. (Beerse, Belgium). Methamphetamine was donated by the Brazilian Federal Police. For HPLC analysis, caffeine standard was purchased from Fluka (Munich, Germany) and 5-O-caffeoylquinic acid standard from Sigma-Aldrich (Milano, Italy). HPLC-grade methanol was purchased from Merck (Darmstadt, Germany) and HPLC-grade water was prepared from distilled water using a Milli-Q system (Millipore, Waters, Milford, USA).

Caffeine dosage
Caffeine quanti cation in CE was carried out through phase reversed high-performance liquid chromatography with Diodo Array detector (HPLC/DAD). The HPLC-DAD analyses were conducted on a Hewlett Packard 1090 II equipped with a degasser, an autosampler, and diode array detector. The identi cation of caffeine was based on retention time, UV spectra comparison with a standard and spiking. Quantitation was obtained through a calibration curve made with known concentrations of caffeine. The calibration curve for caffeine (Y= 2935.6 X) was found to be linear with R 2 = 0.99.
The coffee extract was dissolved in water: methanol (80:20) v/v (5 mg / 2mL) and ltered with a 0.45 μm polytetra uoroethylene (PTFE) lter, prior to injection of 31.2 μL into the HPLC system, using a reverse phase, C18, Spherisorb ODS II (Hewlett Packard) column (4.6 × 250 mm, 5 μm), connected to a guard column. Spectral UV data from all peaks were collected in the spectral range 240 -400 nm, and chromatograms were recorded at 330 (caffeoylquinic acid derivatives and feruloylquinic acids derivatives) and 270 nm (caffeine) 21 . The mobile phases consisted of eluent A (0.1% aq. formic acid) and eluent B (methanol), with the following gradient pro le: 0 min -20% B in A; 10 min -30% B in A, 20 min -50% B in A; 30 min -70% B in A; 40 min-90% B in A; 45 min -40% B in A and nally returned to the initial conditions (20% B in A) to re-equilibrate the column prior to another run. The ow rate was kept constant at 0.8 mL/min, and the temperature of the column was maintained at 28°C.

Animals
Male Swiss mice (40 to 50 g) and male Wistar rats (300 to 400 g) 2-5 months old were provided by the bioterium of Psychobiology from Universidade Federal de São Paulo (UNIFESP). Animals were kept in rooms with controlled temperature (23 ± 2ºC) and light/dark cycle of 12 hours with water and food ad libitum. The project was approved by the Ethics Committee of UNIFESP (CEP #1946/06) and we followed the ARRIVE guidelines principles of the use of laboratory animals.
Animals' treatment CE (doses of 100 and 400 mg/kg) and caffeine (31.2 mg/kg -the equivalent caffeine in 400 mg/kg of CE) were solubilized in water and orally administered by gavage in the volume of 1 mL/kg (rats) and 10 mL/kg (mice). The control animals received the vehicle water at the same volume by gavage.

Evaluation of antioxidant capacity
Antioxidant activity was evaluated by the lipid peroxidation inhibition measured by malondialdehyde concentration in rat brain homogenate. Tissue homogenate was prepared with phosphate buffer and centrifuged at 3000 rpm for 15 minutes. The supernatant was collected and then diluted again in phosphate buffer (1:3). Five concentrations of CE were added to each homogenate tube as follows: 4.17 µg/mL; 8.33 µg/mL; 12.50 µg/mL; 16.67 µg/mL; 33.33 µg/mL ( nal concentrations), and then incubated in water bath with thiobarbituric acid for 60 minutes at 37 ºC. The experiment was carried out according to the methodology described by Stocks et al. 22 . The antioxidant activity was determined for each concentration of the extract and the concentration inhibiting 50% of lipid peroxidation (Q 1/2 ) was calculated by linear regression using the log of the concentration and the percentage of inhibition (mean of 4 assays).

Catalepsy induced by haloperidol
Groups of 10-13 mice were acutely or for 15 consecutive days treated by gavage with vehicle (controls), CE at doses of 100 or 400 mg/kg (experimental groups) or caffeine at dose of 31.2 mg/kg (positive control). Thirty minutes after the acute or the last administration of CE, caffeine or vehicle, the animals received haloperidol (5 mg/kg, ip), except the negative control group, which received saline (ip).
Forty-ve minutes after the administration of haloperidol (or saline) the animals were placed with their forepaws positioned on a bar suspended 5 cm above the bench and the catalepsy time (immobility) was recorded 23 . The number of times each animal got off the bar over 10 minutes was recorded with a maximum of 10 times. The catalepsy index was calculated by the rate between the total catalepsy time and the number of times the mice left the bar.
Unilateral lesion induced by 6-OHDA Surgery procedure Male Wistar rats were anesthetized with ketamine (90 mg/kg, ip) and xylazine (5 mg/kg, ip) and placed on a stereotaxic apparatus. After the skin incision and the skull exposition, a hole was performed to the bone using a dentist drill and the neurotoxin 6-OHDA (12 μg/2 μl buffered saline in 0.2% ascorbic acid) was infused with an injection pump 24 in the medial forebrain bundle (MFB) of the right hemisphere (experiment 1) or in the right striatum (experiment 2) in the lesioned group, while sham-operated animals were infused with 2 μl of buffered saline. The coordinates for MFB lesion were: AP, -1.9 mm; ML, -1.9 mm; DV, -7.2 mm of bregma; and for striatum: AP, +0.5 mm; ML, -2.5 mm; DV -4.5 mm; according to the brain atlas Paxinos and Watson 25 . After surgery, dental wax was applied on the skull to close the access hole, the head skin was sutured, and the animals were allowed to recover in individual boxes with food, water and controlled temperature. The experimental design of experiments 1 and 2 is shown on Fig. 1. A pilot study showed that the lesion was effective in inducing neuronal loss in the substantia nigra ( Fig. 1 qualitative analysis). Experiment 1A: Effect of acute treatment with CE on the rotational behavior of rats with lesion on MFB Groups of 10-15 rats lesioned on MFB with 6-OHDA were challenged acutely on the 14 th day with CE (100 or 400 mg/kg), caffeine (31.2 mg/kg) or vehicle by gavage. An extra group of rats that received buffered saline instead of 6-OHDA (sham-operated group) was challenged with CE (400 mg/kg) on the 14 th day.
Immediately after the respective treatments, the animals were evaluated for 60 minutes and the number of contralateral and ipsilateral complete rotations (360 o ) in a cylinder was recorded 26 . On the 15 th day, all animals received apomorphine (0.1 mg/kg, sc), and on the 16 th day, methamphetamine (2.5 mg/kg, ip), and the number of ipsilateral and contralateral rotations was recorded again. Experiment 1B: Effect of repeated treatment with CE on the rotational behavior of rats with lesion on MFB For this experiment, groups of 12-15 rats lesioned on MFB were orally treated for 30 days with CE (100 or 400 mg/kg), caffeine (31.2 mg/kg) or vehicle, by gavage. An extra group of sham-operated rats received vehicle by gavage for the same period. On the 15 th and 29 th days after surgery the animals were evaluated for 60 minutes regarding the number of ipsilateral or contralateral rotations, after administration of apomorphine (0.1 mg/kg, sc). On 16 th and 30 th days, methamphetamine (2.5 mg/kg, ip) was administered and the number of rotations was once again recorded. Experiment 2: Effect of pre-and post-treatment with CE on the rotational behavior of rats with striatal lesion Groups of 11-15 rats were pre-treated for 15 days with CE (100 or 400 mg/kg), caffeine (31.2 mg/kg) or vehicle by gavage. One hour after treatment on day 15 the rats were submitted to the stereotaxic surgery and 6-OHDA (lesioned group) or buffered saline (sham-operated group) was administered into the right striatum. The treatment continued for another 15 days and the animals were challenged with apomorphine (15 th day after surgery) and methamphetamine (16 th day) as previously described. The treatment with CE or caffeine on the days of rotational tests was made after the behavioral evaluation to avoid acute interference with apomorphine or methamphetamine.

Central monoamines dosage by HPLC
After 24 hours of the last behavioral evaluation, the animals from experiments 1B and 2 were euthanized for brain collection, and the brains were frozen at -80 o C. Later, the right (lesioned) and left (control) striatum were dissected, weighed, homogenized, centrifuged and then ltered in nitrocellulose membrane as described by Machado et al. 27 . Precipitates were diluted in a solution of 0.1 N of sodium hydroxide and evaluated for protein concentration using a colorimetric quanti cation kit (Pierce Chemical, Rockford, USA) and supernatants (20 µL) were used for monoamines quanti cation through high-performance liquid chromatography (HPLC) according to method described in details by Machado et al. 27 . Each sample was analyzed in duplicate for concentrations of dopamine (DA), noradrenaline (NE), serotonin (5-HT) and their metabolites dihydroxyphenylacetic acid (DOPAC), homovanillic acid (HVA) and 5hydroxyindolacetic acid (5-HIAA). The recovery rate of the analytes was determined by adding a predetermined amount of internal standard dihydroxybenzylamine for the homogenization of tissue. The peak areas of samples were compared with peak areas of standards of each neurotransmitter or metabolite, which were injected at known concentrations, at intervals between runs.

Statistical analysis
Statistical comparisons were performed using GraphPad Prism 8.3 software. After checking for normality distribution using Shapiro-Wilks test, Kruskal-Wallis test followed by Mann-Whitney was applied to analyze the results of rotational behavior test along with Wilcoxon to compare ipsilateral and contralateral rotations among each group. Monoamines levels were analyzed using Kruskal-Wallis test and One-way ANOVA followed by Tukey to analyze the catalepsy. The results are expressed as mean ± standard error (SEM) and the raw data is available in the Supplementary Data S1. Statistical signi cance was considered at p<0.05.

Caffeine dosage
Chlorogenic acids are a family of esters formed between quinic acid and trans-cinnamic acids, such as caffeic, p-coumaric and ferulic acid, all of which can be esteri ed at one or more of the hydroxyls at positions 1, 3, 4, and 5 of quinic acid to generate a series of positional isomers. In RPHPLC, the isomers of the monocaffeoylquinic acids elute in the order of 3-, 5-, and 4-isomer and for dicaffeoylquinic acids, the elution order is 3,4-isomer, 3,5-isomer, and, nally, 4,5-isomer 18 .
The main constituents found in the CE are listed in the Supplementary Data S2, together with their retention times and maximum absorption wavelength observed in their respective ultraviolet spectrum 28 . Caffeoylquinic acid and feruloylquinic acid derivatives exhibited a UV/vis maximum wavelength absorption between 310 and 330 nm (band I) and a shoulder between 290 and 300 nm (band II), while caffeine exhibited a UV/vis maximum wavelength absorption at 272 nm 17 . Caffeine was the main constituent detected at a wavelength of 272 nm with the retention time of 22.18 minutes, in which the percentage area corresponded to 36%. Through the calibration curve, the caffeine content was calculated as 7.8%.
Caffeoylquinic acid and feruloylquinic acid derivatives were the main constituents detected at a wavelength of 330 nm, but caffeine is hardly detectable due to its low absorption at λ = 330 nm. Caffeoylquinic acid derivatives were the main constituent detected at a wavelength of 330 nm, at 18.19 -19.44 minutes and exhibited a percentage area corresponding to 69% (Supplementary Data S2).

Evaluation of antioxidant capacity
The percentage of inhibition of the spontaneous lipid peroxidation in rat brain homogenate for the different concentrations of CE is shown in Fig. 2. The value of Q 1/2 obtained by linear regression from 4 assays was 8.73 µg/mL.

Catalepsy induced by haloperidol
Haloperidol (5 mg/kg, ip) induced catalepsy on the vehicle control group when compared to the negative control group (Fig. 3), demonstrating that the drug was effective in producing dopaminergic blockade, as expected. The acute administration of caffeine and CE (100 mg/kg) decreased the catalepsy index (Fig.  3a) when compared to the vehicle group after haloperidol injection (F=5.127; p=0.0013 followed by Tukey p=0.0097 for caffeine and p=0.0455 for CE 100 mg/kg). Also, the administration for 15 days with CE (100 and 400 mg/kg) decreased the catalepsy index in comparison to the vehicle group (F=6.923; p=0.0010 followed by Tukey p=0.0020 for CE100 and p=0.0262 for CE 400 mg/kg) (Fig. 3b).
Unilateral lesion induced by 6-OHDA Experiment 1A: Effect of acute treatment with CE on the rotational behavior of rats with lesion on MFB Rats with unilateral lesion on MFB were acutely challenged with vehicle, caffeine or CE (14 th day), apomorphine (15 th day), and methamphetamine (16 th day). The challenge with caffeine and CE 400 mg/kg increased the number of ipsilateral rotations on lesioned rats, compared with lesioned rats treated with vehicle (H=15.104; p=0.004 followed by Mann-Whitney p=0.0218 for caffeine and p=0.0145 for CE 400 mg/kg), but it did not alter the number of contralateral rotations, as shown in Fig. 4a. There was signi cant difference between the number of contralateral and ipsilateral rotations in the group's caffeine (Wilcoxon, p=0.0078) but not in the other groups.
The lesion induced by 6-OHDA was effective in inducing contralateral rotations after apomorphine (U=811; p<0.0001) or ipsilateral rotations after methamphetamine (U=624.5; p<0.0001) administration when compared to sham-operated group, as can be observed in Fig. 4b and Fig. 4c, respectively. Experiment 1B: Effect of repeated treatment with CE on the rotational behavior of rats with lesion on MFB The effect of repeated treatment with CE on the rotational behavior of rats with right lesion of MFB was evaluated after challenge with apomorphine (15 th and 29 th days) and methamphetamine (16 th and 30 rd days). Fig. 5a shows that apomorphine induced contralateral rotations in all MFB experimental lesioned groups when compared to sham-operated group on 15 th day (H=22.545; p=0.0002 followed by Mann-Whitney p<0.05). The same result was observed at ipsilateral rotations induced by methamphetamine on 16 th day (H=22.280; p=0.0002 followed by Mann-Whitney p<0.05) (Fig. 5b). However, neither CE nor caffeine treatment changed the rotational behavior when compared with vehicle control group both after apomorphine and methamphetamine challenge (p>0.05). Similar results were observed on the 29 th day with apomorphine (Fig. 5c) and 30 rd day with methamphetamine (Fig. 5d) and these results were not altered by CE or caffeine treatment (p>0.05).

Experiment 2: Effect of pre-and post-treatment with CE on the rotational behavior of rats with striatal lesion
Rats were pre-treated with CE or caffeine for 15 days before the surgery and then for more 15 days after striatal lesion, prior the rotational evaluation. Typical rotations were observed after apomorphine (0.1 mg/kg, sc) or methamphetamine (2.5 mg/kg, ip) administration on vehicle control rats with striatal lesion, although the statistical analysis has not shown differences when compared with sham-operated group, due to the great variability observed (Fig. 6a and Fig. 6b). On the other hand, statistical difference was observed between ipsilateral and contralateral rotation for the vehicle control group challenged with apomorphine (Wilcoxon p=0.0151). Also, the number of ipsilateral rotations was more than twice that of contralateral rotations for the vehicle control group, but the values were not statistically signi cant (Wilcoxon p=0.2136).
Lesioned rats treated with CE or caffeine did not differ from the sham-operated group as the number of rotations and there was no difference between contralateral and ipsilateral rotations on intra-group comparison ( Fig. 6a and Fig. 6b).
Central monoamines dosage by HPLC Table 1 shows the monoamine levels measured on the right and the left striatum of animals with MFB lesion and non-lesioned rats (sham-operated group). There is an apparent decrease of DA, NE, 5-HT and some of their metabolites on vehicle lesioned group when compared to sham-operated group on right striatum, but the values were not statistically signi cant (p>0.05). Nevertheless, the treatments with CE or caffeine did not change the levels of monoamines. No signi cant effects were observed on the left (nonlesioned) side. Table 2 shows the quanti cation of monoamines in the right (lesioned) and left (control) striatum of rats with striatal lesion and sham-operated animals. No signi cant changes (p>0.05) were observed regarding quanti cation of neurotransmitters or their metabolites in both hemispheres (comparison among groups).

Discussion
Epidemiological studies have shown an inverse correlation between regular coffee intake and development of Parkinson's disease 3,4,6,29 . Experimental studies support the neuroprotective effect of caffeine through its adenosinergic antagonism 11,30,31 and also suggest that other mechanisms might contribute to this protective action, such as the anti-in ammatory effect, inhibition of α-synuclein aggregation and modulation of other pathways [32][33][34] .
It is known that antioxidant compounds can potentially decrease the progression of neurodegenerative diseases such as Alzheimer's disease and PD, in which oxidative stress participates in the induction of cell death process 35,36 . The phytochemical analysis con rmed that the green coffee extract employed in the current study is rich in chlorogenic acids (caffeoylquinic acid and feruloylquinic acid derivatives). The content of caffeine was calculated in 7.8% and used to de ne the dose of caffeine group in the behavioral tests.
The antioxidant activity of the green CE was con rmed in our study using rat brain homogenates by the lipid peroxidation assay. Previous studies compared the antioxidant capacity of green and roasted coffee in different tests. The literature is controversial about the effect of roasting process on the antioxidant capacity of coffee, but some studies suggest that melanoidins and other compounds formed during the roasting also present elevated antioxidant capacity 15,16,34,37,38 . Daglia et al. 38 showed that green coffee presented higher antioxidant activity on beta-carotene-linoleic acid in vitro assay, while roasted coffee was more potent on lipid peroxidation assay. This divergence could be explained by the difference in the species or varieties of coffee, the type of extract, preparation method, roasting time and methodologies used to evaluate the antioxidant activity.
PD is characterized primordially as a motor dysfunction, and catalepsy and akinesia induced by haloperidol have been used as animal models of PD and to study pro-and anti-dopaminergic drugs. Moo-Puc et al. 39 observed an anticataleptic effect in rats treated with low doses of caffeine and suggested that caffeine could be an adjunctive therapy to reduce the doses of anticholinergics drugs in PD. Another study showed that chronic consumption of caffeine on drinking water for 6 months produced perdurable resistance to catalepsy induced by haloperidol in rats 40 . Selective adenosine A 2 antagonists are able to reduce the effect of haloperidol con rming the antagonistic interaction between adenosine A 2A and dopamine D 2 receptors and suggesting the involvement of adenosine A 2 receptors in the mechanisms of catalepsy 41 . The green CE used in our study probably exerts an indirect action on the dopaminergic system, demonstrated by inhibition of haloperidol-induced catalepsy, which was more evident after repeated treatment for 15 days with CE at 400 mg/kg. Such result, combined with experimental studies of caffeine and other coffee components as neuroprotective agents 11,30,31,42−44 support the assumption of bene cial effect of coffee on PD. To evaluate this hypothesis, we produced unilateral lesion on nigrostriatal dopaminergic pathway by injection of 6-OHDA neurotoxin in MFB and striatum. The rotational behavior model after unilateral lesion with 6-OHDA is one of the most used animal models to study PD and drugs with potential neuroprotective effect 26, 45,46 . The unilateral lesion leads to unbalance of dopaminergic neurotransmission between left and right striatum which can be evaluated after challenge with dopaminergic drugs, such as apomorphine and amphetamine. Our data indicate that lesioned animals had the expected rotational behavior when challenged with apomorphine and methamphetamine, in accordance with data from literature 24,47,48 . The fact that animals challenged with CE presented ipsilateral rotations provides evidence of its action on the dopaminergic system, indicating an effect similar to methamphetamine, an indirect agonist. Cauli et al. 49 showed that the administrations of caffeine for two weeks could increase dopamine release in the intact side of the striatum, consistent with the ipsilateral direction of the rotational behavior. Moreover, it has been shown that caffeine does not bind to dopamine receptors, suggesting that this action has occurred by indirect mechanism, such as adenosinergic modulation. Furthermore, other coffee constituents would also be responsible for the observed effect.
The effects of the repeated treatment with CE before the challenge with apomorphine and methamphetamine were also analyzed. The treatment started 24 hours after the MFB injury with 6-OHDA and the animals were treated for 30 days with the objective of evaluating whether the treatments could exert a neuroprotective effect and avoid the asymmetry on the rotational test. There was no difference in the type and number of rotations among different experimental groups, including the caffeine group. This result does not corroborate with a previous study in which caffeine (15 mg/kg, ip) given acutely partially prevented the impairment of contralateral forepaw stepping in rats with MFB unilateral lesion induced by 6-OHDA 50 . These authors also found a similar effect after acute injection of caffeine at dorsal striatum and external globus pallidus (20-40 µg), but in both studies the authors evaluated only the behavioral response 20 minutes before the test, while the lesion itself was not quanti ed. Bové et al. 51 discuss that the dose and number of sites of 6-OHDA may produce different effects, making di cult to compare studies using different protocols. The route of administration, dose and time of treatment with the experimental drug are also fundamental to achieve the drug e cacy.
We also evaluated the effect of CE and caffeine in the rotational behavior of rats with 6-OHDA unilateral lesion in the striatum. The animals from the control lesioned-group presented a suggestive increase in the contralateral and ipsilateral rotations (after apomorphine and methamphetamine challenge, respectively) compared to sham-operated group, although the statistical analysis did not indicate signi cant differences. The intra-group comparison showed that the vehicle control group presented difference between ipsilateral and contralateral rotations after apomorphine challenge. On the other hand, the rats treated with CE or caffeine showed behavior similar to those of sham-operated group, where the number of ipsilateral and contralateral rotations was equivalent. This data may suggest that the treatment was able to compensate the unbalance between right and left striatum by indirect mechanisms, which would indicate some level of neuroprotection. This result is in agreement with Machado-Filho et al. 32 which employed the same experimental model and observed that caffeine (10-20 mg/kg) given orally for two weeks moderately reduced the contralateral rotations induced by apomorphine. The content of caffeine in our study was 7.8 and 31.2 mg/kg in CE 100 and CE 400 mg/kg, respectively. The neuroprotective effect of caffeine was also observed in studies using the neurotoxin 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) 30,31 .
The contents of monoamines and their metabolites were measured in the left and right striatum of the rats with lesion on right MFB (study 1) and right striatum (study 2). In the rst study the levels of DA and NE were moderately reduced in the lesioned side in the vehicle-control group when compared to the shamoperated group (decrease of 52.1 and 44.5%, respectively), although the difference was not signi cant. However, the levels of neurotransmitters or metabolites of rats treated with CE and caffeine did not differ from the vehicle control group. This suggest that the treatment was not e cient in protecting nigrostriatal neurons from 6-OHDA.
Regarding the second experiment, the levels of DA, NE, 5-HT and its metabolites seemed decreased on lesioned control group (vehicle) in comparison with the sham-operated group, but there was no signi cant difference. The levels of DA and NE were reduced in 59.1 and 61.3%, respectively. Moreover, the experimental groups treated with CE or caffeine showed levels of neurotransmitters and metabolites similar to the vehicle, indicating that the treatments did not prevent the reduction of monoamines in the lesioned striatum. This result is in contrast with the behavioral data, but it may suggest that the effect of CE and caffeine on the rotational behavior is not dependent on the dopamine level and neuronal integrity. The decrease of NE levels in both experiments may be explained because we did not give a preadministration of desipramine or other NE transporter inhibitor to the animals 24 . As result, our protocol allowed the uptake of 6-OHDA neurotoxin both by dopaminergic and by noradrenergic neurons causing non selective oxidative stress and neuronal death of catecholaminergic neurons. This protocol reproduces an advanced stage of PD in which there is pronounced loss of NA neurons, but it has the disadvantage of producing some interference in the behavioral tests.
Several studies have showed that caffeine is able to reduce the contralateral rotations induced by apomorphine, the lesion size and can attenuate the dopamine depletion after 6-OHDA injury 10,11,32,50 and can protect the neuronal loss induced by MPTP or other neurotoxins 30,31,52 . Other substances present in green coffee were also evaluated on different cellular and animal models of PD. The chlorogenic acid protected PC12 cells against α-synuclein 42 and several coffee compounds exerted positive effects in cell culture 20 . Yan et al. 44 evaluated the effect of low doses of caffeine and eicosanoyl-5-hydroxytryptamide (EHT) in two PD models: transgenic mice expressing human α-synuclein and striatal injection of αsynuclein pre-formed brils in mice. The combination of the two components for 6 months was effective in normalizing the biochemical and histological parameters analyzed and in promoting improvement in behavioral tasks, while caffeine or EHT alone were ineffective or produced only moderated results 44 . In another study, the EHT showed protective effect against the toxin MPTP in vivo and MPP + in vitro 43 . The treatment with caffeic acid partially inhibited the in ammatory process in mice that received MPTP, evaluated by the level and expression of pro-in ammatory cytokines 53 . Other studies employing decaffeinated coffee extracts also observed protective effects in cell culture and animal models, in agreement with the hypothesis that other compounds are relevant and occasionally more important than caffeine on the coffee bene cial effects 34,54,55 . Based on the literature and in our results we propose that caffeine and other coffee compounds could act in different targets modulating the dopaminergic neurotransmission and motor function.
An important limitation of our study is that we did not quantify the extension and severity of the lesion induced by 6-OHDA regarding the neuronal death. The neuronal loss on substantia nigra was qualitatively con rmed for rats from experiment 1, which brains were sectioned and submitted to cresyl violet staining (Fig. 1). This data con rms that the protocol produced neuronal death in substantia nigra, but it is not enough to evaluate the coffee effect. In experiments 1B and 2, we measured the levels of monoamine and their metabolites in the striatum as an indirect measure of neuronal death. Despite the high variability found (see Supplementary Data S1), it was evident that the lesion protocol produced marked reduction in the levels of dopamine in the striatum of the lesioned hemisphere in control animals and CE was not able to avoid the reduction of dopamine level. Also, the lack of statistical difference in the rotational behavior and monoamine dosages among the experimental groups contrast to the observational data obtained. Deumens et al. 56 argue that one of the limitations on 6-OHDA lesion model is the great variability observed among the lesioned animals, what makes it di cult to nd statistical differences. The rotation model induced by unilateral lesion with neurotoxins, although widely used, have a certain complexity regarding the effects observed. Different and even opposite results may be observed depending on the neurotoxin used, the local of injury, and the extent of neuronal death 45 . Future studies should investigate the effect of green extract and its isolated components on lesion size after 6-OHDA or other neurotoxins employing different concentrations of extract and toxin and exploring different pathways.
The results found in our study are in line with the hypothesis that CE components were able to modulate the dopaminergic transmission by indirect pathway, observed by the inhibition of haloperidol induced catalepsy in mice and blockade of contralateral rotation after apomorphine challenge in rats with unilateral lesion of striatum. Taken together, our results suggest a pro-dopaminergic action of the CE, but its mechanism remains unclear and insu cient to predict whether unroasted coffee would indeed have a neuroprotective effect on PD. Table 1 Monoamines dosage (HPLC) in the right and left striatum of animals with lesion in the right side of medial forebrain bundle (MFB) and in sham-operated animals. Data are expressed as mean ± standard error of mean (n = 4-7)  Table 2 Monoamines dosage (HPLC) in the right and left striatum of animals with lesion in the right side of striatum and in sham-operated animals. Data are expressed as mean ± standard error of mean (n = 7-8)   Linear regression to determine the percentage of inhibition of spontaneous lipid peroxidation in homogenate of the rat brain by coffee extract (CE). The dotted curves indicate the 95% con dence interval. The Q1/2 was calculated as 8.73 µg/mL (n= 4 assays).