From the Cover: Harmane-Induced Selective Dopaminergic Neurotoxicity in Caenorhabditis elegans

Abstract

Parkinson’s disease (PD) is a debilitating neurodegenerative disease. Although numerous exposures have been linked to PD etiology, causative factors for most cases remain largely unknown. Emerging data on the neurotoxicity of heterocyclic amines suggest that this class of compounds should be examined for relevance to PD. Here, using Caenorhabditis elegans as a model system, we tested whether harmane exposure produced selective toxicity to dopamine neurons that is potentially relevant to PD. Harmane is a known tremorigenic β-carboline (a type of heterocyclic amine) found in cooked meat, roasted coffee beans, and tobacco. Thus, this compound represents a potentially important exposure. In the nematode model, we observed dopaminergic neurons to be selectively vulnerable, showing significant loss in terms of structure and function at lower doses than other neuronal populations. In examining mechanisms of toxicity, we observed significant harmane-induced decreases in mitochondrial viability and increased reactive oxygen species levels. Blocking transport through the dopamine transporter (DAT) was not neuroprotective, suggesting that harmane is unlikely to enter the cell through DAT. However, a mitochondrial complex I activator did partially ameliorate neurodegeneration. Further, mitochondrial complex I activator treatment reduced harmane-induced dopamine depletion, measured by the 1-nonanol assay. In summary, we have shown that harmane exposure in C. elegans produces selective dopaminergic neurotoxicity that may bear relevance to PD, and that neurotoxicity may be mediated through mitochondrial mechanisms.

Parkinson’s disease (PD) is a debilitating neurodegenerative disease. The pathological hallmarks are the loss of nigral neurons and aggregation of α-synuclein in inclusions known as Lewy Bodies in surviving dopaminergic (DA) neurons (Spillantini et al., 1997). Although numerous genetic links have been identified and a number of environmental exposures have been linked to etiology, the majority of the PD cases are sporadic (Cannon and Greenamyre, 2011). Given that available treatment options do not slow disease progression, it is critical to identify and reduce potentially causative exposures. Dietary exposures have, to date, received far less attention than other classes of compounds, such as pesticides. Dietary compounds represent a potentially very common and modifiable exposure. Should dietary factors be identified that contribute to disease etiology, a reduction in intake could potentially reduce PD incidence (Agim and Cannon, 2015). To that end, emerging data suggest that heterocyclic amines (HCAs) should be examined as possible DA neurotoxins (Griggs et al., 2014; Louis et al., 2014). HCAs are found in many dietary components, especially in charred meat. Interestingly, elimination of dietary red meat has been reported to improve motor function in PD patients (Coimbra and Junqueira, 2003). Thus, this class of compounds deserves attention with respect to potential neurotoxicity.

To date, some HCAs have been examined for potential relevance to PD. For example, we found that 2-amino-1-methyl-6-phenylimidazo[4, 5-b]pyridine (PhIP) is selectively neurotoxic to DA neurons in primary rat midbrain cultures (Griggs et al., 2014). Further, 3-amino-1, 4-dimethyl-5 H-pyrido[4, 3-b]indole (Trp-P-1) and 3-amino-1-methyl-5 H-pyrido[4, 3-b]indole (Trp-P-2) were found to interfere with dopamine metabolism (Kojima et al., 1990; Maruyama et al., 1994).

Here, we chose to examine a specific HCA, 1-methyl-9 H-pyrido[3, 4-b]indole (harmane). Harmane is a HCA belonging to the β-carboline subclass. It is a known tremorigen found in cooked meat, fish, coffee and tobacco (Skog et al., 1997, 1998). Our focus on harmane was informed by published data that higher plasma harmane levels have been found in PD patients compared with controls (Louis et al., 2014). Prior to this study, harmane had primarily been investigated for a role in essential tremor. For example, those suffering from essential tremor are found to have elevated levels of harmane in blood (Louis et al., 2002). Notably, essential tremor patients have an increased risk of PD (Benito-Leon et al., 2009). Thus, data from PD patients and links between PD and essential tremor prompted our study.

Increased human exposure to harmane could be attributed to abundance in diet. Due to its high lipid solubility, harmane tends to accumulate in brain tissue (Zetler et al., 1972). Ostergren et al. (2004) also showed that harmane has high affinity binding to neuromelanin, that is abundant in substantia nigra in humans, possibly indicating that harmane may accumulate in nigral dopamine neurons. Taken together, there is a compelling dataset that suggests harmane-induced neurotoxicity should be evaluated in vivo.

Caenorhabditis elegans is a transparent, nonpathogenic soil nematode. A major advantage of this model system is that numerous strains are available that express fluorescent reporter systems in specific cell populations, where cell damage and loss can easily be assessed. Thus, C. elegans as a model system offers an outstanding opportunity to examine whether neurotoxicity is selective to specific neuronal populations across a number of doses. Using simple nematode model, we have identified harmane as a potential dietary toxicant responsible for DA cell loss through mechanism coinciding with PD pathology.

MATERIALS AND METHODS

Culture and maintenance of strains

C. elegans strains, Bristol N2, BZ555 (egIs1 [dat-1p::GFP]), MT15620 (cat-2(n4547)), UA57(baIs4 [dat-1p::GFP + dat-1p::CAT-2]), CZ1632 (juIs76 [unc-25p::GFP + lin-15(+)] II), GR1366 (mgIs42 [tph-1::GFP + rol-6(su1006)]), LX929 (vsIs48 [unc-17::GFP]), and Escherichia coli OP50, were procured from Caenorhabditis Genetics Centre, (University of Minnesota, Minnesota), grown on Nematode growth medium (NGM) and cultured at 22°C. A synchronized population of worms was obtained by sodium hypochlorite treatment.

Treatment of worms

L1 staged worms were treated with different concentrations of harmane (Sigma Aldrich, St. Louis, MO, USA, 103276) in liquid culture as described by Pu and Le (2008) supplemented with E. coli OP50 (Pu and Le, 2008). 100 mM stocks of harmane were prepared in DMSO and diluted further to doses ranging from 100 to 1000 µM. The dose range was based on established C. elegans PD models (e.g. MPP⁺; Pu and Le, 2008). Fresh stocks <7-days old were used for each experiment. 70 L1 staged worms were added to 200 μl suspension in 24-well plates and incubated at 22°C. Since harmane is a hazardous substance, care was taken to avoid skin contact.

Given the paucity of data on harmane in in vivo models, we examined the literature on neurotoxicant PD models in C. elegans, in choosing an initial dose range. We focused more on established MPP⁺ models versus 6-OHDA, because the reported log P of MPP⁺ (2.7) is far closer to that of harmane (3.6) versus 6-OHDA (0.2) (PubChem, 2017a,b,c). In worms, it is well known that far higher doses are typically required than in cell culture due to the nematode cuticle which exhibits strong barrier between worm and surroundings (Page and Johnstone, 2007). Although the cuticle is likely less of a barrier for nonpolar compounds, it is worth noting that the examination of lethality for 21 compounds showed that the majority of compounds tested produced LD50s with little variability. Those that are pH sensitive or highly polar did exhibit marked increases in LD50s (Li et al., 2013). In considering dose, we also reviewed suggestions for acute (12–24 h) exposure ranges in C. elegans. Suggested dose ranges were informed by correlation analysis between nematode, mouse, and rat data. Further, the suggested dose ranges were based on globally harmonized system of toxicity classification (Li et al., 2013). For harmane, the calculated suggested molarity dose ranges for lethality in C. elegans would be approximately 1.2–250 mM (GHS, 2007; Li et al., 2013; Sigg et al., 1964). Given that we were interested in neurotoxicity, we tested exposures at far lower doses.

Based on pilot studies and data included here, mechanistic studies were conducted in 2 sets: 1 high dose (650 µM harmane) for 48 h and 1 low dose (500 µM harmane) for 72 h. The rationale for 2 doses was to determine if potentially protective regimens would be effective at different doses and time-points. To determine whether mechanistic conclusions were consistent across exposure and dose, shorter, higher dose studies were also performed (48 h at 650 µM).

Several chemical treatments were utilized to test potential mechanisms of toxicity and protection. We aimed to determine whether the use mitochondrial activators such as D, L 3-hydroxy butyric acid (HBA), N-acetyl-L-cysteine (NAC), and riboflavin (RB) could prevent harmane-induced alterations in mitochondrial viability. Although HBA activates complex I via complex II (Tieu et al., 2003), riboflavin activates complex I cooperatively with complex IV (Grad and Lemire, 2006). In contrast, NAC, a well-known antioxidant is also generalized activator of mitochondrial complexes I, IV, and V (Cocco et al., 2005; Kamboj and Sandhir, 2011; Miquel et al., 1995; Soiferman et al., 2014). For treatment with HBA, a 5 M stock of sodium salt of HBA (Acros organics, NJ, USA, 150834), dissolved in sterile distilled water was prepared. The worms were subjected to different doses ranging from 50 to 200 mM (for both 48 and 72-h time-points) treatment in liquid culture. For treatment with RB (Acros organics, NJ, USA, 132350250), a 100 μg/ml stock was prepared in 10% DMSO. The worms were subjected to doses of 1, 2, and 3 μg/ml (Grad and Lemire, 2006). For treatment with NAC (Acros organics, NJ, USA, 160280250), a 500 mM stock was prepared in sterile distilled water. The worms were subjected to doses of 5, 10, and 15 mM (Oh et al., 2015). Similarly, for treatment with bupropion HCl (Alfa aesar, Haverhill, MA, USA, J61105), 1 mM stock was prepared in sterile distilled water, with testing at doses of 20–80 μm (Felton and Johnson, 2014).

Neurodegeneration assay

In order to determine the effect of harmane on different classes of neurons, L1 worms expressing green fluorescence protein (GFP) in DA, serotonergic, Gamma-amino butyric acid (GABA)ergic and cholinergic neurons were subjected to different doses of harmane (100–500 μM) for 48 h at 22 °C and evaluated using morphological markers of neurodegeneration that have been repeatedly utilized in nematode models (Alexander et al., 2014). Worms exhibiting neuropathological alterations such as loss, breakage in dendrites and loss of soma were considered as affected and percentage of affected worms was calculated. For detailed studies in DA neurons, neurodegeneration was characterized and quantified as described by Yao et al. (2010). Briefly, treated worms were washed 3 times using M9 buffer, anesthetized using 100 mM sodium azide and visualized using the fluorescence microscope (Olympus BX-53). The number of neurons was counted for minimum of 20 worms (8 dopamine neurons per worm) per group. The percentage of morphologically intact neurons was calculated and plotted against its respective neurons type in control. Care was taken to keep the worm number constant and fresh stocks (no older than 1 week) of harmane were used for all the studies.

Dihydroethidium staining for ROS

Dihydroethidium (DHE) staining for reactive oxygen species (ROS; primarily as a detector of superoxide), was performed as described by Chikka et al. (2016), with slight modifications (Chikka et al., 2016). Briefly, worms were washed 3 times with M9 buffer followed by 2 washes with phosphate buffered saline (PBS). Approximately 50 worms in 100 μl of PBS were mixed with equal amount of 12 μM DHE (Millipore, Burlington, MA, USA, 309800) dissolved in DMSO (final concentration = 0.01%). After 40 min of incubation in dark, worms were washed with PBS, anesthetized using 100 mM sodium azide and imaged using fluorescence microscope (Olympus BX-53). Approximately 20 animals per group were analyzed semi-quantitatively using Image J, excluding the gut region from the region of interest (ROI) so as to eliminate the nonspecific fluorescence.

Mitochondria staining

MitoTracker staining was performed according to the manufacturer protocol with slight modifications. Briefly, MitoTracker Red CM-H2XRos (Life technologies, Carlsbad, CA, USA, 7513) was mixed with E. coli OP50 before seeding it to the NGM plates to check the healthy mitochondrial staining. MitoTracker Red stock solution was prepared by dissolving 50 µg MitoTracker Red CM-H2XRos in 100 µl of DMSO. A working concentration of 4.72 µM MitoTracker Red was fed to the worms by mixing it with E. coli OP50. The synchronous embryos were transferred onto the MitoTracker Red containing plates and grown for 48 h at 22 °C. Worms were washed off using M9 buffer and kept in OP50 solution for 30 min and again washed 3 times with M9 buffer. Worms were anesthetized using 100 mM sodium azide and observed using fluorescence microscope (Olympus BX-53). Approximately 20 animals per group were analyzed semi-quantitatively using Image J. In order to eliminate nonspecific fluorescence, the gut region was excluded from ROI analysis.

Nonanol repulsion assay

To indirectly estimate changes in dopamine levels, a dopamine-dependent assay that measures 1-nonanol based repulsive behavior (Bargmann et al., 1993) was performed, which is an established technique in this model (Kaur et al., 2012). Dopamine levels are related to motivation, recognition, reward, memory, adaptation, hormonal regulation, and motor control in worms (Felton and Johnson, 2014). Any changes in dopamine levels alter the worm behavior towards attractants and repellents (Ward, 1973). Worms with normal levels of dopamine exhibit optimum levels of repulsive behavior towards 1-nonanol, while worms with decreased dopamine levels exhibit extended time to exhibit the repulsive; increases in dopamine, correspondingly decreases repulsion time (Kaur et al., 2012; Srivastava et al., 2017). The nonanol repulsion assay for indirect quantification of dopamine levels was conducted as previously described (Kaur et al., 2012). Briefly, treated worms were washed 3 times with M9 buffer. Worms were placed on NGM plates. The poking lash dipped in 1-nonanol (Acros Organic, NJ, USA, AC157471000) was placed near the head region of the worms, while taking care not to touch the worms. Time taken for the worms to exhibit the repulsive behavior was calculated using a stop watch. Any worms coming in contact with the poking lash or 1-nonnaol were not counted as a part of study. To further validate the relevance of the assay endpoint to dopamine levels, we also exposed worms to the DAT inhibitor bupropion HCl and performed the assay in worms overexpressing cat-2 (encodes tyrosine hydroxylase) and cat-2 mutants.

Thrashing assay

The thrashing assay examines motility and locomotion defects in worms. Thrashing assay was performed as described by Lee et al. (2008), with slight modifications (Lee et al., 2008). Briefly, worms from the treatment plates were washed 3 times using M9 buffer. The worms were then transferred to a drop of M9 buffer. After allowing adaptation for 1 min, the worms were scored for body bends for 30 s, where 1 complete sinusoidal bend was scored as 1 movement.

Statistical analysis

Statistical analysis was performed using GraphPad PRISM, Version 7.02 (1992–2016 GraphPad Software, Inc., La Jolla, CA, USA). Each experiment was repeated at least 3 times and normalized to untreated groups. Analysis of variance (ANOVA), followed by Tukey’s or Dunnett’s (where comparison was only to control) post hoc tests were utilized. For all experiments, p < .05 was deemed to be statistically significant.

RESULTS

DA Neurons Exhibit Enhanced Vulnerability to Harmane

In order to determine the specific neuronal populations that were selectively vulnerable to harmane, we quantified the effect of different doses of harmane (100–500 μM) in worms expressing GFP specific to DA, serotonergic, GABAergic and cholinergic neurons (BZ555, GR1366, CZ1632, and LX929, respectively) after 48 h of exposure. We observed that DA neurons were most vulnerable to harmane-induced neurodegeneration; exhibiting neurodegeneration at doses as low as 100 μM harmane, where other neuronal subtypes showed intact morphology (Figure 1). In contrast, higher doses were required to produce statistically significant neurodegeneration in serotonergic, GABAergic and cholinergic neurons. DA neurons exhibited a significant decrease in percentage of worms with intact neuronal morphology at 100 µM (77.50 ± 1.11, mean ± SEM, p < .01), 250 µM (25 ± 14.43, p < .001) and 500 µM (0 ± 0.00, p < .001; n = 6 repeats/group) (Figs. 1A and 1B). Serotonergic neurons exhibited a significant decrease in the percentage of worms with intact neuronal morphology at 250 µM (68.33 ± 1.66, p < .001) and 500 µM (0 ± 0.00, p < .001) (Figs. 1C and 1D) (n = 3 repeats/group). Damage to GABAergic neurons was primarily limited to inter-neuronal commissures; where at neurodegeneration (as percentage of animals affected) was evident at 250 µM (76.66 ± 3.33, p < .05) and 500 µM (15 ± 17.63, p < .001) (Figs. 1E and 1F) (n = 3 repeats/group). Cholinergic neurons, however, were found to be least susceptible to harmane-induced neurodegeneration, where neurodegeneration was evident only at 500 μM (10 ± 5.77, p < .001) (Figs. 1G and 1H) (n = 3 repeats/group). Calculated IC50s are as follows: DA neurons (115 µM); GABAergic neurons (290 µM); serotonergic neurons (389 µM); cholinergic neurons (492 µM). Here, the “inhibition” refers to inhibiting normal neuronal morphology across the total population for a specific neuronal population. Thus, sensitivity to harmane-induced neurodegeneration was deemed to be DA > GABAergic ∼ serotoninergic > cholinergic. The distinction between serotonergic and GABAergic sensitivity was dependent on whether an individual dose was examined or the calculated IC50.

Figure 1.

DA neurons are selectively sensitive to harmane exposure. Worms were treated with harmane for 48 h and neurodegeneration was assessed. DA neurons (Strain: BZ555) exhibited heightened sensitivity to harmane, with neurodegeneration observed at doses as low as 100 μM (A, B). Exposing worms to 100, 250, and 500 μM harmane produced quantifiable neurodegeneration (n = 6). Other neuron types ie, serotonergic (Strain: GR1366; C, D), GABAergic (Strain: CZ1632; E, F) and cholinergic (Strain: LX929; G, H) were comparatively less susceptible to harmane-induced neurodegeneration. The 3 other neuronal cell populations examined were devoid of any neurodegeneration at 100 μM. Serotonergic neurons exhibited significant neurodegeneration at 250 and 500 μM (n = 3). GABAergic neurons showed affliction in neuronal damage at 250 and 500 μM, with damage majorly restricted to inter-neuronal commissures (n = 3). Cholinergic neurons were found to be least susceptible, displaying neuronal damage only at 500 μM (n = 3). Data presented as mean ± SEM. Percentage of intact neurons was calculated by counting the total number of neurons per worm for 20 animals in each experimental group. Data analyzed by one-way ANOVA followed by Dunnett’s post hoc test. *p < .05, **p < .01, ***p < .001.

Harmane Induces Dose-Dependent Neurodegeneration of DA Neurons

C.elegans (hermaphrodite) possesses 8 DA neurons. Of these 8 neurons, there are 3 DA subpopulations: 4 cephalic sensilla (CEP), 2 anterior deirid (ADE) in the head region and 2 posterior deirid (PDE) in the tail region (Figure 2A). After determining increased sensitivity of DA neurons to harmane in worms expressing GFP tagged to dopamine transporter (DAT), we studied the effects of different doses of harmane (100–1000 μM) in total DA neurons and in subpopulations treated for 48 h. In comparison to control, the worms exhibited loss of DA neurons in a dose-dependent manner (Figure 2B). Harmane exposure resulted in various morphological changes such as breakage, swelling, or complete loss of soma and dendrites.

Figure 2.

Harmane treatment produces dose-dependent neurodegeneration of DA neurons. Representative images of DA neurons in C. elegans show 4 CEP, 2 ADE in head region along with 2 PDE neurons in tail region. Tail ray neurons are present in males only (A). Treatment of worms with harmane for 48 h produces dose-dependent degeneration of DA neurons exhibiting morphological changes such as breakage or loss of dendrites (arrows), swelling (filled arrowheads) and loss of soma (open arrows) (B). The percentage of neuronal loss was calculated specific to DA neuronal subtype: total (C), CEP (D), ADE (E), and PDE (F). Data presented as mean ± SEM. Percentage of intact neurons was calculated by counting the total number of neurons per worm for 20 animals in each experimental group. Data analyzed by one-way ANOVA followed by Tukey’s post hoc test. **p < .01, ***p < .001 (n = 3). Scale bar represents 20 μm.

In terms of total DA neurons, a significant decrease in the percentage of total neurons with typical morphology was observed in worms treated with 250 (81.85 ± 1.25, p < .01), 500 (36.66 ± 3.00, p < .001), 750 (30.62 ± 5.20, p < .001) and 1000 μM (1.25 ± 1.25, p < .001) harmane as compared with control (Figure 2C). ADE neurons exhibited significant decrease in percentage of intact DA neurons at 250 (83.33 ± 2.20, p < .01), 500 (27.5 ± 3.81, p < .001), 750 (13.33 ± 3.63, p < .001), and 1000 μM (0.83 ± 0.83, p < .001) (Figure 2E). PDE neurons exhibited considerable decrease in percentage of intact neurons at 250 (58.33 ± 11.67, p < .01), 500 (2.5 ± 2.5, p < .001), 750 (0.83 ± 0.83, p < .001) and complete loss at 1000 μM (Figure 2F). CEP neurons exhibited least susceptibility to harmane exposure exhibiting considerable decrease in percentage of intact neurons at 500 (58.33 ± 3.25, p < .001), 750 (54.16 ± 8.73, p < .001) and 1000 μM (2.08 ± 2.08, p < .001) (Figure 2D) (n = 3 repeats/group for all groups in this Figure). Taken together, our results suggest that PDE neurons are most susceptible to harmane, followed by ADE and CEP neurons.

Harmane Exposures Increase the Levels of ROS

DHE staining was used to evaluate ROS levels (primarily as a measure of superoxide) following 48 h of harmane exposure (Figure 3A). Changes in ROS levels were not detectable at 100 and 250 μM, however, a significant increase in ROS levels was observed at 750 μM (3.40 ± 0.74, p < .05) as compared with that of untreated worms (1.00 ± 0.59) (Figure 3B). The results imply that at higher doses harmane exposure leads to increased ROS (n = 3 repeats/group for all groups in this Figure).

Figure 3.

Harmane-induced neurochemical changes. Harmane treatment (48 h) produces ROS, reduces mitochondrial viability, and alters dopamine levels. Treatment of worms with harmane produced elevated ROS levels (A). Statistically significant changes were detectable at 750 µM (B). Harmane treatment led to decreased mitochondrial viability (C). Statistically significant decreases were quantifiable at doses 250 µM and above (D). Harmane treatment at 50–500 µM produced alterations in DA function as measured through nonanol repulsion assay (E). Lower doses of harmane (50 and 100 µM) significantly elevated dopamine responsiveness in terms of reduced repulsion time, whereas higher doses (250 and 500 µM) exhibited significant delay in responsiveness, indicative of dopamine depletion (E). Data presented as mean ± SEM. Repulsion time was calculated for 18 animals in each experimental group. Worms were analyzed semiquantitatively for approximately 20 animals per treatment group. For quantification of fluorescence images (A, C), the gut was excluded from the ROI so as to eliminate nonspecific fluorescence from the study. Data analyzed by one-way ANOVA followed by Tukey’s post hoc test. *p < .05, **p < .01, ***p < .001 (n = 3). Scale bar represents 50 and 20 μm for (A) and (C), respectively.

Harmane Treatment Decreases Mitochondrial Viability

The effects of 48 h of harmane exposure on mitochondria proton gradients were evaluated using a reduced form of MitoTracker stain, which stains only mitochondria that have an intact proton gradient, deemed to be “viable”. We observed a dose-dependent loss of mitochondria with intact proton gradients in worms exposed to harmane (Figure 3C). Semi-quantitative analysis revealed a significant loss of mitochondria with intact proton gradients in worms treated with 100 μM (0.34 ± 0.04, p < .001), 250 μM (0.14 ± 0.02, p < 0.001), 500 μM (0.05 ± 0.00, p < .001) and 750 μM (0.10 ± 0.00, p < .001) as compared with that of control (1.00 ± 0.00) (Figure 3D) (n = 3 repeats/group for all groups in this Figure), implying that harmane leads to loss of mitochondrial viability.

Harmane Exposure Leads to Decrease in Dopamine-Dependent Response Towards 1-Nonanol

Dopamine modulates various behavioral functions in C. elegans, any alteration in levels of dopamine leads to altered response to 1-nonanol (Kaur et al., 2012). This assay has been extensively used in this system to estimate dopamine levels. To further validate the relationship between the endpoint and dopamine levels, we also showed that DAT inhibition, a strategy to increase synaptic dopamine, lowered repulsion time, indicative of increased dopamine levels (Supplementary Figure 1) (n = 3 repeats/group for all groups in this Figure). We studied the effect on chemo-repulsive behavior in worms exposed to harmane in doses ranging from 50 to 500 μM. In comparison to untreated worms (1.00 ± 0.00), we observed that the lower doses ie. 50 μM (0.61 ± 0.03, p < .01) and 100 μM (0.61 ± 0.03, p < .001) resulted in statistically decreased repulsion time (indicative of higher dopamine levels), whereas higher doses, 250 μM (1.54 ± 0.03) and 500 μM (2.41 ± 0.06) showed significantly increased repulsion time (indicative of lower dopamine levels). These results indicate that the harmane exposure at or above 250 μM results in decreased dopamine levels (Figure 3E).

Neuroprotection Studies on Harmane-Induced DA Neurotoxicity

In general, mechanistic studies aimed at protection showed that harmane toxicity was generally more amenable to modulation under a lower dose (500 μM), longer exposure time (72 h). Thus, mechanistic studies were primarily conducted under this regimen (Figs. 4–6). Nonetheless, for consistency with toxicity experiments and to also evaluate mechanistic modulation under higher-dose (650 μM), shorter exposure time (48 h), and mechanistic studies were also conducted under this regimen (Supplementary Figs. 2–4).

Figure 4.

Harmane induced DA neurotoxicity is unlikely to be mediated through DAT uptake. Lower-dose, longer harmane treatment times (500 µM for 3 days) also produced neurodegeneration that was unaffected by blocking DAT (A). The percentage of neuronal loss was calculated with respect to DA neuronal subtype: total (B), CEP (C), ADE (D), and PDE (E). Data presented as mean ± SEM. Percentage of intact neurons was calculated by counting the total number of neurons per worm for 20 animals in each experimental group. Data analyzed by one-way ANOVA followed by Tukey’s post hoc test (n = 3/group). Scale bar represents 20 μm.

Inhibition of DAT Does Not Alter Harmane-Induced Neuronal Loss

In order to investigate if harmane-induced neurodegeneration is mediated through DAT uptake, a common entry mechanism for DA neurotoxicants, we co-treated harmane-exposed worms with bupropion HCl (BP), a known DAT inhibitor. The studies were divided into 2 groups, by exposure 48 or 72 h and 500 or 650 μM harmane. In 72 h studies, we did not show detectable changes induced by BP treatment in neuronal loss (Figure 4A) for total, CEP, ADE and PDE neurons (Figs. 4B–E) (n = 3 repeats/group for all groups in this Figure). Similarly, in 48-h studies, we did not observe any alterations in harmane-induced neurodegeneration upon treatment with BP (20, 40, and 80 μM) (Supplementary Figure 2) (n = 3 repeats/group for all groups in this Figure). Similarly, the above results imply that harmane-induced neuronal loss is not mediated through DAT uptake.

Mitochondrial Complex I Modulation Partially-Ameliorates Neurotoxicity

To determine if mitochondrial complex I may modulate harmane neurotoxicity, we assessed the effect of mitochondrial complex I activator, HBA, on harmane-induced neurodegeneration. Similar to the studies on DAT inhibitor, the study design employed both 72-h/low dose and 48-h/high dose studies with doses of HBA ranging from 50 to 200 mM. In pilot studies, worms were exposed to HBA alone at concentrations up to 200 mM, which did not produce any evidence of neurodegeneration (data not shown).

In 72-h studies at a lower dose of harmane, the rescuing effect of HBA was relatively more pronounced (Figure 5A). In terms of total neurons, a significant increase in percentage of intact neurons was observed in harmane treated worms (500 μM), when subjected to treatment with 100 mM (58.33 ± 6.61, p < .05), 150 mM (61.66 ± 3.23, p < .01) and 200 mM HBA (64.16 ± 5.20, p < .01) as compared with 500 μM harmane alone (35.20 ± 3.68) (Figure 5B). In DA subpopulations, HBA was not protective in PDE neurons (Figure 5E). However, HBA partially ameliorated neurodegeneration in other subpopulations. In CEP neurons, HBA treatment produced an increase in the percentage of intact neurons at 100 mM (79.16 ± 9.79, p < .05), 150 mM (85.00 ± 3.30, p < .01) and 200 mM (89.58 ± 5.41, p < .01) as compared with harmane 500 μM alone (52.91 ± 4.80) (Figure 5C). In ADE neurons, HBA treatment increased the percentage of intact neurons at 100 mM (75.00 ± 8.03, p < .05), 150 mM (74.16 ± 8.81, p < .05) and 200 mM (77.50 ± 10.00, p < .05) as compared with harmane 500 μM alone (35.00 ± 5.77) (Figure 5D) (n = 3 repeats/group for all groups in this Figure). The results above indicate that mitochondrial complex I activator treatment provides protection from harmane-induced DA neurodegeneration. In 48-h studies, we found that HBA treatment partially rescued harmane-induced neurodegeneration, to a somewhat lesser extent than the lower dose, longer term exposure regimen (Supplementary Figure 3) (n = 3 repeats/group for all groups in this Figure).

Figure 5.

HBA treatment provides protection from harmane-induced neurodegeneration. Neuroprotection was evaluated at multiple doses. Treatment with HBA, a mitochondrial complex I activator treatment resulted in a significant decrease in DA neurodegeneration 500 µM harmane, over a longer exposure time (72 h) (A). Here, the percentage of neuronal loss was also calculated with respect to DA neuronal subtype: total (B), CEP (C), ADE (D), and PDE (E). Data are presented as mean ± SEM. The percentage of intact neurons was calculated by counting the total number of neurons per worm for 20 animals in each experimental group. Data analyzed by one-way ANOVA followed by Tukey’s post hoc test. *p < .05, **p < .01 (n = 3). Scale bar represents 20 μm.

Mechanisms of Harmane-Induced Alterations of Mitochondrial Viability

To confirm that HBA treatment improved mitochondrial function and that neuroprotection was likely to be mediated through this mechanism, we assessed the effect of HBA on mitochondria with intact proton gradients using MitoTracker at 100-200 mM (Figure 6A). In comparison to untreated worms (1.00 ± 0.00), we observed a significant increase in mitochondria with intact proton gradients in worms treated with 100 mM (2.13 ± 0.16, p < .001) and 150 mM (1.81 ± 0.05, p < 0.001) HBA, while 200 mM HBA did not produce significant differences from control (Figure 6B) (n = 3 repeats/group for all groups in this Figure), suggesting that HBA improves mitochondrial viability, particularly at the lower two doses tested.

Figure 6.

Treatment of harmane-exposed worms with HBA and NAC improves mitochondrial viability. HBA treatment led to significant increases in mitochondrial viability (A). At 100 and 150 mM doses, increased viability was quantified, while at a higher dose (200 mM) changes in viability were not detectable (B). At a lower dose (500 μM), longer exposure time (72 h), a modest but significant increase was detected upon treatment with HBA (C, D). At a lower dose (500 μM), longer exposure time (72 h), treatment with RB was devoid of any effect on mitochondrial viability (E, F). At a lower dose (500 μM), longer exposure time (72 h), a significant dose-dependent increase was detected upon treatment with NAC (G, H). Data presented as mean ± SEM. Mitochondrial viability was semi quantitatively analyzed through Image J for approximately 20 animals per group. The gut was excluded from the ROI so as to eliminate nonspecific fluorescence from the study. Data analyzed by one-way ANOVA followed by Tukey’s post hoc test. *p < .05, ***p < .001 (n = 3). Scale bar represents 20 μm.

The effect of HBA on mitochondria with intact proton gradients was then tested in harmane-treated worms. As with the previous experiments, we performed MitoTracker staining for low (500 μM) and high doses (650 μM, Supplementary Figs. 4A and 4B) (n = 3 repeats/group for all groups in this Figure) of harmane. Here, detectable changes were only observed in the low dose group (Figure 6C). The effect on mitochondria with intact proton gradients in worms treated with 150 mM HBA (0.13 ± 0.00, p < .05) was significantly enhanced in comparison to that of worms treated with 500 μM harmane (0.08 ± 0.00) (Figure 6D). RB treatment did not modulate mitochondrial viability in worms exposed to harmane (Figs. 6E and 6F). However, NAC treatment provided significant protection from loss of mitochondrial proton gradient in worms treated with 500 μM harmane at doses 10 mM (0.22 ± 0.019, p < .05) and 15 mM (0.36 ± 0.023, p < .001) as compared with that of worms treated with harmane (0.12 ± 0.00) (Figure 6G and 6H). The above results indicate additional contribution of complex V towards amelioration of mitochondrial viability in worms challenged with harmane.

HBA Enhances Dopamine-Dependent Response Towards 1-Nonanol in Untreated and Harmane-Treated Worms

HBA treatment was also evaluated for effects on harmane-induced dopamine-dependent behaviors, again using the 1-nonanol assay. We observed significant enhancement in dopamine responsiveness evident as decreased repulsion time in worms treated with 100 mM (0.73 ± 0.06, p < .01), 150 mM (0.64 ± 0.02, p < .001) and 200 mM (0.67 ± 0.04, p < .01) HBA as compared with repulsion time in control (1.00 ± 0.00) (Figure 7A) (n = 3 repeats/group for all groups in this Figure). Similar to the previous experiments, treatment with lower doses of harmane ie, 100 μM (0.58 ± 0.02, p < .001) and 150 μM (0.78 ± 0.06, p < .05) resulted in a decrease in repulsion time.

Figure 7.

Treatment of harmane-exposed worms with HBA rescues dopamine levels. Treatment with lower dose of harmane (50, 100, and 150 μM) and HBA (100, 150, and 200 mM) exhibited increase in dopamine levels (as indicated by lowered repulsion time) as compared with untreated control; HBA treatment was found to significantly alleviate dopamine levels in worms treated with 250 and 500 μM harmane (A), MT15620 (cat-2 mutant) exhibited lowered dopamine levels, whereas worms overexpressing cat-2 (UA57) exhibited significantly increased dopamine levels (B), Worms treated with HBA were devoid of any alteration in motility as elucidated through thrashing assay (C). Data presented as mean ± SEM. Repulsion time and number of thrashes (with number of thrashes representing 1 complete sinusoidal movement) were calculated for 18 animals in each experimental group. Data analyzed by two-way ANOVA for grouped analysis and by one-way ANOVA followed by Tukey’s post hoc test. *p < .05, **p < .01, ***p < .001 (n = 3).

In worms treated with 250 μM harmane, the repulsion time was increased relative to controls (1.81 ± 0.02). Reduced repulsion times by HBA treatments suggest amelioration of dopamine depletion: 150 mM (0.92 ± 0.03, p < .001) and 200 mM (0.93 ± 0.11, p < .001). Similarly, in worms treated with 500 μM harmane, repulsion times also suggest neuroprotection (2.82 ± 0.04), 100 mM (1.97 ± 0.17, p < .001), 150 mM (1.22 ± 0.02, p < .001) and 200 mM (1.01 ± 0.09, p < .001) HBA (Figure 7A). We also performed this assay in cat-2/TH mutant (MT15620) and worms over-expressing cat-2/TH (UA57) to validate the findings. As expected, we observed a significant increase in repulsion time in cat-2/TH mutant (1.97 ± 0.02, p < 0.001) worms and worms over-expressing cat-2/TH (0.53 ± 0.02, p < .001) exhibited significantly decreased repulsion times compared with that of controls (1.00 ± 0.00) (Figure 7B). The above results established that HBA restores the dopamine function in untreated and harmane-treated worms.

HBA Treatment Does not Alter the Motility in C. elegans

Since muscles are also rich in mitochondria, increased mitochondrial viability is expected to interfere with motility, biasing the results of 1-nonanol assay. Hence, we performed thrashing assay to see if HBA treatment alters the motility in worms. We observed that HBA treatment at 100, 150, and 200 mM failed to produce significant alterations in motility as compared with that of control (Figure 7C), implying that HBA treatment does not alter motility in worms.

DISCUSSION

A large body of data implies a significant role for environmental exposures in the etiology of PD. Yet, many linked compounds are rarely encountered and may not influence a significant number of cases. Here, we aimed to investigate whether harmane, a dietary toxin, may produce selective neurodegeneration in dopamine neurons. Harmane is a toxin formed in heating of biological matter. Although harmane is a known dietary toxin and tremorigen that has been found to be present in increased levels in PD patients (Louis et al., 2014), detailed in vivo neurotoxicity assessments have yet to be conducted. To conduct our studies, we chose the nematode model C. elegans, which has been extensively used in PD research and allowed us to assess selectivity of neurotoxicity across a wide range of doses and treatment times. Here, we have shown that DA neurons were found to be most susceptible to harmane-induced neurodegeneration, with nonDA neurons requiring higher doses to detect neurodegeneration. Dopamine-dependent behaviors were affected by harmane treatment and evidence of increased ROS and decreased mitochondrial viability were observed. Treatment with the mitochondrial complex I activator, HBA was protective, ameliorating morphological changes consistent with neurodegeneration and many neurochemical changes associated with harmane exposure. The findings presented here are expected to set the stage for follow-up pathological and mechanistic studies in higher order species.

C. elegans strains expressing fluorophores in distinct neuronal populations are ideal tools in assessing selective neurotoxicity. Our findings showed that DA neurons are more susceptible to damage by harmane at doses as low as 100 μM; followed by serotonergic/GABAergic and then cholinergic neurons. In performing detailed studies at 100–1000 μM, we observed complete loss of neurons at doses close to 1000 μM.

Interestingly, we also observed differential sensitivity in DA neuronal subpopulations, with PDE neurons being most vulnerable, followed by ADE and CEP neurons. The differential sensitivity order is essentially the reverse of what has been reported in 6-hydroxydopamine (Nass et al., 2002) and MPP⁺ models exhibiting dopamine neuron sensitivity as ADE > CEP > PDE (Wang et al., 2007). Differences in overall and subpopulation sensitivity observed between toxicants could possibly arise from DAT-dependence or strain differences. 6-hydroxydopamine and MPP⁺ uptake into dopamine neurons is mediated through DAT (Cannon and Greenamyre, 2010). Thus, DAT level expression differences and resultant uptake differences occurring in these populations could mediate subpopulation sensitivity for DAT-dependent toxicants. Although DAT inhibition has been shown to prevent MPP⁺-mediated DA loss in C. elegans (Pu and Le, 2008), such inhibition did not modulate neurotoxicity in our studies. Thus, our data suggest that harmane does not enter DA neurons through DAT. Interestingly, DAT inhibition appeared to increase toxicity, although not significantly. It is possible that DAT inhibition may increase the production of unpackaged reactive dopamine and dopamine metabolites, which may be toxic (Jinsmaa et al., 2009). Further work will focus on whether harmane alters dopamine neurotransmission and whether increased oxidized dopamine metabolites may act directly on mitochondria. Relative DAT levels in DA subpopulations have yet to be determined in C. elegans. Thus, it is possible that the relatively high lipophilicity of harmane allows it to enter cells in a DAT-independent manner as our data indicate.

Increased ROS (Dias et al., 2013) and mitochondrial dysfunction (Reddy, 2009; Schon and Manfredi, 2003; Young, 2009) have been repeatedly linked to the pathogenesis in PD. We observed considerable increases in ROS (superoxide) at 750 μM. Mitochondrial viability changes were observed at far lower doses (100 μM) than required to elicit ROS increase, indicating that mitochondrial loss is likely an early event in harmane-induced neurodegeneration.

We observed behavior indicative of increased dopamine levels at low doses and behavior indicative of reduced dopamine levels (increased repulsion time) at higher doses (250 and 500 μM). The nonanol assay for higher doses was not conducted since worms showed a little movement. Interestingly, harmane is known to inhibit monoamine oxidase activity, which could potentially increase DA levels (Herraiz and Chaparro, 2006). Dopamine itself is reactive and neurotoxic (Hastings and Zigmond, 1994). Thus, it is possible that at higher doses, cell loss and eventual dopamine depletion result from dopamine-dependent toxicity.

Mitochondrial complex I dysfunction is a critical pathogenic event in PD, with multiple PD-relevant toxicants acting through this mechanism to lesion dopamine neurons (Betarbet et al., 2000; Cannon and Greenamyre, 2010). Thus, given that our data showed harmane exposure reduced viable mitochondria, we tested whether a complex I activator would be protective. Previous studies on Streptomyces venezuelae-induced neurodegeneration in worms have shown that 50 mM HBA, a mitochondrial complex I activator, prevents DA cell loss (Ray et al., 2014). Thus, we treated worms with HBA at 50–200 mM. We observed significant amelioration of cell loss in response to HBA. The efficacy of HBA varied in terms of doses and neuronal class. In general, HBA produced partial amelioration of harmane-induced neurodegeneration. Amongst the DA neuronal subclasses, PDE neurons did not exhibit significant protection, perhaps because this population was more sensitive to harmane exposure. We also studied the effect of HBA on mitochondrial proton gradients. We observed a significant increase in mitochondria with intact proton gradients at 100 and 150 mM HBA, corroborating our findings on the beneficial effect of HBA. Specifically, we studied the effect of HBA on mitochondrial viability in worms treated with 650 and 500 μM harmane. Although we did not observe significant amelioration in the high dose group (650 μM), a significant increase in mitochondria with intact proton gradients was observed in low dose group (500 μM) when treated with 150 mM HBA. In light of the observed protective action of HBA, a possible explanation for a marginal effect on viable mitochondria could be that the effects of HBA are greater in neuronal mitochondria, rather than total mitochondria. Given that RB treatment did not modulate mitochondrial viability in worms exposed to harmane, but that NAC treatment provided significant protection, it is possible that activation of other complexes may be protective (ie. complex V). These studies, while suggestive of a critical role for mitochondria, especially complex I in harmane-induced neurotoxicity will need to followed by studies in systems amenable to directly detecting effects on mitochondrial complex activity.

The protective effects of HBA in harmane-treated worms were also evident in dopamine-associated behavior. In harmane treated worms, which exhibited increased dopamine-dependent repulsion time (indicative of low dopamine levels), we observed a significant improvement with HBA treatment. We also used cat-2 mutant (MT15620) and worms over expressing cat-2 (UA57) as positive and negative controls respectively to validate the experimental findings by showing that the dopamine system can be modulated and can be assessed for protection and potentiation. Significant elevation and decline in repulsion time in case of MT15620 and UA57, respectively, confirmed our findings and experimental validity. Since the nonanol assay is a behavior assay and is scored on the basis of repulsive behavior, variations in motility could bias the results of the study. Notably muscles are also rich in mitochondria. Thus, any substance enhancing the viability of mitochondria could also positively affect muscular functioning raising a possibility that the response could partially be due to the improved motility. Therefore, we addressed this question by performing the thrashing assay, where we observed that motility in worms treated with HBA was statistically insignificant to untreated worms, confirming that the observed effects in nonanol assay were solely due to improved dopamine levels.

The doses of harmane used in these studies are far higher than that a human would likely be exposed to. For example, we used 100 µM harmane for many studies. In humans, the mean log blood harmane in PD cases has been reported at 0.59 g⁻¹⁰/ml (roughly double controls) (Louis et al., 2014). Brain concentrations are elevated in essential tremor cases and are also roughly 2.5 times that in blood (Louis et al., 2013). Although comparisons to more polar molecules are not directly relevant, it is worth noting that in general, far higher doses are typically required in C. elegans than other in vivo models for most neurotoxicants and our doses are within suggested nematode testing ranges (Li et al., 2013). Further, in mammalian systems, harmane accumulates in the brain with concentrations approximately 6.5 times greater than peripheral tissue in laboratory animals, making brain as a primary target organ (Zetler et al., 1972). Selective accumulation acrosses the blood-brain-barrier into the parenchyma as well as chronic exposures are not possible in C. elegans.

Our studies for the first time show that harmane exposure is selectively neurotoxic to dopamine neurons in C. elegans. Mechanistic studies indicated that cell entry is unlikely through DAT and that the mitochondria may be a primary target. Our results suggest that future studies focused on epidemiological links between harmane and clinical PD, as well as mechanistic studies in higher order species should be conducted to establish the potential relevance of harmane exposure to PD etiology.

SUPPLEMENTARY DATA

Supplementary data are available at Toxicological Sciences online.

FUNDING

This work was supported by the National Institute of Environmental Health Sciences at the National Institutes (R01ES025750 to J.R.C.).