Early motor deficits, sleep dysfunction and reduction in dopaminergic neurons in a PARK7-/- zebrafish larval model of Parkinson’s disease
Nora Solheim, Brígida R. Pinho, Nuno A. S. Oliveira, Leonor P. Lima, Aina Børve, Jorge M. A. Oliveira, Kari E. Fladmark
TL;DR
This study introduces a zebrafish model of Parkinson’s disease that shows early motor and sleep issues, making it useful for drug testing and understanding the disease.
Contribution
The first stable genetic larval zebrafish model of Parkinson’s disease showing both motor and non-motor symptoms.
Findings
PARK7-/- zebrafish larvae show reduced dopaminergic neurons and motor dysfunction.
The model exhibits PD-related sleep disturbances like increased sleep latency and light phase sleepiness.
MPP+ exposure did not replicate the sleep issues seen in the genetic model.
Abstract
Parkinson’s Disease (PD) is the fastest-growing neurological disorder and only symptomatic treatment is available. Zebrafish are ideally suited for high-throughput screening of disease modifying drugs and mechanistic studies. Mutations in PARK7 are associated with early onset familial PD, however also idiopathic PD patients without known mutations in PARK7 show altered subcellular location and levels of its protein product DJ-1 in pathological tissues. Here, we show that PARK7-/- zebrafish already at their larval stage show a correlation between reduced number of dopaminergic neurons and motor dysfunction. Additionally, PD associated prodromal symptoms as reduced sensory function, increased sleep latency and light phase sleepiness were also observed. The 1-methyl-4-phenylpyridinium (MPP+) exposed PD model did not reproduce similar sleep disturbances. This is the first stable genetic…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4- —University of Bergen (incl Haukeland University Hospital)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsParkinson's Disease Mechanisms and Treatments · Zebrafish Biomedical Research Applications · Neurobiology and Insect Physiology Research
Introduction
Parkinson’s Disease (PD) is the fastest growing neurological disorder^1^ and only symptomatic therapy is available. About 10% of PD are caused by an inherited genetic variant, whilst the majority cases of PD are sporadic age-dependent and most possibly the result of a complex combination of/interplay between environmental exposure and genetic risk factors^2^.
PD diagnosis is set when motor-disabilities occur, although there are prodromal symptoms as disturbances in REM sleep, changes in smell and vision, reduced sense of touch, mood disorders and others^3,4^. PD is characterized by a selective loss of dopaminergic cells in the substantia nigra pars compacta. Cellular changes common for both familiar and spontaneous PD are mitochondrial dysfunction, increased oxidative stress and altered immune-related response (reviewed in^5,6^). However, the mechanisms resulting in the selective loss of dopaminergic neurons in PD are still not well understood.
Considering the limited success of existing treatments for PD and the increasing prevalence of the disease, there is an evident need for both novel disease modifying therapies and methods to halt the disease itself. In this context, in vivo models enabling the understanding of early cellular changes leading to disease and progression are indispensable.
Zebrafish serve as an excellent in vivo platform for pharmacological screening as drugs can be added directly to the water. Additionally, zebrafish can be easily genetically modified and in combination with their small and optically transparent larva they are a unique gateway for elucidating pathological mechanisms.
Dopaminergic neurons are detected already at 1 day post fertilization (dpf) in the zebrafish embryo and at 3 dpf the organization of the CNS is already complete^7,8^. Dopaminergic neurons in the zebrafish are also sensitive to oxidative stress induced by mitochondrial directed neurotoxicants^9^. Thus, zebrafish are excellent models for studying pathological mechanisms in PD and directed therapies.
Loss of function mutations in PARK7 are associated with autosomal recessive forms of early-onset PD^10^. Importantly, the PARK7 protein product, DJ-1 is also linked to spontaneous forms of PD, in which increased expression of irreversibly oxidative modified DJ-1 and reduced association to mitochondrial ATP synthase are observed^11,12^. DJ-1 is a multifunctional redox sensitive and neuroprotective protein. It has been shown to modulate oxidative stress, inflammatory response, mitochondrial function, synaptic recycling, autophagy, and proteostasis^13–19^. Zebrafish DJ-1 is highly similar to human DJ-1 and residues known to be functionally important are also found in the zebrafish orthologue^20^. In the zebrafish, the PARK7 transcript is already expressed at the embryonal stage with most prominent expression in the developing brain and gut^20^.
We have previously established a PARK7 CRISPR/cas9-based knockout zebrafish line^21^ and demonstrated a progressive increase in motor and non-motor symptoms at the adult stage^22,23^. Here, we show that this DJ-1 knockout model, in contrast to other genetic PD-related zebrafish models^24–26^, has a reduced number of dopaminergic neurons and starts to develop both motor and motor and non-motor deficits already at its larval stage.
Results
Loss of DJ-1 evoked no noticeable effects on general development and early motor activity of zebrafish embryos
In wild type zebrafish, DJ-1 is expressed throughout development (Fig. 1A and Suppl.Fig. 1). To assess whether a loss of DJ-1 affects embryonic development, we measured morphometric parameters (Fig. 1B-D). As reflected in gross morphology, the loss of DJ-1 in PARK7^-/-^ did not hamper general development as at 5 dpf wild type and knockout larvae did not differ in body length or eye aspect ratios (Fig. 1B-D). The earliest motor activity in zebrafish can be observed at 24 hpf as spontaneous tail coilings^27^. No significant difference was observed in tail coilings comparing wild type and PARK7^-/-^ larvae (Fig. 1E).Fig. 1. Larval gross morphology and embryonic spontaneous tail coilings are not altered in PARK7^-/-^. (A) Western blot shows expression of DJ-1 at 1, 3 and 5 dpf in total lysates from wild type zebrafish and the lack of DJ-1 in PARK7^-/-^ zebrafish. Arrow points to DJ-1 band. Pon-S was used as loading control. (B) Images of 5 dpf wild type and PARK7^-/-^ larvae indicating eye aspect ratio (white, C) and body length (red, D) measurements. (E) Spontaneous coilings of the tail were recorded for 3 min at 28 hpf. T-test of genotypes showed no significance difference (p > 0.05). Graph shows the mean + /- SEM,* n* = 63–66 embryos per condition, sampled from at least five independent breedings.
PARK7-/- larvae have a reduced touch-evoke escape response
Reduced sense of touch is a common and often prodromal symptom in many PD patients^3^. Mitochondrial-directed toxicants affecting dopaminergic neurons have also been shown to impair tactile sensory processing in a mouse model^28^. In view of this, we examined the touch-evoked response in larvae by applying tactile stimulation to the head and tail (Fig. 2). Larvae were exposed from 3–5 dpf to 1-methyl-4-phenylpyridinium (MPP^+^) which selectively targets dopaminergic neurons. Indeed, at 5 dpf MPP^+^ exposure reduced both head and tail response in wild type larvae. More importantly, a significant impairment in responses was also observed in PARK7^-/-^ compared to wild type, even in the absence of MPP^+^. Additionally, loss of DJ-1 resulted in an increased sensitivity towards MPP^+^.Fig. 2PARK7^-/-^ larvae have lower sensorimotor responsiveness compared to wild types. Touch-evoke response of both head (A) and tail (B) show a significant decrease (**p < 0.01, ***p < 0.001; generalised linear model (GLM) with a quasibinomial distribution followed by Tukey-adjusted pairwise comparisons of model-adjusted means to assess the effects of genotype, drug and their interaction) in sensorimotor responses for PARK7^-/-^. MPP^+^ further decreased responsiveness for both wild type and PARK7^-/-^ larvae. Values are the mean ± SEM (n = 48–84) per condition, sampled from 4–7 independent breedings.
PARK7-/- larvae have disturbed activity and sleep patterns, but no impairment in motor endurance
From 5 to 8 dpf larvae behaviour was analysed using a Locomotor Activity Monitor (LAM)^29^, which is an infrared light-sensor based locomotion set-up that monitors activity over a long period, acquiring handleable amount of data. In this way, not only motor-dysfunction can be monitored, but also circadian function and sleep disturbances. The latter being one of the most common non-motor and prodromal symptoms in PD^30^.
As shown in Fig. 3A all larvae independent of genotype or treatment had a typical circadian pattern with high activity in the circadian light period. Activity in this light phase was significantly reduced in both PARK7^-/-^ larvae and MPP^+^-treated wild types compared to untreated wild types (Fig. 3B). Similar to what was observed in the touch-response assay (Fig. 2), MPP^+^ seemed to further decrease activity in the light phase for PARK7^-/-^ larvae, but not in such degree that it was significantly different from MPP^+^ treated wild types. It should be noted that larvae were without exogenous feed in the LAM which might affect the activity in the end of the recorded period.Fig. 3. Activity and sleep in PARK7^-/-^ larvae are altered compared to wild types. Park7^-/-^ and wild type zebrafish larvae were exposed to MPP^+^ or vehicle from 3–5 dpf and thereafter washed and moved into LAM tubes for continuous behaviour analysis under light–dark cycles from 5–8 dpf. (A) Representative behaviour tracing of activity per 60 min. Y-axis shows number of beam crossings + /- SEM (n = 16). (B–G) Quantification of activity and sleep parameters from three individual experiments. Values are the mean ± SEM (n = 48) from three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, different GLMs were created depending on the data distribution and scale followed by Tukey-adjusted comparisons of the model-adjusted means to assess the effects of genotype, drug and their interaction. (B, C) Activity in light (B) dark (C) phases. (D–G) Sleep analysis of LAM data: sleep bout duration (D); sleep latency (E); sleep ratio in light phases (F); sleep ratio in dark phases (G).
Sleep bout durations and sleep latency were analysed from the dark phases. Genotype or MPP^+^ treatment had no statistically significant effect on sleep bout duration (Fig. 3D). Sleep latency, on the other hand, was increased in PARK7^-/-^ larvae throughout the experimental period (Fig. 3E).
No differences were observed in sleep ratios in dark phases, but in light phases, the sleep ratio was increased in PARK7^-/-^(Fig. 3F–G). A similar increase was not observed in MPP^+^ treated wild types.
It should be noted that larvae are not fed or exposed to toxicant between 5 and 8 dpf and that their regenerative capacity might influence the results at the late phase in the LAM.
Motor endurance has been shown to be decreased in mice exposed to the MPP^+^ precursor MPTP^31^. It was therefore of interest to evaluate motor endurance in PARK7^-/-^ larvae. Also, in zebrafish wild type larvae MPP^+^ reduced motor endurance, but in PARK7^-/-^ larvae motor endurance was similar to unexposed wild types (Supplementary Fig. 2A,B). Neither did we observe any degradation in neuromuscular junctions in PARK7^-/-^ larvae (Supplementary Fig. 1C), a feature accompanying reduction in motor endurance in zebrafish models of motor neuron disease^32,33^. The ultrastructure of the mitochondria in the PARK7^-/-^ and wild type muscle and sub-sarcolemma appeared similar (Suppl. Fig. 3) which is compatible with our previous observation that mitochondrial complex I activity is intact at this early stage^21^.
PARK7-/- larvae exhibit a lower number of DC1 diencephalic dopaminergic neurons compared to wild types
Dopaminergic (DA) neurons can be detected already at 18 h post fertilization (hpf) and at 3 dpf the distribution of DA neuronal populations is complete^7,8^. We studied the number and distribution of DA neurons using an antibody detecting tyrosine hydroxylase (TH), a marker for DA neurons (Fig. 4).Fig. 4PARK7^-/-^ larvae have a lower number of dopaminergic neurons compared to wild types. Zebrafish were exposed to 500 µM MPP^+^ or vehicle from 3 to 5 dpf, thereafter fixed, bleached, whole-mount immune-stained using anti-tyrosine hydroxylase and mounted in clearing solution. Larvae were imaged by confocal microscopy. (A) Dopaminergic cell populations at 5 dpf. Cells of interest are encircled with a dotted line. Colours refer to the stack level. (B) Circled cell population from (A) after selecting stack images of interest corresponding to DC1 and DC2 populations. (C) Schematic sagittal section of the 5 dpf brain showing the location of the DC1 and DC2 dopaminergic cell populations from a lateral view. (D) Projections of dopaminergic neurons in the selected DC1 and part of DC2 population from wild type, wild type + MPP^+^, PARK7^-/-^ , and PARK7^-/-^ + MPP^+^ at 5 dpf. Arrows point to the larger pear formed DC2 dopaminergic neurons. (E) The number of dopaminergic cells in the DC1 population at 3, 4, and 5 dpf. Numbers are the mean ± SEM from 3 to 6 (n = 3–6 larvae). **p < 0.01, ***p < 0.001 using Type II ANOVA and Tuckey-adjusted pairwise comparisons. Bars, 50 µm. Dopaminergic cell populations are named according to Rink and Wulliman^34^, which corresponds to MPP^+^ responsive cell populations (5, 6, and 11) as shown in Sallinen et al.^9^.
We measured the number of the diencephalic DA neurons of the posterior tuberculum referred to as DC1^34^ or 5, 6, 11 populations^9^ (Fig. 4A). This is the DA neuronal population comparable to the mammalian substantia nigra and which is selectively targeted by MPP^+^^9^. The DC1 population can easily be distinguished from another nearby DA population (DC2) based on morphology and the coronal level (Fig. 4 B-C). Our results showed a significant reduction in the number of DA neurons in the DC1 population in the PARK7^-/-^ at 5 dpf compared to wild type (Fig. 4D–E). This difference seemed to progress with embryonal/larval aging, since at the earlier 3 and 4 dpf the difference was not significant (Fig. 4E).
As expected, applying MPP^+^ from 3 to 5 dpf decreased the number of DC1 dopaminergic neurons in the wild type larvae. Adding MPP^+^ to the PARK7^-/-^ larvae further decreased the number of DC1 neurons, but only to a level that was comparable to the MPP^+^-treated wild types.
In another zebrafish PD model, the PTEN-induced kinase 1 (PINK1)^-/-^, the larger DC2 dopaminergic neurons appear to be targeted by affecting the neurogenesis of DC2 dopaminergic neurons^35^. We determined the number of neurons in the smaller DC2 population but found no differences between genotypes or the effect of MPP^+^ (Supplementary Fig. 4). Thus, indicating that loss of DJ-1 specifically targets the DA neuronal population comparable to the mammalian substantia nigra.
Discussion
No disease-modifying therapeutics exists for Parkinson’s disease, only symptomatic relieving drugs. Zebrafish, at their larval stage, offer unique possibilities for high-throughput phenotypic analysis and imaging analysis of a well characterized nervous system. Here we describe the behaviour and dopaminergic cell populations at the larval stage of a CRISPR/cas9-based PARK7^-/-^ zebrafish. Our results show that PARK7^-/-^ larvae not only have sensory and motor deficits but also sleep disturbances and a lower number of the dopaminergic neurons comparable to the mammalian substantia nigra. To our knowledge, this combination of PD relevant behaviour and cell specific targeting have not been shown previously in other genetic or toxicant based zebrafish models.
Mutations in PARK7 leading to a dysfunction in its gene product DJ-1 is linked to a rare, familial type of early onset PD^10^. However, even in idiopathic cases of PD, with no apparent mutations in PARK7, post-mortem analysis of brain samples revealed an increase in oxidative cysteine-modified DJ-1^11,36^. Recently, a reduced association of DJ-1 to mitochondrial F1FoATP synthase was shown in sub-cellular compartments (predominantly distal neurites) of dopaminergic neurons in the substantia nigra of post-mortem brains from sporadic PD patients^12^.
In relation to gross morphology and basic body plan, PARK7^-/-^ larvae were indistinguishable from wild types (Fig. 1). Neither was there any difference in spontaneous motor activity at the embryonal stage (24 hpf). At the 5 dpf stage, however, spontaneous motor activity, as measured using the LAM, was significantly reduced in PARK7^-/-^ larvae compared to wild types (Fig. 3). Impaired motor function has also previously been observed in the morpholino-based knockdown PD models for LRRK2^37^ and SNCA^38^, although LRRK2 knockdown induces developmental defects and conflicting evidence exists for its effect on motor activity^39,40^. Mutations in PARK7 have been associated with signs of amyotrophic lateral sclerosis (ALS) pathology in two independent studies^32,33^. However, the PARK7^-/-^ larvae showed no reduction in motor endurance, so they do not seem to display this ALS-associated phenotype (Suppl. Fig. 2)^41^.
In dopamine depleted mice tactile perception is impaired^28^. This dopamine-dependent response seems to be a conserved feature as tyrosine hydroxylase-depleted C.elegans had a dysregulation in tactile plasticity which could be rescued by dopamine^42^. We performed repetitive touches and found that PARK7^-/-^ had a significant lower response to both head and tail touches compared to wild type (Fig. 2). Thus, reduced touch responses in PARK7^-/-^ might be coupled to a dysfunctional dopaminergic system and be relevant to PD pathology, which often includes a reduced sense of touch^3^.
It is well established that mitochondrial dysfunction is strongly implicated in both familial and spontaneous PD. The mitochondrial toxicant MPTP and its oxidized product MPP^+^, which target a specific population of dopaminergic neurons, are therefore commonly used to replicate PD pathology in animal models. In this study, we used MPP^+^ to compare exposed wild type with unexposed PARK7^-/-^ and to evaluate whether loss of DJ-1 renders larvae more sensitive to MPP^+^. Loss of DJ-1 only significantly increased susceptibility to MPP^+^ in the case of lost sensitivity to head and tail touch-evoke responses (Figs. 2, 3 and 4).
In contrast to another stable genetic model of PD, the pten-induced putative kinase 1^-/-^ (PINK1^-/-^)^43^, mitochondrial Complex I activity and mitochondria in the sub-sarcolemmaof PARK7^-/-^ appeared unaffected^21^(Suppl. Fig. 3). Thus, mitochondrial dysfunction in the PARK7^-/-^ seems to develop gradually since both Complex I activity and mitochondrial morphology are affected at a later stage^21,23^.
Sleep disorders are common and evident at the prodromal stage in PD^44^. Rapid eye movement sleep behaviour disorder, excessive daytime sleepiness, and increased sleep latency are among these sleep disturbances^30,45^. Using the LAM we were able to monitor both total day-/nighttime sleep and sleep latency prior to darkness phase. Interestingly, both sleep latency and sleep ratio during daytime were increased in PARK7^-/-^, whilst sleep ratio during night phase remained unaltered (Fig. 3E–G). In contrast to motor activity these sleep disturbances persisted throughout the experimental period (5–8 dpf). MPP^+^-treated wild types did not show a similar pattern of sleep disturbances. Previous reports on exposure of rodents to MPTP have not been able to show a correlation between sleep disruptions and loss of dopaminergic neurons, suggesting that the commonly used MPTP model for PD does not fulfil a comprehensive role as a PD model, specifically in relation to non-motor symptoms^46^.
The pronounced PD-related sleep disorder phenotypes observed in PARK7^-/-^ larvae therefore increases its value as a PD model.
The dopaminergic cell population in the DC1 of zebrafish larva is comparable to the mammalian substantia nigra^8^. In zebrafish, the neurotoxicants MPTP and MPP^+^ selectively target the DC1 dopaminergic cells, leaving other dopaminergic cell populations unaffected^9^. Focusing on the tyrosine hydroxylase positive cells in the DC1 region, we saw that PARK7^-/-^ at 5 dpf, even in the absence of MPP^+^, had less dopaminergic neurons in the DC1 population compared to wild type (Fig. 4). Comparing the number of dopaminergic DC1 neurons between PARK7^-/-^ and wild type at the 3 and 4 dpf stage did not show a significant difference, thus, the reduction in dopaminergic DC1 in PARK7^-/-^ seemed to be a progressive effect. In parallel with the behavioural analysis, the loss of DJ-1 did not appear to make PARK7^-/-^ more sensitive towards MPP^+^ than wild type. Even though MPP^+^ treatment of PARK7^-/-^ larva further decreased their number of dopaminergic DC1 cells, the resulting number of cells was not significantly different from MPP^+^ treated wild types.
In the PINK1^-/- ^larvae a reduced number of dopaminergic neurons is also evident already at the larval stage^25^. Loss of PINK1 impairs neurogenesis in zebrafish, and also in human PINK1-deficient organoids^35^. As shown in Fig. 4, both wild type and PARK7^-/-^ larvae increase their number of dopaminergic DC1 cells from 3–5 dpf, but this increase is depressed in the PARK7^-/-^. Thus, loss of DJ-1 in the PARK7^-/-^ most likely inhibits neurogenesis, as in the PINK1^-/-^ larvae. Like PARK7, PINK1 is also associated with early onset PD and it has been shown that DJ-1 and PINK1 form a ubiquitin E3 ligase complex with Parkin and have a role in oxidative stress protection and promoting unfolded protein degradation^47^.
To summarize, loss of DJ-1 in the PARK7^-/-^ zebrafish results in reduced touch-evoke response, sleep disturbances and locomotor dysfunction together with a reduced number of dopaminergic neurons. This is the first stable genetic model for Parkinson´s disease presenting both motor symptoms, sleep disturbances, and reduction of dopaminergic neurons already at the larval stage^48^. This makes it an attractive model for pharmacological screening and mechanistic studies.
Materials and methods
Zebrafish husbandry
Adult zebrafish were maintained at 26–28 °C with a 14:10 light cycle and were fed twice daily. Embryos and larvae were maintained at 28 °C and raised in E3 media (5 mM NaCl, 0.17 mM KCl, and 0.33 mM MgSO4). Establishment of the PARK7^−/−^ knockout line was approved by the Norwegian National Animal Research Authority (FOTS ID8039 and ID14039)^21^.
Western blot analysis
Embryos and larvae were sampled at 24 hpf, 3 dpf and 5 dpf. Deyolking was performed with 24 hpf and 3 dpf embryos/larvae. Lysates were prepared by suspending in 150 µl homogenization buffer (10 mM K_2_HPO_4_, 10 m MKH_2_PO_4_, 1 mM EDTA, 0.6% CHAPS, 0.2 mM Na_3_VO_4_, 50 mM.
NaF, and protease cocktail (Roche Diagnostics GmbH:11836153001)) and sonication (4 × 5 s) followed by incubation on ice for 20 min. Samples were pelleted at 15,000 × g for 15 min at 4 °C and 25 µg protein from the supernatant was separated by SDS-PAGE and transferred to PVDF membranes using 14 V overnight at 4 °C.
Membranes were blocked in 1% BSA and incubated with anti-DJ-1 (1:3000, Novus Biologicals NB300-270, 1 h) followed by secondary antibody. Ponceau-S was used as a loading control.
Tail coiling
Embryos at 27 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\pm$$\end{document} 1 hpf were transferred to a 96 well-plate containing E3 (6 embryos/well). After a 5 min acclimation period, the embryos were recorded for 3–4 min using a Dino-Eye eyepiece camera (AM-423U). The videos were cut to exactly 3 min and the number of tail-coilings (random movement) were counted manually.
Morphometric analysis of larvae
Larvae were fixed o.n. in 4% PFA/PBS and washed in PBS before being mounted in agar molds. Images were taken with a Nikon SMZ800N and morphometric analysis was performed using ImageJ.
1-methyl-4-phenylpyridinium (MPP+) treatment
At 3 dpf, hatched larvae that did not display any abnormalities were randomly distributed in 12-well plate (10–15 larvae/1 mL/well) and exposed to 500 µM MPP^+^ (D048, Merck KGaA, Darmstadt, Germany). The MPP^+^ solution and vehicle were renewed at 4 dpf by exchanging 400 µl with 500 µl fresh solutions. At 5 dpf, the larvae were transferred to E3 in 6-well plates (1 larva/well) for sensorimotor responses or the LAM system (1 larva/tube) for continuous activity monitoring, without further exposure to MPP^+^.
Sensorimotor responses
In 6-well plates (one larva/well) each larva was gently touched with a micropipette tip, alternating between the head and tail. Immediate swimming was registered as a positive response, whereas no movement was registered as a negative response, yielding a binary result^49^. Each larva was touched ten times at both head and tail, with a 30-s break in between. A minimum of n = 12 larvae, randomly selected from several breedings, were used for each condition.
Locomotor activity monitor (LAM)
Larvae were individually placed in E3 media-filled LAM tubes (3.65 mL) with a 100 μl air bubble inside. They were evenly dispersed and placed in the LAM for continuous activity monitoring until 8 dpf, while maintaining 28 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\pm$$\end{document} 1 °C and a 14 h:10 h light–dark cycle, as previously described^29^. The starting point of the day consisting of 24 h was set to 09.00AM when the lights were turned on. The LAM data were analysed in Rtivity software^50^. Sleep threshold was set to 10 min of inactivity and bout activity threshold to 2 min. Larval activity was measured as the number of infrared beams crossing disturbances per 30 s. Total activity for each day (counts) was analysed. Furthermore, activity (counts), sleep ratio (proportion of time spent sleeping), sleep latency (the mean time between lights turned off and start of the first sleep bout, min) and sleep bout duration (the mean time spent being continuous asleep without interruptions, min) were analysed for each day separated by light/dark phases. Sleep latency and sleep bout duration only contained data from the dark phases. n = 16 larvae for each condition randomly selected from several breeding tanks were used for each experiment.
Motor endurance assessment of zebrafish larvae
For motor endurance assessment, we used a custom gravity-fed counter-current swimming system, capable of inducing the counter-current swimming reflex in zebrafish larvae without immediately overpowering them (manuscript in preparation). 7 dpf zebrafish larvae were moved to a 90 mm dish with chlorine-free water, 30 min prior to the experiment, for acclimation. 5 to 10 larvae were then moved into a 5 mL serological pipette, used as the swimming tunnel, and left to acclimate for 5 min. During this period, the larvae were guided to the middle of the swimming tunnel. Afterwards, water flow was initiated. Initial flow speed was set at 3 cm/s, gradually increasing to 5 cm/s over the course of 10 min. Endurance times were recorded as the length of time that each fish was able to stay in the swimming tunnel, before being pushed out by the water flow.
Whole-mount labelling and image analysis
Larvae were fixed in methanol free 4% PFA/PBS 4 °C o.n. with gentle agitation. Thereafter the larvae were washed in 6 × 15 min in PBS with 0.05% Triton X (PBS-TX) and stored at 4 ºC until further use. Permeabilization was performed by incubation with acetone at -20 ºC o.n. Thereafter larvae were washed in PBS-TX followed by 2 times wash in PBS. Bleaching was done in 3% H_2_O_2_ in 0.89% KOH under bright light (approx. 8 min for 5 dpf larvae). Larvae were then washed 3 × 5 min in PBS-TX and 2 × 15 min PBS-TX. Clearing was performed according to Pende et al.^51^. Larvae were then transferred to blocking solution (5% normal goat, 1% BSA and 0.1% Triton X) for 3 h at RT and incubated for 3 days with either anti-tyrosine hydroxylase (TH) (1:250, Merck Millipore MAB318) or anti-SV2 (1:1000, DSHB SV2-C) in antibody dilution buffer (5% goat serum, 1% BSA and 0.1% Triton-X in PBS) at 4 °C with gentle agitation.. It should be noted that available antibodies detecting zebrafish TH only detect TH1 and not TH2, in which TH1 is the isoform predominantly expressed in the brain^52^. Samples were then washed 6 × 15 min in PBS-Tx and incubated for 2 days at 4 °C with appropriate secondary antibody together with a 1:6000 dilution of DAPI in PBS-Tx. Larvae were then washed 3 × 15 min in PBS-Tx followed by 3 × 15 in PBS. Larvae were then transferred to 50% refractive index solution (50% sucrose, 12% Antipyrin, 8% nicotinamide and 10% trietanolamine)^53^ for at least 2 h at RT followed by at least 30 min in 100% refractive index solution. Larvae were then mounted in 2% low melt agarose on glass slides and imaged on Olympus FLUOVIEW FV3000 confocal laser scanning microscope.
Confocal images were processed in Image J, using the plugin CLAHE to enhance local contrast^54^. The number of DA neurons were calculated by counting tyrosine hydroxylase positive cells in the diencephalic posterior tuberculum DC1 area by moving through the confocal stack from the dorsal side until reaching the easily recognizable larger DA neurons of the DC2^34^. A maximum projection of the described area was generated and the DC1 neurons positioned anterior to the DC2 neurons were counted. The DC1 population of DA neurons includes posterior tuberculi populations 5, 6, and 11 as shown in Sallinen et al.^9^ known to be specifically targeted by MPP^+^.
Ultrastructural analysis
5 dpf larvae were fixed with 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M sodium cacodylate buffer overnight. Samples were then washed with 0.1 M sodium cacodylate buffer prior to post-fixed with 1% osmium tetroxide in 0.1 M sodium cacodylate buffer. Samples were then washed in 0.1 M sodium cacodylate buffer and dehydrating with increasing ethanol solutions. The samples were embedded in Agar 100 resin, sectioned using a microtome (Reichert Ultracut S Ultra microtome, Leica Biosystems, Nussloch, Germany), stained with uranyl/lead, and imaged with a transmission electron microscope (Jeol JEM-1230).
Analysis in R-studio
Statistical analysis was done in R-Studio. The mean ± standard error (SEM) was calculated for each genotype with respective treatment for n larvae from 3 to 5 different breedings. For each dataset, model choice was based on data distribution and scale. t-test or ANOVA was used for normally distributed data. For non-normalized distributed data, generalized linear models (GLM) were created. Analysis was done between genotypes, MPP^+^ exposure, and their interaction.
As sensorimotor response was measured as the positive response in percentage from n = 10 touches, the data was modelled with a generalized linear model using a quasibinomial distribution with logit link to account for overdispersion, followed by Tukey-adjusted pairwise comparisons. Total activity, activity separated by light phases, sleep latency, and sleep bout duration in the locomotor activity monitor were analysed using a generalised linear model with a negative binomial distribution as the data showed overdispersion (variance > mean). Post hoc comparisons of estimated marginal means were Tukey-adjusted. Sleep ratio separated by light phase were analysed using beta regression, followed by post hoc Tukey-adjusted comparisons on the fitted model. Lastly, the number of cells positively stained for tyrosine hydroxylase were assessed with Type II ANOVA, post hoc Tukey-adjusted comparisons. Statistical difference was defined as p-value < 0.05.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary Material 1