Longitudinal Assessment of Skilled Forelimb Motor Impairments in DJ-1 Knockout Rats

Abstract

Background:

DJ-1 knockout (DJ-1 KO) rats exhibit a moderate parkinsonian phenotype, with gross motor deficits and ca. 50% loss of midbrain dopaminergic neurons appearing around 6–8 months of age. Fine motor impairments are often observed in Parkinson’s disease (PD), but skilled motor function in recently developed transgenic rat models of PD is not well characterized.

Objectives:

To assess the longitudinal performance of DJ-1 KO rats on a skilled forelimb reaching task.

Methods:

DJ-1 KO and wild-type (WT) rats were trained from 2–10 months of age on an isometric pullbar task designed to test forelimb strength and coordination. After 36 consecutive weeks of training (ca. 10 months old), task difficulty was then increased to challenge the motor capabilities of the DJ-1 KO rats. Throughout the study, subjects also received weekly assessments of gross locomotor activity in an open field.

Results:

Pull-task performance of the DJ-1 KO rats was impaired compared to WT, with deficits reaching significance around 7–9 months of age. When challenged, DJ-1 KO rats were able to exert increased force on the pullbar but continued to exhibit deficits compared to WT rats. Throughout the study, no differences in distance travelled or rearing frequency were observed in the open field, but DJ-1 KO rats were found to spend significantly more time in the center of the open field than WT rats.

Conclusions:

Using a sensitive, automated assay of forelimb strength and coordination, we find that skilled forelimb motor performance is impaired in DJ-1 KO rats.

INTRODUCTION

Parkinson’s disease (PD) is the most common neurodegenerative movement disorder, affecting an estimated one million people in the United States (Kowal et al., 2013; Rossi et al., 2018). PD is characterized by the progressive death of dopaminergic cells in the substantia nigra pars compacta (SNc), which results in the muscle rigidity and bradykinesia that are the hallmark of this disease. With the recent increase in the accessibility of gene editing technologies, several transgenic rat lines have been developed that model loss-of-function mutations associated with familial forms of PD (Creed & Goldberg, 2018; Dave et al., 2014). While genetic mutations are linked to only 5–10% of PD cases (Balestrino & Schapira, 2020; Tysnes & Storstein, 2017), these preclinical rodent models provide important tools for understanding PD pathophysiology and developing improved treatments.

Among these recently developed rat lines, PARK7/DJ-1 knockout (KO) rats have been shown to exhibit a moderate parkinsonian phenotype (Dave et al., 2014; Giangrasso et al., 2020; Kyser et al., 2019; Sun et al., 2013; Yang et al., 2018). PARK7/DJ-1 was first identified as an oncogene (Nagakubo et al., 1997) and its product, DJ-1, is a redox-responsive protein implicated in a rare form of early onset PD (Abou-sleiman et al., 2003; Bonifati et al., 2003). DJ-1 plays a key role in the maintenance of mitochondrial health through the regulation of oxidative stress. DJ-1 aids in nuclear translocation of reactive oxygen species (ROS) activating proteins (Wang et al., 2011), apoptosis regulation by sequestering and precluding cell death signaling proteins (Junn et al., 2005), regulation of anti-oxidant enzymes (Yan et al., 2015; Zhong & Xu, 2008), and mitochondrial matrix uncoupling under oxidative stress (Guzman et al., 2010). Due to their high basal levels of oxidative stress, SNc dopaminergic neurons are thought to be particularly sensitive to dysregulation of these mitochondrial signaling pathways (Bose & Beal, 2016; Guzman et al., 2010; Subramaniam & Chesselet, 2013; D. James Surmeier et al., 2017; D J Surmeier et al., 2011). PARK7/DJ-1 loss of function hinders the ability of SNc cells to regulate oxidative stress, increasing the probability of SNc cell death and the development of parkinsonian symptoms.

Phenotypic characterization of DJ-1 KO rats has revealed deficits in gross motor function, including reduced rearing, impaired hindlimb grip strength, and gait abnormalities that develop between 6 and 8 months of age (Dave et al., 2014; Kyser et al., 2019). Additional orofacial deficits are observed at even earlier timepoints (Yang et al., 2018) and alterations in metabolic signaling have been reported at 3 months of age (Almikhlafi et al., 2020). Consistent with the development of gross motor deficits, by 8 months of age, DJ-1 KO rats also exhibit ca. 50% loss of tyrosine hydroxylase positive neurons in the SNc and in the locus coeruleus (LC) (Dave et al., 2014; Yang et al., 2018). This loss of catecholaminergic cell bodies is accompanied by a paradoxical increase in striatal dopamine concentration (Dave et al., 2014; Yang et al., 2018), among other neurochemical changes (Creed et al., 2019), which have been hypothesized to reflect a compensatory increase in DA release by remaining nigral cells during the earliest stages of neurodegeneration (Giangrasso et al., 2020; Sun et al., 2013). As deficits in dexterity and fine motor control are commonly observed even in early stages of PD (Rosenblum et al., 2013), we aimed to test whether DJ-1 KO rats exhibit impairments in skilled forelimb motor coordination, which has not been extensively characterized in this rat line.

To characterize skilled motor performance in the DJ-1 KO rats, we trained DJ-1 KO and WT rats on an automated skilled reaching pullbar task previously shown to be a sensitive assay of forelimb strength and coordination (Hays et al., 2013). To examine the development and progression of motor impairments in the DJ-1 KO line, we longitudinally characterized pullbar task performance from 2 to 10 months of age. DJ-1 KO rats were found to exhibit significant performance deficits compared to WT subjects, which began to reach statistical significance between 7 and 9 months of age. After 36 weeks of training, pullbar task difficulty was increased. During this challenge period, we found that DJ-1 KO rats were able to apply increased force to the pullbar, but they continued to exhibit performance impairments compared to WT rats. Our results suggest that DJ-1 KO rats exhibit deficits in fine motor coordination that are not explained by a deficiency in forelimb strength.

METHODS

All procedures were approved by The University of Texas at Dallas Institutional Animal Care and Use Committee in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Animal Subjects

Male DJ-1 KO rats (n = 8, LEH-Park7tm1sage, Horizon) and wild-type Long Evans controls (WT, n = 8, HsdBlu:LE, Envigo), aged eight weeks old at study start, were included in these experiments. Rats were housed in a 12:12 hour light cycle facility (lights on: 6 am) and trained during their light cycle.

Isometric pullbar task

Training on the skilled reaching pull task (Hays et al., 2013) required subjects to reach with their right forelimb and pull an isometric bar placed 1.5 cm outside a MotoTrak (Vulintus Inc, Louisville, CO) training booth (Fig. 1A–B). Rats were required to meet or surpass threshold force on the pullbar within 2 sec from trial initiation to receive a food reward (Fig. 1C–D). Rats began training on the pull task at 2 months of age, and received two 30-minute training sessions daily, 3–5 days per week (sessions/week = 9.5 ± 1.2, mean ± STD). To maintain motivation to perform the task, rats were lightly food restricted throughout training. Prior to the first training session of each week, food was removed from the homecages, and rats were fasted overnight for 12–18 hours. After each daily pull task session, rats received three pellets of standard rat chow (ca. 10–12 g; Labdiet Prolab RMH 1800) upon return to their homecages. Following the last training session of each week, rats were allowed to consume food ad libitum until the fasting period began for the following week. Weights were monitored daily and maintained at or above 95% of subjects’ free feeding body weight.

Figure 1.

DJ-1 KO and WT rats underwent 41 weeks of skilled forelimb motor assessment. (A) Rats were required to reach for and pull an isometric bar with their right forelimb. Pull thresholds (“hit thresholds”) required to receive reward were adapted within each session according to each rat’s performance. (B) Task training consisted of three phases: Acclimation (weeks 1–4), Acquisition (weeks 5–36), and Challenge (weeks 37–41). Acclimation was conducted in several stages, in which the lever was progressively moved to its final position (Pos) 1.5 cm outside the training booth, and in which hit thresholds (Hit Thresh) were progressively increased prior to the longitudinal Acquisition phase (see Methods for detail). During Acquisition, task difficulty was held constant for 32 consecutive weeks, with hit thresholds adjusted adaptively within each session between 100 and 200 g. During Challenge, task difficulty was increased by moving the lever position to 2 cm outside the training booth (requiring a greater reach) and by increasing the upper limit of the hit threshold to 300 g (requiring stronger pull force to receive reward). (C) Sample force traces from a DJ-1 KO subject (DJ2) during week 31 of training. Top: correct (“hit”) trial examples, bottom: incorrect (“miss”) trial examples. In each panel, t = 0 denotes trial initiation, and the horizontal gray line denotes the hit threshold for the trial shown. (D) Example force traces as in (C) for a WT subject (WT7) during week 31 of training.

DJ-1 KO rats began training at 2 m.o., 5 weeks prior to age-matched WT rats. Rats were trained on the pullbar task in several stages (Fig. 1B). Rats first received 3–4 days of Habituation sessions (6.27 ± 1.73 sessions) to acclimate to food delivery at the reward port. During each 30-minute Habituation session, reward pellets (Bio-Serv, Flemington, NJ; part #F0021) were randomly delivered as the rat freely explored the training booth. Acclimation Stage 1 followed, during which subjects began to associate pullbar engagement with reward delivery. During all subsequent Acclimation and Acquisition sessions, trials were automatically initiated if the detected force applied to the pullbar exceeded 3 g. Trials were then terminated upon detection of above-threshold pull force (hit trials), or after 2 seconds if force on the pullbar never exceeded the hit threshold (miss trials). A 1-sec timeout period occurred after trial termination before a new trial could be initiated. For each trial, applied force was continuously sampled at 100 Hz and stored from 1 sec before to 3 sec after trial initiation.

During Acclimation Stage 1 (7.73 ± 4.4 sessions), the pullbar was initially placed 2.0 cm inside the booth, and the pull force threshold required to receive reward was low (3 g). The pullbar was progressively moved back, in 0.5 cm increments every 30 correct trials, from 2.0 cm inside the booth until the final position of 1.5 cm outside the training booth was reached. Subjects were moved to Stage 2 upon completion of at least 30 successful trials with lever position 1.5 cm outside the booth. During Stage 2 (11.5 ± 4.6 sessions, mean ± STD), the lever position was fixed at 1.5 cm outside of the booth and the force threshold required to receive reward was set to a constant 10 g. Stage 3 was introduced once all subjects performed over 100 trials per day and achieved at least 50% correct trials for at least 2 days at Stage 2. In Acclimation Stage 3 (11.7 ± 5.9 sessions, mean ± STD), the adaptive threshold was introduced. During Stage 3, the pull force threshold was initially set to 30 g. After the first 10 trials, the force threshold was adaptively set to the mean of the top 5 maximum pull forces achieved in the last 10 trials, up to a maximum threshold of 120 g. Stage 3 continued until the end of the training week. All rats completed all stages of Habituation and Acclimation in 3–4 weeks (25–37 sessions).

Longitudinal task Acquisition sessions began immediately following Acclimation. During Acquisition, the lever was positioned 1.5 cm outside of the booth and the force threshold was adapted throughout each session. In each Acquisition session, the reward threshold was initially set to 80 g. After the first 10 trials, the reward threshold was set to the mean of the 5 greatest maximum pull forces achieved in the last 10 trials, with lower and upper adaptive threshold limits of 100 and 200 g, respectively. Acquisition data is reported for the age-matched DJ-1 and WT cohorts from weeks 5–36 of training (ca. 3–10 months of age).

At training week 37, the task difficulty was increased during a 5-week Challenge period (training weeks 37–41, ca. 10–11 months old). During Challenge sessions, the upper limit on the adaptive pull force threshold was increased to 300 g and the lever position set to 2.0 cm exterior to the training booth.

Data Analysis

To characterize longitudinal forelimb performance in the pullbar task, we examined four main parameters: number of trials performed, percent correct (rewarded) trials performed, maximum pull force achieved, and average adaptive force thresholds for each session. All pull task results presented in the main text are computed from session-wise averages of these performance parameters. For each subject, number of trials and percent correct performance were calculated for each session, then averaged across all sessions in each week to obtain a weekly average for each subject prior to group-wise comparisons. For analysis of maximum forces achieved and force thresholds, parameters were calculated for each trial, then averaged across trials within each session, then across sessions within each week to obtain weekly data for each subject. Maximum force for each trial was taken as the highest force achieved within the full trial-time window from 1 second before to 3 seconds after trial initiation. For each subject, weight-normalized calories consumed was additionally estimated for each week of training as the average number of correct trials per session x 3.6 calories per pellet delivered divided by the subject’s average body weight during that week.

Two-factor mixed design analysis of variance (ANOVA) was used to compare the effects of genotype and training week for all behavioral parameters of interest. One-way repeated measures ANOVAs were then used to test whether a significant effect of training week existed within each separate genotype cohort. As Mauchly’s test indicated a lack of sphericity in the longitudinal pull task data (p < 0.000 for all parameters), for all repeated-measure effects, Greenhouse-Geisser epsilon was used to correct the degrees of freedom of the F statistic prior to computing p-values. ANOVA results were then corrected for multiple comparisons across the 5 pull task performance parameters using the Holm-Bonferroni method; corrected p-values are reported throughout. Where within-group changes in performance across Acquisition or Challenge phases were of interest, post hoc paired t-tests were used to compare performance during initial versus final training sessions in each phase, as noted in the text. To test for between-group differences in pullbar performance during each week of training, post hoc Student’s t-test comparisons were performed.

To ensure that our findings were not impacted by the choice of averaging procedure, we repeated the statistical analyses of the main output measures (percent correct, maximum force, and pull threshold) using weekly averages computed for each subject across all trials performed in each week (rather than across sessions). Similar results were obtained using both methods, and results using the alternative averaging approach are presented in Supplementary Material.

All statistical analyses were performed in Matlab (Mathworks, Natick, MA), and all results are reported as significant if p < 0.05. Summary data are reported as mean ± standard error of the mean (SEM).

Open Field

Throughout the study, rats also underwent weekly monitoring of locomotor behavior in an open field. Open field data was collected throughout the 41-week study, except during 2 weeks coinciding with an institutional closure (rats continued to receive pullbar training during these weeks). During open field sessions, subjects were placed in a 61 × 61 × 45.7 (L × W × H) cm acrylic box for 30 minutes. Three of the sides of the box were covered with white paper, leaving one side transparent and the top uncovered to allow video monitoring from the side and above, respectively. Video recordings were captured from a Logitech C270 Webcam at 10 frames per second with a 320×240 pixel resolution using VirtualDub software. Custom MATLAB scripts were used to track the location of the rats within the maze offline, to measure each subject’s total distance traveled and the percent time spent in the center of the open field. Rearing events were counted manually by 2 graders blinded to the animal’s genotype. If the initial rearing counts of the 2 graders differed by less than 20%, the total number of rears was estimated to be the average of the two scores (570 of 652 recordings). For sessions in which the initial counts differed by more than 20% (82 of 652 recordings), a third blinded grader was assigned, and all three counts were averaged together. The average percent difference in rearing scores for all videos quantified was 6.77 ± 6.03%.

To characterize gross locomotor function in the open field, total distance traveled, percent time spent in the center of the field, and the number of rears performed were calculated for each weekly 30-minute session. Weekly open field data was then compared between groups and across training weeks using 2-factor mixed ANOVA, followed by one-way repeated measures ANOVAs to test for separate within-group effects of training week. Repeated measures tests were Greenhouse-Geisser corrected, and all ANOVAs were adjusted for multiple comparisons across the 3 open field parameters using the Holm-Bonferroni method. Four weeks of data were excluded from open field ANOVAs due to the institutional closure, which coincided with DJ-1 KO weeks 39 and 40, and WT weeks 34 and 35. Additional data was excluded from total distance and percent time in center analyses if top-view video was not recorded during the open field session (n = 2 sessions total: subject DJ2, weeks 11 & 29), and from rearing analyses if side-view video was not captured (n = 2 sessions total: subject WT4, week 26; subject DJ2, week 29). To test for between-group differences in open field performance during each week of training, post hoc Student’s t-test comparisons were performed. For all statistical tests, significant differences are reported if p < 0.05. All summary data are reported as mean ± standard error of the mean (SEM).

RESULTS

DJ-1 KO and WT rats underwent longitudinal pullbar task training for 41 consecutive weeks, beginning at 2 months of age (Fig. 1A–D). Pull task training consisted of Acclimation (weeks 1–4), Acquisition (weeks 5–36), and Challenge (weeks 37–41) phases (Fig. 1B). The pull task required rats to reach for and pull an isometric bar located outside the training booth with sufficient force to receive food reward (Fig. 1C–D). Force thresholds were adapted within each training session according to the subject’s behavioral performance to maintain motivation to perform the task (see Methods). Throughout the study, rats also received weekly assessments of locomotor activity in an open field (Fig. 2A).

Figure 2.

DJ-1 KO rats did not exhibit gross motor deficits in the open field. (A) Sample paths traveled by one DJ-1 KO subject (DJ3; left/yellow) and one WT subject (WT3; right/blue) during week 20 of training. Traces are from representative subjects in each cohort exhibiting the median distance travelled and percent time spent in the center in week 20. Gray shading denotes the center of the field. (B) No differences were found between DJ-1 KO and WT rats in total distance traveled in the open field across the 41-week study. (C) DJ-1 KO rats spent significantly more time in the center of the open field than WT rats throughout the study. (D) Rearing in the open field decreased in both DJ-1 KO and WT rats across training weeks. In B-D, *: p < 0.05 and **: p < 0.01, between-group t-test comparisons for each week of training; #: p < 0.05 and ##: p < 0.01, within-group one-way ANOVA across training weeks.

DJ-1 KO rats exhibited no impairment in the open field

To assess gross motor activity in the open field, we first examined total distance traveled during each weekly 30-minute session. Two-way mixed ANOVA revealed no significant main effect of genotype, and no significant genotype x week interaction for all 41 weeks of training (Fig. 2B and Table 1), suggesting DJ-1 KO and WT were similarly ambulatory in the open field throughout the study. We did observe a significant main effect of training week, which subsequent one-way repeated measures ANOVAs indicated was significant only within the WT group. Post hoc tests revealed no significant difference in distance traveled between the first 4 weeks of training versus the last 4 weeks for either group (paired t-test weeks 1–4 vs. weeks 38–41, DJ-1 KO: p = 0.399; WT: p = 0.159), suggesting no consistent changes in ambulatory activity occurred across the 41-week study. Weekly between-groups comparisons confirmed that there were no significant differences in total distance traveled between DJ-1 KO and WT rats in any training week (Table S1).

Table 1.

Comparison of gross locomotor activity in the open field and skilled forelimb pull-task performance between DJ-1 and WT rats.

	2-way mixed ANOVA (genotype × week)			1-way ANOVA (week effects, within-group)
	pgenotype (Fgroup) [d.f.]	pweek (Fweek) [d.f.]	pint (Fint) [d.f.]	pdj (Fdj) [d.f.]	pwt (Fwt) [d.f.]
Weight (weeks = 1–41)
Weight	0.027 (6.12) [1,14]	0.000 (296.10) [40,560]	0.000 (18.02) [40,560]	0.000 (74.01) [40,280]	0.000 (302.82) [40,280]
Open Field Locomotion (weeks 1–41)
Total Distance Traveled	0.860 (0.03) [1,14]	0.014 (2.97) [34,476]	0.457 (1.57) [34,476]	0.264 (1.39) [34,238]	0.031 (2.83) [34,238]
% Time in Center	0.011 (12.07) [1,14]	0.045 (2.70) [34,476]	0.237 (1.44) [34,476]	0.306 (1.90) [34,238]	0.009 (4.76) [34,238]
Rearing Count	0.110 (4.39) [1,14]	0.000 (5.70) [34,476]	0.317 (1.51) [34,476]	0.046 (3.49) [34,238]	0.018 (3.71) [34,238]
Pull Task – Acquisition (weeks 5–36)
Percent Correct	0.248 (4.11) [1,14]	0.000 (12.42) [31,434]	0.037 (3.28) [31,434]	0.173 (2.96) [31,217]	0.000 (12.75) [31,217]
Maximum Pull Force	0.375 (1.92) [1,14]	0.000 (8.34) [31,434]	0.043 (3.00) [31,434]	0.200 (1.68) [31,217]	0.001 (10.08) [31,217]
Pull Threshold	0.657 (0.21) [1,14]	0.000 (7.99) [31,434]	0.183 (1.61) [31,434]	0.209 (2.32) [31,217]	0.003 (8.84) [31,217]
Trials Performed	0.005 (17.38) [1,14]	0.000 (6.94) [31,434]	0.001 (5.36) [31,434]	0.005 (6.35) [31,217]	0.002 (5.95) [31,217]
In-task Calories Consumed per Body Weight	0.477 (2.21) [1,14]	0.000 (8.86) [31,434]	0.000 (8.39) [31,434]	0.001 (6.88) [31,217]	0.000 (9.91) [31,217]
Pull Task – Challenge (weeks 37–41)
Percent Correct	0.424 (3.44) [1,14]	0.002 (7.13) [4,56]	0.177 (1.76) [4,56]	0.216 (1.71) [4,28]	0.005 (8.03) [4,28]
Maximum Pull Force	0.304 (3.08) [1,14]	0.000 (22.89) [4,56]	0.017 (4.86) [4,56]	0.001 (12.15) [4,28]	0.001 (15.45) [4,28]
Pull Threshold	0.388 (3.16) [1,14]	0.000 (20.66) [4,56]	0.023 (4.51) [4,56]	0.002 (9.80) [4,28]	0.000 (16.69) [4,28]
Trials Performed	0.376 (191) [1,14]	0.020 (3.64) [4,56]	0.014 (5.23) [4,56]	0.022 (7.23) [4,28]	0.286 (1.35) [4,28]
In-task Calories Consumed per Body Weight	0.904 (0.015) [1,14]	0.000 (9.89) [4,56]	0.001 (7.82) [4,56]	0.007 (9.56) [4,28]	0.005 (7.88) [4,28]

We next examined whether DJ-1 KO rats differed from WT rats in the time spent exploring the center of the open field, which is often used as an assessment of anxiety-like behavior. Two-way mixed ANOVA revealed significant main effects of both genotype and training week, but no significant interaction effect (Fig. 2C and Table 1). One-way repeated measures ANOVA again suggested that the effects of training week were significant only in the WT group, though post hoc comparisons revealed no differences between performance in the first 4 weeks of the study compared to the last 4 weeks for either group (paired t-test weeks 1–4 vs. weeks 38–41, DJ-1 KO: p = 0.172; WT: p = 0.250). Weekly between-groups comparisons revealed that in 28 of 37 weeks tested (75.68%), DJ-1 KO rats spent significantly more time in the center of the open field than WT rats (Table S1), consistent with a previously reported low-anxiety phenotype for this transgenic rat line (Kyser et al., 2019).

We further tested whether DJ-1 KO and WT rats exhibited significant differences in rearing across weeks of training, which is frequently used as a measure of exploratory behavior in the open field. Two-way mixed ANOVA revealed no significant genotype effect, a significant effect of training week, and no significant genotype x week interaction (Fig. 2D and Table 1). Subsequent one-way repeated measures ANOVAs revealed significant effects of training week for both DJ-1 KO and WT groups. Post hoc comparisons confirm that both groups decreased their rearing across weeks (paired t-test weeks 1–4 vs. weeks 38–41, DJ-1 KO: p = 0.002; WT: p = 0.012), consistent with a decline in exploratory behavior with repeated exposure to the testing environment (Rojas-Carvajal et al., 2018). Weekly between-groups comparisons revealed that DJ-1 KO and WT rats exhibited significant differences in rearing in only 9 of 37 weeks tested (24.32%), which were spread throughout the 41 week study (Table S1).

Taken together, our results indicate that, compared to WT subjects, the DJ-1 KO rats in our study exhibited a low-anxiety phenotype that was unaccompanied by significant deficits in gross locomotor output.

DJ-1 KO rats exhibit impaired performance on the pullbar task

Skilled forelimb reaching was longitudinally assessed in DJ-1 KO and WT rats using an automated isometric pullbar task. After pullbar task acclimation, difficulty was held constant throughout 32 weeks of Acquisition training (training weeks 5–36, ca. 3–10 m.o.), with adaptive hit thresholds in each twice-daily 30-min training session ranging between 100 and 200 g. Overall task performance for DJ-1 KO and WT rats was first examined using weekly measures of percent correct performance and maximum pull force applied to the isometric bar. For percent correct performance (Fig. 3A & Supplementary Fig. 1A) we observed no significant main effect of genotype, however, we did observe a significant main effect of training week and a significant training week x genotype interaction (Table 1 and Table S2). Subsequent one-way repeated measures ANOVAs revealed a significant effect of training week on percent correct performance only in the WT cohort (Table 1 and Table S2). Post hoc comparisons indicated that performance improved in both cohorts between the first and last months of Acquisition, though this effect was stronger in the WT group (paired t-tests, percent correct performance in weeks 5–8 vs. weeks 33–36; DJ-1 KO: p = 0.004; WT: p < 0.000). Similar results were observed for maximum pull forces applied to the bar (Fig. 3B and Supplementary Fig. 1B), though statistical effects were generally weaker than for percent correct performance (Table 1 and Table S2; paired t-tests, max pull force in weeks 5–8 vs. weeks 33–36; DJ-1 KO: p = 0.035; WT: p < 0.002). Combined, these results suggest that, over the 32 weeks of Acquisition, WT rats improved their performance on the pullbar task to a greater degree than DJ-1 KO rats.

Figure 3.

DJ-1 KO rats exhibit forelimb motor deficits during pullbar task Acquisition. (A) WT rats exhibit higher percent correct performance than DJ-1 KO rats beginning ca. week 22 of training. (B) Across all trials (hit and miss), WT rats exert higher averaged maximum pull forces than DJ-1 KO rats beginning ca. week 26. Inset: during weeks 26 to 36, DJ-1 KO rats also exhibit a deficit in pull force when only correct (hit) trials are considered. (*: p = 0.043, Student’s t-test). (C) No significant differences were observed in average hit thresholds across Acquisition between DJ-1 KO and WT rats. (D) DJ-1 KO rats performed more trials per session than WT rats during most Acquisition weeks, and particularly after ca. week 22 of training. (E) DJ-1 KO were heavier than WT rats prior to training week 18 (ca. 6.5 m.o.). (F). In-task calorie intake was normalized to each rat’s weight to test whether weight and performance differences could explain differences in task engagement. DJ-1 KO rats consume more calories per gram of body weight than WT rats, particularly after ca. week 22 of training. All panels: *: p < 0.05, **: p < 0.01, ***: p < 0.001, between-group t-test comparisons for each training week; #: p < 0.05, ##: p < 0.01, ###: p < 0.001, within-group one-way ANOVA across training weeks.

Post hoc t-test comparisons revealed statistically significant differences in pull task performance began to emerge between DJ-1 KO and WT groups around week 22 of training (ca. 7.5 m.o.; Fig. 3A–B, Supplementary Fig. 1A–B, Table S3, and Table S4). WT rats performed a significantly higher percentage of correct trials compared to DJ-1 KO rats starting in week 22 and exhibited greater maximum pull forces starting in week 26 (ca. 8.5 m.o.). When only successful trials were considered, DJ-1 KO rats were still seen to exert lower pull forces than WT rats between weeks 26 and 36 of training (Fig. 3B, inset; DJ-1 KO vs. WT: p = 0.043; unpaired two-tail t-test). Throughout Acquisition, mean pull thresholds increased across training weeks, but did not differ significantly between the DJ-1 KO and WT cohorts in any week (Fig. 3C, Table 1, Supplementary Fig. 1C, Table S2, Table S3, and Table S4), indicating that task difficulty was approximately matched across the two groups throughout the study. Taken together, our pull task results are consistent with a mild-to-moderate deficit in forelimb motor coordination in the DJ-1 KO rats that reaches statistical significance around 7–9 months of age.

To assess whether DJ-1 KO rats exhibited reduced motivation to perform the pullbar task, we compared the average number of trials performed per session by DJ-1 KO and WT rats throughout the Acquisition period. Two-way mixed ANOVA revealed significant main effects of genotype and training week, as well as a significant interaction effect (Fig. 3D and Table 1). One-way repeated measures ANOVAs revealed a significant effect of training week for both DJ-1 KO and WT cohorts (Table 1). Post hoc t-tests showed that DJ-1 KO rats performed more trials per session than WT rats in 20 of the 32 weeks of Acquisition (62.5%), and 14 of the last 15 weeks (93.33%), suggesting that, compared to WT rats, DJ-1 KO rats exhibited increased motivation to engage in the pull task as performance deficits emerged (Fig. 3D and Table S5).

We have previously observed that task engagement is highly correlated with rats’ weight and caloric needs (Tseng et al., 2020). We therefore examined whether the difference in trials performed between DJ-1 KO and WT rats could be explained by differences in body weight and percent correct performance. DJ-1 KO rats were significantly heavier than the age-matched WT rats prior to week 18 of training (ca. 5.5 m.o.; Fig. 3E, Table 1, and Table S5). To compare in-task calorie consumption across the two groups, we normalized each subject’s weekly calorie intake (hits x calories/pellet) by its average body weight for each week of Acquisition (Fig. 3F). Two-way mixed ANOVA on normalized calorie intake revealed a significant main effect of training week and a genotype x week effect, but no significant main effect of genotype (Table 1). One-way repeated measures ANOVAs showed a significant effect of training week for both DJ-1 KO and WT rats (Table 1). Post hoc t-test comparisons revealed that DJ-1 KO rats consumed significantly more calories per gram of body weight than did WT rats for 13 of 32 weeks of Acquisition (40.62%), and 10 of the last 15 weeks (66.67%). These results suggest that differences in body weight and task success rate cannot fully explain the increased task engagement of the DJ-1 KO rats once forelimb performance deficits emerge after ca. 7.5 months of age.

DJ-1 KO rats remain impaired during pull task Challenge phase

After 32 weeks of Acquisition, task difficulty was increased during a 5-week Challenge period in which the adaptive pull force threshold in each session ranged between 100 and 300 g. This increase in task difficulty was accompanied by a sharp decrease in percent correct performance for both cohorts between the end of Acquisition and first week of Challenge (Fig 4A and Supplementary Fig. 2A; paired t-test, percent correct performance in last month of Acquisition [weeks 33–36] vs. week 1 of Challenge [week 37]: DJ1-KO: p = 0.001, WT: p < 0.000). Pull task performance subsequently improved throughout the 5-week Challenge period, as assayed by percent correct performance (Fig. 4A and Supplementary Fig. 2A), maximum pull forces applied (Fig. 4B and Supplementary Fig. 2B), and hit thresholds (Fig. 4C and Supplementary Fig. 2C). For overall percent correct performance (Fig. 4A and Supplementary Fig. 2A), two-way mixed ANOVAs revealed a significant effect of training week, and no significant effect of genotype or genotype x week interaction (Table 1 and Table S2). One-way repeated measures ANOVAs showed a significant effect of training week in the WT, but not the DJ-1 KO, cohort (Table 1 and Table S2). Post hoc comparisons confirmed that from week 1 to week 5 of Challenge, only the WT rats showed a significant increase in percent correct performance (paired t-test on percent correct in week 37 vs. week 41: DJ-1 KO: p = 0.141; WT: p = 0.009). Weekly between-group comparisons further suggest that WT rats outperformed DJ-1 KO rats, with differences in percent correct performance reaching statistical significance in Challenge week 4 (Fig. 4A, Supplementary Fig. 2A, Table S3, and Table S4).

Figure 4.

DJ-1 KO rats exert additional pull force during Challenge but forelimb performance remains impaired. (A) Percent correct performance declined in both groups as task difficulty increased from Acquisition to Challenge. Percent correct performance improved over the 5 weeks of Challenge for WT, but not DJ-1 KO, rats. (B-C) Pull forces (B) and hit thresholds (C) increased during the Challenge period for both DJ-1 KO and WT cohorts, though WT rats exerted significantly higher pull forces than DJ-1 KO rats during the final weeks of training. All panels: *: p < 0.05, between-group t-test comparisons for each training week; ##: p < 0.01, ###: p < 0.001, within-group one-way ANOVA across training weeks.

Interestingly, both DJ-1 KO and WT rats were seen to exert increasing force on the pull bar across the Challenge period. For both maximum force applied (Fig. 4B and Supplementary Fig. 2B) and for hit thresholds (Fig. 4C and Supplementary Fig. 2C), two-way mixed ANOVAs revealed a significant main effect of training week, and a significant genotype x week interaction, though no significant main effect of genotype alone (Table 1 and Table S2). One-way repeated measures ANOVAs revealed a significant effect of training week for both groups, for both measures. Post hoc comparisons further confirmed that both groups exhibited increased pull forces from week 1 to week 5 of Challenge (paired t-test on max pull force in week 37 vs. week 41: DJ-1 KO: p = 0.005; WT: p = 0.002), and hit thresholds also increased (paired t-test on hit thresholds in week 37 vs. week 41: DJ-1 KO: p = 0.006; WT: p = 0.001). These results suggest that DJ-1 KO rats were capable of applying additional force to the isometric bar once task difficulty increased, though unlike WT rats, this increase in pull force was not sufficient to translate into an increase in percent correct performance. Weekly between-group comparisons confirmed that forelimb reaching performance remained impaired in DJ-1 KO rats compared to WT rats during the Challenge period, with differences in maximum pull force and hit threshold reaching statistical significance during the final two weeks of training (Fig. 4B–C, Supplementary Fig. 2B–C, Table S3, and Table S4). Significant effects of training week were observed for weight-normalized calorie intake using two-way mixed and one-way repeated measures ANOVAs (Table 1), but post hoc tests revealed no significant between-group differences in any Challenge week (Table S5).

As DJ-1 KO subjects were capable of applying additional force to the bar during the Challenge period, our results suggest that performance deficits exhibited by the DJ-1 KO rats during Acquisition cannot be fully explained by a deficit in forelimb strength.

DISCUSSION

In the current study, we extend the characterization of the recently developed DJ-1 KO rat line by longitudinally assessing forelimb strength and coordination using a skilled reaching isometric pullbar task. We find that around 7–9 months of age, DJ-1 KO rats begin to exhibit statistically significant performance deficits on the pullbar task, suggesting that in addition to well-documented gross motor deficits, these rats also display impairments in skilled forelimb motor control.

By contrast with several prior reports, we did not observe significant differences between DJ-1 KO and WT rats in distance traveled or rearing frequency in the open field (Dave et al., 2014; Giangrasso et al., 2020; Kyser et al., 2019). These disparate results may be due to differences in experimental protocol. Our rats were slightly food deprived and assessed weekly in a larger arena than was used in prior studies. During these weekly assessments, we found that both DJ-1 KO rats and WT subjects exhibited similarly high levels of ambulatory and exploratory behavior, though unsurprisingly, overall activity in both groups declined over time, as subjects grew increasingly acclimated to the testing environment. Our results suggest that DJ-1 KO rats did spend significantly more time in the center of the open field than WT subjects, which is consistent with prior reports of a low-anxiety phenotype in this line (Giangrasso et al., 2020; Kyser et al., 2019).

Forelimb reaching performance in DJ-1 KO and WT rats began to significantly diverge beginning around 7.5 months of age, consistent with prior behavioral studies characterizing gross motor deficits in these rats. We find that from ca. 7.5 to 10 months of age, DJ-1 KO rats consistently apply lower pull forces, and achieve a lower percentage of correct trials, than their WT counterparts. Our finding that DJ-1 KO rats were able to exert additional force on the pullbar during Challenge, coupled with prior studies that report no deficit in forelimb strength for these rats (Dave et al., 2014), are consistent with centrally mediated disruption in forelimb motor coordination. The development of these motor impairments is correlated with previously reported catecholaminergic cell loss in both the SNc and the LC, which has been shown to occur between 6 and 8 months of age. Other alterations in forebrain neurochemistry have been reported for this strain and may additionally contribute to performance deficits in the DJ-1 KO rat line. These include an increase in striatal dopamine and tyrosine hydroxylase levels after ca. 8 months of age (Dave et al., 2014; Giangrasso et al., 2020), an increased evoked release of striatal acetylcholine prior to ca. 8 months of age (Creed et al., 2019), and an increase in serotonin transporter binding in prefrontal cortices at ca. 4 months of age (Giangrasso et al., 2020). Based on the unique neurochemical alterations reported for the DJ-1 KO rat line, it has been hypothesized that behavioral and pathophysiological changes observed in these rats may model the very early stages of DA dysfunction and neurodegeneration that result from DJ-1 loss of function (Guzman et al., 2010; D J Surmeier et al., 2011). Consistent with this idea, we find that DJ-1 KO rats exhibit a stable, mild-to-moderate impairment in forelimb reaching performance, and we extend the longitudinal behavioral characterization of this “early-stage” DJ-1 KO phenotype out to 11 months of age.

We also observe that DJ-1 KO rats performed more trials than WT rats throughout the course of our longitudinal study. Prior to ca. 5 months of age, this difference could be accounted for by a significant difference in body size (Yang et al., 2018), as the heavier DJ-1 KO rats titrated their task engagement to earn equivalent calories per gram of body weight as the WT rats. However, around 7.5 months of age, when motor performance deficits reached statistical significance, DJ-1 KO rats were also found to consume significantly more calories per body weight than WT subjects. The root causes of this difference remain unclear, but our findings are consistent with several possibilities, including a reported involvement of DJ-1 in glucose metabolism (Shi et al., 2015; Wu et al., 2017). Alternatively, increased task engagement might arise from changes in motivation or impulsivity related to alterations in forebrain neuromodulatory signaling that are reported around this time period in the DJ-1 KO rats.

Impairments in manual dexterity and fine motor control are common in PD patients, and often manifest in the earliest stages of the disease (McLennan et al., 1972; Rosenblum et al., 2013), even prior to the tremor, rigidity and bradykinesia that are the defining motor symptoms of PD (Moustafa et al., 2016). Forelimb reaching tasks are often used in preclinical studies to assay analogous forelimb function in rodents. Such tasks offer rapid, high-throughput and sensitive assessments of forelimb dexterity, strength, and coordination, and are commonly employed to quantify motor dysfunction and recovery after stroke and other neural injuries (Hays et al., 2013; Johansen-Berg et al., 2018; Khodaparast et al., 2014). Though less frequently employed to assay parkinsonian phenotypes, several neurotoxin-mediated and transgenic rodent models of PD have been previously shown to exhibit deficits in forelimb reaching (Barnéoud et al., 1996; Klein et al., 2012; Montoya et al., 1991; Vergara-Aragon et al., 2003). Though we did not observe significant gross motor impairments in our DJ-1 KO rats in the open field assay, we nonetheless found that DJ-1 KO rats exhibited significant deficits in the more automated and sensitive skilled reaching pullbar task, consistent with a mild-to-moderate parkinsonian phenotype reported for this strain. Our study thus adds further support for the use of such assays in the quantification of mild or early-stage parkinsonian behavioral phenotypes.

Highlights:

• A 41-week study of forelimb motor coordination was conducted in DJ-1 KO and WT rats

• DJ-1 KO rats exhibited deficits in forelimb coordination compared to WT rats

• WT rats outperformed DJ-1 KO rats beginning around 7–9 months of age

• DJ-1 KO forelimb motor deficits were not explained by a deficit in strength

• DJ-1 KO forelimb deficits were unaccompanied by deficits in gross locomotor function