Manganese-Enhanced Magnetic Resonance Imaging and Studies of Rat Behavior: Transient Motor Deficit in Skilled Reaching, Rears, and Activity in Rats After a Single Dose of MnCl₂

Abstract

Manganese-enhanced magnetic resonance imaging (MEMRI) has been suggested to be a useful tool to visualize and map behavior-relevant neural populations at large scale in freely behaving rodents. A primary concern in MEMRI applications is Mn²⁺ toxicity. Although a few studies have specifically examined toxicity on gross motor behavior, Mn²⁺ toxicity on skilled motor behavior was not explored. Thus, the objective of this study was to combine manganese as a functional contrast agent with comprehensive behavior evaluation. We evaluated Mn²⁺ effect on skilled reach-to-eat action, locomotion, and balance using a single pellet reaching task, activity cage, and cylinder test, respectively. The tests used are sensitive to the pathophysiology of many neurological and neurodegenerative disorders of the motor system. The behavioral testing was done in combination with a moderate dose of manganese. Behavior was studied before and after a single, intravenous infusion of MnCl₂ (48 mg/kg). The rats were imaged at 1, 3, 5, 7, and 14 days following infusion. The results show that MnCl₂ infusion resulted in detectable abnormalities in skilled reaching, locomotion, and balance that recovered within 3 days compared with the infusion of saline. Because some tests and behavioral measures could not detect motor abnormalities of skilled movements, comprehensive evaluation of motor behavior is critical in assessing the effects of MnCl₂. The relaxation mapping results suggest that the transport of Mn²⁺ into the brain is through the choroid plexus-cerebrospinal fluid system with the primary entry point and highest relaxation rates found in the pituitary gland. Relaxation rates in the pituitary gland correlated with measures of motor skill, suggesting that altered motor ability is related to the level of Mn circulating in the brain. Thus, combined MEMRI and behavioral studies that both achieve adequate image enhancement and are also free of motor skills deficits are difficult to achieve using a single systemic dose of MnCl₂.

Introduction

Manganese-enhanced magnetic resonance imaging (MEMRI)1 is an imaging tool for “functional mapping” of behavioral-relevant neural circuits at megascale (across many brain areas) that can be applied to restrained or freely behaving animals.2–7 MEMRI is also reported to reveal brain cytoarchitecture,8 function-specific neural pathways,9–13 and function-relevant neural populations.14,15 The unique property of Mn²⁺ is that it is transferred through voltage-gated Ca²⁺ channels in “excited” neurons leaving a relative accumulation of Mn²⁺ in neural populations that have recently contributed to behavior.16,17 Accumulated Mn²⁺ reduces T₁ and T₂ relaxation time resulting an increase in signal or enhancement in the image. Throughout this article, we refer to image enhancement as the signal (and signal to noise) increase resulting from the accumulation of Mn²⁺. There is a direct relationship between the local concentration of Mn²⁺ and the relaxation rate (R1 = 1/T₁). Thus, quantitative R1 (and R2) maps are proportional to Mn²⁺ concentrations and can visualize behavioral-relevant neural populations.18

A major drawback of MEMRI is Mn²⁺-induced toxicity to visceral organs, eg, liver, heart, lungs, and kidneys,19,20 and the central nervous system.21–27 Although Mn²⁺ at high doses20 is known to cause a Parkinson-like condition called manganism, the minimum dosage at which detectable effects on motor behavior occur is still an open question. With single infusion of Mn²⁺ for the purposes of imaging, many studies report the absence of serious behavioral abnormalities.2,15 However, toxic effects also occur after the infusion of Mn²⁺ as an imaging contrast agent in laboratory animals.4,21,28,29 Mn²⁺ is reported to accumulate in brain structures involved in motor control including striatum and basal ganglia and motor cortex.30,31 There are reports of motor abnormalities32,33 with low chronic ingested or inhaled exposures of Mn²⁺. More detailed behavioral studies have mixed findings. One study shows detectable motor abnormalities in wheel running,4 whereas another reports the absence of abnormalities in swimming.6 A follow-up running wheel study34 shows a lack of abnormalities with the use of a continuous-delivery osmotic pump. There are a growing number of studies35 comparing functional imaging with blood oxygen level–dependent and MEMRI methods.

MEMRI methods have the advantage that the functional task takes place outside of the magnet and is more natural for rodents. With a growing interest in functional methods, a careful investigation of the interaction of the MEMRI methods and behavioral tasks is necessary.

The purpose of this study was to investigate further the potential of combining MEMRI with various measures of motor behavior to determine the time frame of potential toxicity. The rats were given behavioral tests (Figure 1) of skilled hand use in reaching for food,36,37 activity in an open field,38 and balance in a cylinder test.39 A moderate imaging dose was used, and behavior was tested before and after MnCl₂ infusion. The present sensitive measures of fine motor skills and balance have not been previously reported in conjunction with MnCl₂ infusion.

Figure 1.

Experimental timeline (upper panel). The infusions (MnCl₂ or saline) took place at day 0 as indicated by the arrow. Behavioral apparatus (bottom panel)—single pellet reaching task evaluates skilled forelimb use, activity cage assesses locomotion and vertical exploratory activity/balancing ability, cylinder task assesses balancing ability. MRI indicates magnetic resonance imaging. P indicates pellet.

Materials and Methods

A total of 30 female, young adult (90–100 days old) Long-Evans hooded rats, from both the University of Lethbridge animal colony and Charles River Breeders (Lachine, Quebec), were used. The rats were assigned to naïve (n = 6), saline (n = 9), and MnCl₂ (n = 15) groups. Some rats in the MnCl₂ group (n = 11) received 48 mg/kg (1.11 mL), whereas others received (n = 4) 95 mg/kg (2.22 mL) of MnCl₂ solution. Saline group received an equivalent volume of saline (1.11 mL). Naïve rats were not exposed to any injections. The experiments were conducted in compliance with the guidelines and approval of the “University of Lethbridge Animal Care Committee” (Institutional Ethical Committee). The “University of Lethbridge Animal Care Committee” follows the standards for the ethical use and care of animals in science of the “Canadian Council for Animal Care” which reflect international guidelines. The rats were housed in Plexiglas cages (36-cm long, 20-cm wide, and 21-cm deep) with sawdust bedding. Each cage contained groups of 2 or 3 rats. The cages resided in a colony room that was maintained on a 12 hours light/12 hours dark cycle (08:00–20:00 hours) with controlled temperature and humidity. Behavioral testing took place between 09:00 and 13:00 hours, in experimental rooms with controlled temperature and humidity.

The timeline of behavioral testing, infusions (MnCl₂ or saline), and magnetic resonance imaging (MRI) is shown in Figure 1.

Skilled reach-to-eat movement

The procedures used to acquire and analyze the single pellet reaching data follow those of Moon et al.40 The elements that are relevant to this work are highlighted.

Single pellet reaching box

Single pellet reaching boxes37 were constructed from clear Plexiglas (35 cm × 14 cm × 45 cm). The center of the front wall of the box had a vertical slot 1-cm wide that extends from 2 cm above the floor to a height of 15 cm. On the outside of the wall, in front of the slot, a 2-cm-wide shelf was mounted 3 cm above the floor. Two indentations to hold the food targets were located 2 cm from the inside of the wall and aligned with the edges of the slot (Figure 1, bottom panel).

Feeding and food restriction

The rats were motivated to reach by gradually restricting their food intake such that they maintained a body weight between 90% and 95% of normal. This was achieved by allowing access to a fixed daily ration (19–20 g per rat per day) of Purina rat chow prior to and during training. During the week prior to training, each rat received 20 dustless precision pellets of 45 mg (product #F0021; BioServe Inc., Frenchtown, NJ, USA) each day 1 hour prior to the daily ration. This familiarized the rats with the food used as the reaching target.

Video recording

Reaching performance was recorded using a Sony DCR-VX1000 3CCD camcorder (shutter speed: 1000/s) under a cold light. A Sony videocassette recorder (DSR-11) was used for subsequent frame-by-frame analysis. Representative still frames (Figure 1, lower left and right) were captured from digital video recordings with Final Cut Pro HD (V.4.5; http://www.apple.com).

Training

The initial pretraining sessions (10 minutes) took place daily. In these sessions, the rats were allowed to retrieve the pellet from the shelf with the mouth and tongue, aimed to familiarize the rat with food being available on the shelf for eating. To encourage the use of a paw, the pellets were moved further away as the animals learned to retrieve them. Once the rats demonstrate a preference for 1 paw, individual pellets were placed into the indentation contralateral to that paw. Because the rats pronate the paw lateral to medial, the pellet placement permits the use of only the preferred paw. During the second week of training, the rats received daily (~10 minutes) sessions of discrete trials. During intertrial intervals, the rats were shaped to leave the slot, walk to the rear wall of the cage, turn, and approach the slot again for the next pellet. In addition, by withholding food pellet on semirandomly selected trials, the rats were taught to sniff the shelf for a pellet and to reach only if a pellet was present. Thus, each rat eventually orients to the food pellet, transports its paw through the slot, grasps the food pellet, and retracts its paw through the opening to release the food into its mouth.41

After 2 to 3 weeks, each rat was given 20 daily trials until group success levels exceeded 50%. The rats then received MnCl₂ and testing of all rats recommenced 24 hours later. After infusion, the rats were retested in reaching for single pellets daily for the first week and then on day 14.

End-point analysis of reaching behaviour

Reaching behavior was analyzed by measures of the following:

1. First trial success. First trial successes were ones in which a rat obtained a food pellet on the first advance of the paw toward the food. First trial success % = (number of pellets obtained on first advance/20)*100.

2. Total success. A successful reach was defined as one in which an animal grasped a food pellet and placed it into its mouth independent of the number of attempts. Success % = (number of pellets obtained/20)*100.

3. Total reaching attempts. Total reaching attempts included all movements of the paw through the slot independent of success.

Movement analysis of reaching behaviour

The biometric measures of reaching action were scored before and after MnCl₂ infusion. Scoring was based on a conceptual framework derived from Eshkol-Wachman movement notation (EWMN).37,42 Eshkol-Wachman movement notation describes the position of individual limbs, the trunk, the snout, and the head in relation to the food pellet with the body treated as a system of articulated axes (ie, limb segments, trunk axis, and snout axis). The following movements were scored36:

1. Orient. The head and the snout are oriented toward the target so that that rat sniffs the target.

2. Limb lift. The paw is raised from the floor until the digits are aligned with the midline of the body.

3. Digits close. The palm is supinated so that it is oriented vertically, and the digits are semiflexed.

4. Aim. The elbow is adducted to the body midline with a movement of the upper arm, whereas the digits remain positioned on the body midline.

5. Advance. The forelimb moves forward through the slot.

6. Digits extend and open. The digits are extended as the limb is advanced and then are opened as the paw is pronated over the food.

7. Pronate. The elbow abducts with a movement of the upper arm pronating the paw over the target in an arpeggio movement.43

8. Grasp. The arm remains still, whereas the digits close to grasp the food and then the paw is extended and raised.

9. Supinate I. The paw is supinated by 90° by abducting the elbow so that the paw holding the food can be withdrawn through the slot.

10. Supinate II. The paw is supinated so that the palm faces the mouth.

11. Release. The food pellet is released into the mouth by opening the digits.

12. Replace. The paw is placed on the floor surface (or sometimes the rat may rear and place the paw on the wall) with all digits extended.

Each of these movement elements was rated using a 3-point scale. A score of “1” indicated that the movement was normal. A score of “1.5” was awarded to a movement that was present but abnormal. A score of “2” meant the movement was absent.

Scoring was achieved by stepping through the video record frame by frame (30 f/s) and applying a score to each of the movements for the first 3 successful reach with first attempt. The movement scores were then averaged for selected reaching actions to calculate the total movement score.

Kinematics of reaching behaviour

MnCl₂ effects on the kinematics of reaching movement were assessed by the time required to execute the individual gestures during the reaching action. Gestures of the forelimb in skilled reaching44,45 were identified by their trajectories, by the pauses, and changes in trajectory that occur between them or by their repetitions:

1. Advance. A gesture in an advance begins with the paw lifted from the floor. The forelimb moves through the slot, to approach the food, and may or may not include a return toward the mouth.

2. Grasp. A grasp begins with the paw moving toward the food target and ends as the digits flex and close in an attempt to grasp the pellet.

3. Withdrawal. A gesture in a withdrawal begins with a movement of the forelimb away from the apparent location of the food to bring the paw to the mouth.

4. Release. A release withdraws the paw from the mouth to return it to the floor or redirects the paw to the advance.

In this study, the time required to execute advance, withdrawal, and release was compared between the groups before and after MnCl₂ infusion.

Motor performance index

Total movement scores were normalized with preinfusion performance (performance index [PI]) to avoid a confounding effect from the baseline (ie., biological) variability in qualitative aspects of reach-to-eat action. The PI was calculated as the postinfusion score divided by the preinfusion score for each rat.40

Activity measures

Daily activity of MnCl₂, saline, and naïve rats was acquired using a VersaMax Animal Activity Monitoring System (AccuScan Instruments, Inc., Columbus, OH, USA; https://www.alnmag.com/company-profiles/accuscan-instrumentsinc). The rat’s spontaneous activity was first examined in the open field (square area with the walls (42.5 cm × 42.5 cm × 30 cm; Figure 1, bottom panel) at 24 hours after infusion (MnCl₂ or saline). During the first postinfusion week, spontaneous activity was measured daily and on day 14 after infusion. The rats were not exposed to the activity box prior to the infusions to ensure that novel environment exploration activity took place. Activity was measured as total distance and the number of rears during each 30-minute session.

Cylinder test

The balancing ability of the rats during vertical activity was examined in the cylinder test39 (Figure 1, bottom panel). Rats were individually placed in a transparent Plexiglas cylinder (20 cm in diameter and 30 cm in height) for 5 minutes. When put in a cylinder, a rat spontaneously rears up to explore the cylinder and the wall of the cylinder. The rat rears either with or without the support of the body by the forepaws. The cylinder was located on a transparent table with a mirror placed directly underneath the table and tilted at an angle to allow filming of the rats from a ventral view. Balancing ability was measured as the number and time of vertical exploratory activity with and without supporting the body by leaning with forepaws on the wall.

Magnetic resonance imaging

MRI contrast agent, dose, and the route of administration

The choice of MnCl₂ dose and the route of administration were based on previous findings, the efficacy of contrast for imaging,46,47 and the least adverse effect on motor behavior and visceral organs.48 Intravenous infusion was used because this delivery leads to rapid transport to the nervous system31,49–51 while minimizing adverse effects on visceral organs.19,20 The dose of MnCl₂ used for all of the rats was 48 mg/kg. The rats were given a single dose of MnCl₂ via a tail vein infusion. For greater image enhancement, a higher dose (95 mg/kg) was attempted on a small number of animals. All of these animals died during or within 2 days of infusion. Because our goal was to achieve image enhancement with a dose that minimally affected behavior, this part of the experiment was stopped. Thus, behavioral testing was only performed on the naïve, saline, and 48 mg/kg MnCl₂–injected rats.

Manganese preparation and infusion

The preparation of the manganese stock solution was based on the procedures outlined in the work by Silva.29 Specifically, NaOH was added to a 100-mM solution of Bicine in double-distilled water (ddH₂O) to adjust the pH to about 8.5. The amount of MnCl₂ powder to achieve a 120-mM solution was then added. This addition lowered the pH to about 6.5. The final pH of 7.4 was obtained by slowly adding drops of a NaOH solution (10 pellets in 10 mL ddH₂O). This procedure resulted in a clear amber colored solution with no visible precipitates.

A syringe pump (Pump 11 Elite; Harvard Apparatus; http://www.harvardapparatus.com) was used to infuse either 1.11 mL (48 mg/kg in 350 g rate) or 2.22 mL (95 mg/kg in 350 g rat) of the solution into the tail vein at a rate of 1.06 mL/h. The delivered dose of 48 mg/kg of MnCl₂ corresponds to 20.8 mg/kg of Mn²⁺ (the exact dose of 47.7 mg/kg has been rounded to 48 mg/kg for brevity). During manganese administration, anesthesia was maintained with about 1% isoflurane anesthesia. During MnCl₂ infusion, Po₂ and respiration, heart rate, and oxygen saturation were carefully monitored using the MouseOx (STARR Life Sciences Corp.) and maintained at 40 to 48 breathes/min, 250 to 320 beats/min, and 96% to 98% of oxygen, respectively. The rectal temperature was monitored and kept constant at 37°C ± 0.5°C using a homeothermic blanket system (www.harvardapparatus.com).

MRI equipment and imaging protocol

The images were acquired using an MRI scanner consisting of a 4.7-T 330-mm bore Oxford magnet (Oxford, UK) and an SMIS MR 5000 (Surrey, UK) console. The imaging gradients were an Oxford self-shielded gradient set capable of 15 G/cm driven by Techron amplifiers (initially, model 7550 upgraded to model 7780). The imaging coil was a homebuilt 8-rung quadrature birdcage design that was 52 mm in diameter and 54 mm in length. The hybrid coupler used to combine the quadrature channels was an Anaren model 10012-3, and the preamplifier was a MiteQ model AU-1466. Sets of crossed diodes and λ/4 cables were inserted between the hybrid and the pre-amp for passive separation of the low power from the high-power radio frequency (RF).52 A 300-W broadband RF amplifier (AMT Model M3206) was used. An MRI-compatible Stereotaxis (Kopf, CA) was used to position the animals with the ear bars passing through holes in the cylinder of the custom-built MRI coil. The imaging protocol for each rat consisted of the following: (1) 3 multislice localizer spin echo [SE] image sets (repetition time [TR]: 600 ms, echo time [TE]: 42 ms, 0.3 mm × 0.3 mm × 1.6 mm) in coronal, axial, and sagittal orientations; (2) a single-slice saturation recovery spin-echo T₁ measurement (8 saturation delays 3000, 2000, 1200, 850, 600, 400, 300, 150; TE: 22 ms, 0.5 mm × 0.5 mm × 1.6 mm) in the sagittal directions; and (3) the same single-slice T₁ measurement in the coronal direction; and (4) a 3-dimensional (3D) T₁-weighted gradient echo (TR: 100 ms, TE: 7 ms, θ: 70°, 0.6 mm × 0.6 mm × 0.6 mm) acquisition sequences. The total imaging time of the protocol was about 1 hour. The sagittal T₁ measurement was prescribed to be along the midline of the brain using the localizer images. The coronal T₁ measurement was acquired on a slice located at the bregma.53 The 3D acquisition used a 12-mm coronal slice centered about the hippocampus. The 3D field of view (FOV) was 76 mm × 76 mm with a 19-mm FOV in the slice direction. The 3D acquisition was 128 × 128 × 32 to give a 0.6-mm isotropic voxel acquisition. Each dimension was zero filled during processing to give 0.3-mm isotropic voxels in the final registered images. Reference samples (ie, phantoms) containing 0.12, 0.11, and 0.10 mM MnCl₂ were placed above the brain for standardization and quality control. The samples were selected to give a T₁ of ~900, 950, and 1000 ms. gas anesthesia (isoflurane) was used during the imaging sessions. T₁ values were calculated for each pixel in the image using home-written software written in IDL (ITT, Boulder, CO, USA) to produce T₁ maps. Regions of interest for various anatomical structures were manually extracted from the T₁ maps with Analyze 11.1 (BIR, USA). Registration was performed with the 3D voxel registration tool of Analyze 11.1.

Statistical analysis

Statistical analysis was performed using SPSS 13. Data were treated as interval data according to the definition of Field and Hole.54 Results for end-point measures (success, success on the first reaching attempt), spontaneous activity, and balancing capability were analyzed using mixed-design repeated measures analysis of variance, “days” used as the within-subjects variable, and “groups” as the between-subjects factor. The Mauchly test and Greenhouse-Geisser corrections were used to correct data sphericity. Comparisons between groups were performed by follow-up Bonferroni test. Dependent (paired-samples) t tests were used to compare pre- and postcontrast reaching performance.

Movement elements were analyzed using nonparametric tests (Mann-Whitney U and Kruskal-Wallis). Wilcoxon signed ranked t test (z statistics) followed by the Bonferroni correction was used for multiple follow-up comparisons between pairs of conditions.

The difference in the relaxation rate in brain regions and pituitary gland between groups was determined by independent t test. The Levene test for equality of variance was run to check the assumption of homogeneity of variance, and the degree of freedom was corrected in accordance with the test result. To assess whether the relaxation rate in pituitary gland and behavior was related, correlative (bivariate) analysis was performed between the relaxation rate and behavioral measures. Moreover, the Pearson (r) and Spearman (ρ) correlative analysis was conducted between relaxation rate and motor performance indices of skilled reaching action. For all analyses, significance was considered to be a P value less than .05.

Results

All rats survived after MnCl₂ infusion through the tail vein at the dose 48 mg/kg. General observation of grooming and home cage activity indicated that these rats tolerated the procedure well. The initial recovery from anesthesia was not noticeably different from the saline-infused group, and the rats were ready for behavioral testing within 24 hours of the MnCl₂ infusion. The data below and following discussion were based on the 48 mg/kg dose of MnCl₂. In the MEMRI literature, systemic doses from 5 to 175 mg/kg29 have successfully been used for MRI image enhancement, and thus, the Mn²⁺ amount is referred to below as a moderate imaging dose.

Skilled reach-to-eat movement

End-point measures

Success

A summary of reaching success scores for naïve, saline, and MnCl₂ groups is illustrated in Figure 2. There was no adverse effect of the moderate dose of MnCl₂ on the success level either acute or in the chronic post-MnCl₂ period. Repeated measures analysis of variance did not give significant difference among groups (F_2,23 = 1.09; P > .05), testing days (F_4.1,94.7 = 1.37) or group by day interaction (F_8.2,94.7 = 0.70; P > .05).

Figure 2.

Successful reaches made by rats (mean ± SEM) in a single pellet reaching task. The infusions (MnCl₂ or saline) took place at day 0 as indicated by the arrow. Note that MnCl₂ infusion slightly reduced success on day 1, but the change is not significant.

Success on the first reaching attempts

There was no difference among groups for the more demanding measure of skilled reach success on the first reach. Repeated measures analysis of variance gave no group (F_2,23 = 0.37; P > .05), day (F_3.8,86.9 = 1.52; P > .05), or day by group effect (F_7.6,86.7 = 1.24; P > .05).

Reaching attempts

Number of reaching attempts was also similar between naïve, saline, and MnCl₂ groups following infusion. There was no group (F_2,23 = 0.58; P > .05), day (F_2.7,56.5 = 1.46; P > .05), or day by group effect (F_4.9,56.5 = 0.86; P > .05).

Biometric measures of reaching movement

Prior to MnCl₂ administration, there was no difference between rats assigned to be naïve, saline, and MnCl₂ groups for any movement elements of reaching (P > .05). Significant abnormal reaching movements were observed 24 hours after MnCl₂ administration in the MnCl₂ group compared with naïve and saline groups (total movement scores: χ²(2) = 13.8; P < .001). The group differences originated mainly from the statistically significant difference in supination II and release (Table 1).

Table 1.

Group differences in movement elements in a single pellet reaching task before and 24 hours after MnCl₂ infusion.

There was a statistically significant difference in skilled paw use after MnCl₂ infusion compared with pre-MnCl₂ behavior, but the difference was observed only at 24 hours after MnCl₂ infusion (Figure 3B). Wilcoxon signed rank tests (followed by the Bonferroni correction) between pairs of conditions (pre-MnCl₂ vs day 1—MnCl₂) yielded statistically significant difference in total movement scores (T = 1, z = −2.85; P < .004). The difference stemmed from the MnCl₂ effect on supination II (T = 1, z = −2.71; P < .007) and release (T = 0, z = −2.67; P < .007) (Figure 3A).

Figure 3.

(A) Movement elements of skilled reaching action a single pellet reaching task before and 24 hours after infusion. Note significant abnormalities on “supination II” and “release.” (B) Total scores of all movement elements of skilled reaching. MnCl₂-induced deficit in skilled reaching was transient and recovered on the second day after MnCl₂ infusion. Lower scores correspond to normal movement; high scores—affected movement. Asterisks denote significant group differences (P < .05).

Kinetics of reaching gestures

MnCl₂ slowed down the reaching gestures, especially the gestures that bring the grasped pellet to the mouth (Table 2).

Table 2.

Duration of the reaching gestures in a single pellet reaching task before and 24 hours after MnCl₂ infusion.

The grasp lasted less than 1 ms both before and after MnCl₂ administration. The comparison of grasp time was not significant; however, these movements are at the limit of the timing sensitivity.

Activity measures

A summary of the MnCl₂ effect on gross motor behavior in activity cages is shown in Figure 4. MnCl₂ infusion significantly reduced both horizontal (the distance) (Figure 4A) and vertical (rearing) activity (Figure 4B). The effect was transient and returned to normal levels by day 4. Repeated measures analysis of variance for the distance moved gave significant day (F_3.7,85.9 = 4.15; P < .05) and day by group effect (F_7.5,85.9 = 2.77; P < .05). Group effect (F_2,23 = 1.89; P > .05) was not significant. The planned comparison yielded statistically significant differences between the MnCl₂ and both saline and naïve groups on days 1 and 3 (P < .05) following MnCl₂ infusion. Repeated measures analysis of variance for rearing activity yielded significant group effect (F_2,23 = 4.55; P < .05), day effect (F_3.5,81.2 = 4.91; P < .05), and day by group effect (F_7.06,81.2 = 2.68; P < .05). Bonferroni test yielded statistically significant difference between MnCl₂ and both saline and naïve groups on days 1 and 3 (P < .05) following MnCl₂ infusion.

Figure 4.

(A) Total locomotion and (B) rearing activity over the 30-minute session in the activity cages (mean ± SEM) after infusion. Note that MnCl₂ induced a transient reduction in motor activity in measures of gross motor behavior. Asterisks denote significant group differences (P < .05).

Cylinder

MnCl₂ significantly reduced vertical activity (F_2,23 = 21.23; P < .001) compared with naïve (P < .001) and saline (P < .001) groups only at 24 hours after infusion. This difference stemmed from the reduction in the number of rears performed without support against the cylinder wall (group effect: F_2,23 = 12.25; P < .001) compared both with naïve (P < .001) and saline groups (P < .001) (Figure 5) and the number of rears with the contact of the wall of the cylinder by the MnCl₂ group compared with naïve (P = .013) group.

Figure 5.

Cylinder task: the total number of unsupported rears over the 5-minute session, 24 hours after infusion. The asterisk denotes a significant difference (P < .05) between groups.

MRI measures

R1 maps and 3D MRI

As was expected based on other literature,29 the moderate dose of MnCl₂ significantly (P < .05) enhanced relaxation rates (1/T₁ or R1) throughout the brain compared with the saline group 24 hours after infusion. Significant increases in relaxation rates were found for up to 5 days in the cortex and up to 3 days in subcortical portions of the limbic system, brain stem, cerebellum, and pituitary gland (P < .05). The calculated R1 relaxation rate has a linear relationship with manganese concentration. Thus, the difference between a MnCl₂ infused and non-MnCl₂–infused R1 map will be directly proportional to Mn concentration and account for the baseline tissue R1. Assuming that the regional relaxivity is unchanged, the Mn²⁺ distribution map will be the result. Figure 6A shows a representative image from the sagittal R1 map of one of the manganese infused rats. Each pixel in the image was fit to extract an R1 map for each rat. The maps were then registered and averaged. Figure 6B displays the difference between the average R1 map for the MnCl₂ group and the average R1 map for the saline group, ie, the “Mn” map. Figures 6C and 6D show a representative coronal map and the averaged coronal “Mn” map. The water samples are seen in Figure 6C but have been cropped in Figure 6D as they were not in identical location between different rats and thus could not be registered. All regions of the brain had significantly (P < .05) higher values of R1. As was expected, the greatest change in R1, and thus the highest concentration of Mn²⁺, was found in the pituitary gland. This location is an entry route to the brain as the pituitary gland lacks a blood-brain barrier (BBB). Manganese was distributed to the other parts of the brain including the motor system. A more detailed distribution was seen in the 3D MRI images. Figure 7 displays 3 orthogonal slices from the group-averaged difference between the 3D images of the MnCl₂ and saline groups. The cerebellum and the olfactory bulbs are not visible as they were not excited by the 12-mm slab excitation. The slab excitation was used to prevent folding in the third dimension while reducing the number of phase-encoding steps and time necessary to achieve isotropic voxels. Each 3D image has been registered. To account for differences in session-to-session amplification and in the slice profile, the intensity was normalized to the intensity of the water sample. The images from each group were then averaged. The selected slices in Figure 7 are from the difference between the average MnCl₂ group and the average saline group. Because the signal intensity in the 3D images is weighted by R1 and not a measurement of R1, the images in Figure 7 are proportional to the “Mn” maps shown in Figure 6 but not the same. The pituitary in all the images was the brightest as it has the highest relaxation rate (shortest T₁). The pituitary gland is not directly involved in motor control; however, it is a direct indicator of the total brain dose available to affect the motor system. Thus, in the correlative analysis with respect to behavioral measures, the R1 relaxation rate from the pituitary gland has been used.

Figure 6.

Representative R1 maps of 1 rat brain 24 hours after infusion. (A) sagittal plane, (C) coronal plane. “Mn distribution” map from the group-averaged difference between MnCl₂-injected and saline-injected rats. (B) sagittal plane, (D) coronal plane. Note the highest signal in the pituitary gland and the crescent-shaped intensity that follows the ventricles in A and B. Map D only shows the coronal brain region as the water samples in C varied in location from animal to animal and did not coregister. The horizontal color bar at the top indicates the colors that correspond to the R1 values with the maximum and minimum values indicated. Note that the gray level has been windowed to best reveal the R1 pattern, and the range of values extends from the black to the black line on the white end of the color bar. The white line on each image indicates 1 cm in distance.

Figure 7.

Three orthogonal views from the 3D group-averaged difference. The crosshairs in the coronal (top left) view indicate the locations of the sagittal (right) and horizontal (bottom) slices. The line in the horizontal slice indicates the location of the coronal image. The signal in the water samples (cropped) has been used to normalize the signal of each rat prior to averaging and subtracting (Mn-Sal) the group. Higher signal intensities are proportional to locations of higher Mn concentration or higher relaxivity. Note that only a section of the sagittal direction is excited by the thick slab select pulse, and thus, only part of the brain is visible. The bright region is the pituitary gland.

Correlative analysis: R1 and behavioral measures

To determine whether MnCl₂ affected the behavioral measures, data from either the MnCl₂ group or combined saline and MnCl₂ groups were subjected to bivariate correlation. The correlation was computed between R1 relaxation rate as an independent measure and the variables assessing skilled and gross motor behavior. The MnCl₂ effect on motor behavior was transient. Therefore, the correlative analysis between various behavioral measures and R1 was performed only 24 hours after MnCl₂ infusion. To account for individual differences in preinfusion performance, the correlation analysis was also performed on the index for total success (success_post/success_pre). The statistical results for the end-point measures in a single pellet reaching task are summarized in Table 3.

Table 3.

Correlation between end-point measures and Mn relaxation rate in a single pellet reaching task.

Higher R1 correlated with the reduction in the success level following MnCl₂ infusion (Figure 8A). Most importantly, the correlation between R1 and the index for total success was also significant when only MnCl₂ group was subjected to bivariate analysis (r = −0.52; P < .05). That is, the rats after MnCl₂ infusion were affected in skilled reaching. Moreover, the reduction in the success level was the result of higher MnCl₂ level entering the brain (Table 3).

Figure 8.

(A) The significant correlation (P < .05) between the relaxation rate (R1) in pituitary gland and success index and (B) significant correlation (P < .05) between the relaxation rate (R1) in pituitary gland and supination II index 24 hours after infusion. Success and supination II scores were normalized relative to the preinfusion performance (performance index).

MnCl₂ infusion worsened the reaching movement elements of advance, supination II, and release (Table 4). Figure 8B shows the results between R1 and the index of supination II. More importantly, there was a significant deficit in pronation along increase in R1 (r = 0.77; P < .01) for the MnCl₂ group alone (Table 4), ie, decreasing pronation performance with increasing manganese concentration.

Table 4.

Correlation between behavioral measures and Mn relaxation rate.

Discussion

Magnetic resonance imaging using MnCl₂ as a contrast agent has been used as a tool for functional imaging of behavior in healthy rodents and rodents following brain injury. Although manganese is known to produce motor abnormalities at high doses, its use as an image enhancement tool for studying neuronal circuits continues to grow. With lower doses, it has been reported that meaningful image enhancement is achievable free of detectable running wheel34 and swimming6 motor abnormalities. This study investigated the effect of a single intravenous infusion of MnCl₂ on behavioral tests of fine motor paw function, balancing in a cylinder task, and open field test of general activity. The general expectation from the MEMRI literature is up to 24 hours of lethargy.29 At 95 mg/kg, this was most certainly the case with the rats not recovering. At the moderate dose of 48 mg/kg, the rats could not be classified as lethargic. Nevertheless, the activity and balance test results showed small but detectable abnormalities that were present more than 24 hours and dissipated more than 3 days. The fact that MnCl₂-induced deficits were transient is consistent with previous findings of a transient Mn²⁺-induced learning deficit in rodents.23

The transient nature of Mn²⁺-induced toxicity is a critical point with respect to longitudinal studies combining MEMRI with investigations on freely behaving animals. Our results demonstrate that careful analysis of behavior is required in assessing impairment. Although end-point measures were spared, more detailed measures on movement elements revealed impairment. Notably, the rats were unable to make rotatory movements at wrist and shoulder when they withdrew the paw with a food pellet to the mouth for eating. They used body movements to compensate for abnormalities, and the compensation allowed them to be as successful as control rats. The rats were also impaired in placing the food item into the mouth and replacing the paw on the substrate after releasing the food. Thus, the results show that after a 48 mg/kg dose of MnCl₂, motor skills are detectably altered up the third day after infusion. For most behavioral studies in which the objective is to image brain changes associated with behavior, this is a large time frame. The results are in agreement with the work by Eschenko et al,4 who detected abnormalities at 80 and 16 mg/kg of MnCl₂. Reports of experiments free of detectable abnormalities34 used slow-release osmotic pumps. Osmotic pumps enable direct brain infusion as an option. This enables a high dose delivery that may reverse the MRI contrast55; however, histology follow-up is necessary to distinguish contrast reversal from dead tissue due to the direct infusion.56 MEMRI dose is a balance between image enhancement and toxicity. Although osmotic pumps may have an advantage from a safe delivery point of view, their size puts a limit on the osmolarity-matched dose that can be delivered.57

Signal enhancement 24 hours after MnCl₂ infusion in brain regions has been suggested to be the result of transport of Mn²⁺ via cerebrospinal fluid (CSF). The choroid plexus8 has been shown as a predominant route of Mn²⁺ transport at high plasma concentration of Mn²⁺.58 The regional distribution of Mn²⁺ in the brain after chronic administration is reported to be quite different with the highest regional concentration observed in the striatum.29,30 In this study, the midline sagittal slice of Figure 6 shows a crescent-shaped region of brightness curving up from just in front of the pituitary gland. The slice thickness of this image is 1.6 mm. By looking at the rat atlas53 and the slices that would cover this region around the midline, this intensity follows the shape of the third ventricle. The slices from the 3D MRI images in Figure 7 show that the manganese-enhanced intensity extends into the hippocampal regions but is not as intense in the striatum. The pattern of enhancement in the images suggests that transport is from the blood supply to the pituitary gland, into the CSF, and then to the regions near the ventricles, ie, the hippocampus. Recently, accumulation of Mn²⁺ was reported in other subcortical structures such as hypothalamus and thalamus where the BBB is weak.59,60

The high MRI signal in the pituitary gland has been previously reported.29,46 The imaging data and previous literature support the hypotheses of the pituitary gland being an integral part of the circulation of Mn²⁺to the brain. Although the pituitary gland is not directly involved in motor control, the R1 measurements from the pituitary gland are indicators of the total brain dose available to affect the motor system. In addition, this measure accounts for any differences in blood volume or experimental errors in the infusion procedure and is an easy feature to describe from in a single brain slice. The most interesting findings of this work are the correlations between R1 in the pituitary gland and multiple measures of fine motor skill performance. For some of the performance measures, a higher value is an indicator of better performance, whereas in others lower values are better. In all cases, the sign of the correlation with R1 was such that higher R1 (more Mn) resulted in poorer performance.

The induced abnormalities in skilled and gross motor behavior are likely attributable to Mn²⁺ accumulations in brain regions contributing to motor behavior. The MRI signal was enhanced in all regions including the cortex, subcortical structures, and cerebellum. All of these brain regions are involved in the motor behavior assessed in this study.61,62 For example, the cerebellum63 and dopamine-rich brain regions of the striatum64 are sites of the accumulation of Mn²⁺, and these structures contribute to skilled reaching, locomotion, and balancing ability. The very specific nature of the motor impairments leads us to believe that they are due to the effects of the Mn²⁺ on the brain.

The 100% mortality rate in the group of rats who received intravenous 95 mg/kg MnCl₂ dose is contrary to the literature. One possible reason for the mortality could be the strain of the rats used (Long Evans).6 The rats used in this study are larger than those generally used in the imaging MEMRI literature and thus requiring longer total infusion times, making the infusion process more challenging. Our observation suggests that tail vein infusion of MnCL₂ in adult rats is a complicated procedure and alternative methods should be employed except in special circumstances. Other complicating factors could be rat gender and route of infusions or the use of Bicine as a buffer. Further investigation is needed to determine factors contributing to mortality.

In conclusion, this study reveals that a single 48 mg/kg MnCl₂ dose produces behavioral changes that last up to 3 days. In assessing possible MnCl₂ toxicity, it is advisable to use multiple measures of behavior and to use assessments of the way that movements are performed in addition to end-point measures. In combined studies of behavior and MEMRI with acute single doses, deficits are likely to occur, and carefully controlled experiments to avoid behavioral confounds are necessary. Achieving adequately enhanced MEMRI images while having rats free of detectable deficits likely requires single systemic doses lower than 48 mg/kg of MnCl₂ or an alternate delivery strategy.