Deficits in acetylcholine homeostasis, receptors and behaviors in choline transporter heterozygous mice


R. D. Blakely, PhD, Suite 7140, MRB III, Center for Molecular Neuroscience, Vanderbilt School of Medicine, Nashville, TN 37242-8548, USA. E-mail:


Cholinergic neurons elaborate a hemicholinium-3 (HC-3) sensitive choline transporter (CHT) that mediates presynaptic, high-affinity choline uptake (HACU) in support of acetylcholine (ACh) synthesis and release. Homozygous deletion of CHT (−/−) is lethal shortly after birth (Ferguson et al. 2004), consistent with CHT as an essential component of cholinergic signaling, but precluding functional analyses of CHT contributions in adult animals. In contrast, CHT+/− mice are viable, fertile and display normal levels of synaptosomal HACU, yet demonstrate reduced CHT protein and increased sensitivity to HC-3, suggestive of underlying cholinergic hypofunction. We find that CHT+/− mice are equivalent to CHT+/+ siblings on measures of motor co-ordination (rotarod), general activity (open field), anxiety (elevated plus maze, light/dark paradigms) and spatial learning and memory (Morris water maze). However, CHT+/− mice display impaired performance as a result of physical challenge in the treadmill paradigm, as well as reduced sensitivity to challenge with the muscarinic receptor antagonist scopolamine in the open field paradigm. These behavioral alterations are accompanied by significantly reduced brain ACh levels, elevated choline levels and brain region-specific decreased expression of M1 and M2 muscarinic acetylcholine receptors. Our studies suggest that CHT hemizygosity results in adequate baseline ACh stores, sufficient to sustain many phenotypes, but normal sensitivities to physical and/or pharmacological challenge require full cholinergic signaling capacity.

Acetylcholine (ACh) is a neurotransmitter of key importance in both central and peripheral nervous systems. Cholinergic neurotransmission maintains vital biological functions such as movement, autonomic regulation and cognitive processes, including attention and learning and memory (Kaneko et al. 2000; Misgeld et al. 2002; Perry et al. 1999; Sarter & Bruno 1997; Winkler et al. 1995), through both nicotinic (Dani 2001) and muscarinic (Wess 2004) receptors. Presynaptic CHT activity is believed to be the rate-limiting step for ACh synthesis and release, particularly under conditions of elevated cholinergic tone (Birks & MacIntosh 1961; Jope 1979; Simon & Kuhar 1975). Consistent with this idea, in vitro inhibition of CHT by the competitive CHT antagonist hemicholinium-3 (HC-3) impairs ACh synthesis and release (Guyenet et al. 1973; Maire & Wurtman 1985; Murrin et al. 1977), whereas in vivo pharmacological challenge with sublethal doses of HC-3 impairs learning and memory as measured in the Morris water maze and passive avoidance tasks (Boccia et al. 2004; Franklin et al. 1986; Hagan et al. 1989). Complete pharmacological blockade of CHT (Schueler 1955) or homozygous deletion of the CHT gene in mice (Ferguson et al. 2004) results in death, associated with reduced ACh release at central and peripheral cholinergic synapses that support respiration, including the neuromuscular junction (NMJ).

We identified the murine CHT (Apparsundaram et al. 2001) with the express purpose of establishing novel in vivo models of cholinergic hypofunction derived from genetically determined changes in CHT protein levels. As indicated above, CHT−/− newborn mice are apneic and hypoxic, largely immobile and die within an hour after birth (Ferguson et al. 2004). In contrast, CHT+/− mice survive, reproduce at wild-type rates and are grossly indistinguishable from CHT+/+ mice. Importantly, CHT+/− mice exhibit levels of forebrain synaptosomal high-affinity choline uptake (HACU) equivalent to that of CHT+/+ littermates, despite a 50% reduction in CHT protein levels (Ferguson et al. 2004). These latter findings are consistent with the existence of efficient, post-translational regulation of HACU (Apparsundaram et al. 2005; Ferguson & Blakely 2004; Ferguson et al. 2003). However, CHT+/− mice appear more sensitive to HC-3 (Ferguson et al. 2004), suggesting the existence of adaptive changes engaged to sustain normal behavior that can be exposed by direct manipulations of cholinergic signaling.

In the experiments described below, we sought to elucidate basal behavioral and biochemical phenotypes, and to explore whether altered sensitivity to physical and pharmacological challenges exists in CHT+/− mice. We establish that whereas CHT hemizygosity does not result in overt behavioral phenotypes, reliance on a single CHT gene leads to significantly reduced bulk ACh brain levels and renders animals vulnerable to behavioral and pharmacological challenges of cholinergically mediated behaviors. Moreover, we identify changes in M1 and M2 muscarinic receptor levels that we hypothesize alter the dynamic range of cholinergic signaling. We discuss our findings in relation to mouse genetic models that induce alterations of other components of cholinergic signaling, as well as loss-of-function CHT alleles evident in humans.



Scopolamine hydrochloride (S-1013) and oxotremorine sesquifumarate salt (O-9126) were obtained from Sigma-Aldrich (St. Louis, MO, USA) and dissolved in sterile saline (0.9% NaCl). Both drugs were injected intraperitoneally at a volume of 10 ml/kg.


All animal procedures were approved by the Vanderbilt University Institutional Animal Care and Use Committee. Mice were housed up to five per cage on a 12-h light/dark cycle (lights on at 0600 h), and behavioral testing was performed during the light part of the cycle. Food (Purina Rodent Chow #5001) and water were provided ad libitum. All mice were back-crossed at least seven generations to the C57BL/6 background except in the rotarod, elevated plus maze and scopolamine-induced locomotion experiments, where CHT+/− mice back-crossed for two generations were used. In all cases, CHT+/+ littermates were used as controls. All behavioral tests were carried out in the Murine Neurobehavioral Laboratory of the Center for Molecular Neuroscience, managed by Dr John D. Allison. Mice naive to behavioral testing were used in each behavioral task. Mice were acclimated to the testing location at least 12 h before the start of behavioral testing.

Irwin screen

At 12 weeks of age, seven mice from each genotype were subjected to a modified Irwin screen (Irwin 1968), to assess basic sensory-motor performance. Specifically, the following factors were evaluated on a numerical scale of 0–3: gross appearance (presence of whiskers, fur, wounds and body weight); behavior in a novel environment (mice were removed from their home cage, placed in a new, clean cage devoid of bedding and evaluated for spontaneous activity, tremor, piloerection, tail elevation, urination and defecation) and reflexes (touch escape, reaching reflex and pinna reflex).

Exploratory locomotor activity

Exploratory locomotor activity was evaluated in 9- to 12-week-old mice (CHT+/−, n= 18; CHT+/+, n= 19) using commercially available activity monitors measuring 27.9 × 27.9 cm (MED Associates, Georgia, VT). Each apparatus contained 16 photocells in each horizontal direction, as well as 16 photocells elevated 4.0 cm to measure rearing. Illumination in the room consisted of standard fluorescent light, and intensity was measured at 630 lumen/m2 (LUX). Mice were placed in the monitors for 30 min and allowed to explore freely. Horizontal beam breaks and rearing were automatically recorded and represented as distance traveled and vertical counts, respectively. Data were analyzed in 5-min time bins.


Motor co-ordination and balance were measured in 3-month-old mice (CHT+/−, n= 7; CHT+/+, n= 5), using a commercially available accelerating rotarod apparatus (Model 7650, Ugo Basile, Napoli, Italy). Mice were placed on the rotating cylinder (3 cm in diameter) and confined to a section approximately 6.0 cm long by gray plastic dividers. The rotational speed of the cylinder was increased from 5 to 40 r.p.m. over a 5-min period. Latency at which mice fell off the rotating cylinder was measured. Each mouse was given three trials per day over a period of 3 days.

Elevated plus maze

Anxiety was evaluated in 2- to 3-month-old mice (n= 5 per genotype), using the elevated plus maze. The homemade plus maze consisted of four arms, approximately 10 × 30 cm, connected in a plus configuration and elevated approximately 40 cm above the floor. Two of the arms had walls, approximately 15 cm high (‘closed arms’), and two walls had no arms (‘open arms’). Illumination in the room consisted of standard fluorescent light, and intensity was measured at 250 LUX. Testing at lower light intensities was also performed, but the results were not different. Mice were placed gently in the center of the four arms at the beginning of the session. The number of entries onto the open arms, and amount of time spent in open arms were recorded.

Light/dark exploration

Anxiety responses were also assessed in 2- to 3-month-old mice (n= 8 per genotype), using the activity monitors (MED Associates, Georgia, VT, USA). Half of the chamber was made opaque with a black acrylic insert, while the other half remained transparent. Photocells recorded the movement of the mice between compartments. Illumination in the room consisted of standard fluorescent light, and intensity was measured at 630 LUX. Mice were placed individually into the dark compartment at the beginning of the session. Total time spent in the dark vs. light compartments were recorded. Each mouse was given one 5-min session.


Eight- to 10-week-old mice (n= 11 per genotype) were run on a motorized, six-lane treadmill (Columbus Instruments, Columbus, OH, USA) equipped with an adjustable-speed belt (6–100 m/min) and an electric shock grid at one end. On Day 1 (training) the mice were exposed to the treadmill without running or shock for 10 min. Mice were then exposed to gradually increasing speeds (10 m/min maximum) for 15–30 min in the presence of electric shock (2 mA, 4 min−1 frequency) activated by physical contact with the shock grid. This trial was used purely for training purposes and to familiarize the mice with the apparatus. On Day 2 (gradual increase in speed) the mice were started at 10 m/min and the speed was increased by 1.5 m/min every 2 min. The maximal speed achieved by each mouse and the maximal time a mouse could keep running were recorded and interpreted as measures of motor fatigue. Exhaustion was defined as resting on the electric grid for more than 15 seconds/min or falling back onto the grid more than 15 times/min. On Day 3 (constant speed) the initial running speed was 7 m/min and was increased every 30 seconds by 2 m/min until a speed of 17 m/min was reached after 2.5 min. At this point the speed was maintained at 17 m/min and the mice were run until exhaustion. To exclude the possibility that the phenotype was due to learning difficulties, one group of mice was tested in the same paradigm on two additional consecutive days, and similar results were obtained (data not shown). As a control for pain sensitivity, we assessed pain thresholds in a separate group of experimentally naive littermates (n= 7 per genotype). For this analysis, we exposed individual mice to shock of increasing current amplitude, with rest periods between shocks of varying amplitudes, and recorded the current amplitudes at which the mouse first jumped and/or vocalized, as a readout of sensitivity to shock.

Scopolamine-induced locomotor activity

Mice were acclimated to the activity monitors (MED Associates, Georgia, VT, USA) for 1 h, removed and injected intraperitoneally (i.p.) with vehicle or scopolamine and returned to the activity monitors for 2 h for activity measurements. Naive mice were used for each drug concentration: n= 8, n= 12, and n= 5 mice from each genotype were used for the 1, 0.5 and 0.1 mg/kg doses, respectively.

Oxotremorine-induced tremor

Mice were taken from their home cages and placed in new, individual (one mouse/cage) transparent cages devoid of bedding and allowed to acclimate for 15 min. Each mouse was injected i.p. with vehicle or oxotremorine and returned to the cage for observation. Naive mice were used for each drug concentration, and the observer was blind to the genotypes: n= 5 mice from each genotype were used for the 0.125 mg/kg and 0.031 mg/kg doses, and n= 17 mice from each genotype were used for the 0.062 mg/kg dose. Tremor was scored every 5 min on a scale of 0–2 (0, no tremor; 1, head and body tremor; 2, continuous whole body tremor). The data were expressed as per cent of the maximal possible tremor (for example a score of 1 was represented as 1 (actual score)/2 (maximum possible score)*100 = %tremor) (Li et al. 2003).

Morris water maze

Spatial learning and memory were evaluated in CHT+/− (n= 11) and CHT+/+ (n= 12) mice using the hidden-platform water maze paradigm. The water maze pool was 92 cm in diameter and filled with water made opaque with nontoxic tempura paint. An acrylic platform (10 × 10 cm) was submerged approximately 0.5 cm below the surface of the water to allow relief from swimming. Mice were given six spaced trials per day. The hidden platform remained in a fixed location throughout the experiment, but starting positions were different on each trial. Rest intervals between the spaced trials were approximately 30 min long. Each trial continued for a maximum of 60 seconds or until the mouse reached the platform, whichever occurred first. If a mouse did not reach the platform in 60 seconds, it was captured by the experimenter and placed gently on the platform for 15 seconds. Sessions were captured by an overhead camera and analyzed in real time using an NIH Image macro on a Macintosh computer. Parameters measured included latency (time to reach platform), swimming velocity and path length. Data were analyzed immediately following each set of trials, and once each group of mice reached the criterion of 8 seconds average latency, they were given a probe test 2 h after their last training session. Both genotypes reached criterion on the same day, and therefore all mice received the same number of trials. In the probe trial, the platform was removed, and mice were allowed 60 seconds to explore the water maze. The amount of time spent in each quadrant of the pool and platform location crossings were measured for the 60-second probe trial. A second probe trial was performed 48 h after the initial probe trial. Following this, mice were re-trained using a new platform location. Platform location was constant across days, and each mouse was run for six spaced trials (six different starting locations) each day until target latency to reach the platform (8 seconds average for each genotype) was reached. Rest intervals between the spaced trials were approximately 30 min long. Probe trials were run 2 and 48 h after the last reverse platform training trial.

Acetylcholine levels, choline acetyltransferase (ChAT) and acetylcholinesterase (AChE) enzyme activities

Acetylcholine levels in brain tissue were quantified by previously described high-performance liquid chromatography electrochemical detection methods (CHT+/−, n= 7; CHT+/+, n= 7) (Vanderbilt Neurochemistry Core Resource) (Damsma et al. 1985). Briefly, animals were rapidly decapitated and brain tissue was dissected on ice following microwaving to inactivate AChE (Bertrand et al. 1994), followed by homogenization in acetonitrile, lipid removal with heptane and vacuum drying. Activity of ChAT in freshly dissected brain tissue was measured by the synthesis of [14C]ACh from [acetyl-1-14C]acetyl-CoA and choline (CHT+/−, n= 4; CHT+/+, n= 4) (Frick et al. 2002). AChE activity was determined by colorimetric measurements following the method of Ellman (CHT+/−, n= 5; CHT+/+, n= 5) (Ellman et al. 1961).

Immunoblot analysis of muscarinic acetylcholine receptor (mAChR) expression

Freshly dissected brain tissue (n= 4 mice per genotype) was frozen in tubes on dry ice and stored at −80°C. As needed, tubes were removed from −80°C and placed on ice. Protease inhibitor (Sigma, St. Louis, MO, USA, #2714) was dissolved in 10 mm Tris and 1 mm EDTA, pH 7.5, and 500 μl were added to each tube. The tissue was homogenized with a Brinkman Polytron PT3000 using a PT-DA 3007/2 tip at 20k rpm for 10 seconds and the tip was rinsed with dH2O between each sample. After determining and adjusting protein concentration, the homogenate was diluted with sample buffer to yield 1% sodium dodecyl sulfate (SDS), 31.25 mm Tris, pH 6.8, 5% glycerol and 200 mm 2-mercaptoethanol. The samples were then resolved by SDS–polyacrylamide gel electrophoresis (PAGE, 10% gel), transferred to polyvinylidine difluoride membranes and subjected to immunoblot analyses with M1 and M2 polyclonal antibodies (Li et al. 2003; Volpicelli-Daley et al. 2003), as well as an EF1a monoclonal antibody as a loading control, with quantitation performed using an Odyssey imager (Li-Cor, Lincoln, NE, USA). Blot densitometry was performed using imagej software (Wayne Rasband, NIH) and the muscarinic receptor band densities were normalized to the EF1a band densities (Volpicelli-Daley et al. 2003).

Statistical analyses

Two-way repeated measures anova and two-tailed, unpaired Student’s t-tests were used to analyze the data. The specific statistical analyses used are noted in the text and legends with respect to individual test design.


CHT+/− mice exhibit wild-type levels of overall locomotor activity, but altered patterns of movement in the open field chamber paradigm

Basal exploratory locomotor activity was evaluated in CHT+/− mice and CHT+/+ siblings in the open field activity monitors. CHT+/− mice consistently traveled shorter distances during their first 30-min exposure to the open field chambers (CHT+/−, 3081 ± 171.8 cm; CHT+/+, 3658 ± 217.7 cm) (Fig. 1A). This difference was statistically significant (F1,175= 4.26; P < 0.05). In preliminary studies, we also observed diminished motor activity upon initial, but not continued, exposure to a running wheel setup in the home cage (data not shown), suggesting a novelty effect may be responsible for the CHT+/− hypolocomotive phenotype. Intriguingly, vertical movement was higher in CHT+/− mice (711.4 ± 79.0 counts) compared to their CHT+/+ siblings (463.5 ± 90.3 counts) (F1,175= 5.04; P < 0.05) (Fig. 1B). The inverse relationship between ambulatory and rearing counts suggested that lower levels of horizontal activity by CHT+/− mice may be due to increased vertical activity, possibly due to increased exploratory drive, rather than impairments in sensory-motor ability. Indeed, when time spent in vertical motion was eliminated from recordings, CHT+/− and CHT+/+ mice showed equivalent distance traveled (Fig. 1C).

Figure 1.

Basal locomotor activity: CHT+/− mice exhibit normal overall activity levels but an increased rearing phenotype in the open field paradigm. CHT+/− mice display reduced horizontal locomotion (a) and increased vertical activity (b) in the novel open field environment compared to CHT+/+ littermates. When horizontal activity is corrected for the additional time spent rearing, activity levels are not significantly different between the genotypes (c). Filled triangles, CHT+/− (n= 18); open squares, CHT+/+ (n= 19). Mean (±SEM) values are indicated.

Basal sensory-motor performance is intact in CHT+/− mice

A modified Irwin screen (Irwin 1968) failed to reveal any difference in basal sensory-motor phenotypes between genotypes (data not shown). Although conducted with a relatively small cohort, results from this screen mirror facets of normal motor and sensory function inherent in basal responses in other tests (see below). Future, more detailed characterizations of sensory and motor function in larger groups of mice may reveal very subtle phenotypes. The mice were next tested in the rotarod paradigm. Both genotypes learned the task at the same rate and demonstrated equivalent performance, as indicated by increases in similar average latencies before falling off the accelerating rod over a period of three consecutive trial days (Day 1 latencies: CHT+/−, 66.4 ± 13.4 seconds vs. CHT+/+, 63.2 ± 22.7 seconds; Day 2 latencies: CHT+/−, 188.6 ± 27.1 seconds vs. CHT+/+, 169.1 ± 39.0 seconds; Day 3 latencies: CHT+/−, 245.6 ± 26.0 seconds vs. CHT+/+, 194.5 ± 35.0 seconds) (Fig. 2A). There was also no significant difference in the amount of time CHT+/+ and CHT+/− littermates were able to hold onto an inverted screen (data not shown). These results, along with performance in open field tests, suggest that a single functional CHT allele is sufficient to support overall motor balance and co-ordination as well as muscle strength.

Figure 2.

CHT+/− basal behavior is normal in sensory-motor and anxiety paradigms. Motor co-ordination, balance and learning are similar between CHT+/− (n= 7) and CHT+/+ (n= 5) littermates in the rotarod paradigm (average latencies of 3 trials/day) (a). Anxiety levels are similar between CHT+/− mice and CHT+/+ littermates as indicated by same amount of time spent in open vs. closed arms in the elevated plus maze (n= 5/genotype) (b), in center vs. periphery in open field/novel environment paradigm (CHT+/−, n= 18; CHT+/+, n= 19) (c), and in light vs. dark section of the light/dark paradigm (CHT+/−, n= 6; CHT+/+, n= 8) (d). Mean (±SEM) values are indicated.

CHT+/− mice do not show elevated anxiety

One possible explanation for the changes in vertical activity observed in the open field paradigm is elevation in anxiety. To test this hypothesis, the distance traveled in the center vs. the periphery of the open field was evaluated in the same open field trials used to quantify locomotor movement. We found no significant genotype difference in the ratio between distance traveled in the center vs. periphery (arbitrary units: CHT+/−, 2.92 ± 0.41; CHT+/+, 2.53 ± 0.34) (Fig. 2C). For an additional evaluation of anxiety, a separate, naive group of mice was given a 5-min trial in the light/dark paradigm. Again, there was no significant difference between genotypes in the amount of time they spent in the light (CHT+/−, 112.8 ± 23.4 seconds; CHT+/+, 99.2 ± 9.0 seconds) vs. the dark (CHT+/−, 186.2 ± 23.4 seconds; CHT+/+, 199.8 ± 9.0 seconds) part of the field and, as expected, both genotypes spent more time in the dark part of the field (Fig. 2D). In a final evaluation of anxiety, mice were tested on the elevated plus maze. Consistent with other anxiety measures noted above, the percent time spent in open arms (CHT+/−, 17.4 ± 5.7%; CHT+/+, 18.5 ± 2.7%) vs. closed arms (CHT+/−, 69.2 ± 5.9%; CHT+/+, 56.1 ± 8.0%) (Fig. 2B) and per cent entries into open (CHT+/−, 25.3 ± 6.7%; CHT+/+, 31.5 ± 7.6%) vs. closed (CHT+/−, 58.7 ± 7.6%; CHT+/+, 60.5 ± 10.8%) arms were not different between the genotypes.

CHT+/− mice display normal performance in the Morris water maze

With findings of normal baseline sensory-motor performance in CHT+/− mice, we next evaluated CHT genotype effects on spatial learning and memory using the Morris water maze (Winkler et al. 1995). CHT+/− mice did not differ from CHT+/+ littermates in any aspect of the task. Both genotypes learned the Morris water maze task equally well as indicated by similar latencies to reach the platform (Fig. 3A) (F1,147= 0.01; P= 0.31) and decreasing latencies over 8 days of training (F7,147= 31.02; P < 0.0001). CHT+/− mice were also unimpaired in acquisition of a novel platform location (F1,60= 0.09; P= 0.91) (Fig. 3B) and had similar swim speeds (data not shown). On the probe trial, both CHT+/− and CHT+/+ mice spent significantly greater periods of time in the target quadrant (CHT+/−, 44.5 ± 1.3% of total time (F3,30= 37.19; P < 0.0001); CHT+/+, 41.3 ± 1.8% of total time (F3,33= 13.59; P < 0.0001)) (Fig. 3C), indicating intact memory for the platform location. Similarly, mice of both genotypes were more likely to swim over the former platform location, compared to equivalent locations in the other three quadrants (CHT+/−, 3.3 ± 0.9 and CHT+/+, 2.2 ± 0.5 number of crossings over former platform location) (Fig. 3D). No differences in any of the above parameters were observed upon repeating the probe trial 48 h after the last training session (data not shown) or upon reversing the platform location and repeating the learning and probe trials. To evaluate the effects of aging on spatial learning, a group of naive 1.5-year-old mice of identical genetic background were tested in the same water maze paradigm using the same test parameters, with results similar to those obtained with younger animals (data not shown).

Figure 3.

CHT+/− basal spatial learning and memory performance is normal in the Morris water maze paradigm. Three-month-old CHT+/− (n= 11) and CHT+/+ (n= 12) mice were run in the hidden-platform version of the Morris water maze. CHT+/− mice did not differ on latency to reach the platform during learning trials with original (a) and reversed (b) platform locations. Both genotypes spent similar amounts of time in the target quadrant on probe trials on the last day of training (c) as well as a delayed probe trial (48 h later, data not shown) and did not differ in number of platform location crossings (d). T, target (platform); O, opposite; R, right; L, left quadrant.

Physical challenge unmasks a locomotor phenotype in CHT+/− mice

The lack of impairment in the above behavioral tasks suggests that cholinergic tone in CHT +/− mice is adequate to support motor function as well as spatial learning and memory assessed under standard measurement conditions. We next sought to introduce graded physical demands that should require a greater extent of cholinergic function and could expose phenotypic deficits. Our hypothesis was that under conditions of sustained physical challenge—such as continuously increasing treadmill speed (gradual challenge) and/or sustained high speed (acute challenge)—motor performance would be interrupted sooner in CHT+/− mice compared to their CHT+/+ littermates. Quantitative predictions were that CHT+/− mice would be unable to reach treadmill speeds as high as those attained by CHT+/+ littermates, and that at a constant treadmill speed, CHT+/− mice would fatigue sooner than CHT+/+ littermates. Both predictions were confirmed experimentally: whereas both genotypes were able to sustain activity on the treadmill initially, the average time spent running in the gradual acceleration version of the task was significantly lower in CHT+/− mice (772 ± 60.6 seconds) compared to CHT+/+ mice (1089 ± 55.8 seconds) (t(12)= 3.85, P < 0.001) (Fig. 4A). In addition, 18% of CHT+/+ mice reached maximal speeds of 25 m/min, while none of the CHT+/− mice exceeded 22 m/min, and only 9% of CHT+/− mice reached 22 m/min vs. 45% of CHT+/+ mice (Fig. 4B).

Figure 4.

Physical challenge unmasks a locomotor phenotype in CHT+/− mice. CHT+/− mice (n= 11) remained on the treadmill for significantly shorter periods of time in both gradual (a,b) and acute (c,d) challenge versions of the treadmill paradigm compared to CHT+/+ littermates (n= 11). In addition, none of the CHT+/− mice reached the maximal speed reached by CHT+/+ littermates (b). Mean (±SEM) values are indicated (*P < 0.01,***P < 0.001).

In the acute challenge paradigm, the average time spent running was again significantly lower in CHT+/− mice (1169 ± 139.5 seconds) as compared to CHT+/+ littermates (1743 ± 146.3 seconds) (t(12)= 2.84, P < 0.01) (Fig. 4C). Whereas both genotypes performed similarly initially (100% of both genotypes remained on the treadmill at 10 min), only 45.4% of CHT+/− mice, compared to 81.8% of CHT+/+ mice remained on the treadmill by 20 min, only 18.2% of CHT+/− mice but 72.7% of CHT+/+ littermates remained by 30 min and 0% of CHT+/− mice vs. 64.6% of CHT+/+ animals remained longer than 31 min (Fig. 4B). It is unlikely that the differences were due to lack of sensory feedback in the mice, as all mice vocalized and/or jumped when exposed to the electric grid, and immediately resumed running either voluntarily or in response to gentle prodding, unless they were completely exhausted. In addition, a separate experimentally naive group of mice was tested for responses to increasing shock amplitudes, and no significant differences in shock sensitivity were evident between genotypes: CHT+/− mice first vocalized at 0.33 ± 0.02 mA of current, while CHT+/+ mice vocalized at 0.35 ± 0.04 mA of current.

Locomotor phenotype in CHT+/− mice is induced by muscarinic challenge

To complement studies of physical stress on cholinergic systems, we investigated pharmacological challenges known to impact central ACh signaling supporting basal ganglia motor circuits. Challenge with muscarinic antagonists has a well-known stimulatory effect on locomotor activity, mediated by increased ACh release (Gerber et al. 2001; Gomeza et al. 1999; Miyakawa et al. 2001; Sipos et al. 1999). We therefore constructed a dose–response profile for CHT+/− mice and CHT+/+ littermates, using open field locomotor behavior as a measure of motor activity. In both genotypes, locomotion was not affected or increased equally at low (0.1 mg/kg) (F1,72= 1.26, P= 0.29) (Fig. 5A) and high (1 mg/kg) (F1,169= 0.26, P= 0.62) (Fig. 5C) doses of scopolamine, respectively. However, at an intermediate dose (0.5 mg/kg), CHT+/− mice were significantly less active than their CHT+/+ littermates (Fig. 5B) (F1,308= 4.94, P < 0.05) (Fig. 5D). The difference in sensitivity to scopolamine was evident in measurements of distance traveled only, as both genotypes displayed equal increases in vertical counts following scopolamine injection (data not shown).

Figure 5.

Locomotor phenotype in CHT+/− mice is induced by muscarinic challenge. In both genotypes, locomotion was not affected or increased equally at low (0.1 mg/kg) (a) and high (1 mg/kg) (c) doses of scopolamine, respectively. However, at an intermediate dose (0.5 mg/kg) (b), CHT+/− (n= 12) mice were significantly less active than CHT+/+ (n= 12) littermates, showing a decreased sensitivity to the muscarinic antagonist scopolamine (d). Arrow = point of drug administration (*P < 0.05).

In contrast to their hyposensitivity to scopolamine challenge, CHT+/− mice displayed a dose-dependent increased sensitivity to the muscarinic agonist oxotremorine. Similar to the effects of scopolamine, oxotremorine induced dose-dependent increases in tremor response of CHT+/− compared to CHT+/+ littermates. There was no difference between the genotypes at low (0.031 mg/kg) and high (0.125 mg/kg) doses of oxotremorine. However, 15 min after injection of an intermediate dose (0.062 mg/kg), CHT+/− mice displayed a higher degree of tremor (29.4 ± 6.1%) compared to CHT+/+ mice (11.8 ± 5.3%) (t(31)= 2.17, P < 0.05 by nonparametric unpaired t-test). This hypersensitivity trend persisted as long as 25 min post-injection (CHT+/−, 41.2 ± 4.8%; CHT+/+, 20.6 ± 6.1%; t(30)= 2.65, P < 0.01 by nonparametric unpaired t-test).

Acetlycholine levels are significantly diminished and muscarinic receptor expression selectively altered in CHT+/− brain

The altered sensitivity to muscarinic agents observed in CHT+/− mice could be due to either genotype-dependent alterations in ACh levels, as dictated by compensatory changes in biosynthetic/catabolic pathways, and/or changes in muscarinic receptor expression. We examined the former possibility by evaluating tissue ACh and choline levels, as well as ChAT and AChE activities in various brain regions. Tissue ACh levels (Fig. 6A) were significantly lower in CHT+/− cortex (CHT+/−, 6.9 ± 1.5; CHT+/+, 15.8 ± 2.3 nmol/g wet tissue), hippocampus (CHT+/−, 9.0 ± 2.1; CHT+/+, 20.6 ± 2.2 nmol/g) and striatum (CHT+/−, 24.1 ± 5.2; CHT+/+, 61.7 ± 8.0 nmol/g) (F1,3931= 16.28, P < 0.001). Conversely, tissue choline levels (Fig. 6B) were significantly elevated in CHT+/− cortex (CHT+/−, 53.0 ± 6.2; CHT+/+, 31.0 ± 2.6 nmol/g wet tissue), hippocampus (CHT+/−, 37.7 ± 5.0; CHT+/+, 27.6 ± 1.4 nmol/g) and striatum (CHT+/−, 66.9 ± 5.0; CHT+/+, 46.9 ± 4.9 nmol/g) (F1,3151= 10.74, P < 0.01). In contrast, there were no significant differences in ChAT activity (cortex, CHT+/−, 52.0 ± 1.9; CHT+/+, 51.3 ± 8.3; hippocampus, CHT+/−, 60.2 ± 4.6; CHT+/+, 61.8 ± 4.4; striatum, CHT+/−, 123.8 ± 27.8; CHT+/+, 130.2 ± 9.9 nmol/h/mg protein) (F1,470= 0.05, P= 0.83) or AChE activity (cortex, CHT+/−, 51.8 ± 4.3; CHT+/+, 51.5 ± 4.3; hippocampus, CHT+/−, 54.7 ± 13.9; CHT+/+, 44.8 ± 7.4; striatum, CHT+/−, 288.0 ± 51.0; CHT+/+, 288.2 ± 16.2 nmol/h/mg protein) (F1,126= 0.13, P= 0.72).

Figure 6.

Tissue acetylcholine (ACh) levels are lower in CHT+/− brain. (a) Tissue ACh levels measured by high-performance liquid chromatography are significantly lower in CHT+/− cortex, hippocampus and striatum, whereas choline levels (b) are elevated in the same regions (CHT+/−, n= 7; CHT+/+, n= 7) (*P < 0.05, **P < 0.01).

To test whether alterations were evident in central muscarinic receptors, we performed Western blotting for the M1 and M2 receptor on brain homogenates from CHT+/− and CHT+/+ littermates, using densitometry to quantify receptor expression in cortical, hippocampal and striatal extracts. We focused on these receptors as M2 receptors represent the predominant presynaptic subtype in cortex and hippocampus, but are also expressed in striatum (Rouse et al. 1997), and M1 receptors are widely distributed and participate in both motor and cognitive circuits in the cortex, hippocampus and striatum (Alcantara et al. 2001; Levey et al. 1991). We found M1 expression to be 25.4% lower in the striatum, 31.3% lower in the hippocampus and 20.8% higher in the cortex of CHT+/− mice compared to CHT+/+ littermates (Fig. 7). This difference reached statistical significance in the striatum (t(6)= 3.05; P < 0.05). M2 receptor expression was 47.7% lower in the striatum, 34.9% lower in the cortex and essentially unchanged in the hippocampi of CHT+/− mice compared to CHT+/+ littermates (Fig. 7). The decreases in M2 expression reached statistical significance in both striatum (t(6)= 2.81; P < 0.05) and cortex (t(6)= 2.91; P < 0.05).

Figure 7.

M1 and M2 muscarinic receptor expression is altered in CHT+/− brain. Immunoblotting shows that M1 receptor expression is significantly downregulated in CHT+/− striatum, whereas M2 receptor expression is decreased in both CHT+/− cortex and striatum. Mean (±SEM) values are indicated (n= 4 mice per genotype) (*P < 0.05).


Mouse models featuring genetically established cholinergic deficits have provided new evidence for a dynamic role of CHT as a critical modulator of cholinergic neurotransmission (Bazalakova & Blakely 2006). For example, CHT protein expression is increased in response to decreased catabolism of ACh in AChE knockout mice (Volpicelli-Daley et al. 2003). Conversely, CHT density and function are upregulated in hypocholinergic states, produced by transgenic overexpression of AChE (Beeri et al. 1997). Finally, upregulation of CHT protein and function sustains wild-type levels of AChE activity, as well as tissue ACh content and depolarization-evoked ACh release, in ChAT heterozygote mice (Brandon et al. 2004). In this context, we hypothesized that partial loss of CHT might represent a unique model of cholinergic hypofunction, one where phenotypes would emerge upon behavioral/pharmacological challenges targeting cholinergic signaling. In support of this idea, we previously reported an increase in the sensitivity to the lethal effects of HC-3 in CHT+/− mice compared to wild-type littermates (Ferguson et al. 2004), despite wild-type levels of synaptosomal HACU. Our current studies show that CHT+/− mice achieve normal performance on multiple sensorimotor, anxiety and cognitive tasks. However, physical and/or pharmacological challenges elicit genotype-sensitive phenotypes, consistent with demand-dependent cholinergic hypofunction supported by diminished tissue ACh levels and altered muscarinic receptor expression.

Cholinergic neurotransmission at the NMJ is essential for movement (Brandon et al. 2003; Misgeld et al. 2002) and ACh modulates extrapyramidal motor function (Borison & Diamond 1987; Kaneko et al. 2000). Therefore, we began by characterizing motor function in CHT+/− mice, using the open field chamber paradigm. Whereas overall levels of activity were not different between the genotypes (Fig. 1C), we did observe an interesting pattern of decreased horizontal (Fig. 1A) but increased vertical movement (Fig. 1B) by CHT+/− mice. This phenotype did not appear to be caused by overt musculoskeletal deficits, as CHT+/− mice travel equivalent distances in the Morris water maze and swim at wild-type velocities (data not shown). Additionally, they perform comparably to CHT+/+ mice at slow speeds and short test intervals on the treadmill (Fig. 4) and are not impaired in the rotarod (Fig. 2A), wire hang and inverted screen tests (data not shown). Normal basal performance by CHT+/− mice in the elevated plus maze (Fig. 2B) and light/dark (Fig. 2D) paradigms likely rules out overt anxiety as an alternative reason for the altered locomotor phenotype. The increased rearing, but preserved overall activity levels, possibly represent an altered cognitive response to a novel environment (Sarter & Bruno 1997). Additional studies are needed to explore this possibility and to discern whether modified cholinergic autonomic activity contributes to this phenotype.

The baseline locomotor data described above not only provide controls for studies of stressed motor function but also allowed us to proceed with cognitive studies that rely on normal motor capacities. Similar to motor activity, cognitive performance in CHT+/− mice appeared grossly normal at baseline in multiple behavioral assays, including spontaneous Y maze and rewarded T maze alternation, conditioned freezing, response acquisition and prepulse inhibition (see Supplementary Material). CHT+/− mice were also unimpaired in the Morris water maze—a hippocampus-dependent spatial learning and memory task that investigators have shown can be sensitive to insults on the cholinergic system, including pharmacological challenges with muscarinic antagonists, aging and anatomical lesions of cholinergic pathways (D’Hooge & De Deyn 2001). However, Morris water maze performance is unaffected in other models of cholinergic dysfunction, including M1 knockouts (Miyakawa et al. 2001) and ChAT heterozygous (ChAT+/−) mice (Brandon et al. 2004). Interestingly, the normal water maze performance of ChAT+/− mice has been attributed to wild-type levels of cholinergic neurotransmission capacity, triggered in part by upregulation of CHT (Brandon et al. 2004). In this context, additional tests, such as those developed in the rat to monitor focused attention (Apparsundaram et al. 2005; Sarter & Bruno 1997), may be useful in assessing CHT contributions to cognitive performance.

Having established normal baseline motor and cognitive performance, we asked whether CHT hemizygosity would render CHT+/− mice susceptible to challenges directed at cholinergic neurotransmission. We chose motor performance as our readout, as we could introduce both physical and pharmacological challenges in multiple paradigms. In treadmill tests of physical strain on locomotion, we found that CHT+/− mice remained on the running belt for significantly shorter periods of time (Fig. 4A,C) and were not able to reach and sustain the higher speeds attained by age, gender and weight-matched CHT+/+ littermates (Fig. 4B). Our previous electrophysiological recordings at the NMJ documented normal baseline ACh stores and release in CHT−/− mice, as measured by excitory postsynaptic potential (EPSP) magnitude (Ferguson et al. 2004). However, over time, the ACh release diminishes to a point where EPSPs as well as miniature end plate potentials (mEPPs) are no longer observable in CHT−/− preparations, whereas CHT+/+ NMJs maintain cholinergic neurotransmission. Although we did not measure neuromuscular signaling per se in these studies, previous studies have shown that reduction in mEPP amplitude, reflecting diminished neurotransmitter release, is significantly greater if the prolonged nerve stimulation takes place in the presence of the specific CHT inhibitor HC-3—a pharmacological parallel to the genetic deficit induced in CHT+/− mice (Jones & Kwanbunbumpen 1970). Therefore, a plausible scenario is that CHT+/− treadmill performance is impaired due to inability to sustain ACh turnover at the NMJ under conditions of high demand.

Central cholinergic pathways support the motor program both at brainstem and basal ganglia levels. A paradigm that allowed us to simultaneously investigate central cholinergic drive of locomotion and specific challenge to the muscarinic receptor component of ACh turnover is scopolamine-driven locomotion in the open field paradigm (Sipos et al. 1999). It is well established that scopolamine acts at presynaptic autoinhibitory muscarinic receptors, resulting in elevated ACh release, a corresponding increase in CHT-mediated choline uptake and hyperlocomotion (Douglas et al. 2001; Parikh et al. 2004). Our hypothesis was that, under sustained stimulation, ACh turnover would not be sustainable at CHT+/− synapses, and therefore CHT+/− mice would lack the scopolamine-induced hyperlocomotion evident in their CHT+/+ littermates. Indeed, while a low dose of scopolamine (0.1 mg/kg) failed to elicit a locomotive effect in either genotype, an intermediate dose (0.5 mg/kg) led to significant increases in distance traveled by CHT+/+ mice, but failed to affect CHT+/− mice. This finding was particularly interesting in the context of our previous findings with AChE−/− mice, which display normal basal locomotion in the open field chambers but are hypersensitive to challenge at this same intermediate dose of scopolamine (0.5 mg/kg) (Volpicelli-Daley et al. 2003).

The reverse response to scopolamine can be explained if we view the AChE−/− condition as a potential hypercholinergic state, in contrast with the hypocholinergic status of CHT+/− mice. Consistent with this idea, CHT expression is approximately 50% higher in AChE−/− mice, but 50% lower in CHT+/− mice (Volpicelli-Daley et al. 2003). It is important, however, to recognize that our current experimental design does not allow us to distinguish pre- vs. postsynaptic muscarinic receptor alterations. The latter distinction could, for example, explain the augmented response to a muscarinic agonist (oxotremorine) in the context of a blunted sensitivity to a muscarinic antagonist (scopolamine) in CHT+/− mice. In addition, these differing responses to muscarinic agents, which have also been observed in a mouse line bred for hypersensitivity to anticholinesterase (Overstreet & Russell 1982), may be rooted in the divergent neuroanatomical circuits employed by locomotor- vs. tremor-testing paradigms (Chintoh et al. 2003) used to evaluate the effects of scopolamine and oxotremorine, respectively.

The normal behavioral phenotype of CHT+/− mice in many tasks parallels the wild-type levels of synaptosomal HACU activity we established previously (Ferguson et al. 2004). However, the altered sensitivity to cholinergic pharmacological (HC-3, scopolamine, oxotremorine) challenges suggests that other determinants of ACh signaling are altered in the model. Accordingly, we found CHT+/− mice to have normal ChAT and AChE activities (see Results) but diminished bulk ACh (Fig. 6A) and increased total choline brain levels (Fig. 6B). Whereas total endogenous ACh levels are significantly lower in the CHT+/− brain, preliminary studies show normal evoked [3H]-ACh release, as measured in an in vitro slice-perfusion paradigm (D. Lund and M. Bazalakova, unpublished data), suggesting that a redistribution and altered regulation of ACh release and reserve pools (Neher 1998) may compensate for reduced overall ACh levels.

Decreased bulk ACh and increased total choline brain levels have been observed in several human pathologies and mouse models, including the PDAPP transgenic mouse model of Alzheimer’s disease (Bales et al. 2006). Elevated choline levels in CHT+/− mice may represent a compensatory attempt to enrich the choline supply pool for ACh synthesis. Unlike the impairments seen in AChE transgenic mice (Meshorer et al. 2005), we have little reason to suspect major blood–brain barrier dysfunction leading to increased choline flux in the CHT+/− brain. Therefore, elevated choline levels likely represent either a reduction in consumption of HACU-derived choline for ACh synthesis or increased membrane phosphatidyl turnover, resulting from an attempt to increase choline reserves available as precursors for ACh synthesis (Lee et al. 1993). The decreased bulk ACh levels, but normal exogenous [3H]-choline conversion and release of [3H]-ACh from slices, suggest a possible redistribution of ACh pools (Birks & MacIntosh 1961), where readily releasable pools are maintained at the expense of reserve ACh stores. These observations would predict our behavioral findings of normal baseline performance, but challenge-induced deficits in cholinergically mediated behaviors.

Functional compensation in CHT+/− mice may in part be achieved through a region-specific downregulation of M1 (striatum) and M2 (striatum and cortex) receptor expression (Fig. 7). The downregulation of muscarinic receptor expression is consistent with a hyposensitivity to scopolamine challenge in CHT+/− mice and is reminiscent of compensatory changes in muscarinic expression and function in another model of cholinergic dysfunction—AChE−/− mice (Volpicelli-Daley et al. 2003)—suggesting that loss of presynaptic ACh synthesis, due to a failure to recover choline from released ACh, rather than receptor hyperstimulation, may drive receptor downregulation in vivo. Additionally, M1 and M2 mAChR downregulation in the CHT hemizygous state contrasts with M2 striatal overexpression in AChE transgenics, where CHT density and CHT-mediated choline uptake are elevated at baseline (Beeri et al. 1997; Erb et al. 2001; Svedberg et al. 2003). We were unable to reliably quantify expression of the M3, M4 and M5 receptors, which may also exhibit CHT genotype-dependent modulation. Regardless, our findings reinforce a reliance on presynaptic ACh synthesis and storage to sustain appropriate levels of muscarinic receptors and to maintain normal cholinergic tone.

In conclusion, we have shown that CHT+/− mice perform comparably to wild-type littermates in motor and cognitive tasks, but display impaired performance when cholinergic neurotransmission is subjected to sustained physical or pharmacological challenge. Our findings are particularly relevant in the context of the human CHT single nucleotide polymorphism recently identified by Okuda and colleagues (Okuda et al. 2002). The I89V substitution occurs in approximately 6% of the Ashkenazi Jewish population and at similar or higher frequencies in North American Caucasian populations (B. English and R.D. Blakely, unpublished findings), and results in a 40–50% reduction in choline uptake measured in a transfected cell culture system (Okuda et al. 2002). Our findings predict phenotypes that emerge only upon cholinergic challenge, such as states of high physical activity or sustained focused attention, or which may arise from other alleles/insults, that lessen the activity of additional components of cholinergic signaling pathways. Future studies in CHT+/− mice can provide insights into the possible impact of CHT I89V and other gene variants on human behavior and physiology, and offer a valuable new model for testing theories and therapies related to cholinergic hypofunction.


The authors would like to thank Tammy Jessen and Angela Steele for general lab management, David R. Lund and Dr Alicia Ruggiero for helpful discussions and Dr John D. Allison for support with behavioral experiments. These studies were supported by the Medical Scientist Training Program at Vanderbilt University School of Medicine (MHB), National Institutes of Health awards R37 MH073159 (RDB) and NS30454 (CJH, AIL), a Zenith Award from the Alzheimer’s Association (RDB), as well as Core Services of the Kennedy Center (P30HD15052) and the Center for Molecular Neuroscience at Vanderbilt University.