Sex‐Specific, Intermediate Behavioral Phenotypes in Heterozygous Dopamine Transporter Mutant DAT T356M Mice

ABSTRACT

Autism spectrum disorder (ASD) is a complex neurodevelopmental condition with both genetic and environmental contributions. Previous work identified a de novo mutation in the dopamine transporter (DAT T356M) in an autism proband that results in profound behavioral changes when expressed homozygously in mice. Since complex human genetics are more likely to be present as heterozygous (single allele) mutations, we characterized mice that were heterozygous for the mutation. Both male and female DAT T356M^+/− mice exhibited hyperactivity but normal habituation to novel environments. The difference in hyperactivity compared to wild‐type littermates was dramatically smaller than previously reported in homozygous animals. Other behavioral alterations were sex‐specific, with only male heterozygous mice exhibiting greater repetitive behaviors and impaired spatial learning in the Barnes maze. Sensorimotor gating measured by prepulse inhibition of the startle response was largely unchanged in both sexes. Motor performance on the rotarod showed opposing effects, with male heterozygotes showing decreased latency to fall while females demonstrated increased latency (i.e., enhanced performance). These findings suggest that even a single copy of the DAT T356M variant can impact behavior in a sex‐specific manner. The identification of intermediate phenotypes makes these mice an appropriate model for future studies examining how environmental factors might interact with genetic susceptibility to influence autism‐relevant behaviors, particularly in the context of dopaminergic dysfunction.

Mice heterozygous for a mutation in the gene encoding the dopamine transporter display intermediate, sex‐specific phenotypic features relative to the homozygous mutant. Such animals hold promise for examining gene × environment interactions.

1. Introduction

Autism is a neurodevelopmental condition characterized by weaknesses in social communication and the presence of restricted interests and repetitive behaviors [1]. Despite its high prevalence, the neurobiological underpinnings of autism have remained elusive. This is likely due to the fact that autism is a spectrum disorder without a single etiology. Indeed, over 1000 genes have now been shown to confer elevated risk for autism [2], suggestive of multiple mechanisms giving rise to the complex constellation of characteristics [3, 4, 5, 6, 7, 8, 9].

Although many of the autism‐related genes noted above appear to be involved in developmental processes such as those giving rise to cortical and synaptic structure and function, others appear to impact neuronal signaling via their effects on neurotransmitter systems and signaling. Although much of this work has focused on major excitatory and inhibitory neurotransmitters such as glutamate and GABA, there is growing evidence for a possible role for neuromodulators, such as dopamine (DA) in some forms of autism [10].

Although classically associated with reward and motivational salience, DA's neuromodulatory role extends to include the shaping of sensory and motor functions and in the signaling of prediction [11, 12]. Hence, and in the context of autism, changes in DA signaling may have a significant impact on a number of the phenotypic features of the disorder, including activity levels, repetitive behaviors, and social function. Indeed, there is a growing body of work linking dopaminergic dysfunction with various features of autism despite a lack of consistency in targets of effects of gene changes within DA signaling pathways [13]. One line of evidence revolves around the dopamine transporter (DAT), which regulates dopamine availability at the synapse. Variants of DAT have been associated with autism, with work identifying a de novo autism spectrum disorder (ASD)‐associated mutation in which there is a threonine‐to‐methionine substitution at site 356 (DAT T356M) [14, 15]. This mutation has been shown to effectively reverse the transporter, resulting in an increase in extracellular DA [16]. In prior work, we have demonstrated that knock‐in mice homozygous for this mutation (DAT T356M) exhibit profound impairments in DA neurotransmission and clearance. In addition, these changes in DA signaling were found to be associated with a number of behaviors that recapitulate various facets of the autism clinical phenotype [17].

It is likely that the complex etiology of autism is not solely the result of the manifold genetic influences that have been shown to confer risk, but also includes environmental factors that may interact with an individual's genetic makeup in a way that elevates autism risk. Gene by environment (GxE) interactions are an active area of autism research, and some of the factors that have been put forth as increasing autism risk include: extreme prematurity and very low birth weight, birth difficulties resulting in a period of hypoxia, prenatal exposure to pesticides and other chemical exposures, and maternal health conditions such as obesity, diabetes and immune dysfunction [18, 19]. In addition, there may be dietary factors that play an important and underappreciated role. DA synthesis is particularly vulnerable to environmental modulation due to its easily oxidizable state and reliance on enzymatic cofactors that are obtained through dietary intake. In this way overexposure to metals such as manganese (Mn²⁺) and iron (Fe²⁺), which are readily obtained from food, and may be at high levels in polluted air or poorly filtered water, or deficiency of key antioxidants such as ascorbic acid (vitamin C), can directly impact the presentation of behavioral phenotypes through modulation of synaptic DA [20, 21, 22].

In such GxE settings, the impact of environmental factors is likely to extend beyond those carrying homozygous variants of autism risk genes to those carrying a single allele. We previously reported robust hyperactivity and impaired DA clearance in homozygous knock in DAT T356M^+/+ mice [17]. The abnormal phenotypes in these animals were sufficiently severe that it is unlikely that additional environmental modifications would further impact these effects. Hence, studies of heterozygous animals with the goal of identifying intermediate phenotypes set the stage for work to examine the role of environmental factors in GxE interactions. In the current study, we set out to lay the foundation for this work by characterizing the behavioral characteristics of heterozygous animals carrying a single copy of the mutant DAT gene. Furthermore, given that ASD is more commonly reported in males, and behavioral profiles can differ according to sex and gender, we specifically sought to test whether the same behavioral phenotypes would be observed in male and female mice according to the genotypes expressed by the mice.

2. Materials and Methods

2.1. Animals

All behavioral experiments were performed in the Vanderbilt University Neurobehavioral Core Facility. All behavioral experiments were performed under a protocol approved by the Vanderbilt University Animal Care and Use Committee. Experimental mice were obtained from DAT T356M^+/− × DAT T356M^WT breeding pairs, and all offspring were able to be used for behavioral studies. Mice were weaned and genotyped at 21–24 days of age from a tail sample. All behavioral tests were performed at approximately the same time of day (08:00–13:00) using mice between the ages of 8–12 weeks. Mice were group housed (2–5 mice per cage) on a 12:12‐h light–dark cycle with food and water available ad libitum. Prior to the first day of testing, all mice were weighed, and their tails were marked using a permanent marker for identification. Mice were transferred from the housing room to testing rooms 30 min prior to the start of each behavioral test to allow for an acclimation period. All testing apparatuses were cleaned with a 10% ethanol solution before and after each trial to minimize odor cues and provide a standardized testing environment. There were 137 total mice that were used for this study across two separate experiments that included different combinations of behavioral tasks (DAT T356M^WT male n = 27; DAT T356M^+/− male n = 17; DAT T356M^WT female n = 29; DAT T356M^+/− female n = 22 in Experiment 1, and DAT T356M^WT male n = 7; DAT T356M^+/− male n = 12; DAT T356M^WT female n = 9; DAT T356M^+/− female n = 14 in Experiment 2). Each experiment population was made up of at least three independent cohorts of animals. In Experiment 1, only one experimental task was performed per day. In Experiment 2, one or two tasks were performed each day, with total testing and handling time limited to less than 90 min per mouse per day (see timeline Figure 1A). Not all mice were used for all studies to better control for potential time of testing effects when working with large groups, but all data collected are included in the analysis unless excluded for specific reasons, as described in Section 3.

FIGURE 1.

DAT T356M^+/− mice exhibit spontaneous hyperactivity in novel environments. (A) Timeline for both behavioral experiments. Locomotor activity testing was repeated in both experiments and presented separately, such that Experiment 2 (shown with pattern‐filled bars) formed an independent replication for results that were established in Experiment 1. Exploratory activity was measured by beam breaks over 60 min. (B–E) Total distance traveled was greater in DAT T356M^+/− mice than in wild‐type littermates, although normal habituation across time was observed in both male (B) and female (C) mice. Increased distance traveled was observed in both Experiment 1 (D) and in the replication cohort in Experiment 2 (E). Average time in the periphery was not significantly different among groups in Experiment 1 (F), whereas in Experiment 2, T356M^+/− mice spent less time in the periphery of the maze (G). In the elevated zero maze, we did not note any significant differences in distance traveled (H) or time spent exploring each area of the maze (I), Experiment 2 only. *, **p < 0.05, 0.01 indicating a significant main effect of genotype. T356M^WT mice are shown in gray bars, T356M^+/− mice are shown in green bars. Data from Experiment 2 have dot patterning in the bars.

2.2. Behavioral Testing

2.2.1. Locomotor Activity

Spontaneous locomotor activity in an open field was measured using 27 × 27 × 20.5 cm chambers (Med Associates) placed within sound‐attenuating boxes (64 × 45 × 42 cm). Spontaneous locomotion was detected via infrared beam disruption to measure horizontal and vertical movements (jumping, rearing). The mice were placed in the activity chambers and allowed to explore freely for 60 min. The percent of time mice spent in the periphery of the open field (52% of total area) vs. the center was also calculated.

2.2.2. Elevated Zero Maze (EZM)

Mice were placed on a standard, white zero maze in a brightly lit room (open arms 490 lx, closed arms 220–255 lx) and allowed to explore for 5 min, while being video recorded from above. Time spent and distance traveled in high‐walled “closed” quadrants were calculated compared to the low‐edged “open” quadrants of the maze.

2.2.3. Spontaneous Behavioral Observations

Mice were placed into clean cages containing only corncob bedding. All activity was recorded using a video camera for 15 min. Videos were coded by observers looking for instances of digging, grooming, rearing, jumping, and climbing behaviors. The observers were blinded to the genotypes when coding. Rearing was defined as temporarily standing on its hind legs, either free‐standing or against the cage wall. Climbing was defined as a mouse using its paws to change its location vertically, with the use of the wire bar lid on top of the cage. Grooming refers to actions with fore or rear paws, including scratching‐like behaviors. Each incidence of digging (moving corn cob bedding with its paws) was recorded. All videos were coded by one or two raters who were previously checked to ascertain high interrater reliability.

2.2.4. Horizontal Beam

To assess agility and coordination, mice were placed in the center of a horizontal dowel (30 cm length and 0.5 cm diameter). Mice were placed at 90° to the dowel such that they had to reorient themselves before they could escape to either of the side platforms. Latency to escape to either side was recorded across three trials with a maximum trial time of 60 s.

2.2.5. Inverted Screen

The inverted screen was made of individual chambers (approx. 10 cm × 10 cm) with a wire mesh floor that mice were easily able to grip. The chambers were gently inverted twice to accustom the mice to the need to grip the floor, and on the third inversion, the apparatus was turned through 180° so that the mice were upside down gripping the wire floor. The time taken to fall (approximately 30 cm) was recorded across three sessions with a maximum trial time of 60 s each.

2.2.6. Tube Test for Social Dominance

The dominance tube was approximately 30 cm long and made of clear acrylic with funnel shapes at either end to facilitate entry. One mouse was held by the tail at each end, and the tail was gently pulled backward to encourage the mice to run forward into the tube. Both mice were released into the tube at the same time, and the time until one mouse backed out of the tube was recorded. The mouse that remained in the tube was designated the winner, and the other mouse was designated the loser. A tie was declared if neither mouse backed out within 180 s. Pairs were made up of one T356M^WT and T356M^+/− from within the same cage to test previously established social dominance relationships.

2.2.7. Three‐Chamber Social Preference

An initial 6‐min habituation trial was conducted with empty cups only present in the arena. The test mouse was returned to its home cage during an inter‐trial interval of approx. 3 min, while the maze was cleaned and a target mouse was placed under one of the cups in the maze. Target mice were naïve, young adult wild‐type mice sex‐matched to the test mouse. The location of the target mouse was balanced across trials. The test mouse was then returned to the central chamber for 8‐min test trial and permitted to explore freely. The location was automatically coded from overhead videos using AnyMaze. Time spent in each maze area (center, mouse, and empty cup zones) was automatically coded.

2.2.8. Prepulse Inhibition (PPI) of the Startle Response

PPI of the acoustic startle response was used as an index of sensorimotor gating by measuring the degree to which a prestimulus (tone) inhibits the subsequent response (startle reflex) to a later, more salient tone. PPI was measured using commercially available startle chambers (Med Associates). Mice were placed in a clear plastic cylinder (~5 cm diameter) within a ventilated, sound‐attenuating enclosure. All acoustic stimuli were presented through a loudspeaker mounted 28 cm above the animal. Each session consisted of 6 trial types, including a control no‐stimulus trial, each presented 9 times in a pseudo‐random order, resulting in a total of 54 trials over a 20‐min period. Startle‐only trials included only the startle stimulus—a 40‐ms duration, 120 dB burst of white noise. Prepulse stimuli trials consisted of 20‐ms bursts of 70, 76, 82, or 88 dB white noise, presented 50 ms before the startle (120 dB) stimulus. Prepulse acoustic stimuli of these intensities do not produce a startle response alone. The intertrial interval varied, ranging from 10 to 20 s between trials. The maximum startle amplitude was recorded during the 65‐ms sampling window. PPI was calculated as the percentage of inhibition of the startle amplitude evoked by the 120 dB pulse alone: ([Response on pulse trial − Response on prepulse trial]/Response of pulse trial × 100).

2.2.9. Rotarod

Rotarod was used to assess motor learning and coordination. The rotarod starting rotation rate was 4 rotations per minute (rpm) and accelerated steadily to 40 rpm over a period of 5 min. The time to fall or to complete one passive rotation (holding onto the rod while it rotates through 360° without falling off) was recorded for each animal with a maximum time of 300 s. Animals completed three trials each day for 3 days.

2.2.10. Barnes Maze

The Barnes maze was used to assess spatial learning as we have previously [23, 24, 25]. The maze consisted of a flat, circular table with 12 holes around the perimeter, arranged similarly to a clock. The table was 92 cm in diameter, and each hole (target/escape hole or non‐target/blocked holes) was 5 cm in diameter. The target location was the same for all mice to avoid possible exploratory bias based on room cues. The task relies on the preference of the mice for dark, enclosed spaces compared to brightly lit, open spaces. Each mouse underwent 4 trials per day, with each trial lasting until the mouse found the target hole or until 180 s had elapsed. The inter‐trial interval was approximately 1 h on task acquisition days. If the mouse did not find the hole within the 3 min, the experimenter guided the mouse to the hole to reinforce the learning process. Trials were recorded from above, allowing tracking of escape paths for the determination of escape strategy. A trial in which the mouse ran either directly to the escape hole or to an adjacent hole, and then turned to the target was considered a Direct strategy. A trial in which the mouse ran to the edge of the platform and then traveled around the perimeter of the maze, where all of the holes were located, until it reached the escape hole was considered a Serial strategy. Trials that did not follow clear strategies were considered Mixed.

Following behavioral testing, mice were euthanized by terminal anesthesia, and their brains were removed. Striatum was dissected out from both hemispheres and used for analysis of neurotransmitter levels and protein expression determination.

2.2.11. HPLC

Striatal tissue samples were homogenized with 0.5 mL of 0.2 M perchloric acid per 100 mg of tissue by hand with a plastic pestle. After denaturing the protein in an ice bath for 30 min, samples were centrifuged at 15,000 rpm for 20 min at 4°C, and the supernatant was removed. A total of 1 M sodium acetate (half of the volume of the sample) was added to each sample to modify the pH to 3.0, and 10 μL of the sample was injected for analysis using the Amuza HTEC‐600 with an Eicompak SC‐50DS column (ID 3.0 × 100 mm). The mobile phase was 83% 0.1 M citrate‐acetate buffer and 17% methanol with 190 mg/L sodium octansulfonate and 5 mg/L EDTA‐2Na.

2.2.12. Western Blot

Protein was extracted from striatal tissue and quantified by western blot as described in DiCarlo et al. [17]. Western blot analysis was performed on striatum tissue lysates using the following primary antibodies: anti‐ERK1/2 (p44/42 MAPK) (Cell Signaling Technology, #9102), phospho‐ERK1/2 (Thr202/Tyr204) (Cell Signaling Technology, #4370), tyrosine hydroxylase (TH) (Cell Signaling Technology, #2792), phospho‐TH (Ser31) (Cell Signaling Technology, #3370) and beta actin (C4) (sc‐47,778 Santa Cruz Biotechnology). Membranes were incubated overnight at 4°C with primary antibodies according to the manufacturers' recommended dilutions, followed by appropriate HRP‐conjugated secondary antibodies (Anti‐Rabbit IgG (H + L), HRP Conjugate and Anti‐Mouse IgG (H + L), HRP Conjugate). Signal detection was carried out using enhanced chemiluminescence.

2.2.13. Statistics

All statistical analyses were conducted using GraphPad Prism software (version 10.1.0). Data are presented as mean ± SEM with individual data points shown where possible. Differences are considered statistically significant at p < 0.05. We were interested in whether the same genotype differences would be seen in both male and female mice, given that the original proband from whom the mutation was identified was male, and also that autism is more often diagnosed in males. Univariate ANOVA was conducted with sex and genotype as the independent variables, followed by Šidák's multiple comparison test to compare the effect of genotype within each sex. For data including repeated measures (e.g., rotarod and Barnes maze), analyses were therefore first performed on all data using a 2 (genotype) × 2 (sex) ANOVA with additional repeated measures of time as appropriate. Where there was a main effect of sex, or a sex × genotype interaction, data were then split and analyzed separately according to sex to provide better statistical power for analyses, particularly given uneven group numbers, to test our hypothesis that behavioral phenotypes associated with the genotype may differ according to sex.

3. Results

3.1. Locomotor Activity

One of the most striking abnormal phenotypes reported in homozygous DAT T356M^+/+ mice was extreme hyperactivity that was a three‐ to fourfold increase compared to wild‐type mice over an hour with only limited habituation to the novel environment [17]. Across the full 60‐min testing period, heterozygous DAT T356M^+/− mice also traveled further than their wild‐type littermates (F _{1, 91} = 6.56, p = 0.012; approximately 15% increase in females and 30% increase in males) (Figure 1A–C). Overall, females also traveled further than males (F _{1, 91} = 5.22, p = 0.025). We also examined change over time to understand whether heterozygous animals also showed the same limited habituation as previously reported in their homozygous littermates. Both male and female heterozygous mice decreased activity over the 60‐min interval and were traveling similar low distances by the final 5‐min time bin (time × sex × genotype F _{11, 1001} = 3.63, p < 0.001) (Figure 1A,B).

Average time spent in the periphery of the chambers was used as an index of anxiety‐like behavior. Although DAT T356M^+/− mice, and particularly the males, spent the least time in the periphery compared to other groups (which could indicate some level of behavioral disinhibition), the difference was not significant (Fs_1,91 < 3.24, ps > 0.075) (Figure 1E).

Since the phenotype of increased activity has been shown to be strongly related to poor DA uptake due to alterations in DAT, and was the most important and consistent behavioral finding in these mice, we sought to confirm the hyperactivity in a second, independent group of mice (Experiment 2). We observed even greater increases in activity in the DAT T356M^+/− mice in this group (approximately 35%–45% compared to DAT T356M^WT F _{1, 38} = 11.71, p = 0.0015; Figure 1D). However, we did not observe the same difference according to sex (F _{1, 38} = 1.73, p = 0.20, sex × genotype F _{1, 38} = 0.01, p = 0.93).

The trend toward decreased time spent in the periphery of the locomotor activity chambers was even more pronounced in the replication cohort with less periphery time in DAT T356M^+/− mice (F _{1, 38} = 7.05, p = 0.012; Figure 1G). This effect was similar in both male and female mice (Fs < 1.71, ps > 0.20).

In order to further investigate anxiety‐like behaviors in our second cohort of mice, they were tested in the EZM. The distance traveled in the maze was not greater in T356M^+/− mice compared to littermates (F _{1, 38} = 2.16, p = 0.15; Figure 1H) with no additional differences according to sex (ps > 0.63). Time spent in the closed arms of the maze was used as an index of anxiety‐like behavior and was slightly increased in T356M^+/− mice, although this difference was not statistically significant (F _{1, 38} = 3.16, p = 0.08; Figure 1I). There were no additional differences according to sex (Fs < 1.25, ps > 0.27).

A second hallmark feature of homozygous T356M^+/+ mice was an approximate twofold increase in repetitive rearing behavior [17]. To examine this, as well as other spontaneous behaviors in heterozygous animals, we quantified spontaneous exploratory behaviors across 15‐min sessions in a novel environment (Figure 2). All data were first checked for statistical outliers (Grubbs analysis ROUT Q = 0.1%). Using this approach, for analysis of rearing behavior, one female T356M^+/− mouse was removed (126 rearing bouts), and for analysis of grooming behavior, one male T356M^+/− mouse was removed (24 grooming bouts). We observed a much smaller increase in rearing in heterozygous T356M^+/− mice when compared with the original homozygous animals [17], and these changes were limited to male mice (genotype F _{1, 65} = 2.37, p = 0.13; sex F _{1, 65} = 2.44, p = 0.12; interaction F _{1, 65} = 5.58, p = 0.021) (Figure 2A). In contrast, self‐grooming behaviors were decreased in male heterozygous T356M^+/− mice (genotype F _{1, 65} = 0.59, p = 0.45; sex F _{1, 65} = 0.077, p = 0.78; interaction F _{1, 65} = 6.90, p = 0.011) (Figure 2B). Climbing on the cage lid and digging in bedding did not differ between the groups (F < 2.06, ps > 0.16) (Figure 2C,D). Very few jumping behaviors were noted in the 15‐min scoring sessions—1 out of 16 male and 1 out of 24 female wild‐type mice had one jump counted each, and none of the 14 male heterozygous T356M^+/− mice jumped. Three out of 16 female heterozygous T356M^+/− mice jumped 3, 6, and 12 times, respectively.

FIGURE 2.

DAT T356M^+/− mice exhibit spontaneous hyperactivity in novel environments. Spontaneous behaviors in a novel environment were coded by investigators blinded to genotype over 15 min, including (A) Rearing, (B) Grooming, (C) Climbing, and (D) Digging. N = 16 male, 24 female T356M^WT and 14 male, 16 female T356M^+/−. *p < 0.05 Fisher's LSD multiple comparisons test following a significant Sex × Genotype interaction. Images illustrating mouse behavior generated using BioRender.

3.2. PPI of the Startle Response

We first sought to confirm that baseline startle response was similar among the groups. The increased startle response to the 120 dB tone over blank trials was clear (trial type F _{1, 74} = 717.2, p < 0.001); however, the overall startle response did differ between males and females (sex F _{1, 74} = 7.73, p = 0.007). Consequently, the data were split and analyzed according to sex. Male mice showed the expected startle response to the tone (trial type F _{1, 34} = 251.5, p < 0.001) with no effect of genotype (Fs < 0.31, ps > 0.52) (Figure 3A). In contrast, although female mice showed a startle response (trial type F _{1, 40} = 528.6, p < 0.001), this was modestly decreased in T356M^+/− mice compared to wild‐type littermates (genotype F _{1, 40} = 5.76, p = 0.021, interaction F _{1, 40} = 0.07, p = 0.79) (Figure 3B).

FIGURE 3.

Similar sensorimotor gating (prepulse inhibition [PPI]) in DAT T356M^+/− and wild‐type mice. Mice of both genotypes showed a normal startle response to the 120 dB tone (Experiment 1 (A and B), Experiment 2 (E and F)), although startle magnitude was slightly decreased in female T356M^+/− mice in Experiment 1. Percent PPI to the 120 dB tone following prepulse stimuli (76–88 dB) increased according to the prepulse level but was similar between sexes and genotypes (C and D, G–H). Data shown are raw values in arbitrary units for startle response. Any negative prepulse values were converted to 0 for analysis. (A–D) N = 22 male, 23 female T356M^WT and 14 male, 19 female T356M^+/−. (E–H) N = 9 male, 7 female T356M^WT and 14 male, 13 female T356M^+/−. ****p < 0.0001 main effect of dB level during startle trials, ** uncorrected Fisher's LSD as marked, following significant dB level × genotype interaction.

To assess PPI, we analyzed the 120 dB tone plus prepulse (70, 76, 82, or 88 dB) trials separately, with data calculated from each mouse's own baseline startle data. If mice had trials with negative numbers, then these trials were replaced with 0 to indicate no PPI had occurred. Both male and female mice showed greater PPI with increasing PPIs as expected (male F _{3, 96} = 5.99, p = 0.003; female F _{3, 108} = 14.05, p < 0.001) with no additional effects of genotype for either sex (Fs < 2.26, ps > 0.085) (Figure 3C,D).

3.3. Social Behaviors

We also tested a range of tasks linked to social function in the mice. In Experiment 1, only a subset of mice underwent social preference testing. Data for mice from both experiments that were tested for social interaction preferences are combined for analysis. During the initial 6‐min habituation trial in the three‐chamber task, the slight increase in activity in the T356M^+/− mice was not significant (F _{1, 38} = 2.16, p = 0.15, Figure 4A), and neither were there differences according to sex (Fs < 1.64, ps > 0.21). The majority of mice in each group showed a strong preference for spending time proximal to the novel mouse rather than in the empty side of the chamber during the test trial (T356M^WT male t(17) = 3.88, p = 0.0012; T356M^+/− male t(13) = 2.93, p = 0.012; T356M^WT female t(13) = 2.54, p = 0.025; T356M^+/− female t(17) = 2.81, p = 0.012, Figure 4B). In both WT groups, only one mouse spent more time exploring near the empty cup, while in both T356M^+/− groups, three mice per group exhibited the reverse preference pattern. These differences were not sufficient to impact the data for the group as a whole.

FIGURE 4.

T356M^+/− mice have normal social behaviors. Distance traveled during the habituation trial prior to social preference testing (A) was not different among groups, and all groups showed a preference for exploration near the novel mouse during the social preference test (B). The number (percent) of trials in which the T356M^+/− mice backed out of the tube (T356M^WT “wins”) was not significantly higher than when wild‐type mice “lost” the tube test (C). Number of marbles buried did not differ among groups (D). Social preference and tube testing were performed in mice from both Experiments 1 (subset) and 2; N = 18 male, 14 female T356M^WT and 14 male, 19 female T356M^+/−. Marble burying was performed in Experiment 2 only. ^# p < 0.05 main effect of genotype; *p < 0.05 Šídák's multiple comparisons test following significant genotype × day interaction. (E and F) N = 9 male, 7 female T356M^WT and 12 male, 14 female T356M^+/−.

We also tested for within‐cage social‐dominance relationships using the tube test. As a group, wild‐type mice of both genotypes won more than 50% of the bouts against T356M^+/− littermates; however, this difference was not significant for either sex (male Mann–Whitney U = 36, p = 0.36, female Mann–Whitney U = 82, p = 0.34; Figure 4C).

In Experiment 2, we also assessed marble burying and found no differences according to sex or genotype in the numbers of marbles either partially or completely buried (Fs < 1.50, ps > 0.23, Figure 4D).

3.4. Rotarod

Homozygous T356M^+/+ mice demonstrated superior rotarod performance to wild‐type littermates, possibly because chronically increased activity had altered muscle tone and coordination [17]. In the current study of heterozygous and wild‐type animals, all mice showed learning on the rotarod with increased latency to fall or first rotation on the pole (day F _{3, 146} = 8.64, p = 0.0005). The rate of learning and final fall latency differed according to sex and genotype (sex F _{1, 73} = 4.43, p = 0.039, sex × day × genotype F _{2, 146} = 5.39, p = 0.0055), so data were split and analyzed according to sex. Motor learning was observed in the male mice, but this differed according to genotype, with faster latencies to fall or rotate in the T356M^+/− heterozygous animals (day × genotype F _{2, 64} = 3.68, p = 0.031; Figure 5A). This effect was particularly evident on the final day of training (p < 0.05). All female mice exhibited motor learning over the 3‐day training period (F _{2, 123} = 4.82, p = 0.0096) with no differences according to genotype (Fs < 1.01, ps > 0.37; Figure 5B). While time to first rotation or fall may better represent fine motor learning on the rotarod, maximum time to fall (or end of trial at 300 s) may provide a better representation of overall muscular strength or stamina. As above, data were split by sex following initial analyses (sex F _{1, 73} = 7.41, p = 0.0081, sex × genotype F _{1, 73} = 5.058, p = 0.028). For male mice, maximum trial time was similar across days in both genotypes (Fs < 2.79, ps > 0.10; Figure 5C). In contrast, female T356M^+/− mice were able to remain on the rotating rod for longer than their wild‐type littermates (genotype F _{1, 123} = 4.66, p = 0.033, day F _{2, 123} = 0.96, p = 0.91, interaction F _{2, 123} = 0.49, p = 0.6; Figure 5D).

FIGURE 5.

T356M^+/− showed sex‐dependent genotype differences in the rotarod test. Time to first performance error on the rotarod (fall or rotation) is shown for Experiment 1 in male (A) and female (B) mice. Maximum trial time (300 s or time to fall) is also shown for male (C) and female (D) mice. Latency to reach the escape platform from the horizontal beam (E) and latency to fall from the inverted screen (F) did not differ for mice in Experiment 2. (A–D) N = 18 male, 27 female T356M^WT and 16 male, 16 female T356M^+/−. ^# p < 0.05 main effect of genotype; *p < 0.05 Šídák's multiple comparisons test following significant genotype × day interaction. (E and F) N = 9 male, 7 female T356M^WT and 12 male, 14 female T356M^+/−.

To establish whether differences in rotarod performance were due to gross differences in strength and/or agility, we tested mice from the second replication cohort (Experiment 2) on horizontal beam and inverted screen tasks. We did not observe any differences according to sex or genotype on these tasks (Fs < 2.17, ps > 0.15).

3.5. Barnes Maze

The Barnes maze task was performed to assess differences in spatial learning. For all animals, there was strong evidence of learning across the 5 days of training as shown by decreasing path lengths (F _{4, 200} = 91.5, p < 0.001). A main effect of genotype (F _{1, 50} = 5.39, p = 0.024) suggested that WT mice solved the maze more efficiently than T356M^+/− mice, and this effect appeared to be driven primarily by the male mice (sex × genotype F _{1, 50} = 10.50, p = 0.0021; Figure 6A,B). When learning was analyzed separately in males and females, we observed that all female mice showed decreasing path lengths over the course of training (F _{4, 92} = 43.33, p < 0.001) and there were no differences according to genotype (Fs < 0.69, ps > 0.446). In contrast, male T356M^+/− mice improved over time but had longer path lengths across the duration of training when compared with their WT littermates (day F _{4, 108} = 48.87, p < 0.001; genotype F _{1, 27} = 12.52, p = 0.0015; day × genotype F _{4, 108} = 2.65, p = 0.037; Figure 6A). The difference was significant on days 1 and 2 of training (ps < 0.018), but not on the final 3 days (ps > 0.38, Šídák's multiple comparisons test).

FIGURE 6.

Male T356M^+/− mice show less effective learning strategies in the Barnes maze. (A and B) Average path length (distance traveled) over the four daily trials is shown for the Barnes maze in male (A) and female (B) mice. (C–F) Primary latency (C and D time in seconds to first locate the escape hole), and escape latency (E and F) time in seconds to enter the escape hole in male (C and E) and female (D and F) mice. Percent use of different escape strategies are also shown for male T356M^WT(Gi), male T356M^+/− (Gii), female T356M^WT (Hi), and female T356M^+/− mice (Hii) for direct strategy (mouse runs directly to the escape hole or adjacent location, blue), serial strategy (mouse runs to the edge and circumnavigates until the escape hole is located, green) or mixed strategy (no clear pattern is observed, yellow). N = 17 male, 13 female T356M^WT and 12 male, 12 female T356M^+/−. ^a p < 0.05 Šídák's multiple comparisons test following significant genotype × day interaction; ^# p < 0.05 main effect of genotype on strategy use; **, ***, ****p < 0.01, p < 0.001, p < 0.0001, paired t‐test.

Path length was used as the primary outcome to avoid artificially increased escape latencies caused by mice that locate the escape hole and pause on the edge prior to entering the hole and terminating the trial, or differences in running speed. These differences were also accounted for in differences in primary latency—time to first reach the target hole [23], and escape latency—time to enter the escape hole (or end of the trial). During task acquisition, all mice showed learning through increased latency to reach the target hole location across the 5 days of training, although the rate of learning differed slightly according to group (day F _{4, 200} = 36.73, p < 0.0001; day × sex × genotype F _{4, 200} = 3.07; p = 0.018; Figure 6C,D). Follow‐up analyses in males and females separately confirmed learning across days in both groups (Fs > 18.95, ps < 0.001) and did not establish further differences according to genotype (Fs < 3.78, ps > 0.11). When time to enter the escape hole was considered, a distinct pattern of differences emerged. While all groups showed decreased escape latencies over days (day F _{4, 200} = 43.48, p < 0.0001; Figure 6E,F), the pattern of learning differed among groups (sex × genotype F _{1, 50} = 5.99, p = 0.018; day × sex × genotype F _{4, 200} = 3.97, p = 0.004). Separate analyses in males and females indicated that male T356M^+/− mice took longer than wild‐type littermates to enter the escape hole on days 1–3 of acquisition (genotype F _{1, 27} = 3.30, p = 0.081; day F _{4, 108} = 17.52, p < 0.0001; day × genotype F _{4, 108} = 2.73, p = 0.033; Figure 6C). All female mice exhibited learning with decreased latencies over time (day F _{4, 92} = 28.58, p < 0.0001; Figure 6D,F). Escape latencies did not vary significantly in female mice according to genotype on any of the training days (Fs < 2.75, ps > 0.05).

These differences in learning were reflected in the search strategies used to locate the platform, which were similar among female mice but differed in male mice (Direct search: Day F _{4, 200} = 27.59, p < 0.001, sex × genotype F _{1, 50} = 4.86, p = 0.0321). When strategy data were analyzed according to sex, it confirmed that female mice all increased their use of direct strategies at a similar rate (day F _{4, 115} = 21.84, p < 0.001, genotype F _{1, 115} = 1.37, p = 0.24, interaction F _{4, 115} = 0.55, p = 0.70). Use of direct strategies also increased in male mice (day F _{4, 135} = 8.62, p < 0.001); however, T356M^+/− mice were less likely to use a direct strategy than wild‐type littermates (genotype F _{1, 135} = 7.57, p = 0.0067, interaction F _{4, 135} = 1.047, p = 0.39; Figure 6G,H).

Distance traveled during the 3‐min probe trial was similar among groups with no differences according to sex or genotype (Fs < 2.49, ps > 0.12; Figure 6I). All groups also showed a significant preference for the target area over non‐target areas indicating they had learned and retained the memory of where the escape hole was expected to be located (paired t‐tests for time in target hole surround vs. average of non‐target hole surrounds ts > 5.19, ps < 0.0002 [Figure 6J], and number of entries into target vs. average non‐target surrounds ts > 2.97, ps < 0.009 [Figure 6K]).

3.6. Biochemistry

In the previously described DAT T356M^+/+ homozygous knock‐in animals, data suggested increased DA metabolism due to decreased reuptake of the released DA. In contrast, we observed no significant differences according to genotype in tissue (cortex/striatum) levels of DA, serotonin, or norepinephrine, nor of any of the DA metabolites measured (DOPAC, 3‐MT, and HVA) (main effects of genotype and genotype × sex interactions ps > 0.14; Table 1). We also confirmed there were no differences in the ratios between DA and each metabolite (ps > 0.07, data not shown). Although there was a trend toward higher DA levels in female mice (sex F _{(1, 34)} = 3.51, p = 0.07), this difference was not significant.

TABLE 1.

Bioamine measurements.

	Sex	T356MWT	T356M+/−	2 × 2 ANOVA results
		Mean ± SD (ng/mg)	Mean ± SD (ng/mg)
Dopamine	Male	434.32 ± 417.2	486.55 ± 350.7	Interaction F (1, 34) = 0.0017, p = 0.97
	Female	686.21 ± 428.8	727.55 ± 397.4	Sex F (1, 34) = 3.51, p = 0.07
				Genotype F (1, 34) = 0.13, p = 0.72
DOPAC	Male	58.96 ± 64.5	50.33 ± 25.34	Interaction F (1, 34) = 0.0034, p = 0.95
	Female	76.89 ± 40.9	70.14 ± 56.3	Sex F (1, 34) = 1.37, p = 0.25
				Genotype F (1, 34) = 0.23, p = 0.64
HVA	Male	12.88 ± 6.02	10.70 ± 4.6	Interaction F (1, 34) = 1.55, p = 0.22
	Female	13.77 ± 9.23	18.31 ± 10.6	Sex F (1, 34) = 2.49, p = 0.12
				Genotype F (1, 34) = 0.19, p = 0.66
3‐MT	Male	14.17 ± 7.36	14.42 ± 9.9	Interaction F (1, 33) = 0.31, p = 0.58
	Female	15.65 ± 5.31	19.06 ± 10.6	Sex F (1, 33) = 1.18, p = 0.29
				Genotype F (1, 33) = 0.42, p = 0.52
Serotonin	Male	30.91 ± 13.9	33.61 ± 16.3	Interaction F (1, 34) = 0.72, p = 0.40
	Female	31.29 ± 15.6	25.18 ± 17.6	Sex F (1, 34) = 0.60, p = 0.44
				Genotype F (1, 34) = 0.11, p = 0.74
Norepinephrine	Male	25.90 ± 9.8	19.26 ± 9.5	Interaction F (1, 34) = 1.22, p = 0.28
	Female	18.76 ± 5.9	17.70 ± 5.7	Sex F (1, 34) = 2.96, p = 0.09
				Genotype F (1, 34) = 2.33, p = 0.14

Note: Data shown are mean ± standard deviation. Statistics presented are from 2 genotype × 2 sex ANOVA.

Since similar levels of neurotransmitters could reflect a homeostatic response to genotype‐driven changes, we also measured protein levels of tyrosine hydroxylase (TH) and extracellular signal‐related kinase (ERK) and their phosphorylated forms (pTH and pERK), which are each important in the DA synthesis pathway. In DAT T356M^+/+ homozygous knock‐in animals, there was decreased pTH and pERK relative to the pan form of each protein, suggesting more was in the active unphosphorylated state. In the present study, we observed very few differences in TH or ERK expression or phosphorylation according to genotype (Fs < 1.07, ps > 0.31). The one noted difference was higher levels of phosphorylated TH in the striatum, reflecting a potential increase in DA metabolism in females (F _{(1, 27)} = 5.488, p = 0.027, data not shown).

4. Discussion

Overall, the results of the current study support the presence of a complex intermediate phenotype in mice heterozygous for the T356M mutation in the DAT. As previously reported, mice homozygous for this mutation show a host of behavioral changes that recapitulate a number of characteristics seen in autistic individuals, including hyperactivity, enhanced repetitive behaviors, and altered social interactions [17]. In the current work, heterozygous mice exhibited a subset of these phenotypes, with the most robust finding being hyperactivity at levels intermediate between wild‐type mice and those previously reported in homozygous animals. As this prior work showed major changes in striatal DA neurotransmission in homozygous animals, the speculation is that striatal DA signaling is altered to a lesser extent in heterozygous animals, hence giving rise to less hyperactivity. In support of this, our biochemical analyses showed relatively similar levels of DA, DA metabolites, and other monoaminergic compounds.

A second behavioral feature of homozygous mutant animals was the presence of elevated repetitive behaviors, a core feature of autism. Intriguingly, our analysis of heterozygous mice only revealed enhanced repetitive behaviors (i.e., rearing) in male mice, a finding again consistent with an intermediate phenotype. As autism incidence is currently three‐ to fourfold higher in males vs. females, such a sex‐specific difference in male mutant animals is quite interesting. Indeed, a similar sex‐specific pattern was seen in the Barnes maze performance, a measure of spatial learning. Here, male animals were slower to learn the spatial pattern of the maze, which appears to be a result of taking longer to learn the more efficient direct strategies to locate the target.

As interesting as the pattern of differences was in the heterozygous animals, the behavioral tests showed little difference from wild‐type animals. Notably, many of these were also tests that showed little difference in homozygous animals, including in measures of strength, motor coordination, learning, and anxiety. Such concordance provides further support for the intermediate phenotype being strongly tied to changes in striatal dopaminergic signaling.

Studies of heterozygous animals in mutant models that exhibit intermediate phenotypes when compared with the homozygous animals are of enormous value for a number of reasons. Perhaps foremost among these, particularly in models with strong clinical relevance (such as this DAT mutant derived from a human autism proband), is the insight they can bring to human behavior and diagnoses. It must be acknowledged that all diagnoses are ultimately dictated by exceeding some threshold for traits that make up the condition, and that many individuals may fall very close to (but not over) the diagnostic border. Indeed, it has been argued that many autism traits are highly prevalent in the general population [26, 27, 28]. Hence, those heterozygous for a given mutation (such as DAT 356M) may have traits (e.g., some degree of hyperactivity, repetitive behaviors) present in the autism diagnostic constellation but not of a number or degree necessary for formal diagnosis. Indeed, taking such a traits‐based perspective may ultimately provide greater insight into the underlying molecular, cellular, and circuit changes that give rise to autism, a view captured in the Research Domain Criteria (RDoC) framework [29, 30].

A second important reason for studying heterozygous animals is related to the possibility of the insights they can provide concerning gene × environment interactions. Thus, someone carrying a single copy of the relevant mutation may, if reared within certain environmental conditions, manifest a clinical diagnosis because environmental factors alter gene expression. Such a scenario is particularly relevant for autism as, although mutations in well over 1000 genes have been shown to confer elevated risk [2], many of these mutations are likely to be strongly influenced by environmental factors. DA regulation in humans is exquisitely sensitive to exogenous environmental toxicants, both naturally occurring and man‐made [31]. Widespread exposure to several classes of pesticides (including rotenone and paraquat) through food and water supply, therapeutic and recreational drugs, plus a range of other industrial chemicals impacts populations worldwide, albeit at different levels depending on location and local regulations [32]. Heavy metals such as iron and manganese also have direct impacts on redox balance and oxidative stress to indirectly as well as directly impact DA synthesis and regulation [33]. Products of environmental dopaminergic dysregulation are often reported as resulting in neurodegenerative disease, such as manganism, and the mechanisms underlying potential impacts on developmental disorders are less well defined [34].

Several examples highlight such gene × environment interactions in mouse models associated with autism. In Cntnap2 mutant animals, exposure of pregnant dams to poly I:C, a paradigm meant to mimic viral infection, resulted in offspring with marked increases in repetitive behaviors and impulsivity [35]. Interestingly, in animals heterozygous for Cntnap2, in utero administration of valproic acid reversed social deficits seen in non‐treated animals, showing that environmental exposure can counter the impact of carrying a single copy of the mutant gene [36]. In a study examining BTBR mice, an idiopathic model of autism, maternal age was found to contribute to an elevated likelihood for the manifestation of autism‐like behaviors, including abnormal social and repetitive behaviors [37]. Environmental enrichment has been shown to lessen autism‐like traits in FMR1 mice as well as normalize synaptic development [38].

Collectively, these studies and the current work advocate for future work in animal models heterozygous for mutations that confer elevated genetic risk for autism and in which these animals show an intermediate pattern of phenotypic changes. In addition to characterizing the specific behavioral features observed in these animals (as well as in their homozygous counterparts), challenging these animals with various environmental manipulations could shed important light on gene × environment interactions that are likely to play a major role in autism [39]. These environmental manipulations could be carried out pre‐, peri‐, and postnatally and could include changes in dietary (e.g., adding or withholding certain compounds), sensory (e.g., enhancing or restricting stimuli), and/or social (e.g., alterations in rearing/housing, drug administration) factors.

Conflicts of Interest

The authors declare no conflicts of interest.

Harris E., Paffenroth K. C., Tienda A. A., Harrison F. E., and Wallace M. T., “Sex‐Specific, Intermediate Behavioral Phenotypes in Heterozygous Dopamine Transporter Mutant DAT T356M Mice,” Genes, Brain and Behavior 24, no. 6 (2025): e70041, 10.1111/gbb.70041.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.