Preclinical Models of Attention Deficit Hyperactivity Disorder: Neurobiology,Drug Discovery,and Beyond

A Tribute

This paper is dedicated to the memory of Joseph Biederman. The research described here is a testament to Joe’s enthusiasm, ardent support, and his unique ability to translate clinical observations into preclinical research questions. It is a summary of outcomes from a nearly two-decade long scientific journey that was possible only because of Joe’s remarkable ability to persuade, persist and periodically proselytize until ideas matured into research programs that exceeded Joe’s own exacting standards.

ADHD affects nearly 10% of children and 5% of adults worldwide (Faraone, Biederman, & Mick, 2006; Faraone, Biederman, Spencer, et al., 2006; G. Xu et al., 2018). As with all other psychiatric diagnoses, diagnosis of ADHD is based on clinical evaluation of symptoms rather than specific genetic or biochemical diagnostic criteria. The key symptoms are inattention, impulsivity, and hyperactivity, with a wide variation in symptoms depending on age, sex, comorbid conditions, and context (Faraone, Childress et al., 2021). Treatment is directed at symptoms and includes pharmacological as well as behavioral approaches.

Compelling evidence points to changes in neurotransmitter signaling mechanisms, brain structure and network function as the neurobiological basis of the core symptoms of ADHD. Progress in research on ADHD is dominated by findings in clinical studies of functional and structural imaging of the human brain. Several preclinical models of ADHD have been developed (Kantak, 2022; McCarthy et al., 2022; Regan et al., 2022; Russell et al., 2005; Sagvolden & Johansen, 2012), and the models fall into two broad classes: (1) Models constructed based on specific genetic or environmental contributors to ADHD, and (2) Opportunistic models that share one or more behavioral traits that are consistent with the core symptoms of ADHD, although the models were not constructed based on known etiologies of ADHD. Both classes have contributed novel insights into the neurobiology of ADHD.

Among the most valuable preclinical approaches to examine ADHD neurobiology are those that draw from clinical expertise. It is in this area that Dr. Joseph Biederman’s contributions stand out. He had a remarkable ability to translate his observations in the clinic and his vast expertise in pediatric psychopharmacology into hypothesis-driven science amenable to testing with rigorous preclinical research. The following is an overview of research using prenatal and early postnatal nicotine exposure mouse models of ADHD that was inspired, motivated, and guided by Dr. Biederman’s insight and enthusiasm.

A Mouse Model of Developmental Nicotine Exposure

Animals models of human disorders can be valuable research tools when they fulfill construct, face, and predictive validities. The etiology of ADHD is complex and involves interactions of multiple genes and environmental contributors (Faraone, Banaschewski et al., 2021). Lack of well-defined and discrete etiologic factors makes it difficult to achieve construct validity in animal models of ADHD. One option to get around this challenge is to construct models based on known environmental contributors. As established by Biederman and others, cigarette smoking during pregnancy is a major environmental contributor to the etiology of ADHD. It is associated with a significant increase in ADHD risk for the offspring (Biederman et al., 2012; Cornelius & Day, 2009; Milberger et al., 1996; Nilsen & Tulve, 2020).

The introduction of electronic cigarettes and the phenomenon of “vaping” has added to public health concerns regarding use of nicotine containing products by pregnant women (Froggatt et al., 2020). Nearly 10% of women report using e-cigarettes shortly before pregnancy, 7% around pregnancy, and 1% in the third trimester alone (Adams et al., 2013; Kapaya et al., 2019; Mark et al., 2015; Whittington et al., 2018; T. Xu et al., 2013). The phenomenon of dual use (traditional and e-cigarettes) is 38% in the 3 months before pregnancy, 8% in the final trimester, and 12% in the 2 to 6 months after delivery. Many factors, including a false sense of “safety” promote e-cigarette use. For example, 45% of pregnant women reported that they thought e-cigarettes might help quit smoking, and another 45% viewed the practice as less harmful than traditional cigarettes. Unfortunately, the perceptions of “safety” may be shared by healthcare professional as well (Kandra et al., 2014).

Based on these observations, we developed mouse models of prenatal and early postnatal nicotine exposure as preclinical models of ADHD. We describe below an evaluation of the models to establish their construct, face and predictive validities, based on behavioral, neuroanatomical, and pharmacological findings.

Construct Validity: Rodent models of nicotine exposure use inhalation, oral, intravenous, intraperitoneal, or subcutaneous (osmotic pump) routes of nicotine administration. Among these, inhalation of cigarette smoke or electronic cigarette aerosol most closely mimics human nicotine exposure. However, we chose to deliver nicotine to the mice in their drinking water, because it is likely the least stressful experience for the mouse (Wickström, 2007), especially for pregnant and nursing mice, and certainly a convenient method for the experimenter. Nicotine delivery via the drinking water to pregnant mice produces sustained increases in nicotine levels in the maternal and fetal circulation. Interestingly, although maternal serum nicotine levels fluctuate during cigarette smoking and vaping, fetal nicotine levels remain relatively constant, which supports construct validity of the oral nicotine delivery model (Centner et al., 2020; Wickström, 2007). Moreover, the sustained and relatively constant nicotine exposure in the oral nicotine delivery mouse model mimics nicotine exposure via tobacco chewing and nicotine replacement therapy (Pauly & Slotkin, 2008; Slotkin, 2008).

Female mice were provided with drinking water containing nicotine (200 µg/ml). In some studies, the artificial sweetener saccharin was added to the nicotine-containing drinking water (2% saccharin) to mask the bitter taste of nicotine. We included two sets of controls: mice consuming drinking water containing saccharin only, and plain drinking water (without nicotine or saccharin). The inclusion of the two control groups facilitated discrimination between the effects of saccharin and nicotine. Following 2 to 3 weeks of daily exposure of breeding age (6–8-week-old) female mice to the different drinking waters, the mice were bred with male mice that had been maintained on plain drinking water (i.e., nicotine naïve) (Figure 1).

OPEN IN VIEWER

Figure 1.

Prenatal and early postnatal nicotine exposure mouse models of ADHD. We exposed female mice from the C57BL/6 or Swiss Webster (SW) strains to drinking water containing nicotine (200 µg/ml). Following 2 to 3 weeks of acclimation, the mice were bred with male mice that had been raised on plain drinking water. The nicotine exposure continued throughout pregnancy. In the prenatal exposure mouse model, on the day of birth, the litter was cross fostered to nicotine naïve dams, and the nicotine exposure ended. In the pre- and early postnatal exposure model, the offspring were not cross fostered, and the nicotine exposure continued until weaning. Upon weaning the mice received plain drinking water. In both the models, when the offspring reached approximately 2 to 3 months of age, behavioral, neuroanatomical, and pharmacological assays were performed to validate the mouse model. Upon validation, the models were used for drug discovery, examination of the interactions between untreated ADHD and repeated mild traumatic brain injury (mTBI) and transgenerational transmission of ADHD-associated traits. A parallel set of female mice were raised on plain drinking water as controls. In some studies, the nicotine containing drinking water was sweetened with 2% saccharin, and a separate set of female mice were raised on drinking water containing 2% saccharin alone, to control for the effects of saccharin.

In our initial studies, we used a prenatal nicotine exposure model (Figure 1). In this model, the nicotine exposure of the females began prior to breeding to help the mice to acclimate to drinking nicotine-containing water prior to breeding. The exposure continued throughout pregnancy and ceased on the day of parturition, when the offspring from each dam were cross fostered to dams that had littered on the same day as the biological dams, and that had been maintained on plain drinking water (Zhu et al., 2012, 2014, 2017). The foster dams continued to receive plain drinking water. The cross fostering was performed to preempt potential poor maternal care in the nicotine-exposed mice (Pauly et al., 2004; Schneider et al., 2011; Vaglenova et al., 2004). However, offspring in all three drinking water groups were cross fostered to eliminate cross fostering stress as an unaccounted variable in the study. Thus, in the prenatal nicotine exposure model, the offspring were exposed to nicotine + saccharin or saccharin alone via the mothers from conception until birth (Figure 1), which is roughly equivalent to the first two trimesters of human pregnancy (Clancy, Finlay et al., 2007; Clancy, Kersh et al., 2007; Semple et al., 2013).

As the research continued, the model was modified to mimic maternal cigarette smoking and vaping during the full 3 trimesters of human pregnancy. To do this, we extended the nicotine exposure to include the initial 3-week postpartum period (pre-weaning period for the offspring), which corresponded to the third trimester of human pregnancy (Figure 1). In this extended exposure model, the offspring in all three drinking water groups remained with the biological mothers (were not cross fostered) from birth until weaning, when the offspring were approximately 3-weeks of age (M. M. Martin et al., 2020; L. Zhang et al., 2018, 2021, 2021, 2022). The offspring received plain drinking water from weaning onward.

Mice from the inbred strain C57BL/6 were used in many of our studies (L. Zhang et al., 2018, 2021, 2021, 2022; Zhu et al., 2017). However, in studies that examined the GABA neurotransmitter system, we used the Swiss Webster outbred strain of mice (M. M. Martin et al., 2020). This mouse line was engineered to express the GAD67-GFP transgene, which imparted green fluorescence to all GABA neurons, making analysis of GABA neurons in histological sections of the brain more convenient by not requiring immunohistochemistry (M. M. Martin et al., 2020; McCarthy et al., 2011, 2022; McCarthy & Bhide, 2012).

In our later studies, we discontinued the use of saccharin to sweeten the nicotine containing water, because we discovered that consumption of the nicotine containing drinking water without saccharin was not significantly different from consumption of plain drinking water (M. M. Martin et al., 2020; McCarthy et al., 2018).

We found that the serum cotinine levels in female mice receiving the nicotine-containing drinking water (200 µg/ml) daily were approximately 100 ng/ml (M. M. Martin et al., 2020), which corresponds to the serum cotinine levels in other nicotine exposure rodent models (Pauly et al., 2004) as well as in moderate to heavy cigarette smokers (Benowitz et al., 2009).

Face Validity: We used behavioral, neuroanatomical, neuroimaging, and pharmacological methods to assess face validity of our mouse models.

Behavioral analyses: Male and female C57BL/6 offspring from the prenatal nicotine + saccharin drinking water group showed significant increases in spontaneous locomotor activity, a proxy for hyperactivity in humans, which was analyzed at hourly intervals over a 20-hr period in the home cage of the mouse (Zhu et al., 2012). In addition, male and female mice from this group showed significant attention deficit in the object based attention assay, and males but not females showed significant motor impulsivity (cliff avoidance reflex assay) and working memory deficit (Y-maze assay, Zhu et al., 2012). The mice from the saccharin-only group did not show significant changes in any of the behaviors, demonstrating that the behavioral effects were due to the nicotine exposure (Zhu et al., 2012).

The extended pre- and early postnatal (until weaning) exposure C57BL/6 mouse model did not show significant changes in spontaneous locomotor activity or motor impulsivity in male or female offspring from the nicotine + saccharin group (L. Zhang et al., 2018). However, we observed significant deficits in attention and working memory in male (but not female) offspring from this group (L. Zhang et al., 2018, 2021). The saccharin-only exposure did not produce significant behavioral changes.

Thus, comparison of the behavioral changes between the prenatal only exposure and prenatal + early postnatal exposure models showed that the duration of the nicotine exposure influenced the behavioral phenotypes. In fact, the extended exposure impacted fewer traits compared to the prenatal only exposure. The reason for the variability remains unclear. The timing of the nicotine exposure vis a vis the timing of the developmental events (neurogenesis, gliogenesis, neuronal migration etc.) may impact the behavioral outcomes (Alkam, Kim, Hiramatsu, et al., 2013; Alkam, Kim, Mamiya, et al., 2013; McCarthy et al., 2022). Another possibility is that in our models, the extended exposure may have triggered adaptive responses such that the postnatal exposure may have mitigated some of the effects of the prenatal exposure (review in McCarthy et al., 2022; Polli & Kohlmeier, 2020; L. Zhang et al., 2018).

The pre- and early postnatal extended nicotine exposure model using the Swiss Webster strain of mouse did not show significant effects of the nicotine exposure on spontaneous locomotor activity or working memory (M. M. Martin et al., 2020). However, in this model we observed increased approach behavior in an approach-avoidance assay (elevated plus maze), demonstrating risk-taking behavior as a result of the nicotine exposure. This behavior was not present in the C57BL/6 pre- and early postnatal nicotine exposure model. These data demonstrate that the behavioral changes produced by the nicotine exposure of the developing brain are influenced by the background strain of the mouse (Polli & Kohlmeier, 2020).

Neuroanatomy: We found that the male C57BL/6 offspring from the prenatal nicotine + saccharin group showed significant reductions in the volume of the medial prefrontal cortex (Zhu et al., 2012). Interestingly, the Swiss webster mouse model of pre- and early postnatal nicotine exposure did not show these structural changes (M. M. Martin et al., 2020), once again underscoring the role of the strain differences. A diffusion tensor imaging study of the C57BL/6 offspring from the pre- and early postnatal extended nicotine exposure model showed significant reductions in microstructural integrity (i.e., reduced fractional anisotropy) in the medial prefrontal cortex, dorsal striatum and ansiform lobule of the cerebellum in male mice (females were not analyzed [McCarthy et al., 2022]). These structural changes are consistent with structural changes reported in the brains of treatment naïve ADHD subjects (Makris et al., 2009, 2015; Seidman et al., 2006, 2011; P. Shaw et al., 2006; M. Shaw et al., 2012).

Neurotransmitters: We found significant reductions in the tissue content of dopamine and its synaptic turn-over in the frontal cortex of male C57BL/6 mice in the prenatal nicotine + saccharin group (Zhu et al., 2012). Female mice were not examined. In the pre- and early postnatal extended nicotine exposure model, in vivo microdialysis in the frontal cortex of awake, freely-moving C57BL/6 mice showed a significant reduction in dopamine release at baseline in male mice from the nicotine + saccharin group (L. Zhang et al., 2021). Female mice were not examined. Extended nicotine exposure of Swiss Webster mice showed significant reductions in the GABA -to-projection (glutamatergic) neuron ratio in the prefrontal cortex and medial prefrontal cortex in males and females (M. M. Martin et al., 2020), suggesting downregulation of inhibitory neurotransmission. Thus, nicotine exposure of the developing brain produced significant effects on the dopamine and GABA neurotransmitter systems.

Predictive validity: Initially, we examined whether the stimulant compound methylphenidate mitigated the locomotor activity, motor impulsivity, attention deficit and working memory deficit in the nicotine + saccharin exposed group and increased the dopamine content in the frontal cortex to the levels observed in the control groups.

Responses to methylphenidate: We established that 0.75 mg/kg methylphenidate oral dose was the mouse equivalent of the therapeutic dose of methylphenidate used in the treatment of ADHD, based on a comparison of plasma concentrations of D-methylphenidate (Balcioglu et al., 2009). Specifically, a single oral administration of 0.75 mg/kg methylphenidate produced plasma D-methylphenidate levels of 6 to 10 ng/ml within 15 min of administration, which is the plasma concentration of D-methylphenidate achieved in ADHD patients (Kuczenski & Segal, 2005; Patrick & Markowitz, 1997; Swanson & Volkow, 2002). Next, we examined the effects of a single administration of 0.75 mg/kg methylphenidate (intraperitoneal [i.p.]) on behavioral parameters in the C57BL/6 prenatal and pre- and early postnatal extended nicotine exposure models. We found that the methylphenidate administration alleviated the hyperactivity, reduced the motor impulsivity, and restored attention and working memory to control levels (L. Zhang et al., 2021; Zhu et al., 2012, 2017). In addition, the methylphenidate administration increased dopamine tissue content in the prefrontal cortex of prenatal exposure model by nearly three fold, and dopamine and noradrenaline release by nearly 50% and 20%, respectively measured by in vivo microdialysis in the frontal cortex of the pre- and early postnatal extended nicotine exposure model (L. Zhang et al., 2021; Zhu et al., 2017). Interestingly, the methylphenidate-induced increase in dopamine release was temporally coincident with improvements in attention and working memory (L. Zhang et al., 2021) consistent with the “therapeutic” effects of methylphenidate.

Having established that the mouse models responded to methylphenidate in a manner that was consistent with the responses expected in an ADHD preclinical model, we used the model to search for novel compounds with more favorable therapeutic characteristics. A review of the literature showed that the kappa opioid system increased catecholamine release in the frontal cortex. Therefore, we postulated that the selective kappa opioid receptor (KOR) antagonist norbinaltorphimine (norBNI) would increase dopamine release in the frontal cortex and produce behavioral responses comparable to those produced by methylphenidate. We rested our hypothesis with a view toward investigating if norBNI could be a potential non-stimulant treatment for ADHD.

Mouse Models and Drug Discovery: A Novel Non-stimulant Treatment for ADHD

The pre- and early postnatal nicotine exposure C57BL/6 mouse model was used to evaluate the effects of the selective, long-acting, KOR antagonist norBNI on behavioral traits and frontal cortical neurotransmitter release (L. Zhang et al., 2021). KORs are G-protein coupled receptors activated by the endogenous neuropeptide dynorphin (DePaoli et al., 1994; Georges et al., 1998; Liu-Chen, 2004; Simonin et al., 1995; Yasuda et al., 1993). KORs are presynaptic to midbrain dopamine and noradrenaline neuron axon terminals in the frontal cortex, where the KORs exert negative feedback on synaptic dopamine and noradrenaline release (Berger et al., 2006; Fuentealba et al., 2006; Margolis et al., 2006; Werling et al., 1987). Therefore, selective antagonism of the KORs can be expected to increase frontal cortical dopamine and noradrenaline synaptic release which, in turn, might alleviate the behavioral deficits observed in this mouse model (L. Zhang et al., 2021; Zhu et al., 2017).

Our studies found that a single administration of norBNI (20 mg/kg; i.p.) increased the synaptic release of dopamine and noradrenaline in the frontal cortex of male mice (L. Zhang et al., 2021). Female mice were not evaluated. The increase commenced 2-hr following the norBNI administration and lasted for 5 to 6 hr. In addition, the norBNI administration restored working memory and attention in male mice from the nicotine + saccharin group to the levels observed in the control group (L. Zhang et al., 2021).

A head-to-head comparison between norBNI (20 mg/kg; i.p.) and methylphenidate (0.75 mg/kg; i.p.) showed that the onset of the effects of norBNI on neurotransmitter release and behavior lagged those of methylphenidate by approximately 2 hr, and lasted 2 to 3 times longer (L. Zhang et al., 2021). The long-lasting effects of norBNI are consistent with its long-acting receptor antagonism (Bruchas et al., 2007; Melief et al., 2011). The peak levels of dopamine in the frontal cortex achieved by the single administration of norBNI were nearly two fold higher than those achieved by a single methylphenidate administration. However, the improvements in object based attention and working memory produced by the two compounds were comparable to each other. The norBNI administration did not produce significant changes in dopamine tissue content in the ventral striatum, a key component of the brain’s reward circuitry, offering preliminary evidence that norBNI may have very low abuse potential (L. Zhang et al., 2021).

In summary, we demonstrated that mouse models of prenatal and early postnatal nicotine exposure carry significant construct, face, and predictive validities. The behavioral traits were dependent on the duration of the nicotine exposure and the background strain of the mouse. The mouse models demonstrated that the nicotine exposure produced changes in dopamine, noradrenaline, and GABA neurotransmitter systems. The mouse models facilitated identification of the KOR antagonist norBNI as a potential non-stimulant compound for the treatment of ADHD.

Next, we used the mouse models to investigate the potential impact of untreated ADHD on behavioral outcomes following traumatic brain injury.

Preclinical Investigations of Synergistic Interactions Between ADHD and Repeated, Mild Traumatic Brain Injury

Biederman and others reported that ADHD is a risk factor for traumatic brain injury (Adeyemo et al., 2014; Alosco et al., 2014; Biederman et al., 2015; Cook et al., 2020; Iverson et al., 2016), and that untreated ADHD could negatively impact the behavioral outcomes and recovery following repeated episodes of mild traumatic brain injury (mTBI) such as those suffered by athletes (Biederman et al., 2015; Cook et al., 2020; Iverson et al., 2016, 2020; Kaye et al., 2019). We used our C57BL/6 pre- and early postnatal extended nicotine exposure mouse model to test the hypothesis that untreated ADHD contributed to poor behavioral outcomes following repeated mTBI.

We subjected mice to repeated closed head mTBI once daily for 5 consecutive days to mimic repeated concussions (L. Zhang et al., 2021). Male mice from the pre- and postnatal nicotine + saccharin, saccharin-only and plain drinking water groups were assigned to the mTBI or sham (anesthesia only) groups. Object based attention, novel object recognition memory, spatial working memory, and depression-like behavior were analyzed 1 day and 2 weeks following repeated mTBI. Repeated mTBI produced a transient (present at 4-days but absent at 17-days) attention deficit in the control groups but did not exacerbate the deficits present at baseline (prior to mTBI) in the nicotine + saccharin group (L. Zhang et al., 2021). The mTBI-associated attention deficit in the control groups observed at 4-days was comparable to the attention deficit in the nicotine + saccharin group at 4- and 17-days. Moreover, there was a transient increase in depression-like behavior in the nicotine + saccharin group subjected to the repeated mTBI, suggesting a synergistic effect (L. Zhang et al., 2021). These findings suggested that untreated ADHD may be a risk factor for transient depression following repeated mTBI and that repeated mTBI may be a risk factor for transient attention deficit.

In a second set of studies, we performed repeated closed head mTBI in awake, unanesthetized male C57BL/6 mice from the pre- and early postnatal nicotine exposure model to model concussions in humans. Anesthesia as a variable was eliminated in this study. The mTBI was repeated three times daily for 7 days and was “more severe” than in the previous study. The mice in the nicotine + saccharin group took longer to regain consciousness following the repeated mTBI and showed transient novelty-seeking and depression-like behaviors (L. Zhang et al., 2022). As was the case in the anesthetized mouse study, the mice in the control group showed a transient attention deficit following the repetitive mTBI. However, the attention deficit in the control group in this study persisted longer than that in the previous study following the final episode of mTBI. It was observed at 5- and 18-days compared to only at 4-days in the previous study. The magnitude of the deficit was comparable between the control and nicotine +s saccharin groups in the present study as well. Collectively, our findings from the mouse models support the notion that untreated ADHD may be a risk factor for poor cognitive outcomes following concussions.

Transgenerational Transmission of ADHD-Associated Traits in the Mouse Models

Cigarette smoking during pregnancy and the early postpartum period continues to be a major public health concern. Nearly 7% of pregnant women report smoking cigarettes, and 16% do so in the initial postpartum period (Tong, Dietz, Farr, et al., 2013; Tong, Dietz, Morrow, et al., 2013). With the availability of e-cigarettes, nicotine use overall has increased significantly (US-DHSS, 2016).

Accumulating evidence from multiple fields of biology demonstrates that environmental exposures can impact the directly exposed individuals as well as their descendants in multiple generations. In other words, the impact of the environmental exposures can manifest as traits transmitted transgenerationally along maternal and paternal lines of descent, and these environment-induced traits can linger in multiple generations even after the termination of the environmental exposures (Bell & Hellmann, 2019; Bohacek & Mansuy, 2015; Donelan et al., 2020; Goldberg & Gould, 2019; Maurer et al., 2022; McCarthy & Bhide, 2021; Nilsson et al., 2018; Skinner, 2014; Vassoler et al., 2014). Under this scenario, maternal smoking or vaping during pregnancy could produce adverse effects not only for the mother and her children, but also in multiple generations of her descendants. Since cigarette smoking during pregnancy was highly prevalent in the 1950s and 1960s, and vaping during pregnancy is on the rise, evidence in support of transgenerational transmission of traits such as hyperactivity and attention defect from the prenatally nicotine exposed individuals could expand our understanding of the environmental contributors to ADHD in the current and future generation.

We were among the first to demonstrate that the effects of nicotine exposure during the prenatal period were not limited to the directly exposed generation but were transmitted to multiple generations descending from the directly exposed generation (Zhu et al., 2014). In our prenatal nicotine + saccharin exposure C57BL/6 mouse model we showed that locomotor hyperactivity was manifested in at least two generations descending from the prenatally nicotine exposed mice, and that the transgenerational transmission occurred via the maternal line of descent (Figure 2; Zhu et al., 2014). Moreover, the hyperactivity in each generation was alleviated following a single administration of methylphenidate (0.75 mg/kg, i.p.) demonstrating that the dopamine neurotransmitter mechanisms impacted by the prenatal nicotine exposure were transmitted transgenerationally as well (Zhu et al., 2014). Since then, a number of studies in human subjects (Golding et al., 2017, 2020) as well as animal models (Buck et al., 2019; Liu et al., 2020; Renaud & Fountain, 2016; Yohn et al., 2015) have shown transgenerational transmission of traits following in utero nicotine exposure.

OPEN IN VIEWER

Figure 2.

Transgenerational transmission of ADHD-associated traits in mouse models of prenatal (A) and paternal (B) nicotine exposure. In the prenatal paradigm (A), female in C57BL/6 mice were maintained on nicotine-containing (200 µg/ml) drinking water sweetened with 2% saccharin. Following 2 to 3 weeks of acclimation, the mice were bred with nicotine-naïve males. The nicotine treatment continued throughout pregnancy. On the day of birth, the offspring were cross fostered to dams maintained on plain drinking water. The nicotine-exposed female dams represent the founder F0 generation, and the prenatally nicotine exposed offspring the F1 generation. Male and female mice in the F1 generation were hyperactive. F1 mice were mated with nicotine naïve partners (only F1 female is shown here) to produce the F2 generation. The F2 mice derived from F1 female founders (but not male F1 founders) were hyperactive. The F2 mice were mated with nicotine naïve partners (again, only F2 female is shown here) to produce the F3 generation. The F3 mice derived from F2 females were hyperactive. Separate sets of F1, F2, and F3 mice derived from the saccharin-only and plain drinking water (neither nicotine nor saccharin) exposed F0 females were studied in parallel. In the paternal exposure paradigm (B) adult male in C57BL/6 mice were exposed to drinking water containing nicotine (200 µg/ml) for 12-weeks (F0 generation) and bred with nicotine naïve female partners. The male and female offspring in the F1 generation displayed reversal learning deficits. Male and female F1 mice were bred with nicotine naïve partners to produce the F2 generation (only F1 female is shown). Mice in the F2 generation also showed reversal learning deficits. F1, F2, and F3 generations were studied in the prenatal exposure paradigm (A) whereas only F1 and F2 generations were analyzed in the paternal exposure paradigm (B). In the prenatal paradigm, the F1 germ cells were directly exposed to nicotine in utero, which meant that the hyperactivity observed in the F2 generation may have been the result of the direct effects of nicotine on the founder F1 germ cells. However, the germ cells of the F2 founders were not directly exposed to nicotine at any time. Therefore, the hyperactivity in the F3 generation is transgenerational—inherited in the absence of direct exposure of the immediate founder’s somatic or germ cells to nicotine. In the paternal exposure paradigm, the F0 germ cells were directly exposed to nicotine, but the F1 germ cells were not. Therefore, the reversal learning deficit in the F2 generation was due to transgenerational transmission. Hyperactivity was the only trait analyzed in the prenatal paradigm (A). Although several other parameters were analyzed in the paternal exposure paradigm, only the reversal learning deficit is shown (B) for the sake of simplicity. Nicotine exposed mice and their descendants are shaded gray.

Next, we examined whether transgenerational transmission of the effects of nicotine exposure occurred along the paternal line of descent as well. We reported that paternal cigarette smoking around the time of conception of the offspring increased the risk for ADHD in the offspring, even though the mothers had not smoked cigarettes (Biederman et al., 2020). Our findings were consistent with others in the clinical literature (Altink et al., 2009; Joya et al., 2014). Based on these findings, we developed a mouse model of paternal nicotine exposure (Figure 2) to further characterize transgenerational transmission of ADHD associated traits along the paternal line of descent. In addition, the mouse model offered us the opportunity to examine nicotine-induced epigenetic changes in the DNA of spermatozoa of the founder males as candidate mechanisms of the heritability of traits.

We exposed male C57BL/6 mice to nicotine (200 μg/mL) in drinking water for 12-weeks and bred the mice with females consuming plain drinking water (McCarthy et al., 2018). Male and female offspring (F1 generation) were bred with nicotine-naïve partners to produce the F2 generation (Figure 2). We found that male and female mice in the F1 generation derived from the nicotine-exposed males showed significant increases in spontaneous locomotor activity and significant deficits in reversal learning. The male F1 mice also showed significant deficits in attention, brain monoamine content, and dopamine receptor mRNA expression (McCarthy et al., 2018). Male mice in the F2 generation derived from paternally nicotine exposed female F1 mice showed significant deficits in reversal learning. These data demonstrated that nicotine exposure of male mice produced behavioral changes consistent with ADHD in at least two generations of descendants. Analysis of epigenetic changes in the spermatozoa of the nicotine-exposed male founders (F0) showed significant changes in global DNA methylation and DNA methylation at promoter regions of the dopamine D2 receptor gene, suggesting that nicotine-induced epigenetic changes in the spermatozoa are associated with heritability of the traits.

Our findings are consistent with other reports in the literature on heritable effects of paternal nicotine exposure on offspring behavior in multiple generations that are associated with nicotine-induced epigenetic changes in the spermatozoa of the founder males in animal models (Hawkey et al., 2019; Maurer et al., 2022; McCarthy et al., 2022; McCarthy & Bhide, 2021; Wang et al., 2022; Zeid et al., 2021; Zeid & Gould, 2023; M. Zhang et al., 2020).

Mouse Models and Drug Discovery: An Abuse Deterrent Formulation of Methylphenidate

In this section, we describe our research on pharmacological interactions between opioid receptors and methylphenidate, which in mouse models (not mouse models of ADHD) and led to clinical trials of abuse deterrent formulations of methylphenidate for the treatment of ADHD.

Methylphenidate is among the most frequently prescribed stimulants for treatment of ADHD in children and adults. Although safety and efficacy of methylphenidate in pediatric and adult ADHD patients is well established concerns persist about its potential negative side effects—especially abuse, misuse, diversion, and addiction (Brown et al., 2005; Butler et al., 2021; McCabe et al., 2023). The concerns about methylphenidate abuse potential, addiction and diversion often play a role in the decision of patients, parents, or physicians who opt against treatment. Since the prevalence of substance use disorders is high among individuals with ADHD, clinicians are confronted with the dilemma of having to prescribe a treatment that carries abuse potential to ADHD patients. As a result, it is estimated that only 25% of adults with an ADHD diagnosis seek treatment and approximately 40% of pediatric ADHD is believed to be untreated (Huntley et al., 2012; Rasmussen & Levander, 2009).

Stimulants such as methylphenidate and amphetamine produce their therapeutic effects by augmenting synaptic release of dopamine or blocking its reuptake at the synapse (Arnsten, 2006; Kuczenski & Segal, 2001; Schiffer et al., 2006; Volkow et al., 2005), and thereby alleviate the hypodopaminergic state associated with ADHD. However, supra-therapeutic doses of methylphenidate (e.g., ten times the prescribed dose) results in the subjective feeling of euphoria or “liking” an effect linked to abuse and addiction (Kollins et al., 2001; W. R. Martin et al., 1971; Parasrampuria et al., 2007; Spencer et al., 2006). The euphoria is due to the rapid and robust elevation of dopamine (and to some extent also noradrenaline) in the brain. To counter such rapid elevation, sustained or extended release stimulants formulations were developed (e.g., Concerta^®, Metadate^®, Ritalin LA^®). Clinical studies show lower euphoria with extended-release formulations, suggesting less abuse potential of these formulations (Parasrampuria et al., 2007; Spencer et al., 2006). However, this pharmaceutical approach has not reduced stimulant abuse because most abuse occurs via tampering and use of crushed preparations (Bright, 2008; Ciccarone, 2011; Teter et al., 2006), can negate the slow-release mechanisms in some formulations. Moreover, the immediate-release formulations are less expensive than the slow-release formulations; a factor that may be contributing to the prevalence of the immediate-release formulations.

Another approach toward addressing the abuse potential of stimulant medications is to use non-stimulants such as alpha 2 adrenergic receptor agonists. However, these drugs produce side effects including sedation and cardio-vascular complications. More significantly, the adrenergic receptor agonists (e.g., guanfacine) are often used along with the stimulants to achieve the full therapeutic benefit (Arnsten & Jin, 2012; Biederman et al., 2008; Wilens et al., 2017).

We used a mouse model to develop an abuse-deterrence strategy for immediate-release methylphenidate that was based on central pharmacological mechanisms rather than pharmaceutical formulation techniques. We established that a supratherapeutic dose of methylphenidate (7.5 mg/kg; 10 times greater than the therapeutic equivalent dose) produced reinforcing effects (a proxy for addiction) in a conditioned place preference assay in male C57BL/6 mice (Zhu et al., 2011). The therapeutic equivalent dose did not produce this effect. In addition, receptor-G protein coupling assays showed that the 7.5 mg/kg dose and not the 0.75 mg/kg dose activated the µ opioid receptor in the ventral striatum, a key center in the brain’s reward circuits (Trigo et al., 2010; Zhu et al., 2011). The µ opioid receptor activation in reward circuits of the brain is a mechanistic substrate for reinforcement (Trigo et al., 2010).

These observations raised the possibility that pharmacological antagonism of the µ opioid receptor could mitigate the reinforcing effects of supratherapeutic doses of methylphenidate. We found that administration of naltrexone (1, 5 or 10 mg/kg; i.p.), a mixed opioid receptor antagonist, mitigated the reinforcing effects of 7.5 mg/kg methylphenidate in the conditioned place preference assay, and prevented the activation of the µ opioid receptor (10 mg/kg, i.p. naltrexone) in the ventral striatum (Zhu et al., 2011).

Based on these preclinical data, we performed a 6-week randomized placebo-controlled clinical trial of naltrexone (50 mg/day) in adults with ADHD receiving a long-acting formulation of methylphenidate (Spheroidal Oral Drug Absorption System, titrated over 3 weeks to a dose of 1 mg/kg/day). The subjects were preselected for having experienced euphoria with an oral test dose of 60 mg immediate release methylphenidate. We found that naltrexone (50 mg) significantly reduced the euphoric effects of 60 mg immediate-release methylphenidate (“Liking Effect” on the Drug Rating Questionnaire, Subject Version) in the heightened-risk titration phase during the initial 3 weeks of the trial compared to placebo (Spencer, Bhide, Zhu, Faraone, Fitzgerald, Yule, Uchida, Spencer, Hall, Koster, Feinberg, et al., 2018). Comparable findings of were reported for a combination of naltrexone and amphetamine (Jayaram-Lindström et al., 2008; Jayaram-Lindström et al., 2004). We did not examine the effects of naltrexone on abuse liability of the long-acting formulation of methylphenidate.

In addition, we found that co-administration of naltrexone (50 mg) did not interfere with the clinical benefits of the long-acting formulation of methylphenidate (based on Adult ADHD Investigator Symptom Report Scale) throughout the 6-week trial (Spencer, Bhide, Zhu, Faraone, Fitzgerald, Yule, Uchida, Spencer, Hall, Koster, & Biederman, 2018). These findings demonstrate that a combination of naltrexone and immediate release methylphenidate results in significant reduction in the abuse liability of the immediate release methylphenidate without a reduction in its therapeutic benefit. Moreover, these observations illustrate successful clinical validation of a discovery made using a preclinical model and underscore the potential for preclinical models to contribute to ADHD drug discovery (Bhide et al., 2021).

Conclusion

The research described here demonstrates that mouse models of developmental nicotine exposure can be valuable tools for research in ADHD. The validity of preclinical models of ADHD is not foolproof because of the heterogeneous etiology and symptomatology of ADHD. Despite these limitations, the mouse models uniquely accomplish hypothesis-driven mechanistic research, which can contribute to significant advances in our understanding of the neurobiology of ADHD and facilitate drug discovery. The research was possible because of Dr. Biederman’s unyielding enthusiasm, ardent support, and his unique ability to translate observations in the clinic into hypothesis-driven preclinical research, and his ability to keep us motivated throughout a nearly two decades long, remarkably rewarding journey.