The D2-family receptor agonist bromocriptine but, not nicotine, reverses NMDA receptor antagonist-induced working memory deficits in the radial arm maze in mice

Abstract

Hypofunction of the NMDA receptor (NMDAr) may underlie cognitive deficits associated with schizophrenia and other psychiatric conditions including working memory (WM) impairments. Given that these deficits link closely to functional outcome, treatments remediating such deficits require identification. NMDAr hypofunction can be modeled via treatment with the antagonist MK-801. Hence, the present study determined whether cholinergic or dopaminergic agonists attenuate MK-801-induced WM deficits in mice. WM was assessed in male C57BL/6 mice trained on an automated 12-arm radial arm maze (RAM) paradigm, wherein rewards were delivered after the first but, not after subsequent entries into WM arms (8/12) and never delivered for entries into reference memory (RM) arms (4/12). Mice were then treated with MK-801 (vehicle or 0.3 mg/kg) and nicotine (vehicle, 0.03 or 0.30 mg/kg) in a cross-over design. After a 2-week washout, mice were then retested with MK-801 and the dopamine D₂-family receptor agonist bromocriptine (vehicle, 3 or 10 mg/kg). In both experiments, MK-801 reduced WM span and increased RM and WM error rates. Nicotine did not attenuate these deficits. In contrast, a bromocriptine/MK-801 interaction was observed on WM error rate, where bromocriptine attenuated MK-801 induced deficits without affecting MK-801-induced RM errors. Additionally, bromocriptine produced the main effect of slowing latency to collect rewards. Hence, while NMDAr hypofunction-induced deficits in WM was unaffected by nicotine, it was remediated by treatment with the dopamine D₂-family agonist bromocriptine. Future studies should determine whether selective activation of dopamine D₂, D₃, or D₄ receptors remediate this NMDAr hypofunction-induced WM deficit.

Introduction

Hypofunction of the n-methyl-d-aspartate glutamate receptor (NMDAr) has long been implicated in the etiology of schizophrenia (Olney, Newcomer, & Farber, 1999). The association between NMDAr hypofunction and schizophrenia emerged from the observation that NMDAr antagonist drugs, such as phencyclidine and ketamine, produce a range of effects similar to symptoms observed in schizophrenia (Javitt and Zukin, 1991, Krystal et al., 1994, Lahti et al., 1995). Perturbations of NMDAr expression have also been observed in the brain of afflicted individuals (Kristiansen, Huerta, Beneyto, & Meador-Woodruff, 2007) and many susceptibility genes identified for schizophrenia have direct or downstream effects on NMDAr expression/function (Harrison & Weinberger, 2005). These observations led to the suggestion that NMDAr hypofunction represents a convergence point for the progression and symptoms of schizophrenia, such as cognitive deficits (Snyder & Gao, 2013). Dysregulation of NMDAr expression or function does not appear to be limited to patients with schizophrenia however, as NMDAr hypofunction has also been implicated in the pathogenesis of attention deficit hyperactivity disorder (ADHD) (Chang, Lane, & Tsai, 2014) and bipolar disorder (Fountoulakis, 2012). Identifying treatments that remediate NMDAr hypofunction deficits in cognition could therefore prove useful for treating numerous disorders.

Consistent with the potential link between NMDAr hypofunction and cognition, visual and spatial working memory task performance is disrupted by NMDA antagonist drugs. In rats, radial arm maze (RAM) performance is disrupted by NMDAr blockade produced by MK-801 (Caramanos and Shapiro, 1994, Levin et al., 1998, Levin et al., 2014, Ward et al., 1990), CPP (Ward et al., 1990), APV/AP5 (Caramanos and Shapiro, 1994, Smith-Roe et al., 1999, Yamada et al., 2015), and phencyclidine (Kesner, Hardy, & Novak, 1983). These effects occur after training to stability and in RAM acquisition (Caramanos and Shapiro, 1994, Keseberg and Schmidt, 1993, Malenfant et al., 1991). In mice, similar RAM impairments were observed with MK-801 (Mutlu et al., 2012) and CPP (Wilson, Puolivali, Heikkinen, & Riekkinen, 1999). The use of mice enables evaluation of effects of genetic manipulation on performance, with evidence that mice lacking the 2A subunit NMDAr (NR2A-/-) exhibit impaired WM performance on the RAM (Bannerman et al., 2008). Similarly, impaired performance is also observed in NR1 subunit knockdowns (Dzirasa et al., 2009, Niewoehner et al., 2007). Hence, reduced function of the NMDAr in mice negatively impacts WM performance in the RAM.

Acute or chronic administration of NMDAr antagonists also impairs performance on non-spatial behavioral tasks of animal cognition. Non-competitive antagonists including MK-801, phencyclidine, and ketamine, and the competitive antagonist CPP, all produce substantial impairment of working memory (WM) as measured by the odor span task in rats (Davies et al., 2013, Galizio et al., 2016, Galizio et al., 2013, MacQueen et al., 2011, MacQueen et al., 2016, Mathews et al., 2018). These effects are restricted to the primary working memory (WM) measures of the task while sparing long-term odor memory (reference memory; RM). The odor span procedure has also been adapted for mice (Young et al., 2007, Young et al., 2009) and enhanced performance is observed in subjects genetically modified to overexpress the NR2B subunit of the NMDAr in forebrain (Cui et al., 2011), and mice treated with nicotine (Young et al., 2007). Hence, NMDAr hypofunction reliably impairs WM span capacity, consistent with WM deficits seen in patients with schizophrenia (Kern et al., 2011, Lett et al., 2014), bipolar disorder (Soraggi-Frez, Santos, Albuquerque, & Malloy-Diniz, 2017) and ADHD (Martinussen & Tannock, 2006).

Patients with schizophrenia exhibit WM deficits that are specific to span capacity as opposed to maintenance (Gold et al., 2018, Gold et al., 2010). Notably, high rates of tobacco smoking are observed in psychiatric conditions (2–4 times greater than in the general population; Jackson et al., 2015, McClernon and Kollins, 2008, Winterer, 2010). Importantly, acute nicotine administration enhances cognitive performance among healthy non-smokers (Heishman, Snyder, & Henningfield, 1993) including improved fine motor ability, sustained attention accuracy and reaction time, orienting attention reaction time, short-term episodic memory, and working memory reaction time (Heishman, Kleykamp, & Singleton, 2010). Nicotine reliably improves WM performance in rats (Levin, McClernon, & Rezvani, 2006), in tasks including the RAM (Levin et al., 2002, Levin et al., 1992, Levin et al., 2005, Levin et al., 1996, Levin and Torry, 1996), and remediates WM deficits of mice of caspase 3 over-expressing mice (Young et al., 2007) Thus, higher smoking rates in psychiatric populations may reflect increased motivation to smoke for self-medication purposes (Evans & Drobes, 2009).

Nicotine has been shown to attenuate systemic NMDAr antagonist-induced WM deficits in rats performing the RAM (Levin et al., 1998), but not after intra-amygdala infusion of MK-801 (May-Simera & Levin, 2003), or following intra-hippocampal infusion of MK-801 (Levin, Sledge, Baruah, & Addy, 2003). The nicotinic agonist anabasine (a constituent of tobacco smoke) reversed MK-801 induced RAM deficits (Levin et al., 2014), as did the α4β2 nicotinic receptor antagonist DhβE, the non-selective nicotinic antagonist mecamylamine, and the mixed agonist/antagonist sazetidine-A (Burke, Heshmati, Kholdebarin, & Levin, 2014). Thus, the mechanism by which nicotinic compounds serve to augment WM impairments produced by NMDAr blockade remain unclear. Identifying such effects in mice would enable combinations of genetic techniques to more specifically identify receptors underlying such effects, but these drugs have rarely been tested in mice.

The RAM procedure has however, been used repeatedly to examine the effects of dopaminergic manipulations (Levin, 1988). Destruction of dopaminergic neurons within the limbic system grossly impaired RAM performance (Simon et al., 1986, Taghzouti et al., 1986). While it is reasonable to hypothesize then that augmentation of dopaminergic transmission might serve to facilitate RAM performance, trials with dopaminergic agonists have largely failed to support this proposition. Multiple trials with amphetamine have observed either no consistent effect, or impairments (particularly when delays are utilized), in both rats (Beatty et al., 1984, Buresova and Bures, 1982, Eckerman et al., 1980) and mice (Bruto & Anisman, 1983). Similarly, the broad range dopamine agonist apomorphine failed to demonstrate modulation of RAM performance (Buresova & Bures, 1982). More recently, bromocriptine, a dopaminergic agonist with preferential D₂ receptor activity enhanced RAM performance in mice. Similar improvement in RAM accuracy was not observed with the D₁ receptor agonist SKF 38393, suggesting that selective activation of D₂-family receptors may be required for facilitation of RAM performance with dopaminergic agonists (Tarantino, Sharp, Geyer, Meves, & Young, 2011). Critically, such findings are consistent with bromocriptine-induced improvement in human spatial WM span capacity in healthy participants (Gibbs and D'Esposito, 2005, Mehta et al., 2001). However, interest in the use of such compounds to reverse the cognitive impairments associated with schizophrenia has been limited given that to-date the efficacy of antipsychotic medications have been tied to the D₂ blocking properties of these drugs.

The present study sought to determine the effects of acute nicotine on RAM performance when given in combination with MK-801 in C57BL/6J mice. We hypothesized that MK-801 would impair RAM performance, as has been previously observed with the BALB/cByJ strain (Mutlu et al., 2012), and deficits in accuracy would be ameliorated by nicotine as has been reported in rats (Levin et al., 1998). Given that bromocriptine, a potent dopamine D₂-family receptor agonist, improved mouse RAM performance (Tarantino et al., 2011), we also tested whether bromocriptine could remediate MK-801-induced RAM deficits.