Progressive impairments in executive function in the APP/PS1 model of Alzheimer's disease as measured by translatable touchscreen testing

Highlights

• Rewarded executive function was assessed in APP/PS1 mice using touchscreens

• APP/PS1 mice showed a subtle extinction deficit at middle-age

• Old APP/PS1 mice showed a profound behavioural flexibility deficit

• Genotype did not affect retinal health as measured by OCT or ERG, but age did

• This is the first study to show deficits in rewarded extinction or reversal tasks in the APP/PS1 mice

Abstract

Executive function deficits in Alzheimer's disease (AD) occur early in disease progression and may be predictive of cognitive decline. However, no preclinical studies have identified deficits in rewarded executive function in the commonly used APPSwe/PS1∆E9 (APP/PS1) mouse model. To address this, we assessed 12-26 month old APP/PS1 mice on rewarded reversal and/or extinction tasks. 16-month-old, but not 13- or 26-month-old, APP/PS1 mice showed an attenuated rate of extinction. Reversal deficits were seen in 22-month-old, but not 13-month-old APP/PS1 animals. We then confirmed that impairments in reversal were unrelated to previously reported visual impairments in both AD mouse models and humans. Age, but not genotype, had a significant effect on markers of retinal health, indicating the deficits seen in APP/PS1 mice were directly related to cognition. This is the first characterisation of rewarded executive function in APP/PS1 mice, and has great potential to facilitate translation from preclinical models to the clinic.

Introduction

Alzheimer's disease (AD) is one of the biggest health challenges worldwide, with approximately 50 million people currently diagnosed and rates of disease expected to rise in the next 30 years (Alzheimer's Disease International, 2015). Alzheimer's disease is characterised by amyloid plaque build-up, tau hyperphosphorylation, neurodegeneration and cognitive decline (Jack et al., 2013). Due to this increasing burden of disease, AD has been the subject of intense research for many years. In this time, numerous animal models of AD have been generated, with many potential treatments rescuing cognitive symptoms in a wide variety of models. However, all these treatments have failed to halt progression of the disease in clinical trials. Most preclinical animal studies focus on the assessment of hippocampal deficits as an outcome of treatment efficacy. In the clinic however, other early cognitive changes are seen in AD patients including those in executive function. A greater focus on executive function in preclinical animal models is warranted in order to improve the predictive power of these models to determine how effective treatments will be in the clinic.

Patients with mild cognitive impairment (MCI) and early AD patients show executive function deficits on a wide variety of tasks including Tower of London, Stroop Colour-Word Test, Ravens Coloured Progressive Matrices, Ruff Figural Fluency Test, Colour Trails Test Part B, category fluency, backwards digit span, and the Modified Card Sorting Task (Andriuta et al., 2018; Baudic et al., 2006; Huang et al., 2017; Ramanan et al., 2017). In these tasks, patients show attenuated task acquisition and performance, indicatative of prefrontal cortex (PFC) dysfunction (Levy-Gigi et al., 2011). More importantly, executive dysfunction, especially on the Trailmaking Part B test, is predictive of cognitive decline (Mez et al., 2013) and MCI to AD conversion (Albert et al., 2001; Ewers et al., 2012; Gomar et al., 2011; Huang et al., 2017; Johnson et al., 2009). As executive function deficits occur early and, importantly, have the potential to predict cognitive decline and MCI to AD conversion, these symptoms are of great clinical interest but, to date, have not been the focus of testing treatment efficacy in preclinical animal models.

Amyloid-driven mouse models of AD have been shown to have multiple executive function deficits early in their progression. Some models have shown deficits in reversal learning, which assesses a facet of executive function called behavioural flexibility (Chudasama, 2011). Briefly, the animal must first learn one association or rule (i.e. the location of platform in a pool or a stimulus-reward pairing), which is then reversed (i.e. the platform is moved to a different location or a different stimulus is now rewarded). Many AD models have shown reversal impairments in spatial maze-based tasks like Morris water maze (Baruch et al., 2015; Hooijmans et al., 2009; Jankowsky et al., 2005; Marchese et al., 2014; Musilli et al., 2013; Papadopoulos et al., 2013; Stover and Brown, 2012), T water maze (Dong et al., 2005; Filali et al., 2012, 2011; Filali and Lalonde, 2009), cheeseboard (Cheng et al., 2013) and Barnes maze (Stover et al., 2015) as well as olfactory (Girard et al., 2014, Guérin et al., 2009), digging (Shirey et al., 2009; Zhuo et al., 2008) and place preference (Masuda et al., 2016) reversal tasks. AD models have also shown deficits in instrumental extinction, another test of executive function, where the animal must learn to stop responding to a conditioned stimulus (i.e. a stimuli that previously indicated a reward or punishment). Extinction tasks are similar to reversal tasks in that an animal must learn to adjust a stimulus/outcome association, but without having to learn a new rule. AD models show enhanced extinction in rewarded tasks (Romberg et al., 2013) and a mix of attenuated and enhanced extinction in aversive tasks like fear conditioning (Bonardi et al., 2011; Cheng et al., 2019) and conditioned taste aversion (Hanna et al., 2009; Janus et al., 2004; Pardon et al., 2009; Ramírez-Lugo et al., 2009; Rattray et al., 2010). However, the vast majority of tasks applied to AD mouse models either rely on aversive environments or are ethologically relevant to mice; thus they are not analogous to tasks performed clinically, where the patient is responding to stimuli on paper or a screen in a calm environment. As such, we need a paradigm that can assess executive function in non-aversive, human relevant conditions.

Touchscreen testing is a relatively new paradigm to behaviourally test AD mouse models with great translational potential (Shepherd et al., 2016). To date, 3 AD mouse models, the TgCRND8, APPPS1-21 and APPSwDI/Nos2-/- mice, have undergone characterisation of executive function using reversal and extinction paradigms in touchscreens. Contrary to predictions from previously used aversive tasks, TgCRND8 mice show both accelerated reversal and extinction in touchscreens compared to WT mice. This is not due to an accelerated ‘forgetting’ of the first stimulus/reward association, as TgCRND8 mice show intact retention memory for the initial stimulus (Romberg et al., 2013). Conversely, APPSwDI/Nos2-/- mice showed an attenuation of reversal in the touchscreen testing paradigm (Piiponniemi et al., 2017). The APPPS1-21 model showed reduced accuracy in reversal compared to their WT counterparts at 6 and 9 months (Van den Broeck et al., 2019). However, the APPPS1-21 animals performed significantly less trials throughout reversal, thus it is not possible to ascertain if this is a learning or motivational deficit. These conflicting and unclear results highlight the need for more studies focusing on executive function in AD mouse models.

One of the most commonly characterised AD mouse models, the APP/PS1 mouse has not yet been characterised on touchscreen-based executive function tasks. Given previous reports showing slower acquisition of reversal Morris water maze (MWM) at 8 months of age (Hooijmans et al., 2009; Jankowsky et al., 2005; Stover and Brown, 2012) and left-right discrimination learning at 6-9 months old (Filali et al., 2011, 2009) in APP/PS1 mice, we aimed to assess whether reversal learning was impaired in this mouse model using the pairwise discrimination and reversal task. Conflicting findings exist for extinction learning in APP/PS1 mice, with reports of both attenuated (Bonardi et al., 2011) and enhanced extinction to aversive extinction tasks (Cheng et al., 2019; Ramírez-Lugo et al., 2009). When assessed in positively rewarded extinction tasks at 4-8 months of age, APP/PS1 mice did not demonstrate any deficits (Bonardi et al., 2011; Cheng et al., 2013). To resolve this conflict, we set out to test if APP/PS1 mice would show deficits in reward-based executive function tasks using the rodent touchscreen paradigm. We chose to assess 12-month-old and 24-month-old animals on reversal and extinction learning, and 12, 16 and 26-month-old mice in extinction. We chose 12 months of age as this is the best-characterised age in APP/PS1 mice and we wanted this study to be comparable to as many previous studies as possible, and a 24-month-old age group would confirm if a deficit would ever develop in reversal learning. We added the 16-month-old group into extinction as 26-month-old animals struggled to perform the task. We hypothesised that APP/PS1 mice would develop executive function deficits in both appetitive reversal and extinction tasks.