Early development of social deficits in APP and APP-PS1 mice

Abstract

Mimicking relevant behavioral features of the human pathology is one of the most important challenges for animal models of neurological disorders including Alzheimer disease (AD). Indeed, the most popular genetic AD mouse lines bearing mutations of the amyloid precursor protein (APP) and presenilin 1 genes (PS1), often fail to present robust cognitive deficits or show them only at very advanced ages. It is therefore crucial to identify AD-like behavioral alterations which may reliably reflect the early stages of the pathology, thus permitting tests of more efficient early therapeutic interventions. Here, we demonstrated the very early expression of noncognitive AD-like symptoms, i.e., deficits in social interest, interaction and communication, in APP and APP-PS1 transgenic mice. Conversely, other noncognitive behaviors (sensori-motor gating) as well as cognitive abilities (spontaneous alternation) were unaltered in AD transgenics. Our data suggest that social deficits precede other neuropsychiatric and cognitive AD-like symptoms and can be employed as early markers of AD pathology in genetic mouse models.

Introduction

Genetic mouse models represent nowadays the most valuable tool for advancing our knowledge on Alzheimer disease (AD). The best known AD models are undoubtedly the transgenic lines harboring the Swedish (K670N:M671L) mutation of the amyloid precursor protein (APP), either alone (the Tg2576 or APP mouse) (Hsiao et al., 1996) or in combination with mutations in Presenilin1 (PS1; the APP-PS1 mouse) (Holcomb et al., 1998). These mouse lines are able to model the main brain hallmarks of AD and their age-related progression. In the APP mouse, brain soluble (40 and 42) beta-amyloid levels increase by 6 months of age while neuritic plaques are reliably observed in cortical and hippocampal areas around 9–10 months of age (Hsiao et al., 1996, Irizarry et al., 1997, Terai et al., 2001). In the APP-PS1 mouse, increased amyloid brain levels have been described already at 3 months, and a robust plaque deposition can be detected at 6 months (Howlett et al., 2004, Lagadec et al., 2010, Trinchese et al., 2004, Yu et al., 2009).

In contrast to the brain pathology, the behavioral AD-like phenotype has proven to be much more difficult to be modeled in these mouse lines (Dodart et al., 2002). A major part of research efforts have traditionally focused on the cognitive characterization of AD genetic mouse models, due to the central role of memory deficits in the expression and progression of AD. The results obtained were highly controversial: while some studies reported the presence of memory deficits at 10–12 months in the APP model (Holcomb et al., 1998, Holcomb et al., 1999, Hsiao et al., 1996, Lee et al., 2004, Ognibene et al., 2005, Rustay et al., 2010) and at 6–7 months of age in the APP-PS1 line (Arendash et al., 2001, Filali and Lalonde, 2009, Holcomb et al., 1999, Howlett et al., 2004, Lagadec et al., 2010, Malm et al., 2007, Trinchese et al., 2004), others failed to detect robust cognitive alterations in these transgenic mice at these or even much more advanced ages (Bizon et al., 2007, Deacon et al., 2008, Ewers et al., 2006, Filali et al., 2009, Holcomb et al., 1999, Howlett et al., 2004, King and Arendash, 2002, King et al., 1999, Middei et al., 2004). These discrepancies stretch over several cognitive domains, including spatial memory (Arendash et al., 2001, Bizon et al., 2007, Holcomb et al., 1999, Hsiao et al., 1996, King and Arendash, 2002, Lee et al., 2004, Malm et al., 2007, Trinchese et al., 2004), spontaneous alternation (Arendash et al., 2001, Deacon et al., 2008, Filali and Lalonde, 2009, Holcomb et al., 1998, Holcomb et al., 1999, King and Arendash, 2002, Ognibene et al., 2005, Rustay et al., 2010), object recognition (Howlett et al., 2004, Ognibene et al., 2005), and associative learning (Deacon et al., 2008, Ewers et al., 2006, Filali and Lalonde, 2009, King and Arendash, 2002, Lee et al., 2004, Middei et al., 2004, Ognibene et al., 2005, Rustay et al., 2010). Apart from the inconsistencies among studies, this body of data demonstrates a relatively late onset of cognitive abnormalities in these two AD models, which does not reflect the gradual progression of AD-relevant brain changes, which start much earlier in both mouse lines.

Hence, there is an increasing need to find additional behavioral markers of AD pathology, which may be more reliable and accurate in modeling the disease progression in genetic mouse models than the cognitive deficits (Chung and Cummings, 2000). In particular, identifying AD-like behavioral abnormalities already at the early stages of AD pathology, i.e., before extensive plaque deposition, is of critical importance for testing the efficacy of potential early pharmacological and nonpharmacological treatments (Knopman, 1998). For all these reasons, growing attention has recently been devoted to the noncognitive characterization of genetic mouse models of AD, focusing on those behavioral alterations which may parallel specific symptoms observed in AD patients.

Indeed, multiple neuropsychiatric symptoms other than memory deficits are commonly detected in AD patients. The most prominent include abnormalities in emotionality/activity, social interaction/communication, and sensorimotor abilities, e.g., prepulse inhibition (PPI) of the acoustic startle response (Aalten et al., 2003, Chiu et al., 2004, Frisoni et al., 1999, Helkala et al., 1989, Hope et al., 1997, Jessen et al., 2001, Lawlor and Bhriain, 2001, Loewenstein et al., 2003, Loewenstein et al., 2004, Mohs et al., 2000, Rankin et al., 2008, Spalletta et al., 2004, Ueki et al., 2006). These noncognitive alterations have been detected also at very early ages, though their precise relationship to the progression of the disease is still a matter of discussion (Chung and Cummings, 2000, Morris et al., 1989). Noncognitive AD symptoms have been described to appear in a fluctuating manner across the course of the disease, and it has been hypothesized that they may develop independently of cognitive deficits (Mohs et al., 2000, Spalletta et al., 2004). Furthermore, recent studies have suggested that differences in progression characteristics may exist even among noncognitive AD symptoms (Jicha and Carr, 2010, Mohs et al., 2000), some specific alterations, e.g., PPI deficits, apparently being unrelated to other neuropsychiatric deficits (Hejl et al., 2004). In conclusion, the extended neuropsychiatric syndrome characterizing AD is receiving increasing interest, not the least because of its dramatic impact on the life quality of patients and caregivers (Knopman, 1998).

A few studies have evaluated the occurrence of noncognitive AD-like symptoms in APP and APP-PS1 models and its age-dependency. Alterations in emotionality and activity have been evaluated in both transgenic lines at multiple ages, and were observed after 5 months of age (Arendash et al., 2001, Lalonde et al., 2003, Lee et al., 2004, Ognibene et al., 2005, Pugh et al., 2007, Rustay et al., 2010). Conversely, PPI has been assessed in these AD models only at advanced ages (12–18 months), also leading to conflicting outcomes (Ewers et al., 2006, Gruart et al., 2008). Surprisingly, deficits in social interaction and communication have not been investigated very extensively in these two popular mouse models, with the exception of social memory (Deacon et al., 2009) and aggression (Pugh et al., 2007).

In the present study, we therefore evaluated whether noncognitive (abnormalities in social interaction/ultrasonic communication and PPI deficits) and cognitive (reduced spontaneous alternation in the Y-maze) AD-like symptoms were present in the APP and APP-PS1 mouse models at 6 months of age, i.e., at the early phases of AD pathology (Experiment 1). Because the results obtained in Experiment 1 clearly showed marked noncognitive alterations in both transgenic lines, we investigated in Experiment 2 whether these behavioral deficits were already present at an even earlier stage of AD pathology, when plaque deposition was mainly absent in both lines. To this end, we performed the same behavioral tests as in Experiment 1 in both genetic models in an independent cohort of animals at 3 months of age. Indeed, we have recently confirmed that the ages of 3 and 6 months represent two important time points in AD progression in these transgenic mice (Lagadec et al., 2010): at 3 months APP-PS1 mice showed an increase in brain levels of soluble beta amyloid that was observed in APP animals only at 6 months, i.e., when APP-PS1 mice already begin to show robust brain plaque deposition (Lagadec et al., 2010).