Beyond memory impairment: the complex phenotypic landscape of Alzheimer’s disease

Abstract

Neuropsychiatric symptoms (NPS) in Alzheimer’s disease (AD) constitute multifaceted behavioral manifestations that reflect processes of emotional regulation, thinking and social behavior. They are as prevalent in AD as cognitive impairment and develop independently during the progression of neurodegeneration. Albeit clinically challenging, AD clinical research to date has been focused mostly on cognitive decline. Here, we summarize emerging literature on the prevalence, time course and the underlying genetic, molecular, and pathological mechanisms related to NPS in AD. Overall, we propose that NPS constitute a cluster of core symptoms in AD and understanding their neurobiology can lead to a more holistic approach to AD research, paving the way for more accurate diagnostic tests and personalized treatments embracing the goals of precision medicine.

The need to understand the pathobiology of NPS in AD

Precision medicine aims to incorporate lifestyle, clinical and genetics data to tailor diagnosis, prevention and treatment of specific disorders to the needs of the individual patient [1]. To realize this goal in the treatment of Alzheimer’s disease (AD) dementia, there is a pressing need to better understand the biological basis of some of the most debilitating symptoms of the disease, namely neuropsychiatric symptoms (NPS). While AD is generally associated with gradually progressive, multi-domain, cognitive decline that usually affects memory and language, NPS are prevalent and important facets of the illness. NPS are sometimes referred to as behavioral and psychological symptoms of dementia, [2] are common in patients with AD and are associated with accelerated cognitive impairment, higher caregiver burden, higher rate of institutionalization and earlier deaths [3]. To date, there is no clear understanding of whether NPS are independent entities, albeit highly comorbid with AD, or if they constitute a manifestation of the underlying AD pathology [4] Here, we outline the NPS most commonly reported with AD, discuss their various characteristics and implications for treatment, and make the case that precision medicine in AD requires a holistic understanding of its clinical manifestations, including the contribution from NPS.

Clinical overview of Neuropsychiatric Symptoms in AD

The main NPS in patients with AD are: apathy (with a predicted overall prevalence of 49%), followed by depression (42%), aggression/agitation (40%), anxiety (39%), sleep disturbance (39%), irritability (36%), appetite disturbance (34%), aberrant motor behavior (32%), delusions (31%), disinhibition (17%), hallucinations (16%) and, finally, euphoria (7%) (see also Table 1) [5]. There is a significant degree of heterogeneity in the clinical presentation of NPS in AD. That is, different clusters of NPS manifest in different individuals and at different timepoints along the disease’s course. Moreover, discrepancies exist regarding the prevalence of NPS among studies and this may be partly explained by heterogeneity in the timeframe of occurrence of the symptoms, which is supported by studies showing that some symptoms, such as hallucinations and delusions, are less frequent in prodromal and early AD but increase in prevalence with progression of the disease [6]. Additionally, inherent differences in study design can also contribute to discrepancies observed in rates of NPS in AD. Most studies require recollection of behaviors that occurred between visits as reported by caregivers, meaning that the frequency of the NPS will reflect the frequency of the visits, and that the apparent incidence of the NPS will be higher with monthly assessments than with annual. Furthermore, there is recall bias with recency effects as well as additional recall bias stemming from the severity of the event. To better assess the frequency and potential impact of NPS in AD, more longitudinal studies and improved screening tools are needed. In the following section, we provide a clinical overview for the NPS that are common in AD, and, in Table 1, summarize the definition, prevalence, onset and association of NPS with cognitive function.

Table 1. Summary of neuropsychiatric symptoms in Alzheimer’s disease.

Neuropsychiatric symptom	Definition	Prevalencea	Onset	Associations with cognitive function
Apathy	Loss of emotional reactivity and motivation in the absence of sadness	49%	Early phases, even before clinical AD diagnosis	Executive dysfunction
Depression	Manifestations of low affect such as tearfulness, lack of energy, suicidal behavior or ideation, guilt or sense of burden to family or caregivers	42%	Early phases, even before clinical AD diagnosis	Executive and memory dysfunction
Agitation	Inappropriate verbal, vocal or motor activity that is not explained by acutely altered mental status (delirium)	40%	Moderate severity/middle phase	Visuospatial, executive and conceptualization dysfunction
Anxiety	Uneasiness, internal turbulence, feeling of impending doom and also anxiety disorders such as panic disorder and generalized anxiety disorder (GAD)	39%	Early phases, even before clinical AD diagnosis	No associations
Sleep disturbance	Insomnia, sleep-wake cycle reversal	39%	Early phases, even before clinical AD diagnosis	Memory disorders
Irritability	Inappropriately increased reactivity to environmental stimuli (i.e. low threshold for anger and frustration)	36%	Early phases, even before clinical AD diagnosis	No associations
Appetite disturbance	Anorexia and appetite loss	34%	Early phase	No associations
Aberrant motor behavior	Excessive amount of purposeless and repetitive activity (i.e. fidgeting, rummaging through drawers, dressing/undressing and wandering)	32%	Moderate severity/middle phase	Decline in verbal fluency tests
Delusions	Persistent false belief based on incorrect inferences about external reality, resistant to persuasion or contrary evidence, and not attributable to social or cultural mores	31%	Moderate severity/middle phase	Executive dysfunction, reasoning
Disinhibition	Behaviors that are socially inappropriate such as impulsivity, excessive jocularity and hypersexuality	17%	Mostly late phase	Executive dysfunction
Hallucinations	Perceptions in the absence of external stimulus (mostly visual)	16%	Middle to late phase	Visuospatial dysfunction
Euphoria/Elation	Abnormally elevated mood	7%	Moderate severity/middle phase	No associations

Timeline of NPS with severity and progression of AD

Our understanding of the temporal relationship between NPS and disease progression is still incomplete. Nevertheless, the majority of NPS tend to appear during the preclinical or mild cognitive impairment (MCI) phase of AD, [6], (Table 1), and the term mild behavioral impairment has been developed to describe these early onset behavioral manifestations at the preclinical stage of AD that have been shown to increase disease progression. [7]

Although the progress of cognitive symptoms appears to be gradual and more linear across the disease course [2], NPS are fairly unstable and have a more episodic character [8]. Generally, the first NPS to appear are apathy, depression, anxiety, irritability, and sleep disturbances [9], many of which are often apparent before AD diagnosis. Conversely, the prevalence of aberrant motor behavior, delusions and hallucinations increase in more advanced disease stages [10]. It has been reported that apathy demonstrates a continuous course and is quite persistent, while anxiety and depression have been shown to be moderately stable along the course of the disease [11]. On the other hand, delusions, hallucinations, appetite, and sleep disturbances show less persistence [11]. In addition, some NPS demonstrate a symptom specific association with cognitive decline, where specific symptoms tend to be associated with specific types of cognitive impairment (Table 1).

There is evidence that NPS, particularly affective symptoms like apathy and depression, might precede the onset of cognitive impairment and, in this context, could be an early manifestation of dementia, facilitating their use as a means to identify at-risk patients, especially during the (MCI) stage [9]. Similarly, psychotic symptoms identify a subgroup of AD subjects with a more rapidly progressive cognitive and functional deterioration [12]. To better decipher the relationship of the NPS to core AD pathobiology, in the next sections we present findings from studies of genetic and molecular factors implicated in both NPS and AD. Further elucidation of the complex interplay between the mechanisms linking these two could provide valuable insights towards the design of more effective therapeutic interventions and may lead to better overall nosological understanding of the disease.

Genetic and molecular mechanisms related to NPS

NPS in AD share some clinical features with serious mental illness (SMI), defined as any mental health condition that severely impairs daily functioning [13]. For example, delusions and hallucinations are among the core symptoms of schizophrenia, while euphoria/elation is seen in bipolar I disorder (BD), and depression in major depressive disorder (MDD). Indeed, a systematic review and meta-analysis that evaluated the risk of dementia in patients with a history of BD indicates a significant association between BD and the risk of AD [14], and data from genome wide association studies (GWAS; see Glossary) suggest a significant overlap in the genetic basis of these two diseases [15]. One hypothesis is that NPS develop in AD patients, in part due to shared genetic factors among AD and SMIs, though this relationship may not be straightforward. For example, the risk of psychosis in AD has been genetically correlated with depression, but not with schizophrenia [16]. The molecular pathways shared between AD and SMI implicate gene expression signatures, causal proteins, protein-protein interactions and biological pathways, the majority of which implicate synaptic pathology and immune system activation [17]. In the following sections, we discuss recent findings from genetics and molecular studies that support shared mechanisms among AD and SMI related to AD NPS, which we also summarize in Figure 1.

Figure 1. Summary of the genetic and molecular mechanisms shared between AD and SMI.

A. Shared genetic architecture between AD and MDD, and between AD and psychosis. B. Shared gene expression signatures in the superior temporal gyrus (STG), the dentate gyrus and the hippocampus implicate the autophagy regulation system in AD and schizophrenia (SCZ); shared gene expression signatures implicate microglial response in AD and BD; shared gene expression signatures implicate immune response and endocytosis regulation in AD and MDD. C. Shared causal proteins between AD and schizophrenia (ACE, ADAM10, CCDC6, DOC2A) and between AD and BD (CCDC6 and MTSS1L). D. Shared protein-protein interactions between AD causal proteins (STX6, LACTB, CCDC6) and SMI causal proteins. Figure was created with Biorender (www.biorender.com).

Shared genetic architecture between AD and major depressive disorder

Most variants associated with increased risk for brain disease reside in non-coding regions of the genome and are thought to exert their effects by disrupting the spatiotemporal regulation of gene expression [18]. These risk loci are frequently large and often contain multiple implicated single nucleotide polymorphisms (SNPs) due to local linkage disequilibrium (LD) structure [18]. Although challenging, identification of the causative SNPs is an important step towards identifying affected genes and pathways, and in furthering our understanding of the genetic basis of disease. The LD score regression approach [19] has been used to reveal positive genetic correlations between AD and MDD, BD, post-traumatic stress disorder (PTSD), and alcoholism [17]. It is worth noting that genetic correlation may arise from pleiotropy (i.e., genes independently affecting both MDD and AD), or from the causal effect of MDD on AD, or vice versa. To this end, a study using 115 MDD GWAS SNPs [20] found a significant causal effect of MDD on AD [21]. Conversely, no significant effect was found when the probability of a causal effect of AD on MDD was examined using 61 AD GWAS-significant SNPs [21],[22]. Moreover, no evidence of horizontal pleiotropy was found among AD and MDD [21]. Together, these findings are consistent with the notion that SMI has some shared genetic basis with AD and suggest a potential causal role for MDD in AD.

This raised the question of whether genetic risk for MDD is associated with AD endophenotypes, or vice versa. Examining the relationship between MDD polygenic risk score (PRS) and the rate of cognitive change over time revealed a significant association between higher MDD PRS and faster decline of episodic memory [21]. A similar study examined the relationship between AD PRS and depression among participants with normal cognitive performance [23] and found that those with higher AD PRS were more likely to experience clinically significant depression at, or after, age 50, in a manner that is not explained by apolipoprotein E (APOE) risk alleles [23].

Shared genetic architecture between AD and psychosis

Multiple studies have demonstrated that the risk for psychosis after onset of AD is familial, with an estimated heritability of 61% [12], [24]. This finding has two important implications: the risk of psychosis in AD is partly driven by genetic variation, and psychosis in AD arises from a distinctive underlying neurobiology and is not just the result of the serendipitous location of accumulated neuropathological lesions. This latter implication is addressed in subsequent sections.

To date, numerous studies have examined genetic correlates of psychosis in AD [25]. However, these early studies do not meet current rigorous standards for appropriate sample size, interrogation of the genome with sufficient depth, control for participant ancestry, and correction for genome-wide hypothesis testing. Only recently has the first study been performed to address these concerns and provide unbiased identification of the association of psychosis in AD with common genetic variation and insights into its genetic architecture. Several key findings emerged from a genome-wide meta-analysis of 12,317 AD subjects, of whom 5,445 had been diagnosed with psychosis [16]. It established that a significant portion of the heritability of psychosis in AD is derived from common genetic variation. Associations at two loci, one in ENPP6 (best SNP rs9994623) and one spanning the 3’-UTR of an alternatively spliced transcript of SUMF1 (best SNP rs201109606), reached genome-wide significance. Using gene-based analysis, a significant association of psychosis in AD with APOE, due to the APOE risk haplotype ε4, was also found, as was a positive genetic correlation with genetic risk for AD beyond the effect of ε4. In keeping with known clinical correlates of psychosis in AD, negative genetic correlations with cognitive and educational attainment, and a positive genetic correlation with depressive symptoms, were also found [16]. Interestingly, there was no overall genetic correlation of psychosis in AD with schizophrenia. Nevertheless, a subset of SNPs likely to affect risk for both psychosis in AD and schizophrenia were identified, and these await further evaluation in future, better powered, studies.

Molecular changes in NPS and AD

Gene expression

Several studies have surveyed shared gene expression changes in AD with SMI, such as schizophrenia, BD, and MDD. While etiology and neuropathology are dissimilar between AD and SMI, these studies suggest that they share some region-specific gene expression signatures. Among the multiple brain regions examined, Brodmann area 22, the superior temporal gyrus, showed the largest overlap in dysregulation between AD and schizophrenia [26]. The superior temporal gyrus may be particularly vulnerable to schizophrenia [27], [28], as well as to neuritic plaque and neurofibrillary tangle pathology in AD [29]. Commonly dysregulated gene signatures implicated the impairment of autophagy, which is well documented for both AD [30] and schizophrenia [31], [32]. Differential gene expression analysis of BD [33] discovered genes previously associated with AD, such as A2M, TREM2, CTSB, and PLCG2. Moreover, genes dysregulated in BD were enriched in immune activation and inflammatory response pathways from a microglia-specific module previously implicated in AD [34]. Another study, focused on the conditional association of late-onset AD with MDD, identified several common genes, including BIN1, PICALM, and PSMC3, further implicating the immune response as well as the regulation of endocytosis [35].

Shared causal proteins

To expand on the evidence of a shared genetic and molecular basis among SMI and AD, a proteome-wide association study (PWAS) [36], followed by Mendelian randomization [37] and colocalization analysis [38], identified cis-regulated brain proteins that are consistent with a causal role in SMI and AD or related dementia (ADRD).[17] Additionally, the approach also identified trans-regulated proteins that are consistent with a causal role in each of these brain conditions. Interestingly, 13 putatively causal proteins common to the SMI and ADRD groups were found [17], two of which were shared between MDD and Parkinson’s disease (PD) (ADK and AKT3), two between BD and AD (CCDC6 and MTSS1L), four between schizophrenia and AD (ACE, ADAM10, CCDC6, and DOC2A), two between anxiety and PD (MAPT and STX1B), one between PTSD and PD (MAPT), and one between alcoholism and AD (HSDL1). The study also revealed a positive relationship between the degrees of genetic correlation and percentages of shared casual proteins among SMI and ADRD [17]. Together, these shared causal proteins highlight a shared genetic susceptibility and molecular basis among SMI and ADRD.

Protein-protein interactions between psychiatric and neurodegenerative causal proteins

Thus far, we have considered proteins that are mutually causal in SMI and ADRD as evidence for shared susceptibility. Another way a shared mechanism may arise is through protein-protein interaction (PPI) between causal proteins. One study tested the hypothesis that there is a higher level of PPI between the SMI causal proteins and ADRD causal proteins than by chance alone, and found 2.6 fold (n=118) more physical PPIs between the causal SMI and causal ADRD proteins [17]. Among them are a few notable proteins with wide-spread interactions with other causal proteins. For example, PDHA1 (a causal protein in PD) physically interacts with 15 SMI causal proteins; MAPT (another PD causal protein) physically interacts with 11; LACTB (an AD causal protein) interacts with 10; and CCDC6 and STX6 (AD causal proteins), each of which interact with 7 SMI causal proteins [17]. These interactions provide important evidence supporting the hypothesis of shared molecular mechanisms between the psychiatric disorders and ADRD groups.

In summary, these studies leveraged GWAS results and brain transcriptomic and proteomic data to examine shared transcripts and proteins between AD and SMI. Future studies examining genetic and brain transcriptomic and proteomic data in donors with both AD and NPS would enhance our understanding of the underlying mechanisms of NPS in AD.

Shared biological processes between SMI and ADRD

To gain insights into shared molecular processes, gene set enrichment analysis was performed on the shared causal proteins and shared and interacting causal proteins [17]. The 13 causal proteins shared between the SMI and ADRD groups were enriched for calcium ion-regulated exocytosis and localization of proteins to the synapse, while the 118 shared interacting proteins were enriched for a number of processes; including neurotransmitter secretion, vesicle-mediated transport in synapse, synaptic vesicle recycling, SNAP receptor activity, myeloid leukocyte activation, and mitochondrial processes [17]. These shared genetic and molecular properties are consistent with the epidemiologic literature showing that having a psychiatric illness in early or mid-life is associated with up to 4 times higher risk for developing dementia in late-life [39],[40],[41],[42],[43],[44], and that at least 85% of individuals with ADRD experience neuropsychiatric symptoms that overlap with symptoms of these psychiatric disorders [11].

NPS neuropathology

Although AD is considered a clinical syndrome, neuropathology remains the gold standard for a definitive diagnosis [6]. The severity of neuropathology has been associated with the severity of the NPS, a finding that supports the hypothesis that NPS are caused by an underlying pathology [6]. In Table 2 we present the main associations of NPS with neuropathology and in Box 1 we expand further on the biological correlates of psychosis in AD.

Table 2. Summary of neuroimaging, neuropathological and neurotransmitter alterations of specific NPS.

NPS	Neuroimaging	Neuropathology	Neurotransmitters	References
Apathy	Anterior cingulate cortical and subcortical lesions		Low levels of striatal D2 receptors	[9]
Depression	Anterior cingulate cortex Hippocampus Basal Ganglia Corpus Callosum (frontal limbic circuitry and Default Mode Network)	High Aβ burden, tau phosphorylation	Serotonergic neuron loss in raphe nuclei and hippocampus Noradrenergic neuron loss in locus coeruleus, low levels of 5HT1A receptors in hippocampus, lower density of serotonin receptors in basal ganglia and thalamus	[51], [52], [53], [54], [55], [56]
Agitation	Orbitofrontal subcortical circuit	TDP-43, high neurofibrillary burden in orbitofrontal and anterior cingulate cortex	Decreased cholinergic and serotonergic activity in frontal and temporal cortex	[9], [57]
Anxiety	Amygdala, Locus coeruleus, Hypothalamus		Serotonergic neuron loss in raphe nuclei and hippocampus Noradrenergic neuron loss in locus coeruleus	[9], [52], [58], [53], [55]
Sleep disturbance		High Aβ burden, tau phosphorylation and synaptic dysfunction	Serotonergic neuron loss in raphe nuclei and hippocampus Noradrenergic neuron loss in locus coeruleus	[52], [53], [55], [58],
Irritability	Orbitofrontal subcortical circuit			[9]
Delusions	Frontotemporal lobe, left frontal atrophy, hippocampal atrophy (misidentification delusions) and DMN disruption	High tau phosphorylation and synaptic dysruption	Serotonergic neuron loss in raphe nuclei and hippocampus Noradrenergic neuron loss in locus coeruleus, low levels of serotonin in in temporal cortex High levels of D2 receptors in the striatum correlate with delusions and wandering	[9], [52], [53], [55], [57], [58], [59]
Disinhibition	Orbitofrontal subcortical circuit			[9]
Hallucinations	Anterior insula	High tau phosphorylation and synaptic disruption	Serotonergic neuron loss in raphe nuclei and hippocampus Noradrenergic neuron loss in locus coeruleus, low levels of serotonin in temporal cortex Reduced acetylcholinesterase activity	[9], [52], [53], [55], [57], [58],

Box 1. Psychosis in AD.

A tripartite approach has been used to account for the burden of AD and comorbid pathologies in human tissue studies of psychosis in AD [60]. First, psychotic and non-psychotic groups were matched for age, duration of AD, sex, and, as a gross estimate of neuropathologic stage, Braak score. Second, quantitative neuropathologic measures of phosphorylated tau, fibrillar Aβ, and microglial density and activation state were generated for all subjects. Third, to account for correlation amongst the variables, demographic and neuropathologic variables were evaluated in a multiple regression model. Using this approach, it was shown that ~18% of the variance in psychosis status could be explained by 5 neuropathologic variables [60], of which increased PHF-1 tau volume fraction was the single most significant predictor of AD with psychosis (AD+P) (7% of the variance in psychosis status, p = 0.01).

If not primarily explained by neuropathologies, what are the other biologic correlates of psychosis in AD? Multiple neuroimaging studies have provided indirect evidence of excess reduction in synaptic integrity in AD+P in comparison to AD without psychosis (AD-P): reduced perfusion [61],[62],[63], reduced metabolism [64],[65], and lower gray matter density [66]. These impairments have been found throughout lateral neocortical regions - the dorsolateral prefrontal cortex (DLPFC), lateral temporal cortex, and inferior parietal cortex, but not in medial temporal lobe structures [25].

Studies of excess cognitive burden in AD+P have similarly pointed to DLPFC-dependent functions [67], [68], [69]. In contrast, although the hippocampus and other medial temporal lobe structures are affected in AD, they do not differentiate AD+P from AD-P in neuroimaging, postmortem, or cognitive studies [25]. Postmortem studies have also shown indirect evidence of excess synaptic impairment in AD+P relative to AD-P in neocortex: elevated incidences of synaptic membrane breakdown products [70], reduced gray matter levels of a panel of synaptic proteins [60] and reduced synaptic protein mRNA levels [71]. This latter report identified a reduced proportion of excitatory neurons in AD+P relative to AD-P [71]. More recently, the first direct examination of a synaptic component, the post-synaptic density (PSD), has been reported, finding reduced PSD abundance in AD+P relative to AD-P. The PSD proteome signature of psychosis in AD was characterized by reduced levels of a network of kinases and proteins regulating the actin cytoskeleton. Importantly, deficits in PSD abundance and in the PSD proteome were in excess of that accounted for by neuron loss [72].

Neuroimaging of NPS in AD

Neuroimaging studies have demonstrated that cognitive decline and NPS correlate with anatomical and functional alterations in different brain regions, pointing to different underlying etiopathogenic mechanisms [9]. These studies can target specific NPS and help elucidate underlying pathophysiological mechanisms by revealing the brain structures associated with each symptom in living patients [9],[45], (See also Table 2). Intriguingly, most studies implicate the anterior cingulate cortex as a potential central behavioral hub that, when dysregulated, gives rise to NPS and along with neuropathological studies point to three overlapping models for NPS in AD; cortical-cortical circuit disruption, fronto-subcortical circuit disruption and monoaminergic system dysregulation [46] [47]. Neuroinflammation is also a very important contributor to NPS in AD and a cross sectional study identified microglial activation, measured by PET scans, as a main association with NPS severity, and particularly with irritability, highlighting its importance as a biomarker [48]. Another study reported that a variant of AD characterized by the predominance of disinhibition, apathy and executive dysfunction demonstrated tau pathology in the cortices of the orbitofrontal, frontal insular, anterior cingulate and medial prefrontal area [49], while the predominance of tau pathology using PET scans has been further emphasized [50].

Overall, neuroimaging can provide valuable insights into the circuits implicated in NPS in AD. However, further research is needed to show whether the genetic risk factors underlying various NPS are acting through changes in different brain circuits, with the goal to elucidate the complex interplay between brain structure and function in the development of these symptoms.

Concluding remarks and future directions

We attempted to present the common NPS that constitute an important facet of the clinical presentation of AD. Genetic and molecular studies converge on implicating various biological processes which, if integrated harmoniously with complementary studies (e.g. neuroimaging), hold the promise to enhance our understanding of how the underlying neurobiological perturbations affect brain circuits, and how the interplay between different lesions in functionally distinct regions of the brain affecting different circuits correlates to the observed NPS. This will further elucidate the, thus far, enigmatic relationship between NPS and AD. It is still not clear whether NPS are independent entities, or if they are a manifestation of the underlying AD pathology. In any case, they constitute a central component of the clinical manifestations of AD and should be investigated with the same rigor as the hallmark symptom of cognitive decline.

Most studies rely heavily on subjective tests to screen for NPS, while looking for similarities in AD and SMI in different donors. Future studies should focus on the longitudinal assessment of NPS since this will provide a more effective way to capture NPS along the disease course, leading to more accurate estimation of their prevalence and impact. Additionally, the frequency of the surveys is a factor that can introduce biases. For example, most scales rate behaviors in the previous month. How is the caregiver’s ability to recall the patient’s behaviors influenced by the frequency of the surveys, as well as the severity of the symptoms? More frequent symptom surveillance, and more severe symptoms, can lead to more frequent reports, thus, influencing prevalence measures. Moreover, the use of behavior ratings to promote biological understanding is also problematic. Whether a given behavior is reflective of a trait, a state, or some interaction of the trait with a cognitive state (which the data supports for psychosis) remains to be understood. Examining genetics for a given state might be less informative than neuroimaging studies, and better screening tests are also necessary along with study designs that can quantify NPS within different AD/SMI donors and better powered future GWAS analyses (see Outstanding questions). In the dawn of precision psychiatry, the incorporation of all types of data, from clinical to genetic, and from neuroimaging to lifestyle/environmental, will be necessary for a better stratification of patients, adding to the future clinician’s armamentarium the necessary tools for an accurate diagnosis along with precise avenues for treatment and prevention (see Clinician’s corner).

Outstanding questions.

1. What is the exact relation between NPS and AD? Are they independent entities but highly comorbid? Are they a manifestation of the underlying AD pathology?

2. How can we more effectively estimate the NPS prevalence?

3. How can we leverage the shared genetic and molecular pathways between AD and SMI to further understand the etiopathogenesis and improve the treatment of the NPS?

4. How can we integrate genetic/molecular and neuroimaging findings in the generation of an overarching framework that can guide diagnosis and treatment?

5. Are the changes observed in microglial cells associated with any detectable changes in the peripheral blood cells? And if so, how can we discover biomarkers that can capture early disease processes even in the pre-clinical phase of AD?

Clinician’s corner.

The implication of neuroinflammation along with the contribution of neurotransmitter systems paves the way for designing novel therapeutic approaches that take into consideration the complex interplay between neurotransmitters and the immune system of the brain.

Better understanding of the mechanisms implicated in NPS in AD will lead to more precise diagnosis through the discovery of biomarkers enabling early detection of changes in the genetic and molecular level.

A precision medicine focused approach warrants the incorporation of data from multiple interdisciplinary approaches including genetics, molecular biology and neuroimaging.

Highlights.

Neuropsychiatric symptoms (NPS) in Alzheimer’s Disease (AD) are as prevalent as cognitive decline.

The majority of the NPS appear during the preclinical or mild cognitive impairment (MCI) phase.

Genetics, neuropathology and neurochemical dysregulation likely contribute to the NPS in AD.

Genetic predisposition to AD increases the chance of developing depression in cognitively normal persons over 50.

The shared biological processes between serious mental illness (SMI) and AD include the autophagy system, neurotransmitter secretion, vesicle mediated transport in synapse, synaptic vesicle recycling, SNAP receptor activity, myeloid leukocyte activation and mitochondrial processes.

There are shared causal proteins and protein interactions between AD and SMI.

Neuroimaging can help elucidate disruptive pathology in specific circuits.

Glossary

Broadmann area 22

brain area located in the posterior superior temporal gyrus

Endophenotypes

subtypes of behavioral symptoms with a clear genetic connection

Gene set enrichment analysis

method that identifies classes of genes or proteins that are overrepresented in a large set of genes or proteins

Genome wide association studies (GWAS)

observational study of genetic variation in different individuals that infers associations between single nucleotide polymorphisms and a trait

Linkage disequilibrium (LD)

the non-random association of alleles at different loci in a given population

Major depressive disorder (MDD)

a clinical syndrome characterized by persistently low or depressed mood, anhedonia or decreased interest in pleasurable activities, feelings of guilt or worthlessness, lack of energy, poor concentration, appetite changes, psychomotor retardation or agitation, sleep disturbances, or suicidal thoughts

Mendelian randomization

method using measure variation in genes to interrogate the causal effect of an exposure on an outcome

Polygenic risk score (PRS)

score that reflects how likely an individual has a trait based on their genetics

Proteome wide association study (PWAS)

a study used to detect gene-phenotype associations mediated by protein function alterations

Single nucleotide polymorphism (SNP)

a substitution of a single nucleotide at a specific position in the genome

Three prime untranslated region (3’-UTR)

the section of messenger RNA that immediately follows the translation termination codon