New Insights Into Mechanisms of Psychosis in Alzheimer’s Disease

On February 25, at an afternoon symposium, 4 speakers addressed advances in identifying and understanding specific biological causes of psychosis in Alzheimer’s disease patients.

Motor Predictors of Psychosis in Alzheimer ‘s Disease: A 2-year Longitudinal Study

“Approximately 30% of Alz-heimer’s disease (AD) patients present some form of psychosis, typically characterized by hallucinations and delusions,” said Michael Caligiuri, PhD, associate professor, Department of Psychiatry, University of California, San Diego.

Existing research provides little information on new incidence or onset of psychosis, despite data on prevalence. Previous studies have demonstrated that psychosis in AD patients is often associated with increased severity of dementia, greater duration of illness, extra- pyramidal signs (EPS), and more rapid cognitive decline. Research has also shown that AD patients with EPS (e.g., Parkinsonian features of rigidity, bradykinesia and postural abnormalities) exhibit poorer prognosis for survival, low functional capacity, presence of neurobehavioral disturbances, and faster cognitive deterioration.

Data indicate that EPS may be a risk factor for psychosis in AD patients. Findings have demonstrated that AD patients with EPS have twice the frequency of psychotic symptoms. Results from a study, conducted by Rosen and Zubenko, indicated that 60% of patients with EPS eventually developed psychosis (Biol Psychiatry 1991;29:224-232).

Dr. Caligiuri and colleagues at the UCSD Alzheimer’s Disease Research Center studied means of providing earlier, more sensitive detection of EPS movement abnormalities in patients, including those with AD. Benefits of early detection of EPS in AD patients include: earlier identification of patients at risk for neurobehavioral disturbance, and the possibility for a wider range of therapeutic options (e.g., initiation of therapies that are more appropriate at early stages of illness). Additionally, early detection enables patients’ families to better gauge the course of illness and plan for the potential economic burden of psychosis (especially when institutionalization may be necessary).

Two attributes identify EPS as a likely predictor of psychosis in AD. Neuroanatomy illustrates that both EPS and psychosis are likely mediated by similar frontostriatal circuitry. In addition, both EPS signs and psychotic behaviors are believed to stem from dopamine dysregulation.

Dr. Caligiuri and colleagues conducted a 2-year, longitudinal study of 54 patients without baseline psychosis to study EPS as a risk factor for psychosis in AD. EPS was determined based on clinical scores on the Unified Parkinson disease rating scale (UPDRS) and instrumental (iEMG) forearm surface readings for rigidity and bradykinesia.

Psychosis was judged on the basis of three outcome variables: behave-AD subscale scores for delusions and hallucinations, use of neuroleptic medications and institutionalization. Results from this research (publication in preparation) indicated a 36% overall incidence of psychosis in AD patients after two years. More specifically, 80% of patients with iEMG abnormalities for bradykinesia met the criteria for psychosis, compared to 29% and 43% for iEMG measures of rigidity and clinical EPS ratings, respectively.

Findings dovetail with results from previous research conducted by Graybiel and Ragsdale that demonstrated communication between the motor and
limbic systems, via the patch matrix organization in the striatum (Proc Nat Acad Sci 1978;75:5723-5726).

Dr. Caligiuri hypothesized that these findings could relate to dementia of Lewy bodies (DLB) in three possible ways. Alzheimer’s disease may share the “core” features of DLB at some point in the natural history of the illness. Or, AD may evolve into DLB as the disease progresses (perhaps as represented by early bradykinesia signs). A third possibility is that the clinical distinction between AD and DLB may be continuous rather than absolute.

Neuroimaging — Mapping Psychosis in Alzheimer’s Disease

“Traditional functional neuro-imaging techniques like PET or SPECT that have measured metabolism or blood flow within regions of the brain can now go to the next level and tell us what’s going on in the brain,” said David Sultzer, MD, associate professor, Department of Psychiatry, University of California, Los Angeles; director, Geropsychiatry, VA Greater Los Angeles Healthcare System. “Studies can look at neurochemistry, receptor dynamics, regional effects and specific symptoms,” said Dr. Sultzer.

In the past, neuroimaging techniques have been used for a variety of purposes in Alzheimer’s disease (AD), including differential and early diagnosis and prediction of disease progression in people with pre-clinical symptoms.

Structural neuromaging and EEG techniques have been used to explore psychosis in AD in a limited number of (mostly small scale) studies. Results from CT and MRI studies of regional brain volumes demonstrated enlargement of the right lateral ventricle anterior horn (associated with misidentification), enlargement of the right temporal horn (associated with paranoid delusions), and diffuse atrophy (associated with hallucinations). Emerging data from qualitative MRI studies of regional white matter changes indicate right and anterior hemisphere hyperintensities that appear to be associated with delusions in AD. Some studies using conventional EEG techniques have demonstrated increased slow wave frequencies diffusely throughout the brain in AD patients with psychosis.

Dr. Sultzer and colleagues explored the use of functional neuroimaging techniques such as PET and SPECT to study specific pathophysiology, in an effort to determine precise locations of brain changes that occur in AD patients with psychosis.

Advantages of functional neuroimaging include the ability to measure dynamic changes in vivo (versus static CT scans; allowing for incorporation of patient actions and thoughts into images produced) and enhanced sensitivity to clinical states, cognitive condition, or treatment status. In addition, functional techniques provide the ability to reveal relevant neuronal circuits, and the capacity to probe for neurobiological underpinnings (e.g., blood flow, metabolism, neuroreceptor density and neurochemistry kinetics), perhaps allowing for prediction of response to treatment or other changes over time.

Existing research using functional neuroimaging to study psychosis in AD is scarce. While some studies suggested involvement of more than one brain lobe in AD psychosis (the frontal and temporal cortex), others found correlations with only the frontal lobe.

Dr. Sultzer and colleagues used PET imaging to assess metabolic activity in various brain regions, and explore potential associations between regional metabolic rates and delusion scores. Various regions within the frontal cortex and temporal cortex were examined in 25 AD patients with a range of delusion scores. Images produced depicted “holes” where less metabolically active brain regions were located. Results demonstrated apparent associations between mean lobar metabolic rate and delusion scores, particularly for the right frontal lobe (and less so for the left frontal lobe). No significant associations were apparent for other brain lobes.

Regression analysis was used to determine specific areas within the frontal and temporal regions that might be independently associated with delusions. Results indicated strongest associations with the right hemisphere and frontal lobe (particularly the right superior doso-lateral region). Associations were also apparent in the right inferior frontal pole and lateral orbitofrontal region. Simple bivariate correlations indicated associations for additional frontal right hemisphere regions, along with some homologous left hemisphere locations. In general, low metabolic rates in these regions were associated with delusions.

Only one temporal region (the right middle temporal gyrus) appeared to be correlated with delusions, and an inverse relationship (high metabolic rate associated with delusions) was demonstrated there.

Dr. Sultzer hypothesized that a psychosis phenotype in AD may emerge from dysfunction (i.e., metabolic deficits) within specific cortical circuits, probably within the frontal cortex (and likely the right frontal cortex). “Research results suggest that the neuropsychiatric aspects of AD are really part of the illness; it’s the brain talking, it’s not old people with a cognitive disorder who then act funny; there’s a much greater precision to the relationships here,” said Dr. Sultzer.

Additional studies are indicated to further explore specific neuronal networks involved in delusions. Future research will also help understand how interactions among brain regions may contribute to delusions, how brain involvement may vary based on specific “content” of delusions (e.g., paranoid vs. misidentification vs. delusions of fact), and whether brain-behavior relationships are consistent across other dementing and psychotic illnesses.

Clinical implications include enhanced understanding of the course of AD, and improved ability to predict who will develop delusions or other specific cognitive or psychiatric symptoms, who will experience greater morbidity associated with these symptoms, and which patients will respond to treatment.

Post-Mortem Studies of Neuropathology and Neurochemistry in Alzheimer’s
Disease with Psychosis

“Many investigators have been interested in whether or not there is a specific pattern of plaques and tangles that lead to psychosis in Alzheimer’s disease (AD),” said Robert A. Sweet, MD, associate professor, Department of Psychiatry, University of Pittsburgh School of Medicine. “It’s a mixed bag right now; there is no good evidence one way or the other that specific neuropathologic markers are associated with the presence of psychosis in AD patients,” said Dr. Sweet.

Postmortem studies of the potential association between psychotic symptoms in AD and the presence of more severe neuropathologic features (e.g., specific patterns of plaques and tangles) have yielded conflicting results. A 2000 study by Dr. Sweet and colleagues did not reveal any significant relationship between presence of psychosis and specific neuropathologic markers in a number of neocortical and limbic brain regions (International Psychogeriatrics 2000; 12(4): 547-558). Similarly, a 1994 study by Forstl and colleagues demonstrated no significant differences in plaques, neurofibrillary tangle density and neuron density in AD patients with psychosis (Br J Psychiatry 1994;165:53-59).

In contrast, a 1991 study by Zubenko and colleagues associated psychosis in AD with increased density of senile plaques in the prosubiculum as well as increased neurofibrillary tangles in the middle frontal gyrus (Arch Neurol 1991;48:619-624). A larger, more recent study by Farber and colleagues demonstrated a generalized pattern of increased neurofibrillary tangles in the midfrontal, superior temporal and inferior parietal brain regions in AD patients with psychosis (Arch Gen Psychiatry 2000;57:1165-1173).

Results from these studies do not indicate whether observed increases in neurofibrillary tangles in psychotic AD patients are the result of actual numeric increases or a reflection of changes in density resulting from altered (i.e., shrunken) brain volumes in psychotic patients. Future research that accounts for reference volume will help clarify the reason for these discrepant findings.

In a recent postmortem study, Dr. Sweet and colleagues used magnetic resonance spectroscopy (MRS) to explore possible neurochemical causes of psychosis in AD patients. Results from comparisons of 15 non-psychotic and 12 well-matched psychotic subjects demonstrated both significant elevation of GPE (a marker of membrane breakdown) and significant reduction of NAA (a marker of neuronal integrity) in AD psychotics. Other metabolites trended differently for psychotics, but not at significant levels (Figure 1).

More specific regional analysis of NAA decreases indicated most extreme reductions in neocortical areas, with significant decreases in the superior temporal gyrus and inferior parietal cortex. Concentration of GPE was significantly elevated in the inferior parietal and occipital cortex, and levels of GPC (another marker of membrane breakdown, not significantly different overall) were elevated significantly in the occipital cortex. Reductions in concentration of GABA (the primary inhibitory neurotransmitter in the brain) were similarly accentuated in neocortical regions, though no single region demonstrated a significant decrease versus non-psychotic AD subjects (Neurobiology of Aging, In press).

Study results provide evidence for involvement of the neocortex in AD with psychosis, consistent with excess neuronal and/or synaptic degeneration. Further studies are needed to better understand the relationships between the observed neurochemical differences and any structural brain changes in psychotic AD patients. In addition, research comparisons between schizophrenics and psychotic AD subjects may identify common pathways and help understand mechanisms for developing therapies.

“I think there’s promise for postmortem studies of Alzheimer’s psychosis, with the caveat that we need to pay a lot of attention to what region we’re analyzing and how it compares in size in the psychotic and non-psychotic group,” said Dr. Sweet.

³¹P and ¹H MRSI Studies of Psychosis in Alzheimer’s Disease

“Evidence seems to indicate that there are significant alterations going on at a molecular level in psychotic Alzheimer patients,” said Jay W. Pettegrew, MD, professor, Department of Psychiatry and Neurology and Health Services Administration, Western Psychiatric Clinic, University of Pittsburgh School of Medicine; director, Neuro Physics Lab, University of Pittsburgh. “Most of these alterations relate to the turnover of membrane phospholipids, which are vitally important in the function of synapses,” said Dr. Pettegrew.

Dr. Pettegrew and colleagues used non-invasive in vivo spectroscopy to determine whether differences exist at a molecular level in Alzheimer’s disease (AD) patients with and without psychosis (when degree of dementia is constant). Individual voxels were examined in various locations of the brain. A generalized linear model was developed to allow analysis of correlated data arising from 12 regional measurements.

Magnetic resonance spectroscopy results from a pilot study of 25 non-psychotic and 6 psychotic AD patients with delusions but no hallucinations demonstrated that psychotic subjects have significant alterations in several molecular and metabolic measurements that distinguish them from non-psychotic subjects.

Findings indicated substantial elevation in psychotics of membrane phospholipid building blocks in the right dorsal prefontal cortex (seemingly consistent with attempts to make or repair membranes). Psychotics also had increased levels of membrane breakdown products in the right side inferior parietal region, indicating more significant degeneration of the brain at the molecular level. Attempts at brain repair were apparent in the dorsal prefrontal cortex, while degeneration seemed to occur in the right inferior parietal cortex, where the brain was unable to repair itself. (These findings confirm earlier research results indicating atrophy in the inferior parietal cortex and decreased uptake of deoxyglucose.)

Results also demonstrated decreased levels of phosphorylated macromolecules in certain areas of the brain in psychotics, particularly on the right side (right prefrontal and right inferior parietal). Psychotics also exhibited decreased levels of synaptic and transport vesicles in the dorsal prefrontal cortex.

Other molecular level differences in psychotic AD patients included increased N-acetyl aspartate (NAA/phosphocreatine+creatine {PCR+Cr}) ratio in the right prefrontal region, and decreased NAA/PCr+Cr ratio in the right inferior parietal, possibly reflecting an attempt to repair the neuronal membrane and restore levels of NAA molecules. (These findings confirmed previous research that demonstrated elevated phosphocreatine in the right inferior parietal, likely reflecting reduction in high-energy phosphate consumption due to destruction of synaptic activity.)

In addition, increased levels of high-energy metabolites were apparent in the right inferior parietal, and decreased levels of glycosylating molecules were seen in the right prefrontal cortex and right basal ganglia, indicating reduced glycosylation activity in psychotics.

In conclusion, psychotic AD patients exhibited apparent alterations in membrane building blocks, breakdown products, energy metabolism, a putative neuronal marker, and synaptic vesicle content. Changes also were apparent in molecules that likely reflect the complex phosphorylation/dephosphorylation cascade, and the process of adding sugars to these large molecules. Additional research is needed to help confirm hypotheses regarding location of these changes (mostly right side, perhaps anterior) and to better understand why molecular alterations occur and why psychotic symptoms occur in some AD patients and not in others.