Biomarkers in preclinical familial Alzheimer disease

Abstract

Background: Alzheimer disease (AD) is a neurodegenerative disorder, characterized by the accumulation of b-amyloid (Ab) plaques and tangles consisting of hyperphosphorylated tauprotein in the brain. It accounts for 60-70% of dementia cases, making it the most common cause of dementia. In rare cases the disease is inherited in autosomal dominant early onset form caused by mutations in APP, PSEN1 or PSEN2. These familial forms of AD (FAD) allow for studies of the long preclinical stage of the disease and may thereby address unanswered questions about the natural history of AD which can be used to develop optimal tools for early diagnosis and for monitoring treatment response, as well as finding new possible treatment targets. To this end we conducted a prospective study, involving repeated clinical evaluations and collection of biomarkers from asymptomatic carriers of mutations leading to FAD with non-carriers (NC) from the same families as controls. The asymptomatic mutation carriers (MC) are good representatives of the preclinical stage of AD as they will develop symptoms of the disease in the future at an age which can be estimated based on the age at symptom onset in their family members who have already become symptomatic.

Aims: To map biomarker changes in preclinical AD, as well as their temporal trajectories and sequence, through repeated collection and analysis of biomarkers in asymptomatic FAD MC and NC.

Results: There were significant differences in the levels of the cerebrospinal fluid (CSF) biomarkers Ab42, total-tau protein (t-tau) and phosphorylated tau-protein (p-tau), as well as in the Ab42/p-tau ratio when comparing MC to NC, more than 7 years before the expected onset of symptoms in the MC. Ab42 and the Ab42/p-tau ratio were lower in MC than NC, while ttau and p-tau were higher in MC than NC. There was a trend of Ab42 and the Ab42/p-tau ratio decreasing as the onset of symptoms approached in MC, while t-tau and p-tau showed a trend of increasing with approaching symptom onset. On structural magnetic resonance imaging (MRI) of the brain, the MC had reduced volume of the left precuneus, left superior temporal gyrus and left fusiform gyrus, 9 years before the expected symptom onset. However, there was no observable decline in grey matter thickness or volume as the onset of symptoms approached, making the temporality of these changes difficult to assess. In the same group of subjects there was no significant difference on neuropsychological assessments between MC and NC, but a trend of poorer results was observed in the MC regarding immediate memory, episodic memory and attention/executive function. The CSF biomarkers YKL-40, reflecting glial activation, and neurogranin, a synaptic marker, were compared between asymptomatic MC and NC and found not to differ between the groups. A longitudinal study of changes in YKL-40 and neurogranin with approaching symptom onset was also conducted, revealing an increase in YKL-40 in both MC and NC as the age of symptom onset drew nearer, with a steeper increase in MC than NC. No such correlation to years to symptom onset was found for neurogranin. The APP processing products sAPPa, sAPPb, Ab42, Ab40 and Ab38 were compared both between the MC group as a whole and the NC and between subgroups of MC carrying specific mutations and the NC. The whole MC group had lower levels of Ab42, Ab40 and Ab38, as well as a lower Ab42/Ab40 ratio than NC. No significant correlation was observed between any of the aforementioned APP processing products and years to symptom onset in MC. When comparing different MC subgroups to each other, the whole MC group and the NC group, some mutation specific differences in the levels of the APP processing products and their temporality emerged. During the biomarker studies presented above the presence of a statistical outlier came to our attention, an MC carrying the PSEN1 H163Y mutation who had passed the age at symptom onset in his family but displayed no cognitive decline and no abnormalities in CSF biomarkers. This individual had been followed-up within the FAD study for 22 years and had opted for a presymptomatic genetic test, making his mutation status known to him and to the researchers involved in the study. His clinical case was characterized in paper III, with his brother serving as a control. The brother was only one year older than the outlier but had already passed away from AD at the end of the follow-up time, having displayed typical signs and symptoms of the disease in the preceding years.

Conclusions: The study revealed early preclinical changes in CSF biomarkers, reflecting Ab aggregation, glial activation, tau phosphorylation and neurodegeneration, as well as loss of volume in specific areas in the left hemisphere of the brain on structural MRI in asymptomatic carriers of FAD mutations. When assessing the temporality of specific biomarkers in the CSF, Ab42 and the Ab42/p-tau ratio seemed to decrease with approaching symptom onset, while ttau and p-tau increased as symptom onset drew nearer. These results are based on crosssectional data, but only longitudinal studies can properly assess temporal changes, as we did for CSF neurogranin and YKL-40 (with YKL-40 increasing at a faster rate in MC than in NC). However, the overall results give an important indication of the true nature of these preclinical temporal changes. We also observed mutation specific differences in APP processing products in the CSF and characterized a case of reduced penetrance of the PSEN1 H163Y mutation. In conclusion, the study sheds light on preclinical biomarker changes in FAD and the possible sequence of these changes. It also emphasizes the differences in phenotype between specific FAD mutations and the presence of reduced penetrance which affects the estimation of symptom onset in these families and has an impact on genetic counseling and possibly on the design of clinical trials in this population.