In this CSF proteomic study of persons inheriting FAD mutations, most of whom were presymptomatic, we identified known and novel candidate biomarkers. Differences were found in 56 CSF proteins between the MC and NC cohorts based on univariate analysis. The fold changes among the 56 proteins were relatively small, ranging from 1.18 to 1.90 (mean, 1.33). The differences in CSF Aβ42, t-tau, and p-tau181 found in MCs were large and consistent with reported fold changes for these traditional markers in persons with mild cognitive impairment and AD of later onset. In this study, the changes were detected 10 or more years before diagnosis.
Many more of the 56 proteins are upregulated (n=46) than downregulated (n=10) in MCs. The finding of both upregulated and downregulated proteins suggests that our results are not due to nonspecific brain degeneration, degradation of the blood-brain barrier, or contamination from blood. Such global changes would be attenuated by the normalization algorithm in the analysis software.
Fourteen proteins found to be differentially expressed in previous studies were also seen in our study, indicating that identifiable molecular changes occur 10 years before diagnosis. Transthyretin, albumin, and apolipoprotein A15
may also change prior to the appearance of symptoms but were among the abundant proteins depleted from our samples. There was sufficient residual α2
-macroglobulin and complement component 3 after depletion to identify it as differentially expressed, consistent with previous findings.3
Multiple complement components were elevated, consistent with previous studies19
indicating an early role of inflammation in AD. Some of the proteins showing the most robust differences between MCs and NCs (CD99 antigen isoform b, di-N
-acetyl chitobiase, and secreted phosphoprotein 1 isoform b) were unique, and their relationship to AD pathogenesis is unclear.
Numerous peptides derived from APP were detected and quantified. The median concentration change observed in these peptides is 23% upward for the MC cohort. The normal functions of APP are not clear but may be related to neurite growth, synaptogenesis, and synaptic plasticity. The APP peptides from the soluble extracellular region were consistently elevated. Our finding of elevated CSF APP before diagnosis in the study participants with FAD mutations is consistent with models in which Aβ42 aggregates induce a positive feedback cascade that elevates APP production through BACE (beta-site APP cleaving enzyme) activity.20
Both the current and previously reported21
data in the same population determined with use of immunoassays found reduced levels of Aβ42 in the CSF of the MCs, a finding thought to represent selective deposition of Aβ42 in the brain. The short Aβ peptides were likely removed in the proteomic processing used in this study, which uses a 5-kDa filter cutoff prior to tryptic digestion of these proteins.
Secreted phosphoprotein 1 (ie, osteopontin), a cytokine expressed by activated T cells, was elevated in the CSF of MCs. Osteopontin has been found to be elevated in the pyramidal cells in the brains of persons with AD22
and in transgenic mice23
; in addition, there is a single report24
of it being elevated in the CSF of persons with AD. Osteopontin is also elevated in multiple sclerosis,25
where it might serve as a therapeutic target. A similar role might be considered in AD.
Superoxide dismutase 3 was decreased in the CSF of the FAD MC group. The activity of superoxide dismutase has been found26
to be decreased in the CSF of persons with AD. Diminished superoxide dismutase activity could contribute to oxidative stress and therefore propagate disease pathologic characteristics. However, consistent with a proteomic study18
of CSF in postmortem AD cases, we found elevated levels of hemopexin, a protein that binds heme, in presymptomatic MCs, thus preventing its pro-oxidant and proinflammatory effects.27
We also observed changes in several proteins related to excitatory synapses. Neuronal pentraxin 2, neuronal pentraxin receptor, and the AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) 4 glutamate receptor were found to be elevated in presymptomatic MCs. Neuronal pentraxin 2 is a secreted protein homologous to acute-phase reactants and therefore an indicator of inflammation. Neuronal pentraxin receptor was found to be elevated in proteomic studies of CSF in patients with AD,6,17
but NP1 and neuronal pentraxin receptor were decreased in another such study.16
Neuronal pentraxin receptor is a transmembrane protein expressed on neurons and glia that binds AMPA receptors and plays a role in synaptogenesis.28
Elevation of these proteins in CSF could represent synapse turnover or loss occurring early in the course of the disease. Calsyntenin 3, a postsynaptic membrane protein with putative Ca2+
-binding capacity associated with γ-aminobutyric acid–ergic neurons, was decreased in the CSF of MCs in our study. The calsytenins have been linked to APP transport.29
A potential criticism of this study is the apparent lack of control for multiple comparisons. In the present study, P values were computed at the ion levels, whereas the tables are protein-level summaries of peptide results and consistent changes across peptides representing a single protein were required to make the list. Using the current methods in which hundreds of peptides were profiled, no false-positives were found at the protein level in an anonymized comparison of proteomics groups in which identical samples were spiked with differing levels of 12 proteins.
There are many advantages to the study of persons with FAD mutations. Discovery of biomarkers for late-onset neurodegenerative diseases that develop over decades is confounded by the fact that disease-associated markers may also change with normal aging.30
In addition, some healthy control participants in their 70s and 80s may be presymptomatic for neurodegenerative diseases diagnosed after the study. Another advantage of studying persons at risk for FAD is that the population is more homogeneous, enabling smaller cohort sizes, and that control participants and those who are presymptomatic carriers are younger, with fewer age-dependent neurologic and somatic changes. There are, however, potential limitations in the generalizability of findings in FAD to late-onset AD. It is thought that the pathogenesis of FAD is driven by a relative overproduction of Aβ42, whereas late-onset AD may be due to decreased degradation of Aβ. A previous investigation31
in this population revealed a unique pattern of APP-derived peptides in carriers of the A431E PSEN1
mutation that was not present in CSF from persons with late-onset AD.
In summary, although the statistical power of the present study was limited by small numbers, many novel proteomic differences were found in addition to others consistent with prior studies. We identified candidate biomarkers associated with mutation status that suggest changes occurring a decade before clinical dementia, including increases in APP, increases in inflammatory markers, and changes in proteins related to synaptic plasticity. These measures may be useful in diagnosis, patient stratification, and monitoring of response to therapy. Replication in a larger FAD population, longitudinal analysis of the same individuals, and comparison with changes in mild cognitive impairment and AD of late onset would provide further validation of our findings.