Biomarkers help identify people on the path to Alzheimer’s, but so far have not proved useful as surrogate outcome measures for clinical improvement in trials. Because cognitive loss correlates with damage to synapses, synaptic markers might do a better job of tracking clinical outcomes. However, these proteins occur in very low levels in cerebrospinal fluid (CSF), making them difficult to measure. In two recent papers, researchers at the University of Gothenburg, Sweden, described methods for detecting SNAP-25 and neurogranin in the CSF of individual patients. In three small, independent cohorts, levels of these synapse proteins were higher in people with Alzheimer’s dementia than in controls, the scientists reported. Moreover, elevated neurogranin in people who were cognitively impaired predicted progression to dementia, and correlated with a swifter rate of cognitive decline. The results hint that these markers could be useful as both early diagnostics and outcome measures, although this remains to be confirmed in larger studies and clinical trials, noted Kaj Blennow, a senior author on both studies.

Other researchers praised the methodology of the studies and noted that the markers offer promise. “The technical characterization and validation of the methods used is quite exemplary,” Marc Gleichmann at EMD Serono, Billerica, Massachusetts, a division of Merck, wrote to Alzforum (see full comment below). “Both CSF-SNAP25 and CSF-Ng are welcome additions to the existing CSF biomarkers, and further use and characterization is warranted.”

In previous investigations, the Gothenburg group had to pool CSF samples to detect synaptic proteins (see, e.g., Davidsson et al., 1999; Thorsell et al., 2010). To improve sensitivity, the researchers used monoclonal antibodies to purify each synaptic target protein from spinal fluid, then quantified each with mass spectrometry. This refinement allowed the detection of the proteins in CSF samples from individual patients, according to the new papers.

For synaptosomal-associated protein 25 (SNAP-25), first author Ann Brinkmalm compared CSF levels in three separate cohorts that together comprised 36 AD patients and 32 age-matched controls. In every group, patients had significantly higher SNAP-25 on average than controls. One cohort also included seven people with prodromal AD, who displayed intermediate levels of SNAP-25. SNAP-25 levels correlated with CSF total tau and phosphorylated tau, but not with Aβ42, suggesting that the marker tracks neurodegeneration. Among Alzheimer’s patients, SNAP-25 also correlated with the degree of cognitive impairment, as measured by scores on the MMSE.

The results parallel those for neurogranin. Joint first authors Hlin Kvartsberg and Flora Duits compared CSF levels of this postsynaptic marker in three separate cohorts totaling 100 AD patients and 80 controls, and an additional 40 people diagnosed with mild cognitive impairment (MCI). Patients had more neurogranin in the CSF than did controls in all studies. Among the MCI group, those judged to have prodromal AD had neurogranin levels that resembled those in patients, while those with other forms of MCI looked like controls. In addition, MCI patients with high neurogranin levels had 13 times higher odds of progressing to dementia than those without. Higher levels of neurogranin also predicted a steeper rate of decline on the MMSE in the prodromal AD patients. As with SNAP-25, neurogranin levels correlated with CSF total tau and phosphorylated tau, but not with Aβ42.

The presence of synaptic proteins in CSF likely reflects degeneration of synapses, Blennow told Alzforum. Tau, on the other hand, may measure breakdown of axons and cell bodies. “We show that they correlate, but we don’t believe they tell you the same thing,” Blennow noted. The exact staging of these markers will require longitudinal studies, he added. Currently, Blennow and colleagues are measuring neurogranin levels in CSF samples from the Alzheimer’s Disease Neuroimaging Initiative to see if the findings repeat. They are also setting up collaborations to test the synaptic markers’ usefulness as outcome measures in clinical trials.

Commentators suggested other experiments, as well. “It will be interesting to correlate the SNAP-25 and neurogranin makers with non-CSF putative markers of neuronal and synaptic injury, such as hippocampal atrophy or even tau PET imaging. It will also be important to learn whether these synaptic markers identify cognitively normal individuals with preclinical AD,” John Morris at Washington University in St. Louis wrote to Alzforum.

Will the methods for measuring these synaptic markers be practical for widespread use? Most larger Alzheimer’s centers have mass spec equipment and could perform these protocols, Blennow said. In addition, the researchers measured neurogranin using an ELISA assay, and obtained comparable results to the mass spec approach. ELISAs are simpler and more convenient for smaller laboratories, suggesting this biomarker could be widely adopted.—Madolyn Bowman Rogers


  1. In my view, high methodological rigor was applied in both studies. The use of three separate cohorts with good clinical characterization and the technical characterization and validation of the methods used are quite exemplary. I have high confidence that the core findings of both papers are valid, i.e., CSF SNAP-25 and CSF-NG are, on average, increased in AD patients versus controls. The studies also demonstrate that it is possible to use moderate sample sizes and still draw solid conclusions. The higher correlation with CSF-tau and CSF-phospho-tau versus CSF-Aβ supports the hypothesis that both biomarkers are indicators of structural degeneration. How far they are indicators of degeneration specifically at synapses, as opposed to general neuronal degeneration, remains unclear. In this regard it may be of interest to compare the newly characterized biomarkers with CSF-NSE (neuron-specific enolase), a cytosolic protein believed to leak from structurally damaged neurons and reported to also be increased in CSF of AD patients (Schmidt et al., 2014; Palumbo et al., 2008; Bahl et al., 2009; Blennow et al., 1994).

    Both CSF-SNAP25 and CSF-Ng are welcome additions to the existing CSF biomarkers and further use and characterization is warranted. The results of additional, larger, and prospective studies will inform us about the value of both biomarkers in the field of Alzheimer’s disease. Potential uses include clinical studies aimed at reducing neuronal and synaptic degeneration with therapeutic intervention. However, at this point no therapy has conclusively demonstrated reduction of neuronal degeneration or clinical progression in Alzheimer’s disease, and until this has been achieved there is a bit of a Catch-22 for biomarkers aimed at measuring the efficacy of therapeutic interventions.

    Diagnosis and differential diagnosis are other potential uses for these markers and it remains to be seen if the addition of CSF-SNAP25 and CSF-Ng to CSF-(p)tau and CSF-Aβ will significantly increase sensitivity or specificity of AD (differential) diagnosis. The hurdle in this area is quite high, however, and from the papers it is quite clear that normal levels of CSF-SNAP25 or CSF-Ng do not rule out a diagnosis of AD. Also, in the future it may be worthwhile to test the biomarkers in other neurodegenerative diseases and dementia, to better characterize how specific the markers are for AD versus neurodegeneration in general. I see the best potential for these biomarkers in helping to further refine prognosis in individuals with MCI. Depending on the outcome of larger prospective studies with MCI patients, the biomarkers may also be used for selecting clinical trial populations in preventive studies. CSF-Ng can be measured by ELISA, which may facilitate wider use and spread of the method when compared with the more demanding mass spectroscopy methods. Before seeing data from larger MCI cohorts, it seems premature to me to judge which, if either, of the two biomarkers holds more potential.


    . Elevated levels of cerebrospinal fluid neuron-specific enolase (NSE) in Alzheimer's disease. Neurosci Lett. 2014 Jun 6;570:81-5. Epub 2014 Apr 16 PubMed.

    . Cerebrospinal fluid neuron-specific enolase: a further marker of Alzheimer's disease?. Funct Neurol. 2008 Apr-Jun;23(2):93-6. PubMed.

    . The diagnostic efficiency of biomarkers in sporadic Creutzfeldt-Jakob disease compared to Alzheimer's disease. Neurobiol Aging. 2009 Nov;30(11):1834-41. PubMed.

    . Neuron specific enolase in cerebrospinal fluid: a biochemical marker for neuronal degeneration in dementia disorders?. J Neural Transm Park Dis Dement Sect. 1994;8(3):183-91. PubMed.

Make a Comment

To make a comment you must login or register.


Paper Citations

  1. . Identification of synaptic vesicle, pre- and postsynaptic proteins in human cerebrospinal fluid using liquid-phase isoelectric focusing. Electrophoresis. 1999 Mar;20(3):431-7. PubMed.
  2. . Neurogranin in cerebrospinal fluid as a marker of synaptic degeneration in Alzheimer's disease. Brain Res. 2010 Nov 29;1362:13-22. Epub 2010 Sep 25 PubMed.

Further Reading

Primary Papers

  1. . Cerebrospinal fluid levels of the synaptic protein neurogranin correlates with cognitive decline in prodromal Alzheimer's disease. Alzheimers Dement. 2015 Oct;11(10):1180-90. Epub 2014 Dec 19 PubMed.
  2. . SNAP-25 is a promising novel cerebrospinal fluid biomarker for synapse degeneration in Alzheimer's disease. Mol Neurodegener. 2014 Nov 23;9:53. PubMed.