Is there a CSF biomarker profile related to depression in elderly women?

Abstract

In light of our previous observation of higher levels of cerebrospinal fluid (CSF) amyloid beta-42 (Aβ42) and CSF/serum albumin ratio in major depressive disorder (MDD), we analyzed two additional CSF biomarkers reflecting neurodegeneration—neurofilament protein light (NFL) and glial fibrillary acidic protein (GFAp)—in relationship to prevalent geriatric depression. Neuropsychiatric, physical, and lumbar puncture examinations, with DSM-III-R-based depression diagnoses and measurement of CSF levels of NFL and GFAp, were evaluated among a population-based sample of 78 elderly women (mean age, 73.9 ± 3.2 years) without dementia for at least 10 years after CSF collection. Eleven (13.1%) women had MDD, and higher levels of NFL compared with women without depression. A multivariate model including age, NFL, Aβ42 and the CSF/serum albumin ratio showed that each biomarker was independently and positively associated with MDD, and that this biomarker profile explained more variation in the model compared with single or combined biomarkers. A CSF profile with higher levels of NFL, Aβ42, and CSF/serum albumin ratio may indicate neuropathological and vascular events in depression etiology. This contrasts with the well-characterized pattern of low Aβ42, higher CSF/serum albumin ratio, and higher NFL in Alzheimer's disease.

Introduction

Our group previously observed higher levels of cerebrospinal fluid (CSF) amyloid beta-42 (Aβ42) and a higher CSF/serum albumin ratio in elderly women with major depressive disorder (MDD) (Gudmundsson et al., 2007), suggesting a combination of neuropathological and vascular factors in depression that differs from that observed in Alzheimer's disease (AD) (Blennow and Galasko, 2000, Gustafson et al., 2007) and underscoring the complex etiology of depression (Arean and Reynolds, 2005). There are two predominant biological paradigms for the etiology of late-life depression including vascular perturbations and neurodegeneration (Naarding et al., 2005). Since cerebrovascular disease, especially white matter lesions, has been reported in late life depression (Alexopoulos et al., 1997), it has been suggested that depression in the elderly may be related to vascular disease in the brain. However, this hypothesis is not supported by all studies (Breslau et al., 1995, Palsson et al., 2001).

Neurodegenerative factors may also be involved in the etiology of depression, and CSF markers reflecting neurodegeneration may be indicative of changes in the brain accompanying depression. For example, the integrity of neurons and glial cells can be studied indirectly through measurement of neurofilament proteins (Rosengren et al., 1996) and the glial fibrillary acidic protein (GFAp) (Rosengren et al., 1994) in CSF.

Neurofilaments are major components of the neuronal cytoskeleton and are particularly abundant in large myelinated axons, playing a central role in controlling axonal caliber (Hoffman et al., 1984, Yamasaki et al., 1991, Zhu et al., 1997), maturation of regenerating myelinated axons (Zhu et al., 1997) and growth of dendrites (Zhang et al., 2002). The neurofilaments are composed of a triplet protein, of which the light chain (NFL) is the subunit with the lowest molecular mass (Cereseto et al., 2006), and is the essential component of the neurofilament core (Rosengren et al., 1999). NFL is used as a marker of neuronal degeneration in animal brains (Karlsson et al., 1991, Cereseto et al., 2006), and CSF NFL levels are elevated in several human cerebral disorders such as cerebral infarction, multiple sclerosis (Rosengren et al., 1996, Norgren et al., 2003), late onset AD (Rosengren et al., 1999, Norgren et al., 2003), vascular dementia (Rosengren et al., 1996, Rosengren et al., 1999, Norgren et al., 2003), and acute brain trauma such as can be induced by boxing (Zetterberg et al., 2006), indicating structural damage of nerve cells. An association between elevated CSF NFL and the presence of white matter lesions has also been observed (Sjogren et al., 2001). One study has shown elevated levels in patients with major depression (Zachrisson et al., 2000).

GFAp, a monomeric intermediate filament protein, is an important component of the astrocyte cytoskeleton (Aurell et al., 1991). Elevated values of CSF GFAp have been demonstrated in central nervous system (CNS) injury (Aurell et al., 1991, Guéz et al., 2003) and in chronic disorders with astrogliosis (Rosengren et al., 1994, Malmestrom et al., 2003, Norgren et al., 2004). Alterations in brain density of GFAp have also been observed post-mortem in MDD patients (Johnston-Wilson et al., 2000, Muller et al., 2001, Webster et al., 2001, Webster et al., 2005, Davis et al., 2002, Fatemi et al., 2004, Si et al., 2004) as well as in studies using animal models of depression (Leventopoulos et al., 2007). However, others found no relationship between GFAp density and MDD (Miguel-Hidalgo et al., 2000, Damadzic et al., 2001).

In light of our previous findings of higher levels of CSF Aβ42 and a higher CSF/serum albumin ratio in elderly women with MDD (Gudmundsson et al., 2007), we have now analyzed two additional CSF biomarkers associated with neuronal degeneration, NFL and GFAp, in the same sample, with the additional criterion that they remain without dementia for up to 10 years after the lumbar puncture. These results are then combined with previously reported biomarkers observed in the same sample.