Introduction

The number of patients suffering from dementia is constantly rising, in Germany from an estimated 935,000 in the year 2000 to approximately 2.3 million by 2050 (Bickel 2000). Therefore, research into parameters relevant to diagnosis and therapy of dementia is of great interest. Activity of intracellular phospholipases of the A2-type (PLA2), acting as key enzymes of membrane repair and remodeling [“housekeeping” enzymes (McLean et al. 1993)], is an interesting target of this research, as it allows the characterization of biologically active processes at cell membrane level. Generally, PLA2 enzymes catalyze the hydrolysis of the middle ester bond of membrane phospholipids, to which a polyunsaturated fatty acid, in turn acting as a second messenger, is often bound (Six and Dennis 2000). Due to this important cell-metabolic function, enzymes of the PLA2 family are involved in numerous processes known to be disturbed in dementia, for example, exocytosis as one aspect of neurotransmission (Bloch-Shilderman et al. 2002; Ray et al. 1999), generation of acetylcholine (Blusztajn et al. 1987; Farooqui et al. 1992), induction of memory by long term potentiation (Fujita et al. 2001; Wolf et al. 1995), memory processing (Hölscher and Rose 1994; Hölscher et al. 1995; Fujita et al., 2000; Sato et al. 2007; Schaeffer and Gattaz 2005, 2007), maintenance of membrane fluidity with influence on receptor function (Farooqui et al. 2004), and antioxidative defense mechanisms (Farooqui et al. 2000).

The group of phospholipase A2 enzymes constitutes a so called “superfamily”, involving three major groups: (1) secretory (extracellular) Ca2+-dependent PLA2 (sPLA2); (2) cytosolic Ca2+-dependent PLA2 (cPLA2); and (3) intracellular Ca2+-independent PLA2 (iPLA2) (Dennis 1994; Sun et al. 2004; Balsinde et al. 1999; Taketo and Masahiro 2002). Previous studies on changes of intracellular PLA2 activity in cases of dementia focused on platelets or post mortem brain tissue of patients with Alzheimer disease (AD), showing a decrease in enzyme activity in almost all studies (Gattaz et al. 1996b, 2004; Ross et al. 1998; Talbot et al. 2000). This is of interest, as other psychiatric disorders were associated with different PLA2 findings, such as increased PLA2 activity in schizophrenia (Gattaz et al. 1987, 1990, 1995; Lasch et al. 2003; Noponen et al. 1993; Ross et al. 1997, 1999; Smesny et al. 2005; Tavares et al. 2003), and no PLA2 alteration in depression or bipolar disorder (Albers et al. 1993; Gattaz et al. 1987, 1990, 1995; Katila et al. 1997; Noponen et al. 1993; Ross et al. 1999). In schizophrenia, increased PLA2 activity is interpreted as reflecting an ongoing regenerative process compensatory to neurotoxic effects of the acute psychotic state (Law et al. 2006). An understanding of PLA2 decrease in AD is still lacking. Furthermore, studies on specificity of PLA2 alterations among different dementia disorders, including patients with non-Alzheimer dementia, are still not available. To our knowledge, PLA2 activity has also not been investigated in CSF as yet. Therefore, we investigated PLA2 activity for the first time in CSF, including not only patients with AD, but also patients with vascular dementia (VD) and mixed dementia (MD means AD and VD pathology) in order to detect the alterations of PLA2 in the CNS compartment and also in different dementia subtypes.

Methods and materials

Subjects

A total of 101 subjects were screened between 2004 and 2007 for participation in this study. Patients with inflammatory or infectious diseases (n = 17) or those who did not allow a doubtless diagnosis of dementia (n = 22) were excluded, leaving a total of 62 subjects for whom cerebrospinal fluid (CSF) was investigated (demographical data in table 1). All subjects of the patient group were included from consecutive admissions to the geriatric psychiatry unit at the Department of Psychiatry, University of Jena (inclusion criteria: Mini Mental Status Test score ≤23; Clock drawing test score ≥3; see also Table 2). Subjects of the control group were included at the Departments of Neurology and Anesthesiology of the University of Jena. Thus in patients, CSF was taken as part of the routine diagnostics for dementia. In controls, CSF was taken in the context of epidural anesthesia or to exclude neurological disorders. For the healthy control group, we only included subjects found to be free of any CNS inflammatory/infectious disease. While co-variance and correlation analysis to control for effects of age, gender, or medication (statins, acetylsalicylic acid, cholinesterase inhibitors) did not show any influence of these co-variates on the CSF-PLA2-activity, the mean age of patients and controls was significantly different. Neither patients nor controls were given any medication other than mentioned above.

Table 1 Demographical data: groups of patients, patient subgroups [Alzheimer disease (AD), vascular dementia (VD), mixed dementia (MD)] and control subjects (C) in total and separated for gender, mean value of age ± standard deviation
Full size table
Table 2 Diagnostics: neurocognitive test battery, cut-off values used as inclusion criteria, MRI-Scan and routine investigation of blood, CSF parameters and cut-off-values for differentiation between AD/MD and VD
Full size table

CSF samples were acquired in the morning by lumbar puncture using atraumatic cannula, and were immediately divided into two or more sub-samples. One sub-sample was used for routine diagnostics (cell count, protein content, microbiology, tau-Protein, β-Amyloid etc.), the other sub-sample underwent immediate centrifugation to remove cell debris and was stored at −80°C until PLA2 analysis.

To assure the clinical and screening diagnosis of dementia, all patients underwent an extensive diagnostic program (procedures and cut-off values are given in Table 2).

AD and VD were differentiated by standardized criteria taking into account the history, clinical presentation and structural abnormalities [assessed according to the National Institute of Neurological Disorders and Stroke–Association Internationale pour la Recherche et l′Enseignement en Neuroscience (NINDS–ARIEN) criteria (Roman et al. 1993), and the Alzheimer Disease Diagnostic and Treatment Center (ADDTC) criteria, the latter also proposing the definition of “mixed” categories, as used in this study (MD)(Chui et al. 1992)], neurocognitive testing (consortium to establish a registry for Alzheimer’s Disease (CERAD)-series (Morris 1997; Morris et al., 1989), the Nuremberg Aging Inventory (NAI: Nürnberger Altersinventar) (Oswald and Fleischmann 1999), and established diagnostic CSF parameters (phosphorylated Tau-Protein, β-Amyloid 42/40 ratio). Intending to investigate a naturalistic population, a group of patients with MD was established, including patients with features of both AD and VD (e.g. increased Tau-Protein and decreased β-Amyloid-Ratio and SAE, vascular risk factors etc.).

The study was approved by the Research Ethics Committee of Friedrich-Schiller-University Jena. All subjects or their legal guardians gave written informed consent to participate in the study.

Analysis of PLA2 activity

Most of the intracellular PLA2 enzymes investigated here need calcium in micromolar concentrations at most (cPLA2) or are completely independent of calcium (iPLA2). Therefore, according to the most recent genetically defined classification, PLA2 activity investigated in this study comprises most likely activity of group IV and group VI type isoenzymes (Sun et al. 2004). This classification of our target enzyme activity is based on previous methodical investigations (Lasch et al. 2003) and the actual adaptation of the serum PLA2 assay on CSF. This research showed an almost complete (more than 90%) inhibitory effect of calcium ions on enzyme activity and a 70% inhibition of the enzyme activity by bromoenole lactone (BEL), a suicide inhibitor of iPLA2 (Jenkins et al. 2002; Lucas et al. 2005; Song et al. 2006; White and McHowat 2007). The use of selective antibodies could reveal that PLA2 activity in blood serum and CSF results from identical enzyme proteins. Previous research also revealed that our results do not reflect the activity of PAF-Hydrolases, as PAF-Hydrolases do not cleave the used commercial substrate. There was also no reaction with PAF-Hydrolase antibodies.

Thus, the PLA2 subtypes and PLA2 assay were basically the same as has been already established for our investigations in schizophrenia patients (Lasch et al. 2003; Smesny et al. 2005). However, the reaction stock was now adapted to the requirements of measurements in CSF. Briefly, it included 80 μl undiluted CSF, 10 μl HEPES buffer (N-2-hydroxyethylpiperazine-N-2-ethane-sulfonic acid, pH 7.4, 0.4 M) and 10 μl of the commercial fluorogenic substrate NBDC6-HPC (2-(6-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)hexanoyl-1-hexadecanoyl-sn-glycero-3-phosphocholine, Molecular Probes, Europe BV Leiden, The Netherlands). The reaction batch was then incubated for 60 min at a temperature of 37°C. Finally, a separation of the reaction products using thin-layer chromatography and digital image scanning for signal detection followed as described in more detail by Lasch and colleagues (Lasch et al. 2003). Storage time of CSF samples before PLA2 analysis differed between 4 and 73 days (mean ± standard deviation: 38 ± 25). In both patients and controls, there was no association between the storage interval and enzyme activity (patients: r = −2.3, n.s.; controls: r = −1.4, n.s.).

Data analysis

Statistical procedures included analyses of variance for general effects and post hoc tests (significance level α = 0.05). Possible effects of gender and medication were investigated by co-variate analysis and (if possible) subgroup analysis. Possible effects of age or duration of CSF storage were investigated by calculating correlation coefficients. For post hoc analysis, the double sample t-Test was used when variance was equal, and the Welch Test (a more robust version of the t-Test) when variances differed (significance level α = 0.05).

Results

Comparison of groups

Initial ANOVA revealed a main effect of GROUP, indicating differences of CSF-PLA2 activity between patients with AD, VD, MD, and healthy controls (F 3;61 = 3.399; P = 0.024). Post hoc analyses showed significantly smaller PLA2 values in patients with AD (t 27;0.975 = 2.332; P = 0.027), patients with VD (t 27;0.975 = 2.67; P = 0.013, excluding one extreme value), and in patients with MD (t 27;0.975 = 2.575; P = 0.016) as compared to controls (Fig. 1). Also, there was significantly reduced PLA2 activity comparing a merged sample of AD and MD patients with controls (t 23;0.975 = 2.62; P = 0.015).

Fig. 1
figure 1

Specific PLA2-activity in CSF presented in boxplots including the smallest and largest observation, the lower quartile (25%), the higher quartile (75%) and the median value; separated for patient subgroups (AD Alzheimer disease, VD Vascular dementia, MD Mixed dementia) and healthy controls

Full size image

Effects of age and gender

Age: There was no correlation between age and PLA2 values in either the total sample (r = −0.44; n.s.) or healthy subjects (r = −0.13; n.s.).

Gender: ANOVA with GROUP as between-subject factor and GENDER as co-variate revealed no significant effect of the factor GENDER on PLA2 activity (F1;61 = 3.315; n.s.) and no significant GROUP × GENDER interaction (F3;61 = 0.564; n.s.).

Effects of medication

We tested the potential effects of statins, acetylsalicylic acid (max. 100 mg/day), and cholinesterase inhibitors by comparing medicated and non-medicated subjects in each case. ANOVA did not indicate any significant influence of either statins (F 1;55 = 2.317; n.s.), aspirin (F 1;43 = 1.526; n.s.), or cholinesterase inhibitors (F 1;61 = 0.361; n.s.) on PLA2 values.

Discussion

The study revealed two main findings: (1) decreased activity of intracellular PLA2 in CSF of patients with AD. (2) Decreased activity of intracellular PLA2 in CSF of patients with MD and VD. As in previous studies, PLA2 findings were not significantly affected by either age or gender (Smesny et al. 2005), nor by concurrent medication with acetylcholinesterase inhibitors (Gattaz et al. 2004), statins or acetylsalicylic acid (aspirin). There was also no correlation of PLA2 activity with any of the cognitive or routine clinical parameters (among others CSF tau-protein and beta-amyloid).

The first result of decreased PLA2 activity in CSF of patients suffering from AD is in good agreement with results of other groups investigating platelets and postmortem brain tissue of patients with AD (Gattaz et al. 1996a, 2004; Ross et al. 1998; Talbot et al. 2000). This is to our knowledge the first study directly showing PLA2 abnormalities in the CNS compartment of patients suffering from AD. Considering that in total 31 patients (AD + MD) were identified with features of AD, the study also included one of the largest samples being investigated in this field so far.

We are not aware of any study till date, investigating PLA2 activity in dementia other than AD. In our study, decreased PLA2 activity was not exclusively associated with Alzheimer-type pathology. Including patients with non-Alzheimer disease, we were also able to show decreased PLA2 activity in CSF of patients suffering from VD. Decreased PLA2 activity in non-Alzheimer dementia is suggestive of an underlying pathomechanism, common to these different dementias. Indeed, shared pathophysiology of AD and VD has been also repeatedly discussed in the literature (Hentschel et al. 2005; Jellinger 2002; Kalaria and Ballard 1999). Some of them require intact PLA2 function and are therefore of interest here, especially with regards to cholinergic dysfunction, oxidative stress, and disturbances of membrane function.

Wide agreement exists about the substantial role of a cholinergic deficit and a disruption of cholinergic neurotransmission in AD (Gsell et al. 1996, 2004) as well as in VD (Pratt and Perdomo 2002; Tomimoto et al. 2005), representing a major correlate of cognitive deficits (Bierer et al. 1995). The formation of free choline triggered by PLA2-catalyzed hydrolysis of phosphatidylcholine (Blusztajn et al. 1987; Farooqui et al. 1992) is an important molecular pathway for de novo synthesis of acetylcholine in cholinergic neurons. This is in accordance with findings of induction of phosphatidylcholine synthesis followed by increased PLA2 activity (Barbour et al. 1999). Hence, reduced activity of intracellular PLA2 could aggravate or even cause the cholinergic deficit seen in AD and VD. In addition, impaired function of PLA2 enzymes has also general effects on neurotransmission, for example both the cPLA2 and the iPLA2 subtype are involved in exocytosis (Bloch-Shilderman et al. 2002), whereas the Ca2+ activated cPLA2 is involved in the release of neurotransmitters (Ray et al. 1999).

The most striking clinical feature of dementia is memory impairment, often starting with impaired short-term memory function. Changes in short term memory in particular can be traced back to the interference of long-term potentiation (Chapman et al. 1999; Dawson et al. 1999; Morton et al. 2002), which is also impaired in patients with AD (Battaglia et al. 2007). Interestingly, five different animal studies were able to show a direct degradation of memory functions through the intra-cerebral inhibition of PLA2 (Fujita et al. 2000; Hölscher and Rose 1994; Schaeffer and Gattaz 2005, 2007, Sato et al. 2007), presumably also explained by the involvement of PLA2 in long-term potentiation (Fujita et al. 2001; Wolf et al. 1995).

Thirdly, intracellular PLA2 subtypes are crucially involved in membrane repair and remodeling processes. Inhibition of cPLA2 and iPLA2 was shown to result in reduction of membrane fluidity (Schaeffer et al., 2005), which in turn was associated with memory deficits in animal models (Clarke et al. 1999; Hong 1995) and in patients with AD (Eckert et al. 2000; Mecocci et al. 1996, 1997).

Reduced availability of choline-containing compounds, disturbance of exocytosis, interference of long-term potentiation, and impaired membrane fluidity are inter-related through their dependency on intact membrane lipid metabolism. As nervous tissue is naturally highly vulnerable to lipid peroxidation (due to high number of double bounds in membrane phospholipids), one important reason for membrane lipid alterations is increased oxidative stress. Increased lipid peroxidation was repeatedly found in both AD (Mattson 2002; Nunomura et al. 2001; Perry et al. 2002; Sayre et al. 1997), and VD (Ihara et al. 1997; Paragh et al. 2002). Acting as a “housekeeping enzyme” of membrane reconstruction, intracellular PLA2 counteracts increased oxidative stress in hydrolyzing peroxidized fatty acids (Baba et al. 1993; McLean et al. 1993; Salgo et al. 1993; van den Berg et al. 1993). In the case of reduced PLA2 activity, restricted neuronal regeneration mechanisms might promote damages resulting from lipid peroxidation processes.

Our finding of decreased PLA2 activity in both AD and VD indicates PLA2 deregulation, possibly not to be specifically associated with Alzheimer-type pathology. Considering the finding of decreased PLA2 activity as a secondary effect of cellular changes, two different mechanisms are plausible: (1) the functional capacity of PLA2 is exceeded due to increased need for membrane repair processes associated with the disorder. This assumption would also explain the finding of increasing loss of PLA2 function with severity of illness, reported by other groups (Gattaz et al. 1996a, 2004). (2) Decreased activity of PLA2 could also be caused by primary disturbance of enzyme protein function or enzyme regulation. In that case, one would expect decreased PLA2 activity already in the early state of dementia (mild cognitive impairment) or even before clinical manifestation. However, this aspect has not been investigated in detail.

In summary, this first study of intracellular PLA2 activity in CSF corroborates the finding of impaired PLA2 function in AD and extends these to patients with VD. Taken together, our results are likely to reflect an involvement of PLA2 impairment in a variety of pathomechanisms crucial in different dementia subtypes, in which disturbance of intact membrane function appears to be a key mechanism. Membrane function in turn is markedly susceptible to increased oxidative stress, which is counteracted by PLA2. Further studies might characterize the potential of PLA2 activity to serve as marker of the individual capacity to resist to oxidative damage and the related pathology occurring in dementia. This research in the prodromal or early phase of disorder could initiate new preventative, diagnostic, and therapeutic approaches, which, especially against the backdrop of ever increasing proportions of older people in our society, would be of immense importance.