In this study we have shown that levels of carnosinase activity have the potential to differentiate AD from MD with very little overlap between these conditions. As the only known difference between these groups was the presence or absence of cerebrovascular disease, this suggests that vascular pathology may be the reason for lower carnosinase activity. Vascular dementia is associated with cerebrovascular accident, a condition known to lower carnosinase activity [23
]. Furthermore, compared to patients without dementia, carnosinase activity remains significantly decreased three months after a cerebrovascular accident [23
]. Decreased levels have also been found in Parkinson's disease [24
]. In an autopsy series of over 700 Parkinsonism patients, 19% had co-existing cerebrovascular lesions, and 20% of secondary Parkinsonism patients had vascular lesions [34
]. Our current findings allow speculation that carnosinase could differentiate "pure" AD from vascular dementia, which cannot be achieved with either CSF Aβ1–42 or CSF total tau [1
]. Further evidence supporting a vascular role for carnosinase comes from a recent study showing that its substrate carnosine causes vascular relaxation independent of the endothelium [28
Several reasons have been suggested for decreased carnosinase levels in neurological disorders. Wassif et al. [24
] suggested that a disruption of the blood brain barrier (BBB) could result in reduced activity of carnosinase seen in several disorders of the CNS. This idea is supported by the ability of cerebral ischemia to alter the BBB, specifically changing endothelial permeability and basal lamina integrity [35
]. The association of multiple sclerosis with BBB hyperpermeability further supports this idea [36
]. Skoog [37
] found higher CSF/serum albumin ratios in several types of dementia including AD and vascular dementia indicating an impairment of the BBB. Recently, aggregation of Aβ has been associated with damage to the structural microvascular and BBB abnormalities in an AD mouse model [38
]. Damage to carnosinase-producing cells could be another reason for reduced carnosinase activity. In stroke patients, Butterworth et al. [23
] found no relationship between size of the infarct and carnosinase activity. If death of carnosinase producing cells were responsible one would expect more cell death to cause a greater decrease in carnosinase activity.
] suggested that the decreased levels of carnosinase seen during cardiopulmonary bypass surgery were protective as both carnosine and homocarnosine can protect neurons against ischemia and oxidative stress. On the other hand, carnosinase degrades homocarnosine to GABA, an inhibitory neurotransmitter. Lower GABA levels may lead to increased neuronal damage and death. In one study of patients who had suffered a stroke, lower carnosinase activity was correlated with poorer outcome [23
A genetic basis for lower carnosinase activity is also possible. The gene for serum carnosinase is located on 18q22.3 [8
]. A child with serum carnosinase deficiency was found to have a deletion distal to 18q21.3 [39
] and recently a locus on chromosome 18 was identified for familial AD in Caribbean Hispanics [40
]. These two reports, along with a study showing an association between the allelic variation of this gene and carnosinase activity [41
], lead to the speculation that lower carnosinase activity could be genetic. In our study, patients with MD may have always had low carnosinase activity and thus susceptible to dementia. Down-regulation of carnosinase mRNA is an alternative hypothesis for lower carnosinase levels. Gene regulation has been shown for the AD marker neural thread protein (AD7C-NTP). It is up regulated in the brains and CSF of patients with AD compared to controls [42
In our study, carnosinase activity in MD was related to the duration of dementia. This raises the possibility that carnosinase may have a role in monitoring the progression and therapy of patients with MD. Furthermore, patients treated with dementia medications had somewhat higher carnosinase activities in both AD and MD groups (data not shown). Regular exercise was also associated with increased carnosinase activity and regular physical activity has been associated with a decreased risk of vascular dementia in women [43
In contrast to other biomarkers for dementia, which are measured in CSF, carnosinase is measured in the serum, which is advantageous as venipuncture is far less invasive than lumbar puncture. Carnosinase activity in CSF is about 30-fold higher than the normal serum/CSF protein ratio indicative of a predominantly brain-derived source of carnosinase in CSF [24
]. Although studies have not been done to determine the source of serum carnosinase, carnosinase (CN1) is almost exclusively localized to brain tissue.
It is important to comment on the method chosen to measure carnosinase, as there are several methods reported in the literature, making comparisons between studies difficult. Optimization studies for measuring carnosinase activity were previously performed [33
] showing that maximum activity occurs at a pH of 8.5 with Cd2+
as the metal ion cofactor and confirms those of Lenney et al. [31
]. Although Mn2+
is often used as the cofactor carnosinase activity is about 4-fold lower compared to Cd2+
. Carnosine, although not a specific substrate for carnosinase, is preferred over homocarnosine as it is hydrolyzed more rapidly. These conditions essentially prevent any activity from the related cytosolic nonspecific dipeptidase (previously known as tissue carnosinase) from occurring [44
]. Therefore, this assay method achieves high analytical specificity and sensitivity allowing small changes in activity to be detected. These characteristics are beneficial not only for diagnosis but also for monitoring changes in the disease state and treatment.
Since this study was exploratory in nature, designed to see if serum carnosinase activity was altered in dementia, the sample size was small and limited the use of multiple regression analysis to adjust for potential confounding variables. The potential for age, gender, medications and other comorbidities to affect carnosinase activity cannot be ruled out. However, apart from age, univariate analyses for gender, MMSE score and exercise did not change the results. There is limited data in the literature on the effect of age or gender on carnosinase activity [31
]. In general, carnosinase activity is absent or low in newborns and gradually increases until adulthood. Only two small studies examined gender and found no difference [31
]. This study, the first to report carnosinase activity in elderly adults, did not show a difference in activity compared to younger adult populations [31
]. However, a larger group of elderly persons is required to show whether there may be an age or gender association with carnosinase activity.
In this study, there were a few patients with medical conditions that have been associated with lower values of carnosinase activity (e.g., stroke, Parkinson's disease). No patients in the AD and MD groups had any of these conditions. The number of other medical conditions in any group was too few to adjust for using regression analysis, however, the AD group remained significantly different from the MD group if any one of these exclusions was made.
It is also possible that in the absence of post-mortem confirmation of diagnosis, some patients may have been misclassified. Diagnosis was made by mostly clinical assessment and only nine of the twenty patients with dementia had neuroimaging. Neuroimaging does not necessarily improve diagnostic accuracy nor is it required by published criteria for the diagnosis of dementia or AD (though frequently included in the diagnostic evaluation) [47
]. For example, in a study of 501 patients evaluated for dementia, 375 had a CT scan. In these 375 patients, 28% of the diagnoses during life were incorrect when compared with post mortem pathology. In those who did not undergo CT scanning the diagnosis during life was incorrect only 18% of the time [48
], Furthermore, CT scans may both overestimate the presence of strokes [49
] and fail to identify them [50
Carnosinase activity appears to play a role in dementia. Lower carnosinase activity could be a biochemical response for neuroprotection, a genetic variation in the enzyme, or both. Further research is needed to determine whether carnosinase fulfils the potential for a diagnostic and disease-monitoring measure for dementia.