Cerebrospinal fluid/plasma albumin quotient (A), CSF levels of cyclophilin A (B), and CSF levels of active MMP-9 (C) in a population of cognitively normal individuals on 3 different APOE genotypes: APOE2/E3 (n = 11), APOE3/E3 (n = 28), and APOE3/E4 (n = 10). Pearson correlation coefficients (r) were calculated to study the relationship between CSF cyclophilin A (D) and CSF active MMP-9 (E) levels with albumin quotient in individual subjects. D and E, Points represent individual values from 49 subjects. The black bars represent mean values. aP < .01 and bP < .05.
Halliday MR, Pomara N, Sagare AP, Mack WJ, Frangione B, Zlokovic BV. Relationship Between Cyclophilin A Levels and Matrix Metalloproteinase 9 Activity in Cerebrospinal Fluid of Cognitively Normal Apolipoprotein E4 Carriers and Blood-Brain Barrier Breakdown. JAMA Neurol. 2013;70(9):1198-1200. doi:10.1001/jamaneurol.2013.3841
In humans, apolipoprotein E (apoE) has 3 isoforms: apoE2, apoE3, and apoE4. APOE4 is a major genetic risk factor for Alzheimer disease (AD).1 Apolipoprotein E4 has direct effects on the cerebrovascular system, resulting in microvascular lesions and blood-brain barrier (BBB) damage, as recently reviewed.2 Neurovascular dysfunction is also present in cognitively normal APOE4 carriers and individuals with APOE4-associated disorders including AD.1- 3 Moreover, postmortem brain tissue analysis has indicated that BBB breakdown in patients with AD is more pronounced in APOE4 carriers compared with APOE3 or APOE2.4- 6 Our recent studies in transgenic mice have demonstrated that apoE4 leads to BBB breakdown by activating the proinflammatory cyclophilin A (CypA)–matrix metalloproteinase 9 (MMP-9) pathway in brain pericytes, which in turn results in degradation of the BBB tight junctions and basement membrane proteins.7 It has also been shown that apoE4-mediated BBB breakdown leads to secondary neuronal injury and cognitive decline in transgenic mice.7 Apolipoprotein E2 and apoE3 maintained normal BBB integrity in transgenic mice by suppressing the CypA–MMP-9 pathway.7 Here, we studied the cerebrospinal fluid (CSF)/plasma albumin quotient (QAlb), an established marker of BBB breakdown,8 and CypA and active MMP-9 levels in the CSF of cognitively normal individuals with different APOE genotypes to determine whether apoE4-dependent changes in BBB permeability and CypA–MMP-9 pathway as shown in APOE4, but not APOE3 and APOE2 transgenic mice, also occur in humans.
Participants were volunteers who were recruited through advertisements or from the Memory Education and Research Initiative Program at the Nathan S. Kline Institute for Psychiatric Research.9 Participants gave their informed consent to participate in studies approved by the institutional review board of the Nathan S. Kline Institute for Psychiatric Research and New York University School of Medicine. We studied a total of 49 cognitively normal individuals as indicated by the Clinical Dementia Rating score of 0 and Mini-Mental State Examination score of approximately 30. This study did not exclude participants meeting criteria for major depressive disorder because there were no differences in the studied markers of BBB damage in this group compared with controls. The studied individuals represented 3 different APOE genotypes: APOE2/E3 (n = 11), APOE3/E3 (n = 28) and APOE3/E4 (n = 10). Within each genotype, individuals were stratified into 2 age groups—40 through 65 years old and 66 through 85 years old—to control for age-dependent effects (Table). Cerebrospinal fluid and plasma collection and APOE genotyping were performed as described.9 Enzyme-linked immunosorbent assays were used to determine levels of CypA (catalog no. sE90979Hu; USCN Life Science), active MMP-9 (catalog no. 72017; AnaSpec), and albumin (catalog no. E-80AL; Immunology Consultant Laboratories). Data were analyzed by multifactorial analysis of variance with 2 factors (age and APOE genotype), with Bonferroni post hoc tests to adjust for multiple comparisons, and Pearson correlation analysis using Graphpad Prism version 5.0. Analyses were performed by an investigator blinded to the experimental conditions. A P value of less than .05 was considered statistically significant.
Older cognitively normal individuals carrying 1 APOE4 allele compared with younger cognitively normal APOE4 carriers or age-matched APOE4 noncarriers had increased QAlb by approximately 77% and 67%, respectively (P < .01; Figure, A). No age-dependent increase in QAlb was associated with APOE2 or APOE3 alleles. Compared with cognitively normal younger APOE4 carriers or age-matched APOE4 noncarriers, older cognitively normal APOE4 carriers had increased CSF levels of CypA by approximately 190% and 95%, respectively (P < .01; Figure, B) and active MMP-9 by 167% and 110%, respectively (P < .05; Figure, C). No age-dependent changes in CypA or MMP-9 CSF levels were associated with APOE2 or APOE3 alleles. Importantly, increases in QAlb values correlated positively with both CypA and active MMP-9 CSF levels in all studied individuals (r = 0.37, P < .01; and r = 0.45, P < .01, respectively) (Figure, D and E), indicating the greater the increase in CypA and active MMP-9 levels, the greater the magnitude of BBB breakdown assayed by QAlb.
This study showed that APOE4 carriers may be susceptible to an age-dependent BBB breakdown prior to onset of clinical decline as determined by Clinical Dementia Rating and Mini-Mental State Examination scores. Furthermore, these findings are consistent with experimental studies suggesting that apoE4 leads to BBB damage in transgenic mice via activation of the CypA–MMP-9 pathway.7 These findings warrant future longitudinal studies to investigate QAlb and CSF levels of CypA and active MMP-9 in cognitively normal APOE4 carriers as they progress to mild cognitive impairment and eventually AD. With current diagnostic markers, by the time the earliest detectable clinical signs of disease appear, significant brain injury has likely already occurred. Therefore, studying markers of BBB damage along with commonly used β-amyloid 42 and tau CSF levels may contribute to early detection of vascular dysfunction of those at risk for cognitive decline and AD.
Corresponding Author: Berislav V. Zlokovic, MD, PhD, Zilkha Neurogenetic Institute, Keck School of Medicine, University of Southern California, 1501 San Pablo St, ZNI 101, Los Angeles, CA, 90089-2821 (firstname.lastname@example.org).
Author Contributions: Dr Zlokovic had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Pomara, Frangione, Zlokovic.
Acquisition of data: Halliday, Pomara, Sagare.
Analysis and interpretation of data: Halliday, Mack, Zlokovic.
Drafting of the manuscript: Halliday, Zlokovic.
Critical revision of the manuscript for important intellectual content: Pomara, Sagare, Mack, Frangione, Zlokovic.
Statistical analysis: Halliday, Mack
Obtained funding: Pomara, Zlokovic.
Administrative, technical, or material support: Pomara
Study supervision: Pomara, Zlokovic.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was supported by grants R37NS34467 and R37AG23084 to Dr Zlokovic from the National Institutes of Health (NIH) and by grant R01MH-080405 to Dr Pomara from the NIH.
Additional Contributions: We thank Leslie Saint-Louis, MD, for performing lumbar punctures and Antero S. Sarreal, MD, and Raymundo T. Hernando, MD, for their assistance with patient recruitment.