809
Journal of Alzheimer’s Disease 38 (2014) 809–822 DOI 10.3233/JAD-131247 IOS Press
PY
Meta-Analysis of Serum Non-Ceruloplasmin Copper in Alzheimer’s Disease Rosanna Squittia,b,∗ , Ilaria Simonellic,d , Mariacarla Ventrigliaa , Mariacristina Siottoe , Patrizio Pasqualettic,d , Alan Rembachf , James Doecke g and Ashley I. Bushf a Department
of Neuroscience, AFaR - Fatebenefratelli Hospital “San Giovanni Calibita”, Rome, Italy of Neurodegeneration, IRCSS San Raffaele Pisana, Rome, Italy c Department of Neuroimaging, IRCCS San Raffaele Pisana, Rome, Italy d Medical Statistics & Information Technology, AFaR-Fatebenefratelli Hospital “San Giovanni Calibita”, Rome, Italy e Don Carlo Gnocchi Foundation ONLUS, Italy f The Mental Health Research Institute, The University of Melbourne, Victoria, Australia g The Australian e-Health Research Centre, Herston, Queensland, Australia
OR
Accepted 8 August 2013
CO
b Laboratory
AU
TH
Abstract. The fraction of copper not bound to ceruloplasmin seems altered in Alzheimer’s disease (AD). We have addressed this notion evaluating all the studies carried out from 1996 until March 2013 by means of meta-analysis. We performed our analysis on diverse indices evaluating the relationship between copper and ceruloplasmin in general circulation, namely ‘Non-Cp copper’, ‘% Non-Cp copper’, and ‘Adjusted copper’. For Non-Cp copper and % Non-Cp copper, the correct stoichiometry between copper and ceruloplasmin (6–8 atoms of copper for each ceruloplasmin molecule) in healthy controls has been adopted as criterion for the study to be included in the meta-analysis evaluating data with the canonic Walshe’s formula for Non-Cp copper. Copper to ceruloplasmin ratio (Cu:Cp), which is an internal quality control check for ceruloplasmin calibration, was used as an index of the actual stoichiometry in the specimens. Adjusted (Adj-Cp) copper, even though less reliable, was calculated, allowing the evaluation of all the studies selected. An additional meta-analysis of systemic total copper was re-calculated accounting for all the studies carried out from 1983 to March 2013. Ten studies were analyzed in the meta-analysis for Non-Cp copper and % Non-Cp copper reaching a pooled total of 599 AD subjects and 867 controls. For Adj-Cp copper, 14 studies were analyzed with a pooled total of 879 AD and 1,712 controls. 27 studies were considered for systemic total copper meta-analysis, with a pooled total of 1,393 AD and 2,159 controls. All the copper indices analyzed were significantly higher in AD subjects compared to healthy controls. Keywords: Alzheimer’s disease, ceruloplasmin, copper, metal, plasma, serum
INTRODUCTION The Journal of Alzheimer’s Disease has recently welcomed a debate on the robustness of copper as a biomarker of Alzheimer’s disease (AD) (available at http://www.j-alz.com/node/182). The debate focused on the portion of copper not bound to ceruloplasmin, also known as ‘free’ copper or labile ∗ Correspondence to: Rosanna Squitti, PhD, Department of Neuroscience, AFaR - Ospedale Fatebenefratelli, 00186 Rome, Italy. Tel.: +39 06 6837 385; Fax: +39 06 6837 300; E-mail: rosanna.
[email protected].
copper. Besides the academic or theoretical nature of the debate, it is necessary to reach an interdisciplinary consensus regarding the role of copper not bound to ceruloplasmin as a risk factor for AD. Published studies addressing the hypothesis of copper not bound to ceruloplasmin dysfunction in AD have been based on heterogeneous samples, with diverse ethnic and geographical populations, as well as different methodological approaches, which are sensitive to ceruloplasmin calibration. These limitations influence statistical power which impact on the validity and utility of these measures across different research
ISSN 1387-2877/14/$27.50 © 2014 – IOS Press and the authors. All rights reserved
R. Squitti et al. / Meta-Analysis of Non-Cp Copper in AD
cal values of Cu:Cp, which can be effectively measured in specimens of healthy controls, should range between 6–8, even though this ratio can yield diverse values. On this basis, we decided to apply the criterion of having a correct stoichiometry for copper and ceruloplasmin within the 6–8 range for healthy controls, for a study to be included in the meta-analyses of Non-Cp copper [7] and % Non-Cp copper [8, 9] (data calculated with the Washe’s formula). We used the mean value of the Cu:Cp obtained in healthy controls as an index of the actual stoichiometry for copper and ceruloplasmin in those specimens. In other words, we used the Cu:Cp ratio as an index to verify how close the actual stoichiometry measured in the healthy control specimens was to the theoretical correct one. In fact, not fulfilling the criterion of 6–8 copper/ceruloplasmin stoichiometry results in Non-Cp copper in the negative range (for a stoichiometry below 6) and in an overestimation of Non-Cp copper (for a stoichiometry above 8). Caution should be also taken when applying the Walshe’s formula [7] to copper and ceruloplasmin pairs with a stoichiometry below 6.23, since it also yields results in the negative range. Conversely, when estimating copper non-bound to ceruloplasmin with diverse indices, we do not adopt any internal criterion. In fact, several different mathematical approaches using serum copper and ceruloplasmin levels have been proposed as markers of copper status, for example, the Cu:Cp ratio itself [8]. In addition, the ‘% of Non-Cp copper’ and the ‘adjusted copper’ (Adj-Cp copper) [10] formulae have been proposed. Even though they are similar, these indices reflect the diverse biochemical relationship linking copper and ceruloplasmin. Specifically, the Cu:Cp is a valid cross-check of the reliability of the measures of copper and ceruloplasmin obtained in the specimen as detailed above [8]. The Adj-Cp copper is an index of serum copper status clinically equivalent to Non-Cp copper [10]: it adjusts the value of copper concentration to that of ceruloplasmin. This strategy overcomes the issues associated with estimating how many copper atoms are bound to each ceruloplasmin molecule, and consequently, even though less robust, indicates copper not bound to ceruloplasmin status, overcoming the negative values for Non-Cp copper that can be found, as well as age and gender variations in ceruloplasmin concentration. Despite the proposed utility of Cu:Cp, Non-Cp copper, and % Non-Cp copper, as well as Adj-Cp copper for clinical interpretation of a copper dysfunction, these measures suffer largely from the methodological caveats related to the immunological methods used to quantify ceruloplasmin concentration, since
AU
TH
OR
CO
centers. Meta-analysis can help bridge these caveats by comparing all the covariates. Copper is essential for life, but when inappropriately compartmentalized can catalyze oxidative stress, via Fenton and Haber-Weiss chemistry. Text books [1, 2] report that 85–95% of systemic copper is bound to ceruloplasmin (i.e., it is a ceruloplasmin-bound copper) and that the remaining copper, corresponding to 0–1.6 mol/L is loosely bound to and exchangeable among albumin, transcuprin, various peptides, and amino acids and consequently, represents a pool of bioavailable copper for cells and tissues. These values refer to copper not bound to ceruloplasmin measured after an overnight fast, since copper from foods, after digestion through gastrointestinal secretions in the upper gastrointestinal tract reaches the liver via the portal circulation in the form of low-molecularweight-soluble complexes of copper, mainly bound to amino acids, small peptides, micronutrients, and albumin. Copper not bound to ceruloplasmin has been proposed to be a superior marker of serum copper status [3]. It is effective in distinguishing patients with copper metabolic disorders, such as Wilson’s disease, a rare but treatable autosomal recessive disease caused by mutations in the ATP7B gene. Ceruloplasmin binds 6–8 atoms of copper, six structurally mounted in the holo-form of the protein, and an additional two easily exchangeable [4]: from this derives the correct stoichiometry for copper and ceruloplasmin. This evidence has been confirmed using size exclusion inductively coupled mass spectrometry (SE-ICPMS) and from crystal structure analysis [4–6]. Walshe [7] provides an equation for Non-Cp copper calculation (appendix of [7]) based on the measures of copper and ceruloplasmin in serum. Walshe’s formula assumes that the amount of copper in ceruloplasmin is approximately 0.3% g/g of ceruloplasmin, corresponding exactly to 6.23 atoms of copper bound to ceruloplasmin. Through the text we use the term ‘Non-Cp copper’ to referring to the fraction of copper not bound to ceruloplasmin when it is calculated with the canonical Walshe’s formula [7]. Copper:Ceruloplasmin (Cu:Cp) is an additional copper index which serves as a useful internal quality control verification of ceruloplasmin calibration, and provides information about the actual stoichiometry between copper and ceruloplasmin in the specimens. A Cu:Cp value of 6.6 has been advocated as the theoretical optimal ratio for healthy subjects [8]. From the knowledge that the correct stoichiometry for copper and ceruloplasmin is 6–8 atoms of copper per ceruloplasmin molecule [4–6], the plausible theoreti-
PY
810
R. Squitti et al. / Meta-Analysis of Non-Cp Copper in AD
CO
PY
“serum”, metals” and their combinations into PubMed (ncbi.nlm.nih.gov/pubmed/) and selected studies from 1996 to 2013. We also identified other studies using the “Scopus” and “ISI Web of Knowledge” databases. After reviewing the contents of this first selection, we also reviewed their reference lists to search for additional studies via “Google Scholar”. We only considered papers showing comparative analyses of AD and healthy populations, presenting original results, and published in peer-reviewed journals. In all included studies, the severity of the AD subjects’ cognitive decline had been assessed by Mini-Mental State Examination (MMSE, [13]). The primary aim of our meta-analysis was the comparison of the Non-Cp copper values between AD subjects and controls. Our first selection consisted of 18 studies [14–30]. At this point, we further required that articles explicitly report the means of Non-Cp copper values or of copper and ceruloplasmin, together with standard deviations for both patient and control groups. When not provided by the authors, the variance was calculated with standard procedures. Specifically, the variance was calculated with the formula for the calculation of the variance of the linear combination of two variables, which are not independent of each other. From the first selection, four studies from our laboratory were excluded since their patient samples partially overlapped [26, 28–30]. For the studies analyzed with the Walshe’s formula in the Non-Cp copper and % Non-Cp copper meta-analyses, the additional criterion of the correct stoichiometry for copper and ceruloplasmin, i.e., hav-
TH
OR
antibodies against the holo-form of ceruloplasmin (containing 6–8 atoms of copper) cross-react with fragments of the apo-form of the protein, a copper depleted form of ceruloplasmin, which is rapidly cleaved and removed from general circulation. The cross-reaction determines an overestimation of ceruloplasmin concentration [3, 7, 8]. We have previously published meta-analyses, a statistical method that combines the results of different studies, to address the question if total serum copper is changed between AD subjects compared to healthy controls [11, 12]. Here we present meta-analyses of systemic Non-Cp copper, % of Non-Cp copper, and Adj-Cp copper data, clinically equivalent indices of copper not bound to ceruloplasmin, in AD and healthy controls, considering all the studies, evaluating both copper and ceruloplasmin in general circulation, published from 1996 until March 2013, also accounting for the confounding factors described above. Moreover, we have re-run a meta-analysis of systemic total copper in AD [11, 12] adding any newly published studies, and discussed all these results in the context of recent meta-analytic data published on copper concentrations in the AD brain. A critical evaluation based on metaanalytic data about copper and copper not bound to ceruloplasmin in AD, and their tentative elaboration in a holistic perspective is reported herein.
811
AU
METHODS
To identify eligible studies for meta-analysis, we entered the keywords “Alzheimer’s disease”, “copper”, “ceruloplasmin”, “free copper”, “Non-Cp copper”,
Table 1 Demographic data of the eligible studies for meta-analyses References Molaschi et al. [18] Snaedel et al. [21] Squitti et al. [24] Squitti et al. [22] Sedighi et al. [20] Squitti et al. [26] Agarwal et al. [14] Zappasodi et al. [27] Arnal et al. [15] Brewer et al. [16] Squitti et al. [23] Squitti et al. [25] Lopez et al. [17] Rembach et al. [19]
N◦
Gender (%F)
AD Mean age (y)
31 44 47 28 50 51 50 54 110 28 105 93 36 152
100% 72.7% 74.5% 71.4% 48% 78.4% 38% 81.5% 58.2% 46% 78% 77% 55.5% 59.9%
77.2 ± 2.4 74.3 ± 10.7 75.6 ± 7.7 71.4 ± 8.6 76.4 73 ± 8 59.9 ± 11.6 73.7 ± 8.7 74.7 ± 4.1 76.2 ± 11.6 74 ± 8 75.14 ± 8.7 77.7 ± 5.3 77 ± 7.9
MMSE 15.9 (6–28) 18.6 ± 4.7 15.5 ± 6.2 14.3 ± 4.6 19.2 ± 4.2 14.1 ± 7.6 19.5 ± 3.8 16.6 ± 2.3 24 ± 3.91 19.6 ± 4.6 20.7 ± 4.4 19.3 ± 5.2
N◦
Gender (%F)
Healthy controls Mean age (y)
421 44 44 25 50 53 50 20 79 29 100 48 33 716
100% 72.7% 45.5% 68% 50% 66% 34% 65% 48% 69% 57% 48% 63.6% 57.5%
77.6 ± 2.3 74.3 ± 11 71.1 ± 11 70 ± 9.6 67.8 70 ± 10 55.32 ± 11 71.5 ± 9.2 77.8 ± 3.7 68.6 ± 6.3 69 ± 9.7 70.29 ± 9 74 ± 5.0 69 ± 6.8
p value MMSE
28.1 ± 1.3 28.4 ± 1.2 25.8 ± 1.5 28.5 ± 1.2 28 ± 1.7 28.6 ± 0.9 29.8 ± 0.7 25.5 ± 1.3 28.9 ± 1.3 28.9 ± 1.2
nd nd 0.001
