Copper (Cu) and zinc (Zn) are trace elements essential for life and constitute components of numerous enzymes with high importance for survival and function of eukaryotic cells [1
]. Cu and Zn are necessary for cellular metabolism and the antioxidative defense systems [1
]. The regular fetal development and growth critically depend on both Cu and Zn, especially during the maturation of the nervous system [2
]. Similarly, the maturing immune system relies on these trace elements [2
], especially for antibody production (Cu, Zn), function of neutrophils and monocytes (Cu) [6
], the viability, proliferation, and differentiation of cells of both the innate and adaptive immune system (Zn), as well as for the maintenance of the skin and mucosal barriers (Zn) [3
Both a deficiency and an excess of Cu or Zn can cause harm, so the homeostasis of both elements is strictly regulated [1
]. The main source of both micronutrients is dietary intake [1
]. The absorption from the gastrointestinal tract depends on the nutritional form and the micronutrient status of the subject [4
]. The cellular homeostasis is controlled by import and export proteins, cytosolic metallochaperones, glutathione, and metallothioneins [1
]. The latter also acts as a dynamic Cu and Zn pool [3
]. Around 95% of Cu in blood is bound to liver-derived ceruloplasmin (CP), which is used along with serum Cu concentrations as a biomarker of Cu status [4
]. Serum Cu and CP concentrations constitute acute phase reactants [5
]. Cu is mainly stored in liver, secreted as Cu-CP complex into blood, and an excess can be eliminated by biliary excretion [1
]. In contrast to Cu, the protein-mediated transport, storage, and regulated excretion of Zn are more complex and less well understood [3
Cu and Zn deficiencies constitute prevalent and under-diagnosed health risks [1
]. Neonates and especially preterm infants have a notable risk of Cu and Zn deficiency due to their rapid growth and the concomitant increasing requirement for both micronutrients [4
]. On the one hand, Cu or Zn deficiencies impair the immune defense and confer a high susceptibility to infectious diseases [3
]. On the other hand, acute infections cause an increase in serum Cu in the context of an acute phase response [4
] and a decrease in serum Zn due to a redistribution into liver and other tissues [3
]. These two responses to inflammation may feed a vicious cycle of impaired immune defense and higher infection risk, which is of particular importance especially in very vulnerable patients, such as preterm and term neonates with an immature immune system.
Congenital infections and especially neonatal sepsis are a frequent cause for morbidity and mortality of newborns [13
]. Vertical bacteria transmission from mother to child results in early-onset infections, defined by an onset within 72 h postnatum [16
]. The symptoms of a systemic inflammation by newborns are unspecific [13
], and the diagnosis is challenging. Isolation of bacteria from blood cultures of neonates is difficult due to the relatively high blood volumes and long incubation times required, and by the initially low bacteria counts in the majority of affected neonates [13
]. The cytokine interleukin 6 (IL-6) and the acute-phase reactant C-reactive protein (CRP) are important diagnostic markers for early (IL-6) or later (CRP) phases of inflammation [15
]. Despite their usefullness, the current clinical algorithm does not provide a satisfying diagnostic specificity in neonates [14
], implying that additional biomarkers are needed to facilitate the correct diagnosis and to avoid unnecessary drug administrations. No consented diagnostic and therapeutic guidelines for newborn sepsis have been established. The current therapeutic strategy advocates an early application of an empiric antibiotic treatment if an infection is just suspected [17
]. Some antibiotics are applied in terms of an off-label use due to a lack of intervention studies in newborns, raising concerns about their safety [17
]. Despite the supportive and antibiotic treatments, congenital infections may still cause long-term complications such as brain damage and neurological sequelae [18
From animal and clinical studies with adult patients, it is well established that inflammatory cytokines disturb the trace element homeostasis [9
]. Furthermore, the Cu/Zn ratio is altered in certain diseases [19
]. Therefore, we postulated that infections disrupt homeostasis of the trace elements Cu and Zn in neonates.
2. Materials and Methods
2.1. Study Design
The design of this prospective observational case-control study has been described before in the context of the effects of congenital infections on the selenium (Se) status [24
]. Briefly, this explorative study was conducted on the neonatology wards of the Charité-Universitätsmedizin Berlin from February 2013 until April 2014, after clearance with the local ethics committee (approval no.: EA2/092/12). Informed written consent was obtained for each neonate enrolled in the study from the legal guardian(s) prior to analysis. Relevant clinical data were extracted from the electronic and traditional paper-based medical files. Trace elements were analyzed using residual plasma samples from routine laboratory evaluations ordered by the attending physician. Hereby, one sample was collected from each neonate (control or infection) at birth (day 1). The 2nd sample was obtained 48 h later (day 3) only from infected neonates.
2.2. Study Population
Neonates were screened for fulfilling the inclusion criteria of early-onset infection according to published recommendations and the diagnostic steps taken [25
]. In brief, neonates qualifying for the infection group had to exhibit at least one of the following clinical signs: pneumonia, respiratory distress, tachycardia, bradycardia, fever (>38.5 °C), hypothermia (<36.0 °C), irritability, lethargy, hypotonia, poor feeding, increased frequency of apnea, and/or coagulation disorder in combination with laboratory evidence for an inflammation (IL-6 > 100 ng/L, or CRP > 10 mg/L). In neonates with suspected infection, blood cultures were performed prior to antibiotic treatment. Blood cultures were performed with the BacT/ALERT automated system (Organon Teknika, Eppelheim, Germany) in Pedi-BacT pediatric blood culture bottles capable of detecting anaerobic as well as aerobic bacteria. Neonates with suspected early-onset infection were immediately treated with ampicillin and gentamicin for at least 3 days, and a second blood drawing was conducted 48 h later to determine inflammation markers and gentamicin levels. Two residual plasma samples were, thus, available from each of the neonates with suspected infection (day 1 and day 3), and one from the control neonates at time of birth (day 1). The newborns in the control group showed no laboratory evidence for inflammation (IL-6 < 100 ng/L, and/or CRP < 10 mg/L), and were not receiving antibiotic treatment during the hospital stay. Several infants had to be excluded because of birth before 30 weeks of gestation, birth weight below 1000 g, a diagnosed genetic disease, severe congenital malformation, parenteral supplementation with trace elements, or a missing written consent. Details on the neonates enrolled into this study have been published earlier in relation to analyzing their Se status [24
] and are provided below (Table 1
). Neonates with suspected early-onset infection are summarized as the “infection group” and are denoted as “infected neonates” in this scientific report. However, it needs to be pointed out that a suspected early-onset infection in neonates cannot as safely be diagnosed as in adults for a number of reasons including unspecific symptoms, a higher variability of symptoms, and a very limited amount of blood to be analyzed by laboratory tests and blood culture.
2.3. Trace Elements
Total plasma Cu and Zn concentrations were determined by total reflection X-ray fluorescence (TXRF) as described in [24
]. The method was chosen because of the low sample volume requirements, which is necessary particularly for infants, and prospectively because of the short time needed for analysis, which is a prerequisite for considering the technique in routine clinical decision-making. Briefly, 10 µL of plasma samples were diluted with 10 μL of a gallium standard solution (f.c. 550 μg/L, Sigma-Aldrich, Steinheim, Germany) and mixed thoroughly. Duplicates of 8 μL each were applied to ultra clean quartz glass carriers, dried at 37 °C and measured using a TXRF spectrometer (S2 PICOFOX, Bruker nano GmbH, Berlin, Germany) as described in [29
]. The inter-assay coefficient of variation (CV) was less than 10% for both Cu and Zn.
Western blot analysis was performed for assessment of plasma CP levels. Three Western blots were prepared containing plasma samples from neonates of the control group along with samples from infected neonates. An additional Western blot was prepared with samples from the group of infected neonates that were pre-selected with respect to the CRP levels. Plasma was diluted in ultrapure H2O (Biochrom AG, Berlin, Germany) and 4x sample buffer (200 mM Tris-HCl, pH 7.5, 50% glycerin, 4% SDS, 0.04% bromophenol blue, and 125 mM DTT). Plasma proteins were size-fractionated by sodium dodecyl (lauryl) sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and subsequently transferred onto nitrocellulose membranes by semi-dry blotting (Optitran, Schleicher & Schuell, Dassel, Germany). Antibodies against CP (1:2000 dilution, ab19171, Abcam, Cambridge, UK) were used and bands visualized by chemiluminescence (Western-Bright Substrate Sirius, Biozym Scientific, Oldendorf, Germany) using the Fluor Chem FC2 detection system (Biozym Scientific). Quantification of Western blot signals was achieved by using the Java-based image processing program ImageJ (NIH, Bethesda, MD, USA).
2.5. Interleukin 6 and C-reactive Protein
The IL-6 and CRP values were measured by Labor Berlin, Charité Vivantes GmbH, Germany, by standard laboratory analyses as described earlier [24
]. Briefly, IL-6 was measured by an electro-chemical luminescence immunoassay and CRP was determined by a turbidometric assay (COBAS 8000 or COBAS 6000; Roche Diagnostics, Mannheim, Germany).
2.6. Statistical Analysis
Statistical analysis was performed with the Statistical Package for the Social Sciences (SPSS Statistics 21®, IBM, Chicago, IL, USA) and GraphPad Prism (GraphPad Software Inc., San Diego, CA, USA). Normal distribution of values was assessed by the Shapiro-Wilk test. A two-tailed T-test for unpaired or paired variables and the bivariate Pearson correlation test were used for normally distributed values. For not-normally distributed variables, the Mann-Whitney-U-test, the Wilcoxon-test, and the Spearman´s correlation test were used. The quantified Western blot signals were analyzed using nonparametric tests (Mann-Whitney-U-Test, the Wilcoxon-Test, and the Spearman’s correlation test). Linear regression analysis was conducted to specify associations of variables. Multiple logistic regression was performed to evaluate the results in consideration of relevant confounders. Odds ratios where calculated to examine the quality of the Cu/Zn ratio as a biomarker. The results were considered as statistically significant when the p-value was less than 0.05, and differences are marked as follows: p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***). Parametric data are represented as means ± standard deviation (SD) or medians and interquartile range (IQR); median [IQR].
In this study, we evaluate the infection-related differences in plasma Cu and Zn concentrations in preterm and term neonates. Our results indicate that both trace elements increase in plasma with gestational age in control newborns, and that this correlation is lost in infection. Infected neonates show relatively low plasma Zn concentrations at birth, and develop elevated plasma Cu concentrations during infection. As the plasma concentrations of these two trace elements are regulated in opposite directions by infection, we speculated that the Cu/Zn ratio may provide a more robust marker of early-onset infection than either value alone. Our study indicates that the Cu/Zn ratio correlates to the CRP levels determined at day 3 as the established biomarker of infection. Interestingly, this correlation is already shown at day 1. It remains to be tested if it could be of value for the diagnosis of early-onset infections. The Cu/Zn ratio at day 1 may already reflect the severity and predict the potential course of the infection, which then becomes detected later by the elevated CRP levels determined at day 3. It may, thus, constitute a helpful early prognostic biomarker of early-onset infection in term and preterm neonates.
There is evidence that preterm neonates are at especially high risk for Cu and Zn deficiency [4
]. This notion is supported by previous studies reporting that the maternal Cu levels rise with the length of gestation [31
]. Consistent with former findings in neonates [32
], the Cu concentrations of the infants in our study were lower, and Zn was higher than in adults (adult reference intervals according to [11
]; Cu; 10–22 µM, i.e., 635.5–1398.0 μg/L, and Zn; 12–18 µM, i.e., 784.6–1176.8 µg/L). Our data are in line with previous studies in young children, especially with respect to an increase in plasma Cu levels upon infection [33
]. Notably, also in preterm and term infants, a tight correlation of plasma Cu concentrations and plasma CP is reported [34
]. These data along with our findings support the concept that plasma CP may serve as a surrogate marker of plasma Cu concentrations in children. This relation offers the option for bed-side testing of the Cu status by immunological assay procedures, as recently demonstrated by using quantum dots for fast CP quantification [35
]. Unfortunately, no reliable protein biomarker of plasma Zn status is yet at hand, which would enable a fast multiplex bed-side quantification of the Cu/Zn ratio via these surrogate protein biomarkers that can be detected by point-of-care technologies.
Importantly, the inverse regulation of plasma Cu and Zn is a well-established characteristic of infections, and the Cu/Zn ratio has been proven of diagnostic value in a number of human disorders, including pediatric infectious diseases, such as giardiasis or amebiasis [20
] and tuberculosis [23
]. Furthermore, the diagnostic value of the Cu/Zn ratio as a disease marker was also shown in autism, attention-deficit hyperactivity disorder, hypertension, inflammatory, as well as neoplastic diseases [33
]. The quotient was also described as a potential biomarker of inflammation and nutritional status as well as a mortality predictor in elderly people [40
], and as a variable correlating with inflammation, disrupted immune system, and an increased oxidative stress in peritoneal dialysis [41
]. Our data indicate that the Cu/Zn ratio may also be of diagnostic value in neonates with suspected infection, as it was associated with severity of inflammation at an early time point while being independent of gestational age.
4.1. Early-Onset Congenital Infections as Disruptors of the Trace Element Homeostasis
We found significantly lower plasma Zn concentrations in infected neonates compared with that of the control group, which is congruent with former findings in children, adults, and animals suffering from an acute inflammation and/or critical illness [3
]. However, Zn levels were not associated with the early inflammation marker IL-6 or with the late acute phase reactant CRP. This lack of stringent interrelation is in line with studies in adults, where there are only marginal correlations of serum Zn with markers of inflammation [43
]. This may indicate that low plasma Zn may not reflect the severity of the inflammation, i.e., it may not be directly regulated by the cytokines released in response to infection. However, in other studies, a respective correlation of plasma Zn with inflammation markers in critically ill children and adults has been reported [11
]. The lower Zn levels in our infected neonates may, thus, not necessarily be the consequence of the inflammation, but potentially a risk factor for infection [8
]. However, this hypothesis needs to be tested in other clinical trials, as our analysis is an observational study and not designed to identify causal relationships.
The Cu levels increased with gestational age. In general, the fetal hepatic tissue does not efficiently support an incorporation of Cu into the CP apoenzyme [32
]. Thus, the increasing Cu levels with age may reflect the functional maturation of the liver. Due to ethical reasons, only residual plasma samples were available from infected neonates at day 3 but not from control neonates. Such an analysis would shed light on the relative importance of age and infection for the rising Cu concentrations observed in the study. Moreover, the diagnosis of infection in the newborns was based on clinical symptoms in combination with laboratory evidence for an inflammation, and not on positive blood cultures or additional laboratory analysis, which constitutes a general shortcoming of our study.
4.2. The Cu/Zn Ratio as a Potential Biomarker of Early-Onset Congenital Infections
The concentrations of plasma Cu or Zn considered separately did not qualify as useful biomarkers of early-onset infection. The correlation of Cu on day 1 with CRP on day 3 seems to impart a predictive value to the trace element, and the association on day 3 implicates its potential as a clinical marker of disease course. However, Cu levels were not significantly different between control and infected subjects at birth. The strong positive association of CP and CRP on day 3 suggests that CP is related to the severity of infection. CP is described as a positive acute phase reactant in adults [4
]. However, it decreases with time during increasing severity of the inflammation in adults [11
], suggesting that it does not steadily correlate to infection severity. Zn levels were significantly lower in infected neonates compared with that of the control group at birth. However, Zn neither correlated with IL-6 on day 1, nor with CRP on day 3. Considering these interactions, Zn alone does not seem to quality as an appropriate infection marker in neonates.
Nevertheless, the Cu/Zn ratio appears to provide additional information on the possible infection of newborns. This notion needs to be tested in prospective trials with a sufficient number of neonates, as both the infection and control groups were relatively small in our pilot study.