The FASEB Journal article fj.13-228791. Published online May 31, 2013.

The FASEB Journal • Research Communication

Plasma zinc’s alter ego is a low-molecular-weight humoral factor Ou Ou,* Keith Allen-Redpath,† Dagmar Urgast,‡ Margaret-Jane Gordon,* Gill Campbell,* Jörg Feldmann,‡ Graeme F. Nixon,† Claus-Dieter Mayer,§ In-Sook Kwun,储 and John H. Beattie*,1 *Rowett Institute of Nutrition and Health, †School of Medical Sciences, and ‡Trace Element Speciation Laboroatory, University of Aberdeen, Aberdeen, UK; §Biomathematics and Statistics Scotland, Aberdeen, UK; and 储Department of Food Science and Nutrition, Andong National University, Andong, Kyungbook, South Korea Mild dietary zinc deprivation in humans and rodents has little effect on blood plasma zinc levels, and yet cellular consequences of zinc depletion can be detected in vascular and other tissues. We proposed that a zinc-regulated humoral factor might mediate the effects of zinc deprivation. Using a novel approach, primary rat vascular smooth muscle cells (VSMCs) were treated with plasma from zinc-deficient (2500 genes, compared to incubation of cells with zinc-adequate rat plasma. We demonstrated that this effect was caused by a low-molecular-weight (⬃2-kDa) zinc-regulated humoral factor but that changes in gene expression were mostly reversed by adding zinc back to zinc-deficient plasma. Strongly regulated genes were overrepresented in pathways associated with immune function and development. We conclude that zinc deficiency induces the production of a low-molecular-weight humoral factor whose influence on VSMC gene expression is blocked by plasma zinc. This factor is therefore under dual control by zinc.— Ou, O., Allen-Redpath, K., Urgast, D., Gordon, M.-J., Campbell, G., Feldmann, J., Nixon, G. F., Mayer, C.-D., Kwun, I.-S., and Beattie, J. H. Plasma zinc’s alter ego is

ABSTRACT

Abbreviations: AAS, atomic absorption spectrophotometry, ANG II, angiotension II; Ch25h, cholesterol 25-hydroxylase; CORT, corticosterone; CXCL10, chemokine 10; DMEM, Dulbecco’s modified Eagle medium; EZP, exchangeable zinc pool; FBS, fetal bovine serum; PF, pair-fed; PF2, pair-fed 2 ␮M Zn rat plasma; PF9, pair-fed 9 ␮M Zn rat plasma; PF27, pair-fed 27 ␮M Zn rat plasma; PF27r, pair-fed 27 ␮M Zn repleted rat plasma; VCAM-1, vascular cell adhesion molecule 1; VSMC, vascular smooth muscle cell; ZA, zinc adequate; ZA27, zinc-adequate 27 ␮M Zn rat plasma; ZD, zinc deficient; ZD2: zinc-deficient 2 ␮M Zn rat plasma; ZD9, zinc-deficient 9 ␮M Zn rat plasma; ZD9r, zinc-deficient 9 ␮M Zn repleted rat plasma; ZD27, zinc-deficient 27 ␮M Zn rat plasma 0892-6638/13/0027-0001 © FASEB

a low-molecular-weight humoral factor. FASEB J. 27, 000 – 000 (2013). www.fasebj.org Key Words: atherosclerosis 䡠 microarray 䡠 smooth muscle 䡠 zinc deficiency Cellular zinc deficiency has a deleterious effect on cell function due to the importance of zinc as a structural and functional component of many proteins in a wide range of different metabolic and signaling pathways (1, 2). The regulation of cellular zinc kinetics has a profound influence on signaling by zinc (3–5), and the relationship of systemic zinc with this process requires clarification. To maintain cellular zinc status, zinc homeostasis at a whole organism level is well regulated over a wide range of dietary zinc intakes (6). Even when human dietary zinc intakes are at the lower limit of what can be achieved using natural foods, and dietary phytate:zinc molar ratios are raised, blood plasma zinc levels are not profoundly decreased (7–9). In mice, moderate zinc deficiency may significantly deplete liver zinc levels, but plasma zinc levels are unaffected (10). The effect of even moderate zinc deficiency on growth (11), immune function (12, 13), spermatogenesis (10), and cardiovascular disease (14), for example, is nevertheless measureable. Exposure of cells in culture to medium containing zinc levels equivalent to those in plasma from zinc-deficient (ZD) animals and humans has much less effect on cell function and growth than might be expected, given the effects observed at the wholeorganism level. Such effects can only clearly be observed when available zinc is depleted to lower than physiologically relevant levels in the cell culture medium using artificial means, such as Chelex-100 ion exchange medium (15) or by addition to the culture medium of a metal chelator, such as N,N,N=,N=-tetrakis(2pyridylmethyl)ethylenediamine (TPEN), which se1 Correspondence: Rowett Institute of Nutrition and Health, University of Aberdeen, Greenburn Rd., Bucksburn, Aberdeen AB21 9SB, Scotland, UK. E-mail: [email protected] doi: 10.1096/fj.13-228791

1

questers zinc and renders it unavailable for use by the cells (16). The circumstantial evidence suggests that under physiological conditions, reduced supply of zinc from blood plasma in zinc deficiency is not the principal factor influencing cellular function. Other hormones or humoral factors whose expression or secretion is sensitive to zinc status may indirectly play a role in influencing cell function. While the zinc-deficiencystimulated production of systemic hormones, such as cytokines (17) and glucocorticoids (18) have been reported, the idea that the primary effects of zinc deficiency are mediated indirectly and not necessarily as a direct consequence of changes in plasma zinc or the whole-body exchangeable zinc pool size lacks hard evidence. The absence of an explanation for the deleterious consequences of low dietary zinc consumption in terms of changes in plasma zinc remains a conundrum familiar to researchers attempting to show cellular responses to changes in physiological zinc levels in cultured cells. Since zinc has a role in protecting against oxidative (19) and inflammatory (20) stress, we have previously proposed that atherosclerosis, which is promoted by both forms of stress, may be enhanced by zinc deficiency (21). We subsequently noted that dietary zinc deficiency modulates the levels of some aortic proteins related to the phenotype of vascular smooth muscle cells (VSMCs) in rats (22) and that a marginally ZD status increases the development of aortic atherosclerotic plaque in a mouse model of atherosclerosis (14). In the rat studies (22), plasma zinc was unaffected by mild zinc deficiency, and yet changes in the levels of many aortic proteins were detected using proteomic methodology. We hypothesized that a zinc-sensitive humoral factor may be mediating the effects of zinc deficiency, and so we designed a series of cell culture studies, using primary rat VSMCs, to test the presence of such a factor. Our novel approach was to use blood plasma from rats made ZD, along with their zincadequate (ZA) controls, to treat the cultured cells. By adding back zinc to the plasma, we were able to distinguish the effect of zinc independently of any other plasma factors. We used a multistep experimental approach. First, we used microarrays as a sensing technique to detect cellular changes in response to plasma treatments and to identify markers of bioactivity specific to plasma from ZD rats. Second, we manipulated zinc in the plasma used to treat the cells and monitored these markers of bioactivity in order to establish whether zinc could reverse the effects of the plasma treatment. Third, we designed critical studies to distinguish between the effects of zinc (direct effect) and the effects of a zinc-regulated humoral factor (indirect effect). The methodological approach for all studies is summarized in Fig. 1. MATERIALS AND METHODS Quality control All studies were carried out in Lloyds Register Quality Assurance-accredited laboratory facilities. All animal procedures 2

Vol. 27

September 2013

Figure 1. Summary of studies to distinguish between direct and indirect effects of zinc deficiency on rat VSMC gene expression and to investigate the presence and characteristics of a zinc-regulated humoral factor (HF) in the blood plasma of zinc-deficient (ZD) rats. Primary VSMCs were treated with blood plasma from ZD, pair-fed (PF), or zinc-adequate (ZA) rats, and plasma zinc levels were manipulated by adding ZnCl2 or removing zinc using Chelex-100 resin. In this way, the effect of zinc and the HF could be studied independently using transcriptomics and qPCR of gene markers as detectors of HF and zinc bioactivity. Pathway and molecular weight analyses were used to further characterize the HF. were in compliance with UK Home Office animal welfare regulations and approved by the Rowett Institute Research Ethics Committee. Animals were obtained from the InstituteSpecific Pathogen-Free colony unit and maintained in a barrier, access-controlled animal unit under dedicated technical and veterinary supervision. Potential sources of zinc contamination, both in the animal study and laboratory work, were eliminated. Certified standard reference materials, including rye flour and hay (International Atomic Energy Agency, Vienna, Austria) and serum (Seronorm, Billingstad, Norway), were used as standard reference materials to verify sample metal analysis data, and analytical results were within the acceptable reference ranges. Chemicals and reagents were of suitable grade for their intended purpose, and water for analytical applications was of 18 M⍀ quality (Milli-Q Integral 3; Merk-Millipore, Watford, UK). Rat model Thirty male Rowett Hooded Lister rats were reared on a standard rodent chow diet until 8 wk of age and were maintained at 25°C with a 12-h light-dark cycle. Grouphoused rats were then acclimated for 1 wk to a ZA semisynthetic diet (35 mg Zn/kg diet) and a reversed light cycle to facilitate pair feeding in the study. At the end of acclimation, rats were randomly divided into 3 groups (n⫽10), in which mean weights were not significantly different (P⬎0.05). The

The FASEB Journal 䡠 www.fasebj.org

OU ET AL.

animals were individually housed and fed either a ZD diet (⬍1 mg Zn/kg diet) or ZA diet for 2 wk. Zinc deficiency causes reduced food intake due to decreased appetite and/or increased satiety, and so a pair-fed (PF) group was included to control for food intake in the ZD group. The food intake of each ZD rat was measured daily, and the corresponding paired rat in the PF group was given the same amount of food the following day. PF rats were fed in 2 meals, one at the beginning and one near the end of the dark phase of their 12-h light-dark cycle to ensure that they did not go more than 12 h without food and to better control for the feeding pattern of ZD rats, which is periodic (23). With the exception of PF animals, rats were fed ad libitum. Body weight of the rats was measured 3⫻/wk throughout the study. The mean ⫾ se values for body weight and total food intake of the rats at termination are shown in Table 1. At study termination, each animal was deprived of food for 4 h and then anesthetized by controlled inhalation of isoflurane vapor in oxygen gas, using a small-animal anesthesia system and according to UK Home Office regulations. This procedure was carried out in a quiet environment with minimal stress to the animal, and anesthesia was effective within 10 –15 s. Within 2 min of anesthesia, blood was drawn from the abdominal vena cava using heparinized syringes. The blood was centrifuged for 15 min at 2000 g, and plasma was then divided into aliquots and snap-frozen in liquid nitrogen. The entire rat study was repeated 4 times to provide plasma for subsequent cell culture studies described below. Very similar results were obtained using plasma from all 4 studies. Diet and plasma analysis The zinc-sufficient semisynthetic diet was similar to the AIN76A diet, with egg white substituted for casein due to the higher zinc content of the latter protein, and extra biotin added to counteract the presence of avidin in egg albumin preparations. The zinc concentrations in each diet were measured by atomic absorption spectrophotometry (AAS; Unicam Solaar 969; Thermo Fisher Scientific, Hemel Hempstead, UK) following wet ashing using an automated microwave digestion system (Discover SP-D; CEM Microwave Technology Ltd., Buckingham, UK). The experimental diets were shown by analysis to contain 0.56 and 33.59 mg Zn/kg for the ZD and ZA/PF groups, respectively. For plasma zinc analysis, samples were diluted 1:10 with 0.1 M hydrochloric acid, centrifuged at 2500 g, and then analyzed for Zn using AAS. Equal volumes of plasma from the 10 individual rats within a group were pooled for cell treatments in order to average out any variation in individual animal responses to the diets. The pooled plasma samples were reanalyzed for zinc, and the detected levels for ZD, PF, and ZA plasma pools were 9.2, 26.0, and 28.1 ␮M, respectively. Since the mean zinc values for PF and ZA plasmas were not significantly different from each other (P⫽0.195), an intermediate value of 27 ␮M was used for the purpose of clarity in

describing control plasma zinc levels. The plasma samples are therefore referred to as ZD9, PF27, and ZA27. Plasma total corticosterone levels were measured using a Luminex technology system (Luminex B.V., Oosterhout, The Netherlands) and a Millipore Rat Stress Panel (RSH69K; Millipore Corp., St. Charles, MO, USA). Samples were analyzed in duplicate, according to the manufacturer’s instructions. Cells and culture conditions Rat primary VSMCs were prepared as described previously (24), using aorta from ZA adult male Rowett Hooded-Lister rats as the cell source. Cell purity was positively verified by Western blot analysis for ␣-actin, and the absence of endothelial cell contamination was verified by Western blot analysis for von Willebrand factor (data not shown). The cells were cultured in 25-cm2 culture flasks with 20% fetal bovine serum (FBS; Sigma-Aldrich, Dorset, UK) in Dulbecco’s modified Eagle medium (DMEM; Sigma-Aldrich, Gillingham, UK) and incubated at 37°C and 5% CO2. Passage 4 –10 was used for experimental studies, and different clones were prepared from 3 different animals to ensure that results were reproducible across clones. For studies in which zinc was depleted from plasma samples, Chelex-100 ion exchange resin (Bio-Rad Laboratories, Hemel Hempstead, UK) was used to chelate zinc, as described previously (25). Zn, Cu, Ca, and Mg concentrations were analyzed by AAS before and after treatment of plasma with Chelex-100 resin, and Mg, Cu, and Ca (as the chloride salts) were added back to the plasma up to levels detected before Chelex treatment. Zinc was added back as indicated in each study. All studies were performed on VSMCs that had been grown to 95% confluence using DMEM containing 10% FBS (SigmaAldrich). To acclimate cells to a higher level of serum, they were then incubated with 50% FBS for 24 h before experimental treatments. Since FBS contains most of the medium zinc and we did not want to expose the cells to elevated zinc when using the higher serum concentration, the FBS was depleted of zinc using Chelex-100 resin to a final level of 9 ␮M, which was the same concentration as in the ZD9 plasma. Standard culture conditions (37°C and 5% CO2) were used throughout the studies. VSMC transcriptomics The initial primary study objective was to demonstrate that differences in plasma from ZD rats and their controls were sufficient to cause significant changes in VSMC gene expression. Secondary objectives were to evaluate the ontology of altered gene expression and to identify biomarkers associated with zinc deficiency in this cell type. VSMCs were grown in Nunclon Vita MultiDish 6-well culture plates (Fisher Scientific, Loughborough, UK) to 95%

TABLE 1. Body weight, total food intake, IFN␥, and corticosterone levels in plasma from zinc-adequate, pair-fed, and zinc-deficient rats Dietary group and zinc level (mg Zn/kg)

ZA, 33.6 PF, 33.6 ZD, 0.56

Body weight (g)

Food intake (g)

Plasma IFN␥ (pg/ml)

Plasma Cort (ng/ml)

363.1 ⫾ 3.3a 326.0 ⫾ 4.2b 309.1 ⫾ 3.6b

307.0 ⫾ 4.0a 235.0 ⫾ 5.0b 237.0 ⫾ 5.4b

52.2 ⫾ 23.9 36.0 ⫾ 12.5 23.9 ⫾ 6.5

126.5 ⫾ 7.9a 178.0 ⫾ 17.9b 193.2 ⫾ 23.0b

IFN␥, interferon-␥; CORT, corticosterone; ZA, zinc adequate; PF, pair fed; ZD, zinc deficient. Superscripted letters that are different within one column indicate statistical significance (P⬍0.05).

ZINC-REGULATED HUMORAL FACTOR

3

confluence and were then exposed for 24 h to DMEM containing 50% ZD9, PF27, or ZA27 plasma. To study the effect of varying plasma zinc alone, a fourth group was included in which zinc (as the chloride salt) was added to the ZD9 plasma up to the zinc level found in the PF27 plasma. The analyzed plasma zinc level was 26.9 ␮M after zinc addition and is referred to as ZD27. Total cellular RNA from VSMCs was extracted using TRIzol Reagent (Invitrogen, Paisley, UK), according to the manufacturer’s instructions, and further purified using RNeasy mini spin columns (Qiagen, West Sussex, UK). RNA integrity was evaluated using a Bioanalyser (Agilent Technologies, Santa Clara, CA, USA) and 1 ␮g of purified RNA (RIN 8.1 to 9.4 and 260/280 above 1.9) was used for each group. RNA was amplified in 2 steps to cRNA, incorporating Cy-3 and Cy-5 dyes and was purified with RNeasy mini spin columns. Gene expression microarrays were manufactured and supplied by Agilent Technologies, and contained 60-mer probes for 30,367 unique Entrez gene RNAs. Labeled cRNA was hybridized to each array rotating at 65°C for 17 h. After washing, the arrays were scanned using a SureScan high-resolution scanner (Agilent Technologies). Combinations of sample comparisons (including dye-swap) were such that 4 replicate array data were obtained for both ZA27 and PF27 plasma-treated cells, and 6 replicate array data were obtained for both ZD9 and ZD27 plasma-treated cells. qPCR of candidate biomarker genes To validate the microarray data for particular genes and further investigate the influence of zinc deficiency in VSMCs, we utilized qPCR with the TaqMan system and reagents, combined with appropriate gene expression kits (Applied Biosystems, Paisley, UK). RNA samples were reverse-transcribed using a TaqMan reverse transcription kit (N8080234: Applied Biosystems) and levels of target gene mRNA (cxcl10, kit Rn01413889_g1; vcam-1, kit Rn00563627_m1; ch25h, kit Rn01451129_s1) were measured using an ABI 7500 Fast PCR instrument (Applied Biosystems). A reference of 18S rRNA was used to normalize the data (kit 4319413E). Three studies were designed to confirm the microarray data and to distinguish between direct and indirect effects of zinc. Study 1: validation of VSMC transcriptomics To validate the results found in the transcriptomics study, mRNA for cxcl10, vcam-1 and ch25h in 3 different VSMC clones exposed to 50% ZD9, ZD27, PF27, and ZA27 plasmas in DMEM for 24 h were quantified by qPCR analysis. Study 2: does zinc deficient rat plasma perturb VSMC gene expression directly through reduced zinc levels or indirectly via a secondary factor? Regulation of gene expression by addition of zinc to, e.g., ZD9 plasma, as shown in study 1, does not in itself prove that zinc in the medium has a direct effect on the VSMCs, since it may be modulating the effect of secondary factor in the plasma. Depleting zinc from the PF27 plasma down to the level found in the ZD9 plasma would provide more conclusive evidence whether the effect of zinc was direct or indirect. This is referred to as the PF9 plasma. VSMCs were, therefore, exposed to 50% ZD9, ZD27, PF27, and PF9 plasma in DMEM for 24 h, and mRNA levels for cxcl10, vcam-1, and ch25h were quantified by qPCR analysis. 4

Vol. 27

September 2013

Study 3: does plasma zinc modulate changes in VSMC gene expression caused by a zinc-regulated humoral factor? Study 2 indicated the presence of a bioactive zinc-regulated humoral factor in ZD rat plasma, and in order to investigate the efficacy of zinc in modulating its bioactivity, we studied the effect of removing zinc from ZD9 plasma and then adding it back. As a control, the PF27 plasma was treated in the same way. ZD9 and PF27 plasma samples depleted of zinc using Chelex-100 resin were therefore obtained and are referred to as ZD2 and PF2, respectively, since the residual zinc level was 2 ␮M. Since Chelex-100 treatment could be regarded as an independent variable, we controlled for it by adding back zinc to aliquots of ZD2 and PF2 plasma up to their respective natural levels of zinc (9 and 26 ␮M zinc, respectively). These positive controls repleted with zinc are referred to as ZD9 and PF27 repleted (ZD9r and PF27r). VSMCs were, therefore, exposed to 50% ZD9, ZD2, ZD9r, PF27, PF2, and PF27r plasma in DMEM for 24 h, and the cxcl10 mRNA level was quantified by qPCR analysis. Molecular weight determination of the zinc-regulated humoral factor To establish the approximate molecular weight of the humoral factor, plasma samples from ZD and PF rats were separated on a calibrated HiPrep 16/60 Sephacryl S-200 HR chromatography column (GE Healthcare Life Sciences, Chalfont St. Giles, UK) equilibrated with PBS, and 1 ml fractions were collected. Absorbance at 280 nm and metal levels (including zinc) were measured in each fraction by UV spectrophotometry (␮Quant; BioTek, Potton, UK) and ICP-MS (7700x; Agilent Technologies), respectively. Starting with fractions 1–5, each 5 successive fractions from the column were pooled, sterilized using a 0.2-␮m membrane, and added to confluent VSMC cultures to give 50% pooled fraction in DMEM with 10% Chelex-treated FBS (zinc added back up to the ZD9 level). After 24 h of incubation, the cxcl10 mRNA levels were quantified by qPCR analysis in order to detect the presence of the humoral factor. Bioinformatics and statistics The microarray data were analyzed within the statistical programming language R (version 2.15.2) using the Bioconductor library limma (26). Loess normalization was applied to the raw data to remove intensity-dependent dye effects. A linear model with zinc group as factor (with the 4 levels ZD9, PF27, ZA27, and ZD27) was fitted, and the limma-specific moderated F test (27) was applied to obtain overall P values of difference, as well as P values for the pairwise comparisons of interest (ZD9 vs. PF27, ZD9 vs. ZD27). As replication within the microarray experiment was of a technical nature (each pool of RNA was used 4 – 6 times), a multiple testing correction would have not been very meaningful, and instead a P value threshold of 0.001 was combined with a fold-change threshold of 2 to select differentially expressed genes. The lists of differentially expressed genes obtained from this analysis were then processed by Metacore’s Pathway Analysis (Thomson Reuters, New York, NY, USA) to detect the top 48 overrepresented pathways of interest. In addition, reported protein interactions for highly expressed genes were analyzed using String 9.05 (Search Tool for the Retrieval of Interacting Genes/Proteins), which is an online resource of the Novo Nordisk Foundation Center for Protein Research (University of Copenhagen, Copenhagen, Denmark). Statistical comparisons of qPCR data were achieved using a 1-way ANOVA with Fisher’s multiple comparisons (Minitab; http://www.minitab.com).

The FASEB Journal 䡠 www.fasebj.org

OU ET AL.

RESULTS Zinc deficiency and gene expression Compared with cells treated with PF27 plasma, VSMCs treated with ZD9 plasma showed significant changes in the mRNA levels of ⬎2500 genes, with selection criteria of ⱖ2-fold change and P ⬍ 0.001 (Fig. 2A). Adding zinc back to the ZD9 plasma (ZD27) mostly prevented the gene expression changes observed with the ZD9 plasma treatment (Fig. 2A). Inspection of the gene identities revealed that ⬎75% of the same genes were affected by the ZD9 plasma treatment as were affected by adding zinc back (1901⫹222⫽2123 genes in Fig. 2B). A plot of the fold-change for these 2123 genes in the ZD9 vs. PF27 dataset vs. the fold-change of the same genes in the ZD9 vs. ZD27 dataset yielded a linear negative correlation with a slope of ⫺0.99 and a correlation coefficient of 0.996. This indicates that in almost all cases, gene expression affected by ZD9 plasma was completely reversed by addition of zinc. Expression of the same 1901 genes was changed when comparing ZD9 with ZA27, ZD27, and PF27 treatments, and only 117 genes were changed when effectively comparing PF27 with ZA27 at the selected significance and threshold levels (P⬍0.001, ⱖ2-fold difference). This is strong evidence that the food deprivation variable in the ZD and PF rats had only a minor effect on VSMC gene expression compared to the zinc deficiency variable. Pathway analysis revealed that the 2123 genes showing significance (P⬍0.001) in both datasets and the largest fold-changes (2 to ⬎600-fold), were predominantly related to immune function and development (Table 2). The same was true for the 567 genes apparently only affected by zinc added back, but the 433 genes apparently affected only by the ZD plasma were related to the categories cell cycle, cell adhesion, and cytoskeleton remodeling. Some of the more highly regulated of the 2123 genes were interferon related,

Figure 2. A) Number of genes significantly affected (ⱖ2-fold change, P⬍0.001) by the different plasma treatments compared to other treatments. Origin of the plasma used for treatment is indicated (ZD, ZA, or PF), followed by numbers representing the approximate plasma zinc level of 9 or 27 ␮M. B) Venn diagram showing the number of the same and different genes significantly affected (ⱖ2-fold change, P⬍0.001) in the comparisons of ZD9 vs. PF27 and ZD9 vs. ZD27. ZINC-REGULATED HUMORAL FACTOR

such as CXC motif chemokine 10 (Fig. 3). Other highly regulated genes included cholesterol 25-hydroxylase (ch25h: ⫹31-fold) and vascular cell adhesion molecule 1 (vcam-1: ⫺14-fold), and all 3 of these genes were selected for validation of the microarray mRNA data using qPCR. Although fold-differences varied in magnitude between the qPCR analysis and the microarray, the direction of change was the same, and fold-differences were large (Fig. 4). As shown in the transcriptomics study (Fig. 3), qPCR analysis of mRNA after treatment of cells with ZD9 plasma revealed elevated expression of cxcl10 and ch25h and decreased expression of vcam1 compared to PF27 and ZA27 plasma-treated cells (Fig. 4). These effects were completely reversed by addition of zinc to the ZD9 plasma (ZD27). Analysis of the data using principal component analysis revealed a very wide separation of the ZD9 plasma treatment group from the rest in the first principal component (PC1) but also indicated a separation of ZD27 from the control groups (PF27 and ZA27) in principal component 3 (PC3; Fig. 5). Given that zinc-specific effects were revealed in PC1, we hypothesized that the differences observed in PC3 were not zinc-specific, and they may be caused by a zinc-regulated bioactive factor in plasma. Evidence for a zinc-regulated bioactive factor We tested the zinc specificity of the gene expression response to ZD9 plasma by depleting zinc from PF27 plasma down to the same level measured in ZD9 plasma and comparing the responses to both plasmas. We found that depleting zinc from PF plasma (PF9) had no substantial effect on VSMC gene expression (Fig. 6). Removing zinc almost completely from ZD plasma increased the bioactivity of this factor, as indicated by changes in cxcl10 expression, and adding zinc back decreased its bioactivity (Fig. 7). The same trend was found for PF plasma when zinc was removed and then added back, but the magnitude of response was much reduced compared to the ZD plasma treatment (Fig. 7). To investigate the molecular weight of the bioactive factor, we separated ZD and PF plasma on a Sephacryl S-200 HR column and incubated VSMCs with eluent from the column. Absorbance at 280 nm (protein) and calcium levels in column fractions showed little difference between ZD and PF plasma separations, but a substantial decrease in zinc associated with fractions corresponding to albumin (fractions 15–22) was observed in ZD plasma, as compared to PF plasma (Fig. 8A). Bioactivity, as determined by cxcl10 expression was detected in the low molecular weight region (Fig. 8B) corresponding to an Mr of ⬃2000 according to the elution of standard calibration proteins and peptides (data not shown). The low molecular weight of the factor was confirmed by separating low and high Mr components of whole plasma using a 10k MWCO 5

TABLE 2. Top 48 pathways showing overrepresentation of genes significantly (⬎2-fold, P ⬍ 0.001) regulated by ZD9 (compared to PF27) and reversed by addition of zinc to ZD9 plasma (ZD27) Pathway

Immune response IFN-␣/␤ signaling pathway Alternative complement pathway IL-12 signaling pathway IL-1 signaling pathway MIF-mediated glucocorticoid regulation Signaling pathway mediated by IL-6 and IL-1 IL-22 signaling pathway HMGB1/RAGE signaling pathway IL-23 signaling pathway Histamine H1 receptor signaling in immune response HSP60 and HSP70/TLR signaling pathway TLR signaling pathways IL-17 signaling pathways Oncostatin M signaling via MAPK Bacterial infections in normal airways CD40 signaling MIF in innate immunity response Th1 and Th2 cell differentiation IL-12-induced IFN-␥ production Development Angiotensin signaling via STATs G-CSF-induced myeloid differentiation Growth hormone signaling via STATs and PLC/IP3 ERBB-family signaling Angiotensin activation of ERK Angiotensin signaling via PYK2 PEDF signaling Angiotensin signaling via ␤-arrestin EGFR signaling pathway ␤-adrenergic receptors transactivation of EGFR Angiopoietin-Tie2 signaling NOTCH1-mediated pathway for NF-␬B activity modulation Other Cell adhesion: gap junctions Cytoskeleton remodeling: neurofilaments Mucin expression in CF via TLRs, EGFR signaling pathways Cell adhesion: cell-matrix glycoconjugates Mucin expression in CF via IL-6, IL-17 signaling pathways Transcription: role of VDR in regulation of genes involved in osteoporosis Protein folding and maturation: angiotensin system maturation Reproduction:GnRH signaling Bacterial infections in CF airways Chemotaxis: leukocyte chemotaxis Cytokine production by Th17 cells in CF Cell cycle: role of Nek in cell cycle regulation Apoptosis and survival: lymphotoxin-␤ receptor signaling Cytoskeleton remodeling: keratin filaments Proteolysis: Putative SUMO-1 pathway Cardiac hypertrophy: NF-AT signaling in cardiac hypertrophy

P

% Genes

Total genes

1.699E-08 8.628E-09 1.702E-04 4.081E-07 9.642E-04 1.631E-04 4.161E-04 1.203E-04 1.069E-02 1.084E-03 6.422E-04 6.422E-04 3.809E-04 1.141E-02 5.665E-03 2.789E-03 6.764E-02 6.764E-02 1.354E-01

45.8 35.9 30.4 29.5 27.3 26.7 23.5 20.8 20.0 18.8 18.5 18.5 18.3 17.1 16.0 15.4 12.5 12.5 11.1

24 39 23 44 22 30 34 53 25 48 54 54 60 35 50 65 40 40 36

3.992E-05 1.631E-04 8.662E-05 3.599E-05 3.341E-04 4.662E-04 2.827E-04 1.069E-02 1.426E-04 3.652E-03 1.141E-02 3.734E-02

28.1 26.7 25.7 25.6 24.2 20.9 20.4 20.0 19.0 18.9 17.1 14.7

32 30 35 39 33 43 49 25 63 37 35 34

2.249E-05 3.031E-04 1.251E-05 1.724E-04 4.161E-04 1.030E-04 1.084E-03 1.345E-04 1.378E-02 7.902E-03 6.186E-02 9.776E-02 8.007E-02 1.354E-01 2.142E-01 1.485E-01

30.0 28.0 24.0 23.7 23.5 19.7 18.8 18.1 13.8 13.3 12.8 12.5 11.9 11.1 10.3 9.2

30 25 50 38 34 61 48 72 58 75 39 32 42 36 29 65

Percentage of genes data are the proportion of total genes in a pathway that were significantly regulated. On the basis that 2123 out of 30,367 (7%) of unique genes on the array were affected, the significance of a pathway’s response to the treatment is indicated by the magnitude of overrepresentation above 7%, and the statistical significance (P value).

molecular filter. The low Mr filtrate contained 95% of the bioactivity in the ZD plasma. DISCUSSION In this series of studies, we initially showed that blood plasma from ZD rats has a marked effect on gene

6

Vol. 27

September 2013

expression in VSMCs compared to plasma from PF rats. Interpretation of the microarray study and confirming qPCR analysis offered the simple explanation that changes in gene expression could be due to differences in plasma zinc, since the ZD rat plasma zinc level (ZD9) was about one-third of the level in the PF control rat plasma (PF27). However, principal component analysis

The FASEB Journal 䡠 www.fasebj.org

OU ET AL.

Figure 3. Genes showing changes in expression ⬎10-fold when treated with ZD9 plasma, compared to PF27 plasma (solid bars), and ZD27 plasma compared to ZD9 plasma (open bars). Gene expression with a positive fold-change was increased and with a negative fold-change was decreased. Gene IDs and NCBI RefSeq database IDs are indicated and a String network of tentative and confirmed protein associations is shown inset (wider connecting lines indicates higher confidence). Gene names are as follows: Rsad2, viperin; Cxcl10, chemokine (C-X-C motif) ligand 10; Cxcl11, chemokine (C-X-C motif) ligand 11; Ifit3, interferon-induced protein with tetratricopeptide repeats; Oas1h, 2=-5=-oligoadenylate synthetase 1H; Atf3, activating transcription factor 3; Rtp4, receptor (chemosensory) transporter protein 4; Oas1e, 2=-5=-oligoadenylate synthetase 1E; Isg15, ISG15 ubiquitin-like modifier; Irf1, interferon regulatory factor 1; Igtp, interferon gamma induced GTPase; Gbp5, guanylate binding protein 5; Gadd45g, growth arrest and DNA-damageinducible, gamma; Klre1, killer cell lectin-like receptor, family E, member 1; Agt, angiotensinogen; Cd6, T cell differentiation antigen CD6; Ptpn6, protein tyrosine phosphatase, nonreceptor type 6; Gbp2, guanylate binding protein 2, interferon-inducible; Oas1d, 2=-5=-oligoadenylate synthetase 1D; Usp18, Ubiquitin specific peptidase 18; Vcam1, Vascular cell adhesion molecule 1; Ptx3, pentraxin 3, long; Il34, interleukin 34; Vsig8, V-set and immunoglobulin domain containing 8; Phf11, PHD finger protein 11; Cxcl2, chemokine (C-X-C motif) ligand 2; Phf11, PHD finger protein 11; Gzmb, granzyme B; Ccl2, chemokine (C-C motif) ligand 2; Cmpk2, cytidine monophosphate (UMP-CMP) kinase 2, mitochondrial; Cfb, complement factor B; Defb27, defensin beta 27; Arg1, arginase, liver; Klf14, Krüppel-like factor 14; Lyz2, lysozyme 2; Kcnb1, potassium voltage-gated channel, Shab-related subfamily, member 1; Ch25h, cholesterol 25-hydroxylase; Mmd2, monocyte to macrophage differentiation-associated 2; Tsga13, testis-specific, 13; Rasd2, RASD family, member 2; Mogat2, monoacylglycerol O-acyltransferase 2; Fam154a, family with sequence similarity 154, member A; Grm4, glutamate receptor, metabotropic 4; Atp6v0a4, ATPase, H⫹ transporting, lysosomal V0 subunit A4; Scn11a, sodium channel, voltage-gated, type XI, ␣ subunit; Uts2, urotensin 2; Tmprss9, transmembrane protease, serine 9; A4galt, ␣ 1,4-galactosyltransferase; Rec8, REC8 homologue (yeast) of rad21; Slc5a5, solute carrier family 5 (sodium iodide symporter), member 5; Dppa3, developmental pluripotency-associated 3; Il24, interleukin 24; Cacna1s, calcium channel, voltage-dependent, L type, alpha 1S subunit; Spocd1, SPOC domain containing 1; Fer1l4, Fer-1-like 4 (C. elegans); Has2, hyaluronan synthase 2; Adh1, alcohol dehydrogenase 1; Gja5, Gap junction protein, ␣5; Sqstm1, sequestosome 1; Ech1, enoyl CoA hydratase 1, peroxisomal. ZINC-REGULATED HUMORAL FACTOR

7

Figure 4. Relative levels of chemokine-10 (cxcl10), vascular cell adhesion molecule 1 (vcam-1), and cholesterol 25-hydroxylase (ch25h) mRNA in vascular smooth muscle cells exposed to blood plasma from rats that received zinc adequate and zinc deficient diets. See Fig. 1 for treatment identifier explanation. Data are from qPCR analysis and are the means ⫾ se of 3 different studies using VSMC clones from 3 different rats. Letters that differ in each panel indicate statistical significance (P⬍0.05).

Figure 6. Relative levels of cxcl10, vcam-1, and ch25h mRNA in vascular smooth muscle cells exposed to plasma from PF and ZD rats (see Fig. 1 for treatment identifier explanation). Data are from qPCR analysis and are the mean (columns) ⫾ se (error bars) of replicate treatments. Letters that are different in each panel indicate statistical significance (P⬍0.05).

of the microarray data suggested that there was more than one variable affecting the data. Critically, depletion of the PF rat plasma zinc level down to the concentration found in ZD rat plasma, which should have induced a gene expression response in VSMCs if this effect were plasma zinc related, had no influence on marker gene levels (Fig. 6). This demonstrated that gene expression was being regulated not just by zinc, but by some other factor in the ZD rat plasma. However, since adding zinc to the ZD rat plasma (ZD27) abolished the effect of ZD9 on gene expression in VSMCs (Figs. 4 and 6), we proposed that the bioactivity (effect on VSMC gene expression) of the zinc-regulated

humoral factor was also modulated by zinc. To prove this, we almost completely removed zinc from the ZD rat plasma and the VSMC marker gene response was ⬎6-fold enhanced (Fig. 7). Adding zinc back demonstrated the specificity of the response to zinc, and gene expression returned to levels recorded for ZD rat plasma. A reduced but detectable pattern of response for zinc depletion and repletion was also observed in the PF rat plasma-treated cells, which may be a response to decreased plasma zinc or may possibly indicate the presence of the humoral factor in PF plasma, but at much lower levels. Our subsequent quest was to try and identify the humoral factor. We utilized the effect of this factor on expression of a marker gene (cxcl10) in VSMCs to

Figure 5. Principal component analysis of transcriptomics data from vascular smooth muscle cells treated with different rat blood plasma samples (see Fig. 1 for treatment identifier explanation). Principal component 1 (PC1) is plotted against principal component 3 (PC3) and together explained a combined total of 40.4% of the variation in the data.

Figure 7. Relative levels of cxcl10 mRNA in vascular smooth muscle cells exposed to native ZD rat plasma (ZD9), ZD9 plasma with most zinc removed using Chelex-100 resin (ZD2), ZD2 plasma with zinc added back to the ZD9 level (ZD9r), native PF rat plasma (PF27), PF27 plasma with most zinc removed using Chelex-100 resin (PF2) and PF2 plasma with zinc added back to the PF27 level (PF27r). Data are from qPCR analysis and are means ⫾ se of replicate treatments. Letters that are different indicate statistical significance (P⬍0.05).

8

Vol. 27

September 2013

The FASEB Journal 䡠 www.fasebj.org

OU ET AL.

Figure 8. A) Sephacryl S-200 HR separation of blood plasma from ZD (solid symbols) and PF (open symbols) rats, showing profiles for absorbance at 280 nm (blue circles and lines), zinc (red squares and lines), and calcium (green triangles and lines). The calcium peak represents calcium citrate and is effectively the column Vt (total volume). B) Relative levels of cxcl10 after pVSMC treatment with successive 5 ⫻ 1 ml pooled fractions from the Sephacryl S-200 HR separations of ZD and PF blood plasma.

identify its elution position in gel chromatography fractions and subsequently to calculate that its molecular weight was ⬃2 kDa. This is too small to be a protein or micro-RNA and is suggestive of a peptide hormone, although this has yet to be confirmed. Pathway analysis (Table 2) indicated that the factor targets signaling pathways related to immune function and development. Of note was the involvement of immune function-related cytokine pathways, including those related to IL-1, IL-6, IL-12, IL-17, IL-22, IL-23, and particularly IFN. Almost 40% of the genes whose expression was highly regulated by ZD9 plasma (⬎10-fold changed) were related to IFNs ␣, ␤, or ␥ (Fig. 3) and related pathways were some of the major responders to the zinc-regulated humoral factor (Table 2). It is possible that the effects on immune function in response to zinc deficiency are, at least in part, mediated by this factor. IFN␥ was analyzed in the rat plasma samples but showed no significant differences (P⫽0.4) between groups (Table 1), as was also found for plasma IL-1␤, IL-2, IL-6 and TNF-␣ (data not shown). Cytokines in any case have molecular weights that greatly exceed 2 kDa, and the only hypothesis involving IFN that seems compatible and consistent with our data is the production of an active peptide by proteolysis. Such a natural ZINC-REGULATED HUMORAL FACTOR

IFN-derived peptide has not previously been reported, although synthetic mimetic peptides have been generated (28). Natural antimicrobial peptides of the innate immune system (29) are also possible mediators of zinc deficiency affecting immune response pathways. If the humoral factor is a metabolic product of a larger protein or different peptide, another possible candidate is angiotensin II (ANG II, Mr⫽1300), since angiotensin-converting enzyme (ACE) has been shown to be zinc-dependent (30) and ANG II promotes apoptosis through the MAP kinase-Bcl-2/Bax system in a variety of cell types (31). Our studies have shown that zinc deficiency-induced VSMC apoptosis is caused by decreased phosphorylation of BAD through reduced phosphorylation of ERK1/2 (unpublished observations), but we have, as yet, no evidence for ANG II involvement. Evidence that ANG II requires zinc and down-regulation of ZnT3 and ZnT10 to induce senescence of VSMCs would indicate an involvement of zinc in ANG II signaling, and the mechanism proposed involves down-regulation of catalase expression through reduced phosphorylation of ERK1/2 (32). Thus, zinc is proposed to mimic the effect of ANG II rather than protect against it. In our studies, however, the bioactivity of the humoral factor was inversely related to the plasma zinc level, which shows that zinc has an ameliorating effect on the VSMCs. Nonpeptide hormone possibilities for a humoral factor include glucocorticoids, as they have been shown to mediate some effects of zinc deficiency (33, 34). The principal circulating glucocorticoid in rodents is corticosterone (CORT), and our data (Table 1) show that while zinc deficiency increases plasma CORT levels, pair-feeding does the same. In our experience, therefore, raised plasma CORT levels in zinc deficiency are associated more with food deprivation than with lack of zinc. CORT is mostly transported on corticosteroid binding globulin (transcortin, serpin-A6), which has a molecular weight of ⬃45 kDa (35). Free CORT is the active form of this glucocorticoid, but its molecular weight is 346 Da, and both forms are, therefore, well outside the molecular weight range of the humoral factor that we have detected. Therefore, we conclude that the VSMC gene expression effects that we have observed are not likely to be caused by corticosterone or any other steroid hormone. The source of the humoral factor is unknown, but it is likely to be produced within cells in which zinc levels or kinetics are modulated in response to changing dietary zinc intake. The exchangeable zinc pool (EZP) size, which normally constitutes ⬃10% of total body zinc, has been proposed as a marker of zinc status but its use as such in humans remains contentious (9). Nevertheless, it is clear from rodent studies that liver zinc, 80% of which is part of the EZP, decreases in marginal zinc deficiency despite there being no effect on plasma zinc levels (10, 14). Tissues containing part of the EZP that are depleted of zinc during marginal zinc deprivation are, therefore, likely to sense changing status and are, therefore, potential sources of a zinc9

regulated humoral factor. A possible strategy for the body when dietary zinc is limiting is to sacrifice exchangeable liver zinc, in particular, to maintain critical zinc-sensitive functions elsewhere. The generation of a humoral factor when body zinc homeostasis cannot be maintained through regulation of absorption and excretion, may act as a systemic signal, indicating that liver zinc level/EZP size is becoming depleted and that cellular zinc reserves should be conserved. In that regard, the humoral factor could be a useful biomarker of zinc status. The discovery of a humoral factor which mediates the response of tissues to lowered zinc status may provide a mechanism to understand how reduced dietary zinc intake can influence arterial protein expression (22) and accelerate the development of atherosclerotic plaque in a mouse model of atherosclerosis (14). In the present microarray study, pathway analysis of the 1600 significantly regulated genes (P⬍0.001), which were 2–3-fold changed by ZD9 compared to PF27 (therefore, excluding all highly regulated genes), showed a dominance of pathways related to apoptosis and survival, cytoskeletal remodeling, cell adhesion, and cell cycle (44% of the top 50 pathways). Assuming that genes less regulated by the treatment might not be primary targets and might indicate secondary responses, we propose that the affected pathways are consistent with our previous evidence of changing VSMC phenotype in the aorta of ZD rats (22). During the current studies, we also noted changes in cell morphology of VSMCs treated with ZD9, as compared to PF27 and ZA27, but this was not quantified. Our study shows that natural depletion of plasma zinc due to zinc deficiency is not the primary cause of effects on the cells. However, the ameliorating effect of adding back zinc suggests that the action of the humoral factor is blocked by zinc. Whether this occurs through zinc binding to the factor or its cell receptor, or indeed whether the zinc acts on the cell to resist the bioactivity of the factor, is as yet unknown. Plasma zinc levels can be markedly decreased by, for example, inflammation and microbial infection (36 –39), and we predict that this would enhance the bioactivity of a humoral factor produced in response to zinc deficiency, just as we have demonstrated by removing zinc from the ZD9 plasma. Various forms of stress and nutritional deficiencies tend to be found together in human populations, so the potentially synergistic effect of this interaction with regard to zinc status may be worthy of investigation. In summary, we have demonstrated that the blood plasma of ZD rats contains a low-molecular-weight factor whose level is regulated according to the dietary zinc intake of the animal and whose bioactivity is ameliorated by zinc in the plasma. We propose that the dual control of this factor by zinc—its level and its bioactivity—may explain why only relatively small changes in status and plasma zinc may have an effect on vascular function and possibly on the general health of animals. We are currently investigating whether a sim10

Vol. 27

September 2013

ilar factor can be detected in human plasma and whether different cell types are affected in the same way. The work was supported by the Rural and Environment Science and Analytical Services of the Scottish Government to J.H.B, M.J.G., G.C., and C.-D.M., and by the National Research Foundation of Korea (grant NRF 220-2008-1-F00013) to O.O., K.A.-R., D.U., J.H.B., G.F.N., J.F. and I.S.K., with funding in part from the University of Aberdeen.

REFERENCES 1. 2. 3.

4. 5. 6. 7. 8.

9. 10.

11.

12. 13. 14.

15. 16.

17.

18.

Maret, W. (2012) New perspectives of zinc coordination environments in proteins. J. Inorg. Biochem. 111, 110 –116 Fukada, T., Yamasaki, S., Nishida, K., Murakami, M., and Hirano, T. (2011) Zinc homeostasis and signaling in health and diseases: Zinc signaling. J. Biol. Inorg. Chem. 16, 1123–1134 Taylor, K. M., Hiscox, S., Nicholson, R. I., Hogstrand, C., and Kille, P. (2012) Protein kinase CK2 triggers cytosolic zinc signaling pathways by phosphorylation of zinc channel ZIP7. Sci. Signal. 5, ra11 Taylor, K. M., Kille, P., and Hogstrand, C. (2012) Protein kinase CK2 opens the gate for zinc signaling. Cell Cycle 11, 1863–1864 Wilson, M., Hogstrand, C., and Maret, W. (2012) Picomolar concentrations of free zinc(II) ions regulate receptor proteintyrosine phosphatase beta activity. J. Biol. Chem. 287, 9322–9326 King, J. C. (2010) Does zinc absorption reflect zinc status? Int. J. Vitam. Nutr. Res. 80, 300 –306 Kim, J., Paik, H. Y., Joung, H., Woodhouse, L. R., Li, S., and King, J. C. (2007) Effect of dietary phytate on zinc homeostasis in young and elderly Korean women. J. Am. Coll. Nutr. 26, 1–9 Kwon, C. S., Kountouri, A. M., Mayer, C., Gordon, M. J., Kwun, I. S., and Beattie, J. H. (2007) Mononuclear cell metallothionein mRNA levels in human subjects with poor zinc nutrition. Br. J. Nutr. 97, 247–254 Gibson, R. S., Hess, S. Y., Hotz, C., and Brown, K. H. (2008) Indicators of zinc status at the population level: a review of the evidence. Br. J. Nutr. 99(Suppl. 3), S14 –S23 Croxford, T. P., McCormick, N. H., and Kelleher, S. L. (2011) Moderate zinc deficiency reduces testicular Zip6 and Zip10 abundance and impairs spermatogenesis in mice. J. Nutr. 141, 359 –365 Walravens, P. A., Hambidge, K. M., and Koepfer, D. M. (1989) Zinc supplementation in infants with a nutritional pattern of failure to thrive: a double-blind, controlled study. Pediatrics 83, 532–538 Haase, H., Mocchegiani, E., and Rink, L. (2006) Correlation between zinc status and immune function in the elderly. Biogerontology 7, 421–428 Prasad, A. S. (2012) Discovery of human zinc deficiency: 50 years later. J. Trace Elem. Med. Biol. 26, 66 –69 Beattie, J. H., Gordon, M. J., Duthie, S. J., McNeil, C. J., Horgan, G. W., Nixon, G. F., Feldmann, J., and Kwun, I. S. (2012) Suboptimal dietary zinc intake promotes vascular inflammation and atherogenesis in a mouse model of atherosclerosis. Mol. Nutr. Food Res. 56, 1097–1105 Nagata, M., Lonnerdal, B. (2011) Role of zinc in cellular zinc trafficking and mineralization in a murine osteoblast-like cell line. J. Nutr. Biochem. 22, 172–178 Guo, B., Yang, M., Liang, D., Yang, L., Cao, J., and Zhang, L. (2012) Cell apoptosis induced by zinc deficiency in osteoblastic MC3T3-E1 cells via a mitochondrial-mediated pathway. Mol. Cell. Biochem. 361, 209 –216 Bao, B., Prasad, A. S., Beck, F. W., Bao, G. W., Singh, T., Ali, S., and Sarkar, F. H. (2011) Intracellular free zinc up-regulates IFN-gamma and T-beta essential for Th1 differentiation in Con-A stimulated HUT-78 cells. Biochem. Biophys. Res. Commun. 407, 703–707 Takeda, A., Tamano, H., Ogawa, T., Takada, S., Ando, M., Oku, N., and Watanabe, M. (2012) Significance of serum glucocorticoid and chelatable zinc in depression and cognition in zinc deficiency. Behav. Brain Res. 226, 259 –264

The FASEB Journal 䡠 www.fasebj.org

OU ET AL.

19. 20. 21. 22.

23. 24.

25.

26

27. 28. 29.

Eide, D. J. (2011) The oxidative stress of zinc deficiency. Metallomics 3, 1124 –1129 Wong, C. P., Ho, E. (2012) Zinc and its role in age-related inflammation and immune dysfunction. Mol. Nutr. Food Res. 56, 77–87 Beattie, J. H., Kwun, I. S. (2004) Is zinc deficiency a risk factor for atherosclerosis? Br. J. Nutr. 91, 177–181 Beattie, J. H., Gordon, M. J., Rucklidge, G. J., Reid, M. D., Duncan, G. J., Horgan, G. W., Cho, Y. E., and Kwun, I. S. (2008) Aorta protein networks in marginal and acute zinc deficiency. Proteomics 8, 2126 –2135 Reeves, P. G. (2003) Patterns of food intake and self-selection of macronutrients in rats during short-term deprivation of dietary zinc. J. Nutr. Biochem. 14, 232–243 Nixon, G. F., Hamilton, C. A., Wadsworth, R. M., and Reid, J. L. (1993) Comparison of endothelin-1 and noradrenaline stimulated inositol phosphate formation in cultured aortic smooth muscle cells from spontaneously hypertensive and Wistar Kyoto rats. Blood Press. 2, 221–227 Cho, Y. E., Lomeda, R. A., Ryu, S. H., Lee, J. H., Beattie, J. H., and Kwun, I. S. (2007) Cellular Zn depletion by metal ion chelators (TPEN, DTPA and chelex resin) and its application to osteoblastic MC3T3-E1 cells. Nutr. Res. Pract. 1, 29 –35 Smyth, G. K. (2005) Limma: Linear models for microarray data. In Bioinformatics and Computational Biology Solutions Using R and Bioconductor (Gentleman, R., Carey, V., Dudoit, S., Irizarry, R., and Huber, W., eds) pp. 397–420, Springer, New York Smyth, G. K. (2004) Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat. Appl. Genet. Mol. Biol. 3, 3 Johnson, H. M., Ahmed, C. M. (2006) Gamma interferon signaling: insights to development of interferon mimetics. Cell. Mol. Biol. (Noisy-le-grand) 52, 71–76 Radek, K., Gallo, R. (2007) Antimicrobial peptides: natural effectors of the innate immune system. Semin. Immunopathol. 29, 27–43

ZINC-REGULATED HUMORAL FACTOR

30. 31.

32.

33. 34. 35.

36.

37.

38.

39.

Fernley, R. T. (1977) Equine angiotensin converting enzyme: a zinc metalloenzyme. Clin. Exp. Pharmacol. Physiol. 4, 267–281 Horiuchi, M., Akishita, M., and Dzau, V. J. (1998) Molecular and cellular mechanism of angiotensin II-mediated apoptosis. Endocr. Res. 24, 307–314 Patrushev, N., Seidel-Rogol, B., and Salazar, G. (2012) Angiotensin II requires zinc and downregulation of the zinc transporters ZnT3 and ZnT10 to induce senescence of vascular smooth muscle cells. PLoS One 7, e33211 Takeda, A., Tamano, H. (2009) Insight into zinc signaling from dietary zinc deficiency. Brain Res. Rev. 62, 33–44 Fraker, P. J., King, L. E. (2004) Reprogramming of the immune system during zinc deficiency. Annu. Rev. Nutr. 24, 277–298 Breuner, C. W., Orchinik, M. (2002) Plasma binding proteins as mediators of corticosteroid action in vertebrates. J. Endocrinol. 175, 99 –112 Balogh, Z., El-Ghobarey, A. F., Fell, G. S., Brown, D. H., Dunlop, J., and Dick, W. C. (1980) Plasma zinc and its relationship to clinical symptoms and drug treatment in rheumatoid arthritis. Ann. Rheum. Dis. 39, 329 –332 Rofe, A. M., Philcox, J. C., Haynes, D. R., Whitehouse, M. W., and Coyle, P. (1992) Changes in plasma zinc, copper, iron, and hepatic metallothionein in adjuvant-induced arthritis treated with cyclosporin. Biol. Trace. Elem. Res. 34, 237–248 Strand, T. A., Adhikari, R. K., Chandyo, R. K., Sharma, P. R., and Sommerfelt, H. (2004) Predictors of plasma zinc concentrations in children with acute diarrhea. Am. J. Clin. Nutr. 79, 451–456 Mburu, A. S., Thurnham, D. I., Mwaniki, D. L., Muniu, E. M., and Alumasa, F. M. (2010) The influence of inflammation on plasma zinc concentration in apparently healthy, HIV⫹ Kenyan adults and zinc responses after a multi-micronutrient supplement. Eur. J. Clin. Nutr. 64, 510 –517 Received for publication February 15, 2013. Accepted for publication May 20, 2013.

11