View larger version (87K): [in this window] [in a new window] [Download PPT slide]   Figure 2. Cu supplementation–induced recovery of morphological abnormalities in mice subjected to AAC and fed 6 ppm Cu diet for 4 wk. Cardiac morphology changes were observed by cross section (C/S), histopathological changes including necrosis (hematoxylin and eosin [H/E]), fibrosis stained by Sirius red (SR), and increased cell size observed by wheat germ agglutinin (WGA) staining for cell membrane. EM revealed swelling (arrows), structural disruption (asterisks), and membrane rupture (arrowhead) in mitochondria. All of these changes were much more severe at 8 wk but were eliminated by dietary Cu supplementation during the last 4 wk (6/20 ppm) after AAC surgery. Insert in the EM image at 8 wk shows amorphous matrix densities (AMD), a specific severe damage to mitochondria. Bars: 50 µm (H/E, SR, and WGA panels); 1.5 µm (EM panels).

View larger version (37K): [in this window] [in a new window] [Download PPT slide]   Figure 3. Effect of Cu supplementation on AAC-suppressed VEGF, angiogenesis, and cardiac Cu concentrations. (A) Myocardial capillary density detected by CD31 staining. Bars, 50 µm. (B) Quantitative analysis of capillary density using a computer-assisted image analyzer (SigmaScan Pro5.0; SPSS, Inc.). Three fields at 200x magnification per heart were analyzed and six mice in each group were examined for the calculation. (C) Western blot analysis of VEGF levels in the heart tissue. (D) AAC-induced decreases in cardiac Cu concentrations 4 wk after AAC and the effect of Cu supplementation. (E) AAC-induced depression of cardiac CCO activity (CCO/CS ratio) 4 or 8 wk after AAC and the effect of Cu supplementation. *, Significantly different from sham-operated controls (P < 0.05, n = 6). #, Significantly different from AAC-treated mice fed 6 ppm Cu diet for 4 wk (P < 0.05, n = 6).

Anti-VEGF antibody blunts Cu supplementation-induced regression of hypertrophic cardiomyopathy
We investigated whether or not restoration of myocardial VEGF levels and the associated angiogenesis might be involved in dietary Cu supplementation-induced regression of cardiac hypertrophy and dysfunction by using anti-VEGF antibody. At the time when mice with established cardiac hypertrophy and dysfunction were switched from 6 mg Cu/kg diet to 20 mg Cu/kg diet, they were injected intravenously with anti-VEGF antibody every 3 d for 27 d. This treatment completely blunted Cu supplementation-induced regression of hypertrophic cardiomyopathy. The treatment with anti-VEGF antibody affected neither cardiac morphology nor contractile function of the sham controls (Fig. 4). This result thus suggests that dietary Cu supplementation-induced reversal of heart hypertrophy and dysfunction is mediated at least in part by restoration of myocardial VEGF expression and the associated angiogenesis.

View larger version (63K): [in this window] [in a new window] [Download PPT slide]   Figure 4. Effect of anti-VEGF antibody administration on Cu supplementation–induced regression of cardiac morphological changes and contractile dysfunction. Mice were fed 6 mg Cu/kg diet (6 ppm) for 4 wk after AAC surgery and then were changed to 20 mg Cu/kg diet for the last 4 wk in the presence of anti-VEGF antibody administration once every 3 d. All of the parameters and labels are the same as presented in Fig. 1 and 2. *, Significantly different from sham-operated controls (P < 0.05, n = 6). C/S, cross section; H/E, hematoxylin and eosin; SR, Sirius red; WGA, wheat germ agglutinin. Bars, 50 µm.

Cu requirement for HIF-1 activation of VEGF expression

View larger version (23K): [in this window] [in a new window] [Download PPT slide]   Figure 5. Effect of the Cu chelator TEPA on Cu- and IGF-1– induced HIF-1 activation and VEGF expression and the role of HIF-1 in Cu- and IGF-1–induced VEGF expression in cultured human cardiomyocytes. (A) The effect of TEPA on IGF-1– and Cu-induced HIF-1 transcription activity by ELISA assay. (B) The effect of TEPA on IGF-1– and Cu-induced VEGF expression by ELISA assay. Cells were cultured in the presence or absence of TEPA for 24 h before being exposed to IGF-1 or Cu for 16 h. Values are mean ± SD (n = 6), and the experiment was repeated two times with consistent results. *, Significantly different from untreated controls (P < 0.05, n = 6). #, Significantly different from corresponding no-TEPA treated cultures (P < 0.05, n = 6).

View larger version (16K): [in this window] [in a new window] [Download PPT slide]   Figure 6. Effect of excess Cu on TEPA-suppressed VEGF expression induced by Cu. Cells were cultured in the presence of 25 µM TEPA for 24 h before addition of Cu at different concentrations as indicated for 16 h. Values are mean ± SD (n = 6), and the experiment was repeated two times with consistent results. *, Significantly different from untreated controls (P < 0.05, n = 6).

View larger version (36K): [in this window] [in a new window] [Download PPT slide]   Figure 7. Effects of HIF-1 siRNA on IGF-1– or Cu-induced VEGF expression. Western blot analysis shows that treatment with HIF-1 siRNA for 48 h silenced CoCl2-induced HIF-1 expression but mismatched siRNA (MMsiRNA) did not. After being treated with the HIF-1 siRNA for 48 h, the cells were treated with IGF-1 or Cu for 16 h and VEGF levels in culture media were analyzed by ELISA assay and normalized by cell numbers. Values are mean ± SD (n = 6), and the experiment was repeated two times with consistent results. *, Significantly different from mismatched RNA only treated (B) controls (P < 0.05, n = 6). #, Significantly different from corresponding treated cultures (P < 0.05, n = 6).

CCS is involved in the interaction between Cu and HIF-1

View larger version (28K): [in this window] [in a new window] [Download PPT slide]   Figure 8. Effect of CCS on Cu- or IGF-1–induced HIF-1 activation and VEGF expression. (A) Coimmunoprecipitation of CCS and HIF-1. Human cardiomyocytes were treated with CoCl2 for 16 h to stimulate HIF-1 expression, as indicated. The cell extracts were immunoprecipitated by anti-CCS antibody (Santa Cruz Biotechnology, Inc.) and separated on 10% SDS-PAGE gel and immunoblotted with anti–HIF-1 antibody (Santa Cruz Biotechnology, Inc.) and anti-CCS antibody. (B) Effects of CCS siRNA on IGF-1– or Cu-induced HIF-1 activation and VEGF expression. Western blot analysis shows that treatment of the cells with the CCS siRNA for 48 h silenced the CCS expression but mismatched siRNA (MMsiRNA) did not. After treatment with the CCS siRNA for 48 h, the cells were incubated with IGF-1 or Cu for 16 h. HIF-1 transcription activity was assayed using the cell lysate by an ELISA assay and VEGF concentrations in culture media by an ELISA assay and normalized by cell numbers. CCS siRNA significantly suppressed the transcription activity of HIF-1 and VEGF expression. Values are mean ± SD (n = 6), and the experiment was repeated two times with consistent results. *, Significantly different from unstimulated control (P < 0.05, n = 6). #, Significantly different from corresponding mismatched RNA treated cultures (P < 0.05, n = 6).

View larger version (28K): [in this window] [in a new window] [Download PPT slide]   Figure 9. Effects of SOD1 siRNA on IGF-1– or Cu-induced VEGF expression. Treatment with SOD1 siRNA for 48 h silenced the SOD1 expression but mismatched siRNA (MMsiRNA) did not (Western blot). After treatment with the SOD1 siRNA for 48 h, the cells were incubated with IGF-1 or Cu for 16 h. VEGF concentrations in culture media were measured by an ELISA assay and normalized by cell numbers. Values are mean ± SD (n = 6), and the experiment was repeated two times with consistent results.

	DISCUSSION

Top ABSTRACT RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

Regression of preestablished hypertrophic cardiomyopathy has been shown recently using the phosphodiesterase-5A inhibitor, sildenafil (14). Our present results indicate that dietary Cu supplementation should be as effective as oral administration of sildenafil in reversing heart hypertrophy and dysfunction induced by chronic pressure overload. However, the simple and inexpensive dietary Cu supplementation would have obvious advantages over sildenafil. In particular, the level of the supplemented Cu is within the physiologic range and well below the maximum tolerable daily intakes (10 mg/day), so that it does not raise safety concerns and should not have the side effects (e.g., headache, flushing, and vision abnormalities) sometimes associated with sildenafil administration. The level of Cu supplementation used in the present study would be equivalent to 2.9 mg/day for humans, which can be found if conventional multiple mineral and vitamin supplements are used (2 mg Cu/day from the supplement plus RDA of 0.9 mg Cu/day).

Our findings are of clinical interest in multiple ways. (a) Given the high prevalence of hypertrophic cardiomyopathy, amelioration of these conditions by simply increasing dietary intakes of Cu would be practical and very attractive. (b) Marginal dietary Cu restriction in humans, defined as consumption of a diet containing less Cu than the RDA, has been identified from surveys of food consumption (15–17). This would adversely affect those in the population with cardiac complications. Therefore, it appears reasonable to monitor Cu intakes for these populations. (c) The current RDA for Cu has not considered the special case of individuals with cardiac complications. The data obtained here demonstrate that functional deficiency of Cu may occur in these populations and a different RDA for Cu for these populations should be considered. (d) Cu chelation therapy for cancers has recently been in clinical trials (18), but the potential adverse effects of Cu chelation on patients with heart disease need to be evaluated.

An important question is how Cu supplementation causes reversal of hypertrophic cardiomyopathy. We believe that one viable possibility is the effect of Cu on the synthesis of VEGF. A recent study has demonstrated that a lack of coordination between cardiac hypertrophy and angiogenesis (caused by suppression of VEGF activity) is associated with the progression of cardiac hypertrophy to heart failure (12). Acute phase of adaptive cardiac growth is associated with enhanced myocardial VEGF expression, coronary angiogenesis, and preservation of cardiac contractile function. Chronic and pathological hypertrophy is associated with depression of VEGF production and reduced angiogenesis (12). The study has defined that the decrease in VEGF levels is associated with the switching from adaptive to pathological hypertrophy under stress conditions (12). In the present study, we also observed that VEGF production was increased in the early phase, but decreased in the late phase of cardiac hypertrophy induced by sustained pressure overload. However, Cu supplementation restored VEGF production in the late phase of cardiac hypertrophy even in the continued presence of pressure overload.

Cu regulation of VEGF expression has been suggested in several studies (19–22). One study has shown that, at physiologic concentrations, Cu induces the synthesis of VEGF, which promotes angiogenesis in a healing wound (23). In Cu-deficient mouse hearts, generated by dietary Cu restriction, VEGF expression was markedly depressed and Cu repletion of the diet completely restored the expression of VEGF in the heart (22) along with regression of Cu deficiency–induced hypertrophic cardiomyopathy (24). The data presented here further demonstrate the link between Cu and VEGF production and the associated angiogenesis in the heart. Here, we have specifically addressed the question of whether or not VEGF is necessary for the regression of hypertrophic cardiomyopathy. The results clearly show that anti-VEGF antibody-mediated depletion of VEGF prevents Cu supplementation–induced reversal of hypertrophic cardiomyopathy, demonstrating the necessary role of VEGF in the regression.

Is Cu required for VEGF expression in cardiomyocytes? Cu at physiologically relevant levels stimulated VEGF expression, which mimicked the effect of IGF-1 in cultured human cardiomyocytes. The Cu chelator TEPA suppressed not only Cu-stimulated, but also IGF-1–induced VEGF production, indicating that Cu is required for IGF-1–activated VEGF expression. Furthermore, the basal level of VEGF expression was also decreased in the TEPA-treated control cultures, and addition of excess Cu into TEPA treated cultures restored VEGF expression. Therefore, these results strongly indicate that Cu is essential for VEGF expression in cardiomyocytes.

How does Cu regulate VEGF expression? The results showed that HIF-1 was critically involved in the Cu regulation of VEGF expression. The in vitro data that siRNA targeting HIF-1 abolished Cu stimulation of VEGF expression demonstrate an essential role of HIF-1 in Cu-dependent VEGF expression in the heart. In the cell, there is virtually no free Cu (25), so that Cu chaperones direct Cu transport specifically to its target molecules. For instance, CCS has been identified as a Cu chaperone for SOD1 (26). CCS interacts directly with SOD1 for the insertion of ionic Cu into the metalloprotein (25). This efficient Cu delivery mechanism avoids inefficient and potentially dangerous trafficking of Cu within the cell. In this context, the interaction between Cu and HIF-1 could also be mediated by a Cu chaperone such as CCS. In fact, we found that CCS coimmunoprecipitated with HIF-1. This is an important novel observation, as it indicates that CCS may not only transfer Cu ion to SOD1 but to other proteins as well. It also indicates that a direct interaction between CCS and the Cu requiring apoproteins may represent a common mechanism for Cu transfer in the cell. However, it is the Cu requiring apoproteins that determine the functional specificity. In this context, we have excluded the possible involvement of SOD1 in Cu regulation of VEGF by showing that SOD1 gene silencing did not inhibit IGF-1– or Cu-stimulated VEGF expression. An unresolved issue is how Cu interacts with HIF-1 through CCS, a question that will require more extensive chemical and structural studies.

The result obtained from the use of siRNA targeting CCS demonstrated the importance of CCS in the regulation of Cu-activated HIF-1 transcription activity and VEGF expression. The inhibitory effect of CCS gene silencing on IGF-1–induced HIF-1 activation and VEGF expression was as the same as that of Cu chelator, TEPA. Since CCS is a Cu-specific chaperone, the CCS gene silencing result not only provides a mechanistic insight into Cu regulation of VEGF expression but also supports the conclusion obtained from the Cu chelator study for the essentiality of Cu in the regulation of VEGF expression. The CCS gene silencing result also excludes a possible confounding effect of nonspecific metal chelation by TEPA on IGF-1–induced HIF-1 activation and VEGF expression.

This study for the first time presents evidence that Cu is required for HIF-1 activation and VEGF expression in cardiomyocytes. Cu activation of HIF-1 is mediated by CCS, which appears to directly interact with HIF-1. The results provide a scientific rationale for Cu supplementation as a means of causing the regression of hypertrophic cardiomyopathy induced by sustained pressure overload. Should similar effects of Cu supplementation be found in controlled studies in human patients, this will point the way to a simple, nontoxic and extraordinarily economical therapy for hypertrophic cardiomyopathy.

	MATERIALS AND METHODS

Top ABSTRACT RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Animal models.
Male C57BL/6J mice (10–12 wk old; The Jackson Laboratory) were fed standard AIN-93 diet (27) containing 6 mg Cu/kg, which is the adequate level of Cu for mice according to the RDA established by Food and Nutritional Board of the Institute of Medicine (2) and is equivalent to current RDA for humans (0.9 mg Cu/day). Pressure overload was produced by AAC as described previously (28–30). In brief, mice were anesthetized and ventilated with a rodent respirator. Aortic constriction was made through a left thoracotomy by placing a ligature using a 6-0 silk suture securely around the ascending aorta and a 26 1/2– gauge blunted needle and then removing the needle. For the sham operation, the left thoracotomy was made but the aorta was not ligated. The animal survival rate due to the AAC surgery was 80% including postsurgical death. The animal procedure was approved by the Institutional Animal Care and Use Committee, which is certified by the American Association of Accreditation for Laboratory Animal Care.

4 wk after the AAC surgery, some mice were sacrificed for analyses. One group of AAC mice was changed to a diet containing 20 mg Cu/kg, and the other group remained on the 6 mg Cu/kg diet. The supplementation level of Cu in the diet is equivalent to 2.9 mg Cu/day in humans, an intake easily reached if conventional multiple mineral and vitamin supplements are used (2 mg Cu/day plus RDA of 0.9 mg Cu/day). The mice were harvested for the analyses at either 4 or 8 wk after AAC. At the time of harvesting, heart weight to body weight ratio and lung weight to body weight ratio were obtained. Although there were premature deaths of some AAC mice fed 6 mg Cu/kg diet for 8 wk, the number of mice for each parameter measured was six or greater.

Anti-VEGF treatment.
Anti–mouse VEGF antibody (R&D Systems, Inc.), 20 µg in 200 µl PBS, was given every 3 d by intraperitoneal injection to some mice starting when the mice were switched to the 20 mg Cu/kg diet after feeding 6 mg Cu/kg diet for 4 wk in both AAC surgery and sham controls. The treatment lasted for 27 d.

Echocardiography.
Two-dimensional images of the left ventricle were obtained from the apical four- and two-chamber views with a phased-array echocardiography (SONOS 1500 or 2500; Hewlett-Packard, Inc.) and a 2.5-MHz transducer in anesthetized mice. The images were recorded on 1/2-inch videotape for subsequent review and analysis. With the use of a commercially available microcomputer system (GTI; Freeland), the echocardiograms were analyzed qualitatively and quantitatively for the development of regional LV wall motion abnormalities. Quantitative analysis was performed by use of a centerline method that constructs 100 equidistant chords perpendicular to a line centered between digitized LV end- diastolic and end-systolic endocardial borders.

Cu concentrations in the heart.
Total Cu concentrations were determined in the heart using inductively coupled argon plasma emission spectroscopy (model 35608; Thermo ARL-VG Elemental) after lyophilization and digestion of the tissues with nitric acid and hydrogen peroxide (31). To confirm the dietary Cu concentrations, a dry-ashing procedure was used. This procedure included dissolution of the residue in aqua regia and measurement by atomic absorption spectrophotometry (model 503; PerkinElmer).

Assessment of morphology changes.
Hearts were excised, washed with saline solution, and placed in 10% formalin. Hearts were cut transversely close to the apex to visualize the left and right ventricles. Several sections of heart (4–5-µm thick) were prepared and stained with hematoxylin and eosin for histopathology or sirius red for collagen deposition and visualized by light microscope. Capillaries were visualized in frozen left ventricle sections stained with antibodies to endothelial cell adhesion molecule–1 (PECAM-1 or CD31) (Santa Cruz Biotechnology Inc.). The capillary density was assessed by a computer-assisted image analyzer (SigmaScan Pro5.0; SPSS, Inc.). Three fields at 200x magnification per heart were analyzed, and six mice in each group were examined for the calculation. For examination by electron microscopy, a tissue sample preparation procedure described previously (32) was used.

Cell culture studies.
Human cardiac myocytes (HCMs) along with the culture medium were purchased from ScienCell Research Laboratories. The culture medium was composed of basal medium, 5% FBS, 1% cardiac myocyte growth supplement, and 1% penicillin/streptomycin solution (ScienCell). Cells were maintained at 37°C in a 95% humidified air and 5% CO₂. Before each experiment, monolayer HCMs were allowed to grow to 90% confluence, and switched to low-serum medium (basal medium supplemented with 1% FBS) for 24 h, then exposed to 100 ng/ml IGF-1 (Sigma-Aldrich), 25 µM CuSO₄ (Sigma-Aldrich), or 25 µM TEPA (Sigma-Aldrich) in the low-serum medium, alone or in combination. The culture media were collected for VEGF measurement, and cells were harvested and subjected to HIF-1 activity assay at 16 h after the treatment.

Gene silencing.
siRNAs targeting human HIF-1, CCS, and SOD1 and negative mismatched control siRNA were designed and synthesized by Ambion. The siRNA sequences for HIF-1 were sense, GGCCUCUGUGAUGAGGCUUtt and antisense, AAGCCUCAUCACAGAGGCCtt. The siRNA sequences for CCS, sense, GGACCAGAUGGUCUUGGUAtt, and antisense, UACCAAGACCAUCUGGUCCtt. A validated SOD1 siRNA was obtained from the same company (siRNA ID no.: 119429). After the monolayer HCM cultures reached 50% confluence, the cells were transfected with 100 nM annealed siRNAs targeting specific genes or negative mismatched siRNA using a Silencer siRNA transfection kit (Ambion) according to the manufacturer's instruction. At 48 h after the transfection, the culture media were changed to the low-serum medium for 24 hr, and then exposed to 100 ng/ml IGF-1 or 25 µM CuSO₄ alone or in combination. Some cultures were exposed to 150 µM CoCl₂ (Sigma-Aldrich) to increase HIF-1 levels. The culture media were collected for VEGF measurement, and cells were harvested and subjected to HIF-1 analysis at 16 h after the last treatment.

HIF-1 activity assay.
The nuclear extracts of cultured cells were obtained by using a nuclear extract kit (Active Motif) according to the manufacturer's instruction. HIF-1 activity was quantified in 5–10 µg of nuclear extracts by specific binding of HIF-1 to an oligonucleotide containing the HRE from the EPO gene (5'-TACGTGCT-3') using a TransAM HIF-1 transcription factor assay kit (Active Motif) following the manufacturer's instruction. Each sample was measured in duplicate, and the measured activity was normalized to total protein content measured by a Bio-Rad Laboratories protein assay kit.

Quantitative analysis of VEGF.
VEGF in culture media was measured using a Quantikine Human VEGF Immunoassay (R&D Systems) according to the manufacturer's instruction. Each sample was measured in duplicate, and the VEGF concentrations were normalized by the number of cells in each culture sample.

Enzymatic assay for CCO.
Mitochondria were isolated from fresh heart samples using a mitochondrial isolation kit (Pierce Chemical Co.) following the instruction provided. CCO was assayed as described previously (33). In brief, CCO was measured by adding ferrocytochrome c in an isosmotic medium (10 mM KH₂PO₄, pH 6.5, 1 mg/ml bovine albumin, 0.3 M sucrose). The activity was calculated from the rate of decrease in absorbance of ferrocytochrome c at 550 nM ( = 19.1 mM^–1cm^–1) and normalized by mitochondrial protein content. Because accurate estimations of CCO activity could be compromised by variations in mitochondrial yield and integrity, the CCO activity in the present study was normalized using an invariant marker of mitochondrial enzyme activity, citrate synthase. Citrate synthase was assayed by following the reduction of 5,5'-dithio-bis(2-nitrobenzoic acid) (2 mM) in the presence of acetyl-CoA (0.1 mM) and oxalacetic acid (12 mM) in a medium containing 10 mM KH₂PO₄, pH 7.8, 2 mM EDTA, 1 mg/ml BSA, and 0.1% Triton X-100. The change in absorbance at 412 nm was monitored for 30 min (reflecting the formation of thionitrobenzoate). The citrate synthase activity was calculated from the change in absorbance using = 13.6 mM^–1cm^–1, and normalized by mitochondrial protein content.

Immunoprecipitation analysis of CCS and HIF-1.
HCM cell lysates were prepared in ice-cold IP buffer containing 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Triton X-100, 10% glycerol and protease inhibitor cocktail (Sigma-Aldrich). Crude cell lysates containing 300 µg protein were precleared by incubating with protein G–Sepharose 4FF (GE Healthcare) and mixed with 4 µg of rabbit anti-CCS antibody (FL-274; Santa Cruz Biotechnology, Inc.) and incubated at 4°C overnight under rotation. 30 microliters of preequilibrated protein G–Sepharose 4FF were then added and incubated for 1 h at 4°C under rotation. After centrifugation at 1,000 g for 2 min, supernatants were removed, and the beads were washed three times with the IP buffer. The resulting immunocomplex-containing beads were treated with SDS sample buffer and subjected to Western blot analysis.

Western blot analysis of proteins.
The sample preparation from frozen mouse hearts and cultured cells, gel electrophoresis, and transfer to blotting membranes were described previously (34, 35). Nonspecific binding was blocked by using a blocking buffer (5% nonfat dry milk, Tris-buffered saline: 20 mM Tris-HCl, pH 7.5, 500 mM NaCl, 0.5% Tween 20) for 2 h at room temperature, and then the blotting membranes were incubated overnight at 4°C with the following primary antibodies diluted in blocking buffer: rabbit anti–human SOD1 polyclonal antibody (1:1,000 Santa Cruz Biotechnology, Inc.), mouse anti–human HIF-1 monoclonal antibody (1 µg/ml; Sigma-Aldrich), goat anti–mouse VEGF polyclonal antibody (2 µg/ml; R&D Systems), rabbit anti–human CCS (1:2,000; Santa Cruz Biotechnology, Inc.), and mouse anti-GAPDH monoclonal antibody (1:5,000; StressGen Biotechnologies). The membranes were washed in Tris-buffered saline, 1% Tween 20 (TBST) and incubated for 2 h at room temperature with HRP-conjugated rabbit anti–goat IgG and goat anti–mouse or rabbit IgG antibodies (1:20,000; Southern Biotechnology Associates, Inc.) diluted in TBST. The blots were processed and visualized by chemiluminescence (GE Healthcare).

Statistical analysis.
Data were expressed as mean ± SD values and analyzed by ANOVA followed by a Duncan's multiple-range test for further determination of the significance of differences. Differences among groups were considered significant when P < 0.05.

Related Article

Heavy metal for a troubled heart

Hema Bashyam
J. Exp. Med. 2007 204: 455a. [Full Text] [PDF]