Concerns regarding excessive hepatic copper concentrations in dairy cows have increased. The objective of this study was to determine the association of hepatic copper concentrations with evidence of liver disease. Blood and liver samples were collected at the time of slaughter in cull dairy cows (n=100). Liver samples were analyzed for copper using inductively coupled plasma mass spectrometry and crude fat using liquid-liquid extraction and gravimetry. Serum samples were analyzed for glutamate dehydrogenase,
Copper is an essential cofactor in hundreds of enzymatic reactions and is a necessary component of the diet of all species [
Damaged hepatocytes release hepatocellular enzymes into circulation where they can be measured in peripheral blood by routine clinicopathological tests. Serum activities of these enzymes are therefore commonly used as biomarkers for hepatopathies. Clinicopathological analysis for serum hepatocellular leakage enzyme activities is a critical feature in clinical diagnosis of many hepatopathies in a variety of species, including copper accumulation hepatopathies in dogs, sheep, and man [
When copper intake exceeds liver storage and the export capabilities, then free copper is more likely to occur. Free copper has prooxidant potential and in excess can lead to increased oxidative stress through the Fenton reaction [
The prooxidant characteristics of metallic copper may be a link between high hepatic copper concentrations and clinical disease due to increased risk for oxidative stress. Our group has demonstrated that high hepatic copper concentrations occur frequently in dairy cows, but the clinical significance of hepatic copper concentrations in this range is unknown [
Paired liver and blood samples (n=100) from cull dairy cows were collected at the time of slaughter at a West Michigan abattoir. This study was exempted by the Michigan State University Institutional Animal Care and Use Committee. Blood samples were collected in serum separator tubes (BD Vacutainer Serum Separator) and EDTA Tubes, (Becton, Dickinson and Company, Franklin Lakes, NJ 07417) during exsanguination. Serum was separated from the clot within an hour of collection (9.055.8 xG for 10 min) and stored at 4°C until analysis within 96 hours for liver enzyme activity and bile acid concentration. For RONS and AOP analysis, approximately 1 mL of serum was pipetted into cryo-vial tubes. The vials were flash frozen in liquid nitrogen and then later transferred to a -80°C freezer for storage until analysis.
Hepatic copper concentration was determined on fresh tissue but expressed on a dry tissue basis. Tissues were sectioned (1 gm) and digested overnight in a 95°C oven, using approximately 10 times the dry tissue mass of nitric acid. A separate 2-gm section was dried overnight in a 75°C oven to determine the dry matter fraction and calculate the dried tissue mass. The digested samples were diluted with water to 100 times the tissue mass. Elemental analysis was performed using an Agilent Inductively Coupled Plasma Mass Spectrometer (ICP-MS) (Agilent Technologies Inc., Santa Clara CA 95051) [
Animals were stratified based on liver copper concentration and divided into quintiles. Variables associated with oxidative stress were measured in the highest and lowest quintiles (Q5 and Q1). These variables included serum total reactive oxygen and nitrogen species (RONS) and antioxidant potential (AOP).
Blood chemistry profiles were performed by Marshfield Laboratories (
Crude fat analysis was done by solvent extraction using a modified version of a previously published method [
AOP was quantified in serum samples as described previously [
Formalin fixed liver sections in Q1 and Q5 (n=40) were stained with immunohistochemistry (IHC) stains for 4-hydroxynonenal (4HNE) and 3-nitrotyrosine (3NIT) at the Department of Veterinary Pathology, Iowa State University. Tissues were cut into 5 micron sections and placed on glass slides. Slides were baked at 57°C for 30 min. Tissues were deparaffinized in xylene and then rehydrated in alcohol. Tissues were then processed with a commercially available Avidin/Biotin blocking kit (Avidin/Biotin Blocking Kit, Cat # 004303, ThermoFischer Scientific, Waltham, MA 02451). In order to inhibit endogenous peroxidase activity, tissues were incubated in 3% hydrogen peroxide for 2, 10 min applications followed by 3 ultrapure water rinses. For the antigen retrieval process slides were placed in a plastic Coplin jar containing Citra pH 6 antigen retrieval buffer. They were then microwaved on high (full power) until the surface was bubbling. The Coplin jar was then moved to a preheated steamer for 20 min. They were then left at room temperature for 20 min and rinsed twice with PBS. Next, the slides were blocked with 90% NGS/(Tris/PBS/BSA buffer) for 20 min. The primary mouse monoclonal antibody was diluted in Tris/PBS/BSA buffer (4-HNE was diluted 1:50 and 3-NIT was dilution 1:100) (Anti-4-Hydroxynonenal Antibody, Abcam, Cambridge, MA 02139 ) (Anti-3-Nitrotyrosine Antibody, Santa Cruz Technology, Dallas, TX 75220) and incubated for 2 hours followed by 2x PBS rinses, 5 min PBS bath, and 2x PBS rinses. Next, samples were processed with dilute Multilink 1:80 in Tris/PBS/BSA buffer that was applied for 15 min followed by 2 PBS rinses, 5 min PBS bath, and 2 more PBS rinses. Then dilute Horseradish Peroxidase-Streptavidin 1:200 in Tris/ PBS/BSA buffer was applied for 15 minutes followed by 2x PBS rinses, 5 minute PBS bath, and 2x PBS rinses. Subsequently, Nova Red stain was applied for 5 min and rinsed 5 times with ultrapure water. Slides were then transferred into 1/4 strength Shandon’s hematoxylin for 2 min, rinsed in ultrapure water 3 times, placed in Scott’s tap water for 1 min, and then rinsed 3 times in ultrapure water. Slides were then dehydrated with alcohol and Xylene. Coverslip slides were applied with a nonaqueous mounting media (Acrytol Mounting Media, Surgipath, Leica Biosystems, Buffalo Grove, IL 60089) and allowed to dry at room temperature. Formalin fixed liver sections were processed for H&E and rhodanine staining (n=100).
Rhodanine and IHC stained slides were graded in the same manner on a 0-5 scale. A score of 0 showed no IHC or rhodanine staining. A score of 1 was a section that had minimal granules in less than 33% of centrilobular hepatocytes. A score of 2 indicated there was moderate rhodanine or IHC stained granules in less than 50% of centrilobular hepatocytes. A score of 3 indicated there were moderate to large amounts of rhodanine or IHC stained granules in greater than 50% of the centrilobular hepatocytes. A score of 4 had staining in greater than 75% of zone 3 hepatocytes and a score of 5 demonstrated panlobular staining.
H&E slides were all observed for signs of centrilobular necrosis and were scored on relative amount of inflammation and fibrosis. For inflammation a score of 0 meant that there were no visible inflammatory cells, a score of 1 had a mild inflammatory infiltrate with low numbers of lymphocytes, plasma cells, and histiocytes within the portal region, and sections with a score of 2 had severe inflammation with high numbers of the before-mentioned inflammatory cells and neutrophils within the periportal regions. For fibrosis, a score of zero meant that the section had no fibrosis, a score of 1 meant that the section had mild expansion of the portal areas by dense material fibrosis, and sections with a score of 2 demonstrated severe, bridging fibrosis between portal areas.
The associations among continuous variables were examined via an exploratory factor analysis (Proc FACTOR, SAS 9.4). Histograms of the variables were examined visually for their distribution. The serum liver-leakage enzyme activity distributions were skewed to the right in a non-Gaussian pattern, suggesting neither the assumptions of normality nor multivariate normality could be made for these data. Thus, the “unweighted least squares” method of factor extraction, which is robust to nonnormal data distributions [
Differences in copper concentrations across the histologically determined score categories for inflammation, fibrosis, necrosis, and rhodanine staining were evaluated by the Kruskal-Wallis test. Differences in mean values for serum RONS, AOP, and IHC scores in Q1 and Q5 were assessed by the Mann-Whitney
The mean hepatic copper concentration was 496.83
Serum liver-leakage enzyme activity, as well as serum bile acid and hepatic crude fat concentration results, is shown in Table
Liver leakage enzymes, BA, and hepatic CF data.
Mean ± SE | Range | Reference Range | |
---|---|---|---|
GLDH (U/L) | 23.5 ± 1.65 | 4-80 | 6-68 |
SDH (U/L) | 19.66 ± 1.03 | 7-72.6 | 6.6-37.8 |
GGT (U/L) | 28.07 ± 0.96 | 13-68 | 4-41 |
BA ( | 28.12 ± 2.70 | 2-174 | 0-12 |
AST (U/L) | 91.43 ± 6.01 | 32-394 | 48-204 |
Hepatic Crude Fat (%) | 7.96 ± 0.48 | 4.44-33.11 | 3-8 |
Summary of serum liver leakage enzyme activity, bile acid concentration, and hepatic crude fat percentage in cull Holstein dairy cows (n=100).
Pearson’s correlation.
CF | Cu | GGT | AST | GLDH | BA | SDH | |
---|---|---|---|---|---|---|---|
CF | 1.0000 | -0.3053 | -0.0981 | 0.4009 | 0.1443 | 0.0439 | 0.0627 |
0.0020 | 0.3316 | <0.0001 | 0.1521 | 0.6643 | 0.5317 | ||
| |||||||
Cu | -0.3053 | 1.0000 | 0.1211 | -0.1302 | -0.0025 | 0.1719 | 0.1280 |
0.0020 | 0.2301 | 0.1966 | 0.9803 | 0.0872 | 0.2043 | ||
| |||||||
GGT | -0.0981 | 0.1211 | 1.0000 | 0.1371 | 0.4377 | 0.1564 | 0.4023 |
0.3316 | 0.2301 | 0.1739 | <0.0001 | 0.1201 | <0.0001 | ||
| |||||||
AST | 0.4009 | -0.1302 | 0.1371 | 1.0000 | 0.2348 | -0.0236 | 0.2147 |
<0.0001 | 0.1966 | 0.1739 | 0.0187 | 0.8157 | 0.0320 | ||
| |||||||
GLDH | 0.1443 | -0.0025 | 0.4377 | 0.2348 | 1.0000 | 0.2161 | 0.5091 |
0.1521 | 0.9803 | <0.0001 | 0.0187 | 0.0308 | <0.0001 | ||
| |||||||
BA | 0.0439 | 0.1719 | 0.1564 | -0.0236 | 0.2161 | 1.0000 | 0.0603 |
0.6643 | 0.0872 | 0.1201 | 0.8157 | 0.0308 | 0.5512 | ||
| |||||||
SDH | 0.0633 | 0.1280 | 0.4023 | 0.2147 | 0.5091 | 0.0603 | 1.0000 |
0.5317 | 0.2043 | <0.0001 | 0.0320 | <0.0001 | 0.5512 |
Pearson’s correlation of liver leakage enzymes, bile acids, and hepatic crude fat for all of the study samples (n=100). The top number is
Factor analysis of liver-leakage enzymes, bile acids, and hepatic crude fat with an oblique rotation represented in graphic form (n=100). A moderate loading was considered to be 0.4-0.7 and a strong loading was considered >0.7-1.0, in absolute values. The latent variables are represented as (a) hepatocyte health factor and (b) hepatic lipidosis factor.
There was no difference in serum RONS or the Osi ratio between Q1 and Q5. There was, however, a significant increase in serum AOP in Q5 compared to Q1. This may suggest a compensatory response in the redox status of the cows in Q5 (Table
Systemic oxidative stress indices.
Q1 | Q5 | |
---|---|---|
Mean ± SE | Mean ± SE | |
RONS (RFU/ | 34.03 ± 6.48 | 39.73 ± 6.52 |
AOP (TE/ | 16.05 ± 0.50 | 18.75 ± 0.5 |
Osi | 2.17 ± 0.41 | 2.18 ± 0.40 |
Systemic oxidative stress variables for Q1 and Q5 (n=40). Total antioxidant potential (AOP) was higher in Q5 than Q1 (P = 0.0013). No difference was found between Q1 and Q5 for reactive oxygen and nitrogen species (RONS) or oxidant stress index (Osi) (P > 0.05).
IHC staining scores in Q5 were significantly higher than Q1 for both 4HNE and 3NIT (p = 0.0003 and p = 0.0058, respectively) (Table
Immunohistochemistry scores.
Q1 | Q5 | |
---|---|---|
Mean ± SE | Mean ± SE | |
4HNE | 0.85 ± 0.19 | 2.1 ± 0.2 |
3NIT | 0.2 ± 0.12 | 1.03 ± 0.2 |
Immunohistochemistry (IHC) scores for Q1 and Q5 (n=40). IHC slides were scored based on the relative amount of staining from 0 for no staining to 5 for pan lobular staining. Four-Hydroxynonenal (4HNE) staining scores were higher in Q5 than Q1 (p < 0.001) and 3-nitrotyrosine (3NIT) staining scores were higher in Q5 than Q1 (p < 0.01).
The 3-nitrotyrosine (3NIT (a, b)) and 4-hydroxynonenal (4HNE (c, d)) stained slides for two separate samples centered on the zone 3 hepatocytes where copper accumulates first in the bovine liver. Central vein (CV) is labeled and slides are magnified at 20x. Cow 1 (a, c) had a hepatic copper concentration of 1264.27
Hepatic necrosis is a frequent consequence of copper intoxication; however H&E staining revealed no evidence of hepatic necrosis in any of the animals studied. In addition, there was minimal evidence of inflammation or fibrosis. Indeed, there were only 2 cows with fibrosis scores greater than 0 so Kruskal-Wallis could not be utilized. There was no correlation between hepatic copper concentrations and inflammation (P = 0.937). There was a significant positive relationship (r2 = 0.483) between rhodanine score and copper concentration (Figure
Correlation of quantitative copper concentrations (
In the current study we examined the association among multiple variables related to hepatic injury or dysfunction with hepatic copper concentrations. We expected associations, or factors, to emerge that would represent different aspects, types, or dimensions of hepatic injury or dysfunction. Factor analysis is a powerful means of examining and interpreting associations among multivariate data. The correlation matrix of multiple variables forms the basis of the data for factor analysis. The analysis creates unmeasured variables called factors. The strength of association between each measured variable and each factor is defined by the loading parameter, which can vary between -1 and +1 and is interpreted similarly to a correlation coefficient. A relatively high absolute loading of a measured variable on a given factor indicates the strength of its association, positive or negative, with that factor. Of particular interest in the interpretation of factor analysis is the grouping of measured variables among the factors because these groupings define the biological nature of the factor. The factors are sometimes referred to as latent variables, suggesting they are phenomena of interest but are difficult to measure objectively. Further, and pursuant to our major objective, we wish to see if hepatic copper concentration is associated with one or more of these dimensions of hepatic injury or dysfunction. Such an association within the study population would imply that increasing copper concentration might be a risk factor for subclinical hepatic disease.
Factor 1 was termed the “hepatic health” factor as it had strong loadings for serum GLDH and a moderate loading of serum SDH and GGT, well accepted clinicopathological indicators of bovine hepatic health or integrity [
Factor 2 was assigned the term “hepatic lipidosis” factor with strong loadings for liver fat percentage and a moderate loading for serum AST. This is consistent with previous investigations that have observed bovine serum AST activity to be more strongly associated with liver fat concentration than the serum activities of other liver enzymes evaluated in this experiment [
This study demonstrated that hepatic sections in Q5 showed significantly greater oxidative stress as demonstrated by 4HNE and 3NIT IHC staining than those from Q1 (p = 0.0003 and p = 0.0058, respectively). Moreover, IHC staining distribution mirrored that of rhodanine staining within hepatic lobules (Figure
There was a significant increase in AOP in Q5 than Q1. The elevation in AOP may be due to a compensatory increase in thiols following the increase of RONS. This change, however, demonstrates a redox shift in Q5 cows. Compared to results from previous studies, RONS was elevated in both Q1 and Q5 relative to nondiseased cows at dry-off [
In this study a population of cull cows slaughtered at a small commercial abattoir was used as a convenience sample. Cows are culled for many reasons, but an intense period of involuntary culling occurs early in lactation [
The data used to support the findings of this study are available from the corresponding author upon request
The authors declare that they have no conflicts of interest.
The author’s would like to thank Howard and Barbra Stowe for funding the Stowe Endowment that made this nutrition residency and research possible. We would also like to thank the hard work of individuals that contributed to the gathering of samples and analyzing of the data that made this research possible. At the MSU Veterinary Diagnostic Laboratory Cheryl Engfehr and Justin Zyskowski were essential to the nutritional assays. In the MSU Meadow Brook Laboratory, Jeff Gandy and Jennifer De Vries played crucial roles in ensuring the oxidative stress variables were analyzed correctly. Funding for this study was provided by the Michigan State University College of Veterinary Medicine Endowed Research Funds and the Stowe Endowment Fund.