NMR Metabolomic Analysis of Dairy Cows Reveals ... - ACS Publications

Nov 21, 2011 - Wolfgang Junge,. ‡. Georg Thaller,. ‡. Peter J. ... Institute of Functional Genomics, University of Regensburg, Germany. ‡. Insti...
0 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/jpr

NMR Metabolomic Analysis of Dairy Cows Reveals Milk Glycerophosphocholine to Phosphocholine Ratio as Prognostic Biomarker for Risk of Ketosis Matthias S. Klein,† Nina Buttchereit,‡ Sebastian P. Miemczyk,† Ann-Kathrin Immervoll,† Caridad Louis,† Steffi Wiedemann,‡ Wolfgang Junge,‡ Georg Thaller,‡ Peter J. Oefner,† and Wolfram Gronwald*,† † ‡

Institute of Functional Genomics, University of Regensburg, Germany Institute of Animal Breeding and Husbandry, Christian-Albrechts-University, Kiel, Germany

bS Supporting Information ABSTRACT: Ketosis is a common metabolic disease in dairy cows. Diagnostic markers for ketosis such as acetone and beta-hydroxybutyric acid (BHBA) are known, but disease prediction remains an unsolved challenge. Milk is a steadily available biofluid and routinely collected on a daily basis. This high availability makes milk superior to blood or urine samples for diagnostic purposes. In this contribution, we show that high milk glycerophosphocholine (GPC) levels and high ratios of GPC to phosphocholine (PC) allow for the reliable selection of healthy and metabolically stable cows for breeding purposes. Throughout lactation, high GPC values are connected with a low ketosis incidence. During the first month of lactation, molar GPC/PC ratios equal or greater than 2.5 indicate a very low risk for developing ketosis. This threshold was validated for different breeds (Holstein-Friesian, Brown Swiss, and Simmental Fleckvieh) and for animals in different lactations, with observed odds ratios between 1.5 and 2.38. In contrast to acetone and BHBA, these measures are independent of the acute disease status. A possible explanation for the predictive effect is that GPC and PC are measures for the ability to break down phospholipids as a fatty acid source to meet the enhanced energy requirements of early lactation. KEYWORDS: metabolomics, NMR, phosphatidylcholine, production disease, prediction of disease status

’ INTRODUCTION Over the last few decades milk production in dairy cows has dramatically increased. However, this has been accompanied by a substantially raised rate of health and fertility problems.1 Main health issues are the so-called production diseases such as ketosis, mastitis, milk fever, and ruminal acidosis leading to animal suffering, reduced longevity, and economic loss due to reduced milk yield, decreased milk quality, extra costs for drugs and veterinary treatments, additional labor, and involuntary culling.2 The economic costs amount for hundreds of millions of euro per year for France and the United Kingdom alone.3 The enhanced energy requirements during early lactation due to selection of high-yielding cows result in an increased allocation of available energy resources to milk synthesis and less to functions relevant to fertility and fitness.4 If energy requirements exceed energy availability from daily food intake, body fat is mobilized,5 which is a physiological mammalian process. However, excessive mobilization leads to metabolic stress, which is thought to play a major role in the increased occurrence of the aforementioned production diseases. In addition, the presence of health problems has been related to the extent and duration of the energy deficit in early lactation.6 Of the production diseases, ketosis is one of the most common ailments. In milk, ketone bodies like beta-hydroxybutyric acid r 2011 American Chemical Society

(BHBA), acetone, and acetoacetate are accepted biomarkers for subclinical ketotic conditions.7,8 These metabolic biomarkers show elevated levels in the presence of acute ketosis. However, to our knowledge, currently no biomarkers are available that show long-term prognostic potential, which would be of great interest for an early treatment of susceptible animals and for the selection of metabolically stable animals for breeding purposes. Therefore, the detection of prognostic metabolic biomarkers was the aim of this study. Since a milk sample can be collected routinely and noninvasively, it is well suited for metabolic analyses of dairy cows.9 In this study, individual time series analyses of milk constituents were performed for 264 animals from a herd of Holstein-Friesian (HF) cows, reflecting the development of milk metabolites during early and mid lactation. For all animals a complete description of health issues occurring in that period was available. Nuclear magnetic resonance (NMR) spectroscopy is a highperformance tool for the analysis of metabolic disorders using body fluids and tissue extracts, giving a comprehensive overview of the most abundant organic compounds in a sample with a single measurement.10 For the present investigation targeted metabolic profiling, which enables the accurate quantification of Received: October 12, 2011 Published: November 21, 2011 1373

dx.doi.org/10.1021/pr201017n | J. Proteome Res. 2012, 11, 1373–1381

Journal of Proteome Research

ARTICLE

Table 1. Diagnosed Metabolic Disorders (First Lactation Period)

a

Listed is the health status diagnosed within the same week in which the corresponding milk sample was collected. This is also true for the penultimate column corresponding to lactation month six. b Listed is the health status diagnosed at any time point over the whole lactation period irrespectively of time points of milk collection.

biomarkers, was used. Results showed a high predictive potential for two compounds, namely, glycerophosphocholine (GPC) and phosphocholine (PC) that are well-known milk constituents. GPC and PC are part of a phospholipid pathway that is active in many body tissues, including mammary tissue.11 Although GPC and PC levels have been previously assessed using a variety of methods, including high pressure liquid chromatography and gas chromatography mass spectrometry,12 NMR,13 and enzymatic assays,14,15 their relation to production diseases such as ketosis is, to our knowledge, a novel feature. Results were validated on samples obtained from two other breeds, namely, Brown Swiss and Simmental Fleckvieh, indicating the general applicability of the findings.

’ EXPERIMENTAL SECTION Animals and Sample Collection

Milk samples from Holstein-Friesian cows were collected on the Karkendamm research farm in Bimoehlen, Germany. Preselected high yielding cows with an average energy corrected milk yield (ECM) of 32.8 ( 4.7 kg/day were investigated. The animals were offered a total mixed ration (TMR) ad libitum. To all first lactating cows, fixed amounts of concentrates were offered until day in milk (DIM) 180 (average 2.3 ( 0.5 kg). Composition of TMR varied over the observation period. Crude protein, bypass protein, and the NEL of TMR were kept relatively constant. The net energy content for lactation (NEL) of TMR ranged between 6.9 and 7.2 MJ/kg of dry matter. A detailed overview of the feeding practice is given in a previous publication.16 For the first lactation, evening milk samples were collected on a weekly basis during the first five weeks of lactation and an additional milk sample was collected in month six postpartum (mid lactation). This resulted in 1267 milk samples from 264 first lactating animals that were obtained between October 2008 and June 2010. Within the same time period, 232 milk samples from 57 animals in lactation two to seven were collected. At the research farm in Karkendamm, mostly first lactating cows are kept, resulting in a smaller number of available animals in higher lactations. For theses animals, four samples were collected during the first four weeks of lactation plus one additional sample in month six. Note that, due to various reasons (e.g., serious health issues), a complete time series could not be obtained for all cows. For the analysis of blood plasma metabolites, 88 plasma specimens from different cows were collected. Of these, for the investigation

of hydrophilic metabolites, 66 samples were obtained at the day of milk collection within the first five weeks of lactation. Lipophilic plasma metabolites were analyzed from the remaining 22 plasma samples collected at the day of milk collection in the second week of lactation. Plasma samples were taken from the Vena caudalis mediana and collected into Li-Heparin coated tubes. All milk samples were frozen and stored at 20 °C until analysis; plasma samples were frozen and stored at 80 °C. For each animal, all health issues diagnosed during their stay on the farm were recorded (Table 1). Milk yield was recorded on a daily basis. Milk composition was analyzed weekly based on samples from two consecutive milkings. Animals were held with permission of the Christian-Albrechts-University, Kiel, Germany, and were treated in conformance with commonly practiced ethical standards. Additionally, 39 Simmental Fleckvieh and 31 Brown Swiss cows from a previous study were evaluated; for additional information see Klein et al. (2010).9 Note that only animals for which health information was available were included. Sample Preparation and NMR Measurements

Sample preparation and measurements were performed according to established protocols.9,17 Briefly, milk and plasma samples were ultrafiltrated with a cutoff of 10 kDa to remove macromolecules such as proteins and lipids. For the measurement of both milk and hydrophilic plasma metabolites, 400 μL of ultrafiltrate were thoroughly mixed with 200 μL of 0.1 mol/L phosphate buffer, pH 7.4, and 50 μL of 29.02 mmol/L 3-trimethylsilyl-2,2,3,3-tetradeuteropropionate (TSP; Sigma-Aldrich, Taufkirchen, Germany) in deuterium oxide as internal standard. For plasma lipid measurements, plasma was extracted using the following protocol: 100 μL of plasma was added to 500 μL of 2:1 chloroform/methanol mixture and vortexed. Then 100 μL of 0.2 mol/L potassium chloride solution was added, and after vortexing the sample was centrifuged at ambient temperature at 5700g for 10 min. The lower phase was removed. Two washing steps were performed by adding chloroform/methanol mixture and repeating the above procedure. After each washing step, the lower phase was collected. The collected phases were pooled and dried under a stream of nitrogen. The residue was dissolved in 650 μL of deuterated chloroform with octamethylcyclotetrasiloxane (OMS, SigmaAldrich, Taufkirchen, Germany) as internal standard.18 NMR experiments were conducted on an Avance III 600 MHz spectrometer using a triple resonance (1H, 13C, 31P, 2H lock) cryogenic probe equipped with z-gradients (Bruker, Rheinstetten, Germany). 1374

dx.doi.org/10.1021/pr201017n |J. Proteome Res. 2012, 11, 1373–1381

Journal of Proteome Research

ARTICLE

Table 2. Milk Metabolite Concentration Ranges (First Lactation Period)a

Concentrations are given in μmol/L. Min and max refer to the 10% and 90% concentration quantiles, respectively. Abbreviations: BHBA, betahydroxybutyric acid; GP, galactose-1-phosphate; GPC, glycerophosphocholine; NAC, N-acetyl-carbohydrates. The color coding was individually determined for each metabolite with the highest and lowest observed concentrations of a given metabolite corresponding to dark red and dark blue, respectively. b Concentrations determined by 1D NMR. c Concentrations determined by 2D NMR.

a

For each sample, one-dimensional (1D) 1H and two-dimensional (2D) 1H 13C NMR spectra were acquired. All spectra were measured at 298 K, and each sample was allowed to equilibrate for 300 s in the magnet before measurement. The temperature unit was calibrated using a deuterated methanol sample. 1D 1H and 2D 1H 13C HSQC spectra of each sample were automatically collected using the Bruker automated acquisition suite ICONNMR. All spectra were acquired without spinning. 1D 1H NMR spectra were obtained using a 1D nuclear Overhauser enhancement spectroscopy (NOESY) pulse sequence with presaturation during relaxation and mixing time and additional spoil gradients for water suppression. Spectra were automatically Fourier transformed and phase corrected, applying a line broadening of 0.3 Hz and zero filling to 128k points. A flat baseline was obtained by

using the “baseopt” option of TopSpin2.1, which performs a correction of the first points of the FID. For the 2D 1H 13C heteronuclear single quantum coherence (HSQC) spectra, water suppression was achieved using presaturation during the relaxation delay. For initial assignment of metabolites, high-resolution 2D 1H 13C HSQC spectra were collected for milk and plasma. To validate these assignments based on long-range proton-carbon couplings, corresponding high-resolution 2D 1H 13C heteronuclear multiple bond correlation (HMBC) spectra were measured. 2D spectra were semiautomatically processed employing a 90° shifted squared sine-bell window function in both dimensions. For increased resolution in the indirect carbon dimension, the number of data points was doubled prior to Fourier transform using complex forward linear prediction. All 2D spectra were 1375

dx.doi.org/10.1021/pr201017n |J. Proteome Res. 2012, 11, 1373–1381

Journal of Proteome Research

ARTICLE

Figure 1. Histograms of normally and non-normally distributed metabolite concentrations [mmol/L], exemplarily shown for N-acetyl-carbohydrates (NAC, left figure) and GPC (right figure) in lactation month six of the first lactation. One and two Gaussian functions were fitted to the histograms, respectively, to illustrate the different distributions. The frequency values indicate the number of animals contributing to one bar of the shown histograms.

manually phase corrected, and a polynomial baseline correction was applied excluding the region around the water artifact. All 1D and 2D spectra were chemical shift referenced relative to the TSP and OMS, respectively, reference signals. Milk constituents were identified and quantified as described previously.9,17,19 The parameters used can be found in Supporting Information Table S1. Estimation of Breeding Values for Energy Balance and Fatto-Protein Ratio

For the animals of the Karkendamm herd, breeding values for energy balance (EB) and fat-to-protein ratio (FPR) were available. Breeding values are estimates of the additive genetic merit for a particular trait that an animal will pass on to its descendents. Across lactation, daily breeding values can be computed for each animal using random regression technique. As described above, a broad spectrum of traits is routinely recorded at the dairy research farm Karkendamm. Individual feed intake, milk yield, and live weight per day were recorded automatically for lactation day 11 180. EB was calculated as the difference between energy intake and estimated energy requirements for milk output and maintenance as a function of live weight. This was described in detail by Buttchereit et al. (2010).16 For all animals, weekly FPR measurements were available. Animals were discarded from the analysis if the number of observations per trait and cow across the entire period was less than four. Breeding values for EB and FPR, respectively, were estimated based on state of the art methodology using the software package ASReml 3.0.20 In short, the model accounts for the effects day of measurement (test day) and age of calving (21 to 25, 26, 27, 28, or 29 to 38 months). Furthermore, the general pattern of EB and FPR, respectively, was modeled across lactation. In addition, the model accounts for individual deviations from this profile, which represent individual genetic effects (breeding values) and permanent environmental effects. The specific model and technical details are given in the Supporting Information. It should be

noted that breeding values for FPR were multiplied by ( 1) as a low FPR is considered to be preferable. Statistical Tests and Mathematical Software

Distributions were tested for agreement with a normal distribution using the Lilliefors Kolmogorov Smirnov test.21 P-Values were calculated using a two-sided Student’s t test assuming unequal variance. Gaussian distributions were fitted using the mathematical software Origin (OriginLab, Northampton, MA). Odds ratios were calculated as described in the Supporting Information.

’ RESULTS The aim of this study was the identification of metabolic factors indicating an inheritable long-time metabolic stability. In a previous study,9 we could show that different animals of the same breed kept under identical conditions on the same farm coped quite differently with the metabolic stress of early lactation. Therefore, we suspected the presence of genetically different subgroups in the population analyzed here. It can be expected that within these subgroups metabolites connected directly or indirectly to the respective genes are affected and that for some metabolites in each subgroup a distinct distribution of concentrations is observed. In turn, this will lead to non-normally distributed data when considering the whole population, with a distinct two-peak distribution in case of two subgroups. To test for this, we chose milk samples from lactation month six, as the strong concentration fluctuations of milk constituents in the first lactation third have diminished which led to almost stable levels by then. Table 2 shows values for metabolites that could be quantified (concentration values above the individual lower limits of quantification) in at least 20% of samples. As can be seen, many compounds are subject to strong concentration changes during the first weeks of lactation. All quantified milk compounds were tested for agreement with a normal distribution. For most compounds, normal distributions were found. In contrast, for glycerophosphocholine (GPC), fumarate, and 1376

dx.doi.org/10.1021/pr201017n |J. Proteome Res. 2012, 11, 1373–1381

Journal of Proteome Research

Figure 2. Distribution of GPC concentrations [mmol/L] of ketotic and nonketotic cows in month six of the first lactation. The group of ketotic animals contained all animals that were diagnosed at least once with ketosis during the entire lactation period, while the nonketotic animals stayed healthy throughout lactation. Both curves were normalized to a total area of one and were smoothed using a Gaussian kernel density estimator.

oxaloacetate, two-peak distributions were found with significant deviations from normal distributions with p-values of 0.014, 0.0025, and 5.8  10 6 , respectively. In Figure 1, N-acetylcarbohydrate (NAC) and GPC concentrations are shown as examples for normally and non-normally distributed compounds. A Gaussian distribution and a sum of two Gaussian distributions were used to approximate the observed data, respectively. For each of the three non-normally distributed compounds, the population was divided into two subgroups with high respective low values. The threshold was calculated as the interpeak minimum of the fitted sum of Gaussians. To assess the validity of the mentioned group classification, milk samples from other time points were analyzed. Only for GPC the difference in concentration levels was conserved throughout the whole lactation, i.e. animals with high (respectively low) GPC values in mid lactation also had high (respectively low) GPC values in early and late lactation. This indicates that GPC is a robust measure for the underlying, metabolically differing, subgroups. Therefore, the subsequent analyses concentrated on the differences in GPC. For GPC, the calculated threshold was 0.87 mmol/L. Next, it was analyzed whether correlations between the GPC value and disease status were present. Animals were grouped for each disease into two groups: Group 1 contained all animals that were diagnosed at least once considering the whole lactation period with a certain disease, for example, ketosis, while group 2 contained all other animals. Results showed that the subgroup with ketosis had significantly lower GPC values compared to the nonketotic subgroup, with a p-value of 0.036. For the other metabolic diseases shown in Table 1, no significant correlations to GPC were found. The distributions of GPC for ketotic and

ARTICLE

Figure 3. GPC/PC ratios of ketotic and nonketotic cows at week three of the first lactation. The group of ketotic animals contained all animals that were diagnosed at least once with ketosis during the entire lactation period, while the nonketotic animals stayed healthy throughout lactation. Both curves were normalized to a total area of one and were smoothed using a Gaussian kernel density estimator.

nonketotic animals are shown as a density plot in Figure 2. The figure shows that the concentration distributions differ both in range and maximum position between ketotic and nonketotic animals. The “shoulders” left of the maximum of the nonketotic group and right of the maximum of the ketotic group are striking. The first shoulder might be caused by animals with an undiagnosed, subclinical ketosis. The second shoulder might then be due to healthy animals that were erroneously diagnosed with ketosis. It is obvious that the maxima of the ketotic and the nonketotic population correspond to the two peaks of the overall distribution seen in Figure 1. In the group of animals with above-threshold GPC levels, significantly lower phosphocholine (PC) concentrations were observed during the first four lactation weeks, with p-values of 0.013, 2.5  10 5, 9.8  10 3, and 2.6  10 4, for week one, two, three, and four, respectively. Due to the observed interdependence of GPC and PC, the ratio GPC/PC was calculated from the respective molar concentrations. For the GPC/PC ratio, density plots were calculated for ketotic and nonketotic animals according to the definitions given above. As an example, the density plots for lactation week three are shown in Figure 3. The distributions of GPC/PC ratios for ketotic and nonketotic animals are very similar concerning maximum and width, but differ in the right-hand tail for healthy animals. Obviously, animals with high GPC/PC ratios are less prone to ketosis. This effect can be seen in the first four weeks of lactation but vanishes later on. This is due to the fact that GPC and PC levels decrease almost exponentially during the first weeks of lactation, but as PC continues to drop, GPC levels start rising again around week five, approaching the starting concentration by mid lactation (Table 2). 1377

dx.doi.org/10.1021/pr201017n |J. Proteome Res. 2012, 11, 1373–1381

Journal of Proteome Research The observations made so far clearly show that it is possible to define a threshold for the GPC/PC rate above which healthy animals are safely detected. We manually selected a value of 2.5 for GPC/PC to identify healthy animals within the population. Using this threshold, the samples from the first four lactation weeks were classified. Most of the animals that were above threshold dropped below threshold at least once during the first four lactation weeks. To take this into account, the GPC/PC ratio was determined several times during the first four weeks of lactation, and all animals showing ratios above threshold at least once were selected. This resulted in 41 out of 260 animals (16%) with at least one milk sample above threshold. No animal in this subgroup suffered from ketosis at any time during the whole lactation period, compared to an incidence of 6.4% in the subgroup below threshold. This means that during the first few weeks of lactation a safe prognosis for animals that will not develop ketosis later on is possible based on the GPC/ PC ratio. Correspondingly, acetone and BHBA values were lower in the high GPC/PC group during the first four lactation weeks, indicating lower incidence of acute subclinical ketosis in this group. The difference in acetone was significant in weeks one, three, and four, with p-values of 4.6  10 2, 2.1  10 2, and 4.1  10 3, respectively. The difference in BHBA was significant in weeks three and four, with p-values of 4.3  10 2 and 1.0  10 3, respectively. A low ketosis risk in the first lactation is expected to be predictive for the health during the next lactations, as high correlations have been observed between ketosis incidences in different lactations.22 In our data set, disease data from the second lactation was only available for four of the 41 animals above GPC/PC threshold in the first lactation period. None of these four had developed ketosis during the next lactation. Although this number is very low, it supports the idea that the GPC/PC ratio has predictive abilities for following lactations. Milk parameters were compared between the groups above and below GPC/PC threshold. Milk yield, fat, and protein content of each animal were averaged over all measurements for weeks one to five and for month six. For milk yield and protein content no significant differences were observed between the groups at any time point. For fat content, the group above threshold showed lower values at all six time points. This difference was significant at all six time points with p-values of 1.4  10 3, 3.2  10 3 , 1.0  10 2, 7.4  10 4, 4.2  10 5 and 7.1  10 5, respectively. On average, the milk of animals above threshold contained 0.4% less fat than milk from animals below threshold. The concentration values are shown in Supporting Information Table S2. As an alternative for using the GPC/PC ratio, in cases where no values from the first weeks of lactation are available, GPC concentrations from mid lactation were evaluated for their ketosis predictive power. From Figure 2, it is obvious that at high GPC values only healthy individuals are present. As a threshold, we manually chose a value of equal or greater than 1.2 mmol/L. 9.3% of the samples lay above this threshold and no animal in this group had developed ketosis, compared to 7.5% incidence in the group below threshold. Only one animal above the GPC/PC threshold was available from the second lactation. This animal had not developed ketosis during the second lactation either, again hinting at the predictive power of the measurement. Further comparison of Figures 2 and 3 shows that by using the GPC/PC ratio substantially more animals can be detected than by GPC values alone.

ARTICLE

Correlations Between Milk Metabolites and Breeding Values for Energy Balance Values and Fat-to-Protein Ratio

Energy balance (EB) and milk fat-to-protein ratio (FPR) are genetically determined (with low and moderate heritability, respectively) and have been shown to be associated with liability to metabolic disorders, with high EB and low FPR values being favorable.23,24 Milk GPC, PC, and GPC/PC values determined in this study were analyzed for correlations with EB and FPR breeding values determined for the same animals and same lactation days. For EB breeding values, only GPC showed a significant positive correlation (p = 4.1  10 3). For FPR breeding values, all correlations were significant, with p-values of 1.8  10 3, 1.0  10 8, and 5.5  10 8 for PC, GPC, and GPC/PC, respectively. High GPC values and low PC values were connected to high FPR breeding values. Please note that high FPR breeding values indicate low fat-to-protein ratios due to the definition of FPR breeding value given above. FPR breeding values can serve as an indicator whether cows can or cannot adapt to the challenge of early lactation.23 This fits with our finding of GPC and PC values being indicators for susceptibility to ketosis. Plasma Analyses

Next, hydrophilic metabolites in 66 plasma specimens collected within the first five weeks of lactation were analyzed. No PC values could be obtained above the lower limit of quantification. For GPC, values of 5.3 ( 1.9 μmol/L (mean ( standard deviation, N = 23) were observed. However, no significant correlation between blood and milk GPC was observed. As GPC is formed in the breakdown of phosphatidylcholine (PtC), PtC concentrations were estimated for 22 plasma samples obtained in the second week of lactation. For this, the choline content of lipophilic plasma extracts was quantified.25 The observed concentrations ranged between 59.4 and 970 μmol/L. Obtained values were validated by spike-in experiments and extraction of aqueous PtC emulsions. Blood PtC was positively correlated to milk PC and negatively correlated to milk GPC and milk GPC/PC ratios with Spearman correlation coefficients of 0.35, 0.37, and 0.38, respectively. However, these correlations were not significant at the 5% level with p-values of 0.12, 0.10, and 0.092, respectively. Validation of the Results

Next, the results obtained so far were validated on additional groups of animals. As the threshold of 2.5 for the GPC/PC ratio was based only on samples from the Holstein-Friesian herd that were collected during the first lactation, milk samples from higher lactations (lactation two to seven) of the same herd were analyzed as well. Although this group contained only samples from 57 animals, a similar distribution with a tail of healthy animals at high GPC/PC ratios is observed (data not shown). However, one of the samples with a GPC/PC ratio above threshold came from a diseased animal, leading to 5.0% ketotic animals in the group above threshold, compared to 11.0% ketosis incidence in the group below threshold. An odds ratio of 2.38 was found, indicating a more than double ketosis risk in the group below the threshold. The acetone and BHBA values were generally lower in the high GPC/PC group, indicating a lower incidence of acute ketosis, although this difference was not significant. When looking at GPC concentrations at mid lactation, 5.3% of the animals were above the threshold of 1.2 mmol/L. None of these animals developed ketosis during the lactation, compared to 11% incidence in the group below threshold. These findings indicate that the suggested thresholds generally hold true even for higher lactations of the observed breed. 1378

dx.doi.org/10.1021/pr201017n |J. Proteome Res. 2012, 11, 1373–1381

Journal of Proteome Research In a second step, an animal collective presented in an earlier study was re-evaluated.9 For a herd of Brown Swiss cows, 22% of the animals were above the GPC/PC threshold during the first four weeks of lactation. For this herd only severe diseases had been recorded. One of the observed animals was downed (unable to stand) due to ketosis. This animal was in the group below threshold, consistent with the assumption of low ketosis risk in the group above threshold. The subgroup above threshold had significantly lower acetone and BHBA levels with p-values of 0.030 and 0.012, respectively. This indicates a lower incidence of acute subclinical ketosis in the subgroup above the threshold. In a Simmental Fleckvieh herd from the same study, 73% of the animals were above the GPC/PC threshold in the first weeks of lactation. This high number of animals above threshold fits with the finding that this breed has drastically lower acetone and BHBA values in early lactation compared to Brown Swiss and Holstein-Friesian cows, implying considerably lower ketosis susceptibility in this breed.9 Ketosis values were not available for this herd, but animals downed due to metabolic diseases were recorded. The number of animals downed at least once during this lactation was 20% for the subgroup below threshold and 14% in the group above threshold, corresponding to an odds ratio of 1.5. This means that high GPC/PC values are favorable even concerning overall metabolic health. In the Brown Swiss herd, the GPC values at mid and late lactation were above the threshold for 86% of the samples, and in the Simmental Fleckvieh herd 90% were above the GPC threshold in late lactation. These numbers are much higher than those obtained for the corresponding GPC/PC ratios. It seems that the absolute GPC value is varying greatly between different breeds, and that using the GPC/PC ratio corrects for the differences in absolute concentration levels.

’ DISCUSSION GPC is an osmolyte that is accumulated in renal medulla cells to protect the cells against high interstitial salt and urea concentrations present in the kidney.26 GPC is formed in the breakdown of phosphatidylcholine (PtC), the connected synthesis and catabolism pathway is shown in Figure 4.26,27 This pathway is active in many body tissues, including mammary tissue.11 The mammary gland has been shown to break down blood PtC to GPC and free fatty acids to gain fatty acids for the synthesis of milk triglycerides and phospholipids.28 Therefore, the measured milk GPC probably stems from mammary gland breakdown of blood PtC. This assumption is supported by the negative correlation of blood PtC to milk GPC values. As no correlation was found between milk GPC and plasma GPC, the GPC found in milk is probably not GPC absorbed from the bloodstream. In comparison to blood GPC values reported for humans,29 we observed considerably lower values in bovine plasma. The same is true for PtC, where the bovine plasma values of this study are distinctively lower than those previously published on other animals.27,30 As expected from literature values reported for human specimens,28 we were unable to observe bovine plasma PC values above the lower limit of quantification. Changes concerning GPC and PC levels have been reported in a wide range of research. Hypoxia, which is present in many cancers, is known to cause raised PC levels, whereas GPC levels are not affected.31 A change of the GPC/PC ratio has been noticed for cancer and immortalized cell lines of humans. There, in comparison to healthy cells, a decrease in the GPC/PC ratio

ARTICLE

Figure 4. GPC pathway. Shown are main parts of the GPC pathway. Abbreviations: CDP, cytidine diphosphate; Cho, choline; GPC, glycerophosphocholine; PC, phosphocholine; PtC, phosphatidylcholine; CK, choline kinase; CT, cytidylyltransferase; LPL, lysophospholipase; NTE, neuropathy target esterase, PCT, phosphocholine transferase; PLA, phospholipase A2 and A1; PLB, phospholipase B; PLC, phospholipase C; PLD, phospholipase D; PD, glycerophosphocholine phosphodiesterase.

has been observed.27,32,33 In tumors, PtC degradation via phospholipase C (PLC), respectively, phospholipase D (PLD) and subsequent choline kinase (CK) are the main causes of changed PC levels.34 Galons et al. observed an increase of GPC and a decrease in PC in mouse cell lines under a “slow acidosis”, that is, lowering the pH from 7.3 to 6.5 for around 12 h.35 They suggested as an explanation for this so-called “GPC to PC switch”, the activation of phospholipid breakdown as an alternative energy source for the cells, as glycolysis was hampered due to the low pH. The breakdown of PtC via phospholipase A2 (PLA) and lysophospholipase (LPL), respectively, phospholipase B (PLB) subsequently leads to GPC accumulation. Recently, it has been shown that PLB is identical to neuropathy target esterase (NTE).36 In renal medulla, the pathway via PLB/ NTE has been shown to be the dominant way of PtC breakdown for GPC production.26 Regarding the differences observed in milk between ketotic and nonketotic animals, these most closely resemble the changes observed in the “slow acidosis” experiment by Galons et al. However, we do not see any hints for acidic conditions in the investigated bovine plasma samples. The differences observed in this study between healthy animals and animals suffering from ketosis can be summarized as follows: Healthy animals have higher levels of milk GPC and lower levels of milk PC. The animals were divided into two groups according to their GPC/PC ratio. The subgroup with high GPC/PC ratios showed lower plasma PtC levels and lower milk fat levels. Based on these observations, we suggest a possible explanation for the positive effect of high GPC values. Raised GPC/PC ratios and GPC concentrations, respectively, are caused by a higher rate of blood PtC breakdown, possibly due to higher enzyme concentrations or activities of PLA and LPL, respectively, PLB/ NTE. These animals can, therefore, use more blood PtC as a fatty acid source for milk lipids. It has been previously shown that fatty acids derived from blood PtC may be utilized by the mammary gland for milk triglyceride synthesis.27 An alternative source for fatty acids for milk lipid synthesis are free blood fatty acids. Free blood fatty acids are produced by the mobilization of body fat, with excessive body fat mobilization 1379

dx.doi.org/10.1021/pr201017n |J. Proteome Res. 2012, 11, 1373–1381

Journal of Proteome Research being suspected to decrease the food intake of cows.37 In addition, body fat mobilization leads to the formation of ketone bodies such as acetone and BHBA in the liver. Using more PtC for milk lipid synthesis, and thus saving free blood fatty acids, may, therefore, lower the body fat mobilization. Animals with lower fat mobilization will suffer less from the negative side effects such as the decrease in food intake. This, in turn, has positive effects on the energy balance and thus prevents metabolic diseases. Additionally, in animals with high GPC/PC ratios a decreased milk fat content was observed. In our view, these two effects together, namely high blood PtC breakdown as evidenced by high GPC/PC ratios together with lower milk fat content, will lead to a decreased mobilization of body fat. Animals with lower PtC breakdown rates will develop a more negative energy balance and a higher body fat mobilization during the first weeks of lactation, leading to a higher risk of metabolic disorders. The assumption of a constantly higher PtC breakdown in the high GPC/PC group is backed by the fact that GPC is raised not only during the negative energy balance of the first lactation weeks, but throughout the whole lactation period. Selecting animals with high GPC/PC or GPC values should yield animals that can cope better with the negative energy balance of the first lactation weeks and, therefore, are less prone to ketosis. The significant correlations between FPR based breeding values and PC, GPC, and GPC/PC levels indicate that the observed metabolic differences are directly connected to genetic discrepancies and, therefore, a heritability of these effects can be expected.

’ CONCLUSIONS Taking all above results into account, the ratio of glycerophosphocholine to phosphocholine (GPC/PC) during the first four weeks of lactation and the concentration of GPC at mid lactation may serve as markers to identify cows that cope well with metabolic stress. These data suggest that the GPC/PC threshold for determining ketosis risk could be universally applicable, being independent of lactation number, breed, or feeding conditions. The chosen GPC threshold seems to be applicable only for Holstein-Friesian cows, and will need to be individually recalculated for other breeds. To the best of our knowledge, this is the first report of GPC and PC values indicating the ketosis risk in dairy cows. The measurement of GPC and PC is superior to using just thresholds for acetone and BHBA, as suggested for example by Enjalbert et al., as these values can only identify acute ketosis.8 In contrast, the method proposed here allows the early identification of metabolically stable animals that are not prone to ketosis for breeding purposes. ’ ASSOCIATED CONTENT

bS

Supporting Information Additional experimental details and data. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Telephone: +49-941-943-5057. Fax: +49-941-943-5020. E-mail: [email protected].

’ ACKNOWLEDGMENT The authors thank the staff at the Karkendamm research farm for providing milk and plasma specimens and NORD-OST

ARTICLE

GENETIC GmbH & Co. KG (Verden/Aller, Germany) for the kind permission to include their animals in this study. This study was funded by the German Federal Ministry of Education and Research in context of the Fugato-plus MeGA-M framework and by the Bavarian Genome Network (BayGene), Munich, Germany. The authors declare that they have nothing to disclose regarding funding from industry or conflict of interest with respect to this manuscript.

’ REFERENCES (1) Pryce, J. E.; Veerkamp, R. F.; Thompson, R.; Hill, W. G.; Simm, G. Genetic aspects of common health disorders and measures of fertility in Holstein Friesian dairy cattle. Anim. Sci. 1997, 65, 353–360. (2) Ingvartsen, K.; Dewhurst, R.; Friggens, N. On the relationship between lactational performance and health: is it yield or metabolic imbalance that cause production diseases in dairy cattle? A position paper. Livest. Prod. Sci. 2003, 83, 277–308. (3) Bishop, S. C., Axford, R. F. E., Nicholas, F. W., Owen, J. B., Eds. Breeding for Disease Resistance in Farm Animals; CABI: Wallingford, UK, 2009. (4) H€uttmann, H.; Stamer, E.; Junge, W.; Thaller, G.; Kalm, E. Analysis of feed intake and energy balance of high-yielding first lactating Holstein cows with fixed and random regression models. Animal 2009, 3, 181–188. (5) Banos, G.; Coffey, M.; Brotherstone, S. Modeling Daily Energy Balance of Dairy Cows in the First Three Lactations. J. Dairy Sci. 2005, 88, 2226–2237. (6) Collard, B.; Boettcher, P.; Dekkers, J.; Petitclerc, D.; Schaeffer, L. Relationships Between Energy Balance and Health Traits of Dairy Cattle in Early Lactation. J. Dairy Sci. 2000, 83, 2683–2690. (7) Geishauser, T.; Leslie, K.; Tenhag, J.; Bashiri, A. Evaluation of eight cow-side ketone tests in milk for detection of subclinical ketosis in dairy cows. J. Dairy Sci. 2000, 83, 296–299. (8) Enjalbert, F.; Nicot, M. C.; Bayourthe, C.; Moncoulon, R. Ketone bodies in milk and blood of dairy cows: relationship between concentrations and utilization for detection of subclinical ketosis. J. Dairy Sci. 2001, 84, 583–589. (9) Klein, M. S.; Almstetter, M. F.; Schlamberger, G.; N€urnberger, N.; Dettmer, K.; Oefner, P. J.; Meyer, H. H. D.; Wiedemann, S.; Gronwald, W. Nuclear magnetic resonance and mass spectrometrybased milk metabolomics in dairy cows during early and late lactation. J. Dairy Sci. 2010, 93, 1539–1550. (10) Nicholson, J. K.; Lindon, J. C. Systems biology: Metabonomics. Nature 2008, 455, 1054–1056. (11) Holmes-McNary, M. Q.; Cheng, W. L.; Mar, M. H.; Fussell, S.; Zeisel, S. H. Choline and choline esters in human and rat milk and in infant formulas. Am. J. Clin. Nutr. 1996, 64, 572–576. (12) Pomfret, E. A.; daCosta, K. A.; Schurman, L. L.; Zeisel, S. H. Measurement of choline and choline metabolite concentrations using high-pressure liquid chromatography and gas chromatography-mass spectrometry. Anal. Biochem. 1989, 180, 85–90. (13) Govindaraju, V.; Young, K.; Maudsley, A. A. Proton NMR chemical shifts and coupling constants for brain metabolites. NMR Biomed. 2000, 13, 129–153. (14) Chap, H. J.; Moatti, J. P.; Mieusset, R.; Nieto, M.; Laneelle, G.; Bennet, P. J.; Mansat, A.; Pontonnier, F.; Douste-Blazy, L. Simple, rapid enzymatic determination of glycerophosphocholine in human seminal plasma. Clin. Chem. 1988, 34, 106–109. (15) Murray, J. J.; Dinh, T. T.; Truett, A. P.; Kennerly, D. A. Isolation and enzymic assay of choline and phosphocholine present in cell extracts with picomole sensitivity. Biochem. J. 1990, 270, 63–68. (16) Buttchereit, N.; Stamer, E.; Junge, W.; Thaller, G. Evaluation of five lactation curve models fitted for fat:protein ratio of milk and daily energy balance. J. Dairy Sci. 2010, 93, 1702–1712. (17) Gronwald, W.; Klein, M. S.; Kaspar, H.; Fagerer, S. R.; N€urnberger, N.; Dettmer, K.; Bertsch, T.; Oefner, P. J. Urinary metabolite 1380

dx.doi.org/10.1021/pr201017n |J. Proteome Res. 2012, 11, 1373–1381

Journal of Proteome Research

ARTICLE

quantification employing 2D NMR spectroscopy. Anal. Chem. 2008, 80, 9288–9297. (18) Oostendorp, M.; Engelke, U. F.; Willemsen, M. A.; Wevers, R. A. Diagnosing Inborn Errors of Lipid Metabolism with Proton Nuclear Magnetic Resonance Spectroscopy. Clin. Chem. 2006, 52, 1395–1405. (19) Klein, M. S.; Dorn, C.; Saugspier, M.; Hellerbrand, C.; Oefner, P. J.; Gronwald, W. Discrimination of steatosis and NASH in mice using nuclear magnetic resonance spectroscopy. Metabolomics 2011, 7, 237–246. (20) Gilmour, A. R.; Gogel, B. J.; Cullis, B. R.; Thompson, R. ASReml User Guide, Release 3.0; VSN International: Hemel Hempstead, UK, 2009. (21) Lilliefors, H. W. On the Kolmogorov-Smirnov Test for Normality with Mean and Variance Unknown. J. A. S. A. 1967, 62, 399–402. (22) Erb, H.; Grohn, Y. Epidemiology of Metabolic Disorders in the Periparturient Dairy Cow. J. Dairy Sci. 1988, 71, 2557–2571. (23) Buttchereit, N.; Stamer, E.; Junge, W.; Thaller, G. Relationship of energy balance and fat protein ratio of milk to disease liability in dairy cattle. Proc. W.C.G.A.L.P. 2010, 9, 315. (24) Buttchereit, N.; Stamer, E.; Junge, W.; Thaller, G. Short communication: Genetic relationships among daily energy balance, feed intake, body condition score, and fat to protein ratio of milk in dairy cows. J. Dairy Sci. 2011, 94, 1586–1591. (25) Cheung, A. P.; Olson, L. L. 1H-NMR assay of phosphatidylcholine and phosphatidylethanolamine in AL721. J. Pharm. Biomed. 1990, 8, 729–734. (26) Gallazzini, M.; Burg, M. B. What’s New About Osmotic Regulation of Glycerophosphocholine. Physiology 2009, 24, 245–249. (27) Iorio, E.; Mezzanzanica, D.; Alberti, P.; Spadaro, F.; Ramoni, C.; D’Ascenzo, S.; Millimaggi, D.; Pavan, A.; Dolo, V.; Canevari, S.; Podo, F. Alterations of choline phospholipid metabolism in ovarian tumor progression. Cancer Res. 2005, 65, 9369–9376. (28) Easter, D. J.; Patton, S.; McCarthy, R. D. Metabolism of Phospholipid in Mammary Gland: I. The Supply of Phospholipid for Milk Synthesis in the Rat and Goat. Lipids 1970, 6, 844–849. (29) Ilcol, Y. O.; Ozbek, R.; Hamurtekin, E.; Ulus, I. H. Choline status in newborns, infants, children, breast-feeding women, breast-fed infants and human breast milk. J. Nutr. Biochem. 2005, 16, 489–499. (30) Kuksis, A.; Roberts, A.; Thompson, J.; Myher, J.; Geher, K. Plasma phosphatidylcholine/free cholesterol ratio as an indicator for atherosclerosis. Arteriosclerosis 1983, 4, 389–397. (31) Glunde, K.; Shah, T.; Winnard, P. T.; Raman, V.; Takagi, T.; Vesuna, F.; Artemov, D.; Bhujwalla, Z. M. Hypoxia regulates choline kinase expression through hypoxia-inducible factor-1 alpha signaling in a human prostate cancer model. Cancer Res. 2008, 68, 172–180. (32) Podo, F. Tumour phospholipid metabolism. NMR Biomed. 1999, 12, 413–439. (33) Aboagye, E. O.; Bhujwalla, Z. M. Malignant transformation alters membrane choline phospholipid metabolism of human mammary epithelial cells. Cancer Res. 1999, 59, 80–84. (34) Iorio, E.; Ricci, A.; Bagnoli, M.; Pisanu, M. E.; Castellano, G.; Vito, M. D.; Venturini, E.; Glunde, K.; Bhujwalla, Z. M.; Mezzanzanica, D.; Canevari, S.; Podo, F. Activation of phosphatidylcholine cycle enzymes in human epithelial ovarian cancer cells. Cancer Res. 2010, 70, 2126–2135. (35) Galons, J. P.; Job, C.; Gillies, R. J. Increase of GPC levels in cultured mammalian cells during acidosis. A 31P MR spectroscopy study using a continuous bioreactor system. Magn. Reson. Med. 1995, 33, 422–426. (36) Gallazzini, M.; Ferraris, J. D.; Kunin, M.; Morris, R. G.; Burg, M. B. Neuropathy target esterase catalyzes osmoprotective renal synthesis of glycerophosphocholine in response to high NaCl. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 11026–11031. (37) Ingvartsen, K.; Andersen, J. Integration of Metabolism and Intake Regulation: A Review Focusing on Periparturient Animals. J. Dairy Sci. 2000, 83, 1573–1597.

1381

dx.doi.org/10.1021/pr201017n |J. Proteome Res. 2012, 11, 1373–1381