Isotopic Analysis of Calcium in Blood Plasma and Bone from Mouse Samples by Multiple Collector-ICP-Mass Spectrometry

Transcription

1 ANALYTICAL SCIENCES NOVEMBER 2008, VOL The Japan Society for Analytical Chemistry Isotopic Analysis of Calcium in Blood Plasma and Bone from Mouse Samples by Multiple Collector-ICP-Mass Spectrometry Takafumi HIRATA,* Mina TANOSHIMA,* Akinobu SUGA,* Yu-ki TANAKA,* Yuichi NAGATA,* Atsuko SHINOHARA,** and Momoko CHIBA** * Laboratory for Planetary Sciences, Tokyo Institute of Technology, O-okayama, Meguro, Tokyo , Japan ** Department of Epidemiology and Environmental Health, Juntendo University, Hongo, Bunkyo, Tokyo , Japan The biological processing of Ca produces significant stable isotope fractionation. The level of isotopic fractionation can provide key information about the variation in dietary consumption or Ca metabolism. To investigate this, we measured the Ca/ 42 Ca and Ca/ 42 Ca ratios for bone and blood plasma samples collected from mice of various ages using multiple collector-icp-mass spectrometry (MC-ICP-MS). The Ca/ 42 Ca ratio in bones was significantly ( ) lower than the corresponding ratios in the diet, suggesting that Ca was isotopically fractionated during Ca metabolism for bone formation. The resulting Ca/ 42 Ca ratios for blood plasma showed almost identical, or slightly higher, values ( ) than found in a corresponding diet. This indicates that a significant amount of Ca in the blood plasma was from dietary sources. Unlike that discovered for Fe, there were no significant differences in the measured Ca/ 42 Ca ratios between female and male specimens (for either bone or blood plasma samples). Similarity, the Ca/ 42 Ca ratios suggests that there were no significant differences in Ca dietary consumption or Ca metabolism between female and male specimens. In contrast, the Ca/ 42 Ca ratios of blood plasma from mother mice during the lactation period were significantly higher than those for all other adult specimens. This suggests that Ca supplied to infants through lactation was isotopically lighter, and the preferential supply of isotopically lighter Ca resulted in isotopically heavier Ca in blood plasma of mother mice during the lactation period. The data obtained here clearly demonstrate that the Ca isotopic ratio has a potential to become a new tool for evaluating changes in dietary consumption, or Ca metabolism of animals. (Received June 30, 2008; Accepted September 30, 2008; Published November 10, 2008) Introduction Calcium is the most abundant element in living organisms. 1,2 The important biological role of Ca is based on its divalent ionic charge, its adaptable coordination geometry, fast reaction kinetics and modest binding energy in aqueous solutions. 3 Several physiological functions, such as cell permeability and neuronal activities, are critically dependent on the Ca concentration in blood plasma. To maintain these Ca related functions, the concentration of Ca in blood plasma is controlled by homeostatic regulation. The Ca concentration in blood plasma does not reflect the nutritional status of Ca, as is for example the case for Fe, and therefore the adequacy of dietary intake cannot be ascertained via the concentration of inorganic nutrients in food alone. Both the excessive intake and deficiency of inorganic nutrients can cause a declination of the biochemical To whom correspondence should be addressed. hrt1@geo.titech.ac.jp Mina Tanoshima present address: Agilent Technologies International Japan, Ltd., Hachioji site 9-1, Takakura-cho, Hachioji, Tokyo , Japan. Akinobu Suga present address: Nikon Corporation, Development Management Dept., 6-3 Nishi-ohi, Shinagawa, Tokyo , Japan. reactions or a serious disease, and therefore diagnosis of the nutritional status of Ca as well as dietary consumption of Ca are still key issues. The majority of inorganic nutrients are incompletely absorbed, and dietary consumption should also be a function of various parameters, such as bioavailability, turnover rate or body pool size. 4 Hence, the adequacy of dietary intake of inorganic nutrients cannot be ascertained only from concentrations in tissue and biological fluids of the human body. To evaluate the nutritional status of Ca, isotopic tracer experiments have been widely applied. The basic principle of an isotope tracer experiment is to label food with radioisotopes like 45 Ca, 55 Fe or 65 Zn, or adequate spikes enriched in isotopes such as Ca, 58 Fe or 68 Zn. 1,5 7 In the case of tracer experiments, using radioactive isotopes combined with external monitoring with a whole-body gamma-ray counter is now the reference method for measuring the retention time. 6 However, radioisotopes constitute a potential hazard because of internal radioactive exposure. Stable isotope tracers are now increasingly being used in studies of elemental metabolism, bioavailability or toxicity of nutrients, as well as evaluating the elemental turnover time. 7 9 Besides an absence of harmful radiation, this approach has the further advantage of enabling multi-element studies, in which different isotopes can be added to the same meal. 10,11 However, it should be noted that spiking with large amounts of enriched-isotopes will result in a change of the dietary net consumption of inorganic nutrients. Moreover,

2 1502 ANALYTICAL SCIENCES NOVEMBER 2008, VOL. 24 both of these two isotope tracer approaches are based on the assumption that the spiked isotope replaces the native element, and behaves in an identical manner. 1 This indicates that mimicing of the isotopes to the native nutrients must be considered in order to obtain reliable data for elemental metabolism. Faced with this problem, we have developed a new isotope tracer technique based on changes in the isotopic composition of Ca due to a mass-dependent isotopic effect. Hence, no spiking with additional elements, such as radio isotopes or enriched isotopes, was required, and therefore, no mimicing to the native nutrients is required. Moreover, it should be noted that the present stable isotope approach allows regular monitoring of the nutritional status of Ca of the donor without any spiking or addition of isotopes. The measurement of stable isotopes of nitrogen ( 15 N/ 14 N) and carbon ( 13 C/ 12 C) is a useful tool to trace trophic relationships and food webs. 3,12,13 This method assumes that, during the assimilation process, there is an enrichment in heavy isotopes (i.e., 13 C or 15 N) compared to the dietary intake, which is compensated by the preferential excretion of lighter isotopes. 14,15 The isotopic composition of an animals tissue reflects the prey, 16 and different authors have suggested a mean trophic enrichment in stable isotopic ratios of carbon (d 13 C) and nitrogen (d 15 N) of about 0 1 and 3 4, respectively. 14,17,18 Stable isotope abundances of heavy elements are highly constant because of small isotopic effects, mainly due to a smaller relative mass difference. However, recent progress in analytical techniques, including double spike-thermal ionization mass spectrometry (TIMS) or multiple collector- ICP-MS (MC-ICP-MS), enabled us to detect small changes in the isotopic composition of the elements, and a series of pioneering studies revealed that the natural stable isotope fractionations of many elements heavier than S (e.g., Ca, Fe, Cu, Zn) are common on Earth. Mass-dependent fractionation of Ca has been used to understand the formation of marine carbonates, corals or forams, and these isotopic data were employed as a new proxy for marine environments Biological processing including Ca metabolism can cause significant stable isotope fractionation of Ca. 21,22 The ability to quantify the daily consumption using the isotope fractionation of Ca has been demonstrated. 32 The stable isotope study using Ca is now believed to provide various information about the conditions of chemical reactions or the formation sequence of samples. In contrast to the TIMS technique requiring a Ca- 48 Ca double spike, isotope ratio measurements using the MC-ICP-MS technique can be carried out by standard-sample-standard bracketing. Moreover, much higher sample throughput can be achieved by the MC-ICP-MS technique. The main disadvantage of MC-ICP-MS with respect to Ca isotope measurements is the inability to monitor 40 Ca. However, by reducing the plasma energy, the isobaric interference of 40 Ar + on 40 Ca + can be significantly reduced, enabling a direct measurement of the Ca/ 40 Ca isotope ratio. 28 Despite the obvious success in the direct measurement of 40 Ca, we did not monitor 40 Ca throughout this study, since the level of non-spectrometric interferences could become more serious, when adopting the cool plasma technique. Stable isotope investigations of Ca are believed to provide information about the conditions of the chemical reactions or the formation sequence of biologic and geologic materials. 19,21,22,27 In this study, we measured the Ca/ 42 Ca and Ca/ 42 Ca isotopic ratios for bone and blood plasma samples collected from mice of various ages, aiming at the detection of possible changes in the dietary consumption or Ca metabolism of mice. Table 1 Experimental ICP-MS instrumentation and parameters 1) MC-ICPMS Instrument Nu Instruments Nu Plasma 500 RF power 1.35 kw Forward, <3 W Ref. Argon gas flow rates Cooling 14 L min 1 Auxiliary 0.7 L min 1 Nebuliser 0.90 L min 1 Ion energy 4000 V Extraction 2400 V Analysis mode Static Collector Analogue by Faraday Typical transmission 20 V/mg g 1 Monitored isotopes Analyte signals 42 Ca +, Ca +, Ca + (positively charged ions) Sr 2+ correction 84 Sr 2+ (42 u), 86 Sr 2+ ( u), Sr 2+ (.5 u), 88 Sr 2+ ( u) Baseline correction On-peak Baseline Subtraction 2) Desolvating nebuliser Instrument Cetac Aridus Nebuliser Glass Expansion Micrimist 100 Uptake-rate 180 ml min 1 (not pumped) Spray chamber temp. 70 C Desolvator temp. 160 C Sweep gas (Ar) 4 L min 1 3) Sample solution Sample digestion Microwave (Procedure EM-45/A) Analyte solution HNO wt% Instrumentation The multiple collector-icp-mass spectrometer (MC-ICP-MS) used in this study was a Nu Plasma 500 (Nu Instruments, Wrexham, UK). A multiple collector array of Faraday cups allows for the simultaneous detection of 42 Ca, Ca and Ca signals, providing better reproducibility of isotopic ratio measurements. The signal intensities for 40 Ca, 46 Ca and 48 Ca were not monitored because of the relatively large contribution of mass spectrometric interferences by 40 Ar +, 46 Ti + and 48 Ti +. The mass resolution used in this study was M/DM = 400, which was much lower than the mass resolution required for the separation of the mentioned isobaric interferences (M/DM >20000). In the case of Ca isotopic analysis using ICP-MS, the contribution of polyatomic interferences, such as 40 ArH + 2 (50 mv) or 12 CO + 2 (4 mv), can cause serious systematic errors. To minimise the contribution of these polyatomic interferences, a desolvating nebuliser (Aridus, Cetac, Omaha, USA) was employed for sample introduction. With the desolvating nebuliser system, the signal intensities of polyatomic interferences on 42 amu ( 40 ArH 2+ ) could be reduced down to ~1 mv level. Moreover, residual polyatomic interferences were further corrected by on-peak baseline subtraction. The variation in the residual background on 42 amu was < 0.1 mv, and therefore, the resulting contribution onto the Ca signal was much smaller than the <0.2 level. The term measurement used in this publication means 5 cycles of 40 ratio measurements each, using an integration time of 5 s per measurement. The analysis sequence begins with an analysis of the baseline intensity at m/z 42,,.5 and. The resulting baseline values were used for subsequent on-peak baseline corrections of the measured intensities for the sample solutions. The concentration of Ca in the sample solutions was adjusted to

3 ANALYTICAL SCIENCES NOVEMBER 2008, VOL Table 2 Measured Ca/ 42 Ca and Ca/ 42 Ca ratios for NIST SRM 915a Ca/ 42 Ca a Ca/ 42 Ca a Day 1 Run ± ± Run ± ± Run ± ± Run ± ± Run ± ± Run ± ± Mean ± ± %2SD Day 2 Run ± ± Run ± ± Run ± ± Run ± ± Run ± ± Run ± ± Mean ± ± %2SD Day 3 Run ± ± Run ± ± Run ± ± Run ± ± Run ± ± Run ± ± Mean ± ± %2SD Errors were 2-sigma standard deviation. a. No correction for the mass discrimination effect was made mg/g in 0.1 wt% HNO 3, and the resulting typical signal intensity of Ca was ~5 V (equivalent to ~300 Mcps). Details of the instruments and the operational settings are summarised in Table 1. Standardisation The mass discrimination of Ca was typically 5% per mass unit (5%/amu), which was consistent with estimations based on the mass dependency of the mass discrimination effect. 33,34 In this study, no correction for mass discrimination was made on the measured Ca isotope ratios. The reproducibility and repeatability of the Ca/ 42 Ca and Ca/ 42 Ca ratio measurements achieved in this study are given in Table 2. Typical repeatability estimated based on 6 repeated isotopic analyses, were about for both the Ca/ 42 Ca and Ca/ 42 Ca ratios, and the reproducibility, defined as the variation in the measured Ca/ 42 Ca and Ca/ 42 Ca ratios among the different analysis dates was 3 4 for both the Ca/ 42 Ca and Ca/ 42 Ca ratios. The significantly poorer reproducibility obtained in this study was mainly due to different instrumental settings, such as the ion sampling depth, ion lens setting, nebuliser gas flow or sweep gas flow rates. Nevertheless the poorer reproducibility in the Ca/ 42 Ca and Ca/ 42 Ca ratio measurements as well as small changes (<1 ) in the Ca/ 42 Ca and Ca/ 42 Ca ratios can be detected by the standard-sample-standard bracketing technique. 35,36 The relative difference in the isotopic ratios between the sample and standard is expressed as the d-value, defined by I m Ca d m 42 Ca J sample Ca = g 1h 1000( ), (1) I m Ca 42 Ca J standard Fig. 1 Effect of mass spectrometric interference by 86 Sr 2+ and 88 Sr 2+ ion signals onto the (a) Ca/ 42 Ca and (b) Ca/ 42 Ca ratios. Measured Ca/ 42 Ca and Ca/ 42 Ca ratios obtained with Sr free Ca solution are shown by the gray line. Open circles represent the Ca isotope ratios without the m Sr 2+ correction, and the closed-squares represent the Ca isotope ratios with the m Sr 2+ correction (see text). hence, m refers to the mass number of the respective Ca isotopes (m =, ). In this study both the NIST SRM 915a Ca standard and the food for mice were employed as a Ca isotopic standard to calculate the d-value. Correction of Sr interferences In order to obtain precise and reliable isotopic data, Ca must be carefully separated from other matrix components and trace elements (e.g., Na, Mg, K or Sr). 35 Especially, great care must be taken of potential mass-spectrometric interferences on Ca isotopes resulting from doubly-charged Sr ions (i.e., 84 Sr 2+, 86 Sr 2+ and 88 Sr 2+ ). 36 The production rate of doubly-charged ions of Sr (Sr 2+ /Sr + ) under the hot plasma conditions (RF power of 1350 W) was 15 20%, which is consistent with the reported value achieved by a similar instrumental configuration and the operational settings. 35,36 Figure 1 illustrates the contribution of Sr 2+ onto the Ca isotopes, indicating that the presence of relatively small quantities of Sr (Sr/Ca = or 0.1%) can cause a considerable level of analytical error (5 ) in the measured Ca/ 42 Ca and Ca/ 42 Ca ratios. Ca/ 42 Ca initial and Ca/ 42 Ca initial, represented by gray lines Figs. 1a and b, denote the Ca/ 42 Ca and Ca/ 42 Ca ratios obtained from Ca solutions free of Sr. To reduce the contribution of the mass-spectrometric interference resulting from Sr 2+ ions, the signal intensities of 84 Sr 2+, 86 Sr 2+ and 88 Sr 2+ ions were estimated by monitoring the signal intensity of Sr 2+ at m/z.5. For a mathematical correction of the Sr 2+ interferences, the isotopic ratios of 84 Sr/ Sr, 86 Sr/ Sr and 88 Sr/ Sr must be defined. However, this is very difficult because the relative abundance of Sr can vary as a result of the radioactive decay of Rb to Sr. In fact, the Sr/ 88 Sr ratio for sediments and carbonates varies significantly, ranging from to ,38 Faced with this problem, we assumed an Sr/ 86 Sr ratio of in this study, which is a rough average of sediment rocks and modern marine carbonates. Based on this empirical assumption, the signal intensities of m Sr 2+ (m = 84, 86 and 88) can be estimated based on the mass bias factor of the Ca/ 42 Ca ratio using the following exponential law:

4 1504 ANALYTICAL SCIENCES NOVEMBER 2008, VOL. 24 I Ca I m Sr Sr J = I m Sr meas Sr J 42 Ca J meas g true I Ca 42 Ca J true h ln(mm/m) ln(m/m42) In Eq. (2) M m represents the mass of the respective isotopes, and a Ca/ 42 Ca true ratio of was empirically determined for an estimation of the mass bias factor based on Ca isotopes. 39 It should be noted that the mass bias factor for the Ca/ 42 Ca ratio could not be assessed via a Ca solution prepared from a corresponding single-element stock. Single-element Ca solutions usually contain significant amounts of Sr impurities, resulting in Sr 2+ interferences on the measured Ca isotopes. To overcome this, we measured the Ca/ 42 Ca true ratio using a purified, Sr-free Ca solution. The resulting Ca/ 42 Ca ratio was used to estimate the mass bias factor on Ca, and subsequently the signal intensities of m Sr 2+ (m = 84, 86 and 88) were estimated based on the 84 Sr/ Sr, 86 Sr/ Sr and 88 Sr/ Sr ratios and the recorded signal intensity of Sr 2+ at m/z.5 u. This correction is based on the empirical assumption that the mass bias factor on Sr 2+ was identical to that of the Ca isotopes. It is widely accepted that the mass-discrimination effect found on the MC- ICP-MS is a function of the mass of the analytes, 33,34 and therefore, the level of the mass-discrimination effect on Ca should be systematically more serious than that observed on the Sr isotopes. To correct the mass-dependency of the mass discrimination effect, we have defined the modified exponential law for the correction of the mass-discrimination effect. 40,41 Despite the obvious success in obtaining further reliable or consistent isotopic data of the analytes, we could not apply this technique in the present Ca isotope study. Since the Ca isotope ratios for samples can vary due to the mass-dependent isotopic fractionation, the mass bias factor can not be determined only an isotopic ratio measurement of Ca in a sample. Moreover, a direct determination of the mass bias factor for Sr (Sr 2+ ) was also difficult, because of the large contribution of mass spectrometric interferences by Ca onto the doubly-charged Sr signals ( 84 Sr 2+, 86 Sr 2+, 88 Sr 2+ ). These data lead us to conclude that the standard-sample-standard bracketing technique is the most effective approach to detect small changes in the Ca/ 42 Ca and Ca/ 42 Ca ratios. Under the assumption that the mass bias factor on Sr 2+ was identical to that of the Ca isotopes, the signal intensities of the 42 Ca, Ca and Ca isotopes were corrected for the contribution of m Sr 2+ signals based on Eq. (2). The corrected Ca/ 42 Ca and Ca/ 42 Ca ratios are plotted in Fig. 1 as closedsquares. With the correction, the relative deviation from the initial ratios ( Ca/ 42 Ca init and Ca/ 42 Ca init) could be remarkably minimised. It should be noted that a significant level of deviation from the initial ratio at high Sr/Ca ratios (>0.001) still remained. This residual difference could result from the assumed Sr/ 86 Sr ratio (0.708) employed here, or from the difference in the mass bias factor for Ca, which was separately estimated during an individual analytical session. For the spiked-sr (Kanto Chemicals, AA-standard solution), we separately measured the Sr/ 86 Sr ratio ( ± ). 40 Even with this value, there still remained a systematical deviation in the corrected Ca/ 42 Ca and Ca/ 42 Ca ratios from the initial ratios. Moreover, for a solution with an Sr/Ca ratio of 0.001, the changes in the corrected Ca/ 42 Ca and Ca/ 42 Ca ratios did not vary measurably, even with very high Sr/ 86 Sr ratios ( Sr/ 86 Sr = 0.71). These data clearly demonstrate that great care must be given to the difference in the massdiscrimination effect between the Ca isotopes and doubly charged Sr isotopes (Sr 2+ ) to obtain further reliable Ca isotopic data from an analytical solution containing a high level of Sr (2) Table 3 Resulting d Ca NIST915a data for bone and plasma Sample Bone a Plasma a 1. Adult Male ± ± ± ± 0.37 Female ± ± ± ± 0.08 Lactating ± ± Not available 1.13 ± ± 0.12 b 2. Infant 1 week ± ± ± 0.15 b ± ± ± 0.20 b 2 weeks ± ± ± ± ± 0.02 b 0.42 ± 0.14 b 3. Feed c ± ± weeks ± ± 0.10 a. Errors are 2SD calculated based on 5 times repeated analysis. b. Mixture of two mouse samples (see text). c. Repeated twice to test the sample heterogeneity. (Sr/Ca >0.001). Nevertheless, after a correction for Sr 2+ signals, even with the presence of 3 Sr in the Ca solution, the level of the systematical deviation in the measured Ca/ 42 Ca and Ca/ 42 Ca ratios can be almost at an identical level to the measurement precision/reproducibility achieved in this study. Moreover, after the chemical separation procedure described below, the resulting Sr/Ca ratio for the analytical solution was much lower than 10 3, and thus the contribution of Sr would be negligibly small. Samples A series of bone and blood plasma samples were collected from 16 mice of different ages (two adult males, two adult females and nine infants in the first three weeks of life; see Table 3). Among the seven infant mice, two samples collected from them at the age of two weeks were born from the female mice analysed in this study. The absolute amount of Ca collected from individual infants plasma was sometimes smaller than that required for Ca isotopic analysis ( mg). To overcome this limitation, samples from 2 individual infant mice at the same age were pooled. This procedure was followed whenever deemed necessary, except for a specimen at the age of two weeks, stemming from the female specimen analysed. The following samples were pooled as described: (1) 1 week-old samples, No. 1 and No. 2; (2) 1 week-old samples, No. 3 and No. 4; (3) 2-weeks old samples, No. 2 and No. 3; and (4) 3- weeks old samples, No. 1 and No. 2. They were separately pooled into a single solution. After mixing the infants plasma, five sets of Ca isotopic data were obtained. The typical weight of bone material was mg for adults and 5 9 mg for infants. For infant bone samples, the absolute amount of Ca collected from individual thigh bones was large enough for Ca isotopic analysis. In this study, food material, which had been provided for mice from their birth on, were subjected to Ca isotope analysis in order to establish reference isotope ratios of Ca. Sample preparation All anatomy and sample digestion procedures were carried out

5 ANALYTICAL SCIENCES NOVEMBER 2008, VOL. 24 at a biologically controlled-area, and performed based on a guideline for live samples regulated by Juntendo University. Any used apparatus and unused organic samples were discarded after the complete sterilisation procedure. Collected whole blood samples were immediately heparinised, and the resulting sample was then centrifuged at 2500 rpm for 10 min to separate blood plasma and red blood cells. The resulting supernatant was recovered as a blood plasma sample. The typical volume of the blood plasma samples collected from adult mice was >1 ml, and the absolute amount of Ca (>100 mg) was large enough for Ca isotopic analysis. The second sample type analysed in this study were bone samples. Thigh bones were collected from the mouse specimens through anatomical dissection. Collected bone samples were mechanically cleaned using laboratory paper towels in order to remove muscle or soft tissues. Bone marrow was removed by injecting ionised water. Collected bone and blood plasma samples were frozen and stored at 5 C. Food materials were powdered using an agate mortar. Great care had to be taken to avoid contamination during the samplepreparation procedure. In fact, glass-made syringes usually used for blood extraction lead to Ca contamination in the mg range. To minimise Ca contamination, especially from the tools required for the anatomical dissection or sample extraction procedures, polybuthylene bottles, polypropylene tweezers, and polypropylene syringes with stainless-steel needles were used throughout this study. Blood plasma, bone and food samples were decomposed by conventional microwave digestion using acid. A detailed description of the procedure is available in the literature; 42 and only a brief outline of the procedure is described here. About 1 ml of a plasma sample, or 10 mg of bone or feed samples were mixed with 3 ml of conc. HNO 3 in a PFA vial, and thermally treated at 80 C for moderate oxidation of the organic components. This was followed by further oxidation using a mixture of 0.4 ml conc. H 2O 2 and 0.4 ml conc. HClO 4. The resulting blend was finally digested using microwave-based digestion. The dissolved sample solution was evaporated to dryness at 180 C on a hot plate. All reagents used in this study were of analytical or higher grade. Nitric acid, hydrogen peroxide and perchloric acid were TAMA Pure AA-100. Hydrochloric acid was electric grade (EL-grade) from Mitsubishi Chemical Co. Ltd. Deionised water was prepared using a Milli-Q Element (Millpore Corporation, Massachusetts, USA) Chemical separation To minimise both the contribution of the mass spectrometric interferences from Sr ions and non-mass spectrometric interferences by other co-existing elements, Ca was separated using cation-exchange chromatography. 35 The resulting sample cake from evaporation to dryness was re-dissolved in 1.5 M HCl, and the resulting solution was loaded on a column (Polyprep column, Bio-Rad Laboratory, California, USA) containing 1 ml of AG-50W-X12 cation-exchange resin (Bio- Rad Laboratory). After removing of co-existing elements (e.g., Na, Mg or K) by applying 20 ml of 1.5 M HCl to the column, Ca was collected from the fraction between 20 and 40 ml 1.5 M HCl. Strontium did not elute before 40 ml of 1.5 M HCl had passed through the column. The separation was quantitative, because the amount of Sr in the Ca fraction relative to the amount Ca was less than 1. The total recovery of Ca was >99%, suggesting a very small level of isotopic fractionation through chromatographic separation. The Ca fraction was evaporated to dryness on a hot-plate, and the resulting sample cake was dissolved in a 0.1 wt% HNO 3 solution. The total procedural blank for Ca through the samplepreparation procedure was 0.5 mg. Since the typical amount of Ca collected from one mouse sample was ~2000 mg, the contribution of blank Ca was neglected (<0.1%), and thus no correction for the procedural blank Ca was made in this study. Results and Discussion Reproducibility of the measurement The reproducibility of the Ca/ 42 Ca and Ca/ 42 Ca ratio measurements were evaluated by repeated analysis of a purified Ca standard solution (standard solution for atomic-absorption spectroscopy, Kanto Chemicals, Tokyo, Japan) over a fourmonth period. The resulting d Ca and d Ca values for Kanto Chemicals were ± 0.06 (2SD, n = 6) and ± 0.24 (2SD, n = 6), respectively, demonstrating that the analytical precision achieved here is high enough to detect the isotopic fractionation of Ca among organisms. 21,22 The resulting d Ca and d Ca values for Kanto Chemicals showed systematically higher values than those for the NIST SRM 915a standard reference material. This suggests that Ca in high-purity chemical reagents suffered isotopic fractionation through the purification procedure. Food, bone and plasma The resulting d Ca and d Ca data for bone and plasma samples collected from the mice are listed in Table 3. It should be noted that the resulting precision in the d Ca and d Ca values varied significantly. This is mainly due to the changes in the mass bias factor for Ca. Possible sources of the small changes in the mass bias factors on Ca were changes in the ionisation or ion-extraction conditions, including the plasma temperature, electron density or plasma potential. This was especially serious when the desolvating nebuliser (Aridus) was employed for sample introduction. Among the possible causes of changes in the mass bias factor, we believe that the solution uptake-rate can be a major source of analytical error in the Ca isotope ratio meausurements. The basic problem is because precise monitoring of the sample uptake-rate was very difficult. To obtain further precise isotopic data on Ca, great care must be taken concerning any changes in the nebulisation conditions, especially the long-term stability of the solution uptake rate. Nevertheless, the resulting precision or repeatability in the Ca isotopic analysis is high enough to detect the natural variation of the Ca isotope ratios. In order to define the isotopic ratio of Ca in dietary intake, d Ca and d Ca data for food were measured as well (n = 2). In this study, resulting d Ca and d Ca data for the diet were taken as typical reference values for the nutritional intake. To discuss the difference in the Ca/ 42 Ca ratio between the diet and the mouse samples, the d Ca values for all mouse samples were recalculated based on the Ca/ 42 Ca ratio of the diet. The difference in the d Ca values between results based on the NIST SRM 915a (d Ca NIST915a) and the diet (d Ca Diet) was 0.2. The resulting d Ca Diet values for all samples analysed in this study are plotted in Fig. 2. In the case of blood plasma collected from adults mice (n = 4: two females and two males), there were no significant differences to the d Ca Diet case (Fig. 2(a)). This suggests that for these samples a significant portion of Ca was from dietary sources. In contrast, for the bone samples (n = 4: two females and two males), the resulting d Ca Diet values were about 0.5 (0.25 /amu) lighter compared to the diet (P value <4.1%), suggesting that Ca in bones was isotopically fractionated through fractional absorption. Since the amount of dietary

6 1506 ANALYTICAL SCIENCES NOVEMBER 2008, VOL. 24 Fig. 2 Resulting d Ca data for blood plasma and bone samples collected from a series of mice of various ages: (a) adults, (b) infants and (c) adult mice in the lactating period. Note that the Ca/ 42 Ca ratios were normalised by that for dietary food. intake of Ca (~0.1 mg) was significantly smaller than the total amount of Ca in whole body (200 mg) of the analysed mice, the Ca isotopic signature potentially reflects the long-term Ca metabolism of mouse. Lighter Ca isotopes were preferentially absorbed by the mouse body from the food, which is consistent with results of a study performed by Skulan and De Paolo. 21 They measured the Ca/ 40 Ca ratios for bones collected from different animals (horse, chicken, seal and grouper) of various trophic levels, and demonstrated that the Ca/ 40 Ca ratios for bones were 1.3 (0.33 /amu) lighter than the corresponding ratios of dietary food. The level of isotopic fractionation for Ca found in this study (0.25 /amu) was slightly smaller than the value (0.33 /amu) previously reported by Skulan and De Paolo (1999). This difference may result from differences in Ca dietary consumption from food available to the animals. In the case of Fe in human red blood cell samples, it is recognised that there are significant systematical differences in the Fe isotope signatures (i.e., 56 Fe/ 54 Fe and 57 Fe/ 54 Fe ratios) between female and male specimens. 27, This sexual distinction is likely to be a result of the different degrees of Fe absorption. Because of menstrual loss of blood, the absorption efficiency of Fe for female mammal specimens (13%) is significantly higher than that for male mammal specimens (6%). 46 Lower intake efficiency results in an overall larger isotopic fractionation. In contrast to Fe, we could not find any significant differences the in measured Ca isotope ratios between female and male specimens. A lack of sexual distinction suggests that there were no significant differences in net dietary consumption of Ca between female and male. For infant samples, we measured the Ca isotope ratios for bone and blood plasma samples collected from nine mice of various ages (1 3 weeks old). The resulting d Ca NIST915a for infants bone and plasma samples are listed in Table 2, and the calculated d Ca Diet data are plotted in Fig. 2(b). For plasma samples, there are no significant differences in the measured d Ca Diet for plasma between adults and infants (0 0.5 ) (Fig. 2(a)). Moreover, the measured Ca/ 42 Ca ratio for adults and infants did not vary measurably from the Ca/ 42 Ca ratio of the diet. This suggests that a significant portion of the Ca intake was from dietary sources for both the adults and infants. For bone samples, the resulting d Ca Diet data for 1 to 3-week old mice (Fig. 2(b)) were significantly lower than the values for blood plasma (P value <1.8%). The d Ca Diet values of blood plasma and bone samples for most infants showed good agreement with the data for adults (Fig. 2(a)). This suggests that, for the mice of 1-week old or elder, there are no significant differences in the dietary consumption of Ca or Ca metabolism. The Ca isotopic ratios for blood plasma and bone samples collected from lactating mice were also measured (Fig. 2(c)). For bone samples, the resulting d Ca Diet value did not vary measurably from the data for the bones of other adult and infant mice. As pointed out earlier, the resulting d Ca Diet for bone could reflect the long-term Ca metabolism of the mouse, since the amount of dietary intake of Ca (~0.1 mg) was significantly smaller than the total amount of Ca in the whole mouse body (200 mg). This suggests that there would be no variation in the measured Ca/ 42 Ca ratios of bones collected from mother mice during the lactation period. However, this was not the case in the blood plasma samples. The d Ca Diet values for blood plasma (n = 3: samples were from two different mice, and the third set of data was obtained via pooling of first and second mice) were significantly higher than those found for other mice. The very high d Ca Diet values for these blood plasma samples indicate changes in either the dietary Ca consumption or the Ca metabolism of lactating mice. One possible explanation for the high d Ca value is a preferential supply of isotopically lighter Ca into the milk. A preferential supply of isotopically lighter Ca into the milk can result in isotopically heavier Ca in mother mice (Fig. 2(c)), as a mass balance. In fact, for humans, the Ca/ 42 Ca isotope ratio of breast milk was about 0.6 lower compared to Ca in dietary intake. 32 The present Ca isotope results as well as Ca isotope data for the human breast milk 32 lead us to believe that Ca metabolism in mother mice may be changed by the appearance of a new Ca sink through lactation. However, there were no Ca isotopic data for human blood during the lactating period, and therefore this explanation must remain a possibility. Conclusions We have developed a new analytical technique for precise and reproducible Ca/ 42 Ca and Ca/ 42 Ca ratio measurements for plasma and bone samples from mice by means of multiple collector-icp-mass spectrometry (MC-ICP-MS) with foregoing cation-exchange chromatography. The presented analytical technique was applied to detect possible isotopic fractionation of Ca through elemental metabolism. The resulting Ca/ 42 Ca and Ca/ 42 Ca ratios vary significantly from those of dietary food, indicating that Ca was indeed isotopically fractionated

7 ANALYTICAL SCIENCES NOVEMBER 2008, VOL. 24 through metabolism. Unlike the case for the Fe isotopic ratios, we found no differences in the Ca isotopes between female and male mice. The resulting Ca/ 42 Ca ratios for blood plasma collected from mother mice during the lactation period were significantly higher than those for blood plasma from other mice. The higher Ca/ 42 Ca ratio in these types of samples may reflect the contribution of a new Ca sink through lactation. Stable Ca isotope data will provide new information about elemental metabolism in the body, and this can become a new isotope tracer in the field of metallomics. 47 Acknowledgements We are grateful to Drs. Takeshi Ohno, Keita Irisawa (Tokyo Institute of Technology, Japan) and Tadashi Shimamura (Kitazato Univ., Japan) for technical support and their fruitful discussion on the stable isotope study. We also thank Drs. Patrick Galler (LGC Limited, UK) and Nan-Chin Chu (IFREMER, Brest Département Géosciences, France) for critical reading of the manuscript and much scientific advise on this work. This work was partly supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and the Ministry of Agriculture, Forestry and Fisheries of Japan. References 1. B. Sandstrom, Stable Isotopes in Human Nutrition, ed. F. Mellon and B. Sandstrom, 1996, Academic Press, London, R. J. P. Williams and J. J. R. Frausto de Silva, The Natural Selection of the Chemical Elements: The Environment and Life s Chemistry, 1996, Clarendon Press, London, R. J. P. Williams, Calcium Binding Proteins in Normal and Transformed Cells, ed. R. Pochet, D. E. M. Lawson, and C. W. Heizmann, 1989, Plenum, L. Ambard and E. Bedaujard, Arch. Intern. Med., 1904, 1, L. B. Flanagan, J. R. Ehleringer, and D. E. Pataki, Stable Isotopes and Biosphere-Atmosphere Interactions: Processes and Biological Controls, Physiological Ecology, 2005, Elsevier Academic Press (Great Britain), D. J. Hillegonds, R. Fitzgerald, D. Herold, Y. Lin, and J. S. Vogel, J. Accel. Lab. Assoc., 2004, 9, S. Sturup, Anal. Bioanal. Chem., 2004, 378, H. Wanner, H. Gu, B. Hattendorf, D. Gunther, and S. Dorn, Agric. Ecosys. Environ., 2007, 118, K. J. Schulze, K. O. O Brien, E. L. Germain-Lee, D. J. Baer, A. L. R. Leonard, and B. J. Rosenstein, J. Pediatr., 2003, 1, C. Coudary, C. F.-Coudary, M. Rambeau, J. C. Tressol, E. Gueux, A. Mazur, and Y. Rayssiguier, J. Trace Elem. Medic. Biol., 2006, 20, C. M. Weaver, Present Knowledge in Nutrition, ed. B. A. Bowman and R. M. Russel, 8th ed., 2001, Int. Life Sci. Inst. Press, Washington, R. H. Michener and D. M. Schell; Stable isotope tracers in marine aquatic food webs, in Stable Isotopes in Ecology and Environmental Science, ed. K. Lajtha and R. H. Michener, 1994, Blackwell Scientific, London, R. I. E. Newell, N. Marshall, A. Sasekumar, and V. C. Chong, Mar. Biol., 1995, 123, B. J. Peterson and B. Fry, Annu. Rev. Ecol. Syst., 19, 18, P. J. W. Olive, J. K. Pinnegar, N. V. C. Polunin, G. Richards, and R. Welch, J. Anim. Ecol., 2003, 72, E. Dufour and D. Gerdeaux, Cybium, 2001, 25, M. J. De Niro and S. Epstein, Geochim. Cosmochim. Acta, 1978, 42, M. Minagawa and E. Wada, Geochim. Cosmochim. Acta, 1984, 48, A. Heuser, E. Eisenhauer, N. Gussone, B. Bock, B. T. Hansen, and Th. F. Nagler, Int. J. Mass Spectrom., 2002, 220, F. Bohm, N. Gussone, A. Eisenhauer, W.-C. Dullo, S. Reynaud, and A. Paytan, Geochim. Cosmochim. Acta, 2006, 70, J. Skulan and D. J. DePaolo, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, J. D. DePaolo, Rev. Min. Geochem., 2004, 55, T. Hirata, Geochim. Cosmochim. Acta, 1997, 61, A. D. Anbar, Earth Planet. Sci. Lett., 2004, 217, B. L. Beard, C. M. Johnson, J. L. Skulan, K. H. Nealson, L. Cox, and H. Sun, Chem. Geol., 2003, 195,. 26. C. L. Marechal and F. Albarede, Geochim. Cosmochim. Acta, 2002, 66, T. Walczyk and F. Blanckenburg, Int. J. Mass Spectrom., 2005, 242, J. Fietzke, A. Eisenhauer, N. Gussone, B. Bock, V. Liebetrau, Th. F. Nagler, H. J. Spero, J. Bijma, and C. Dullo, Chem. Geol., 2004, 206, N. G. Sime, C. L. De La Rocha, and A. Galy, Earth Planet. Sci. Lett., 2005, 232, N. Gussone, F. Bohm, A. Eisenhauer, M. Dietzel, A. Heuser, B. M. A. Teichert, J. Reitner, G. Worheide, and W.-C. Dullo, Geochim. Cosmochim. Acta, 2005, 69, T. Steuber and D. Buhl, Geochim. Cosmochim. Acta, 2006, 70, N.-C. Chu, G. M. Henderson, N. S. Belshaw, and R. E. M. Hedges, Appl. Geochem., 2006, 21, T. Hirata, Analyst, 1996, 121, C. N. Marechal, P. Telouk, and F. Albarede, Chem. Geol., 1999, 156, M. E. Weiser, D. Buhl, C. Bouman, and J. Schwieters, J. Anal. At. Spectrom., 2004, 19, L. Halicz, A. Galy, N. S. Belshaw, and R. K. O Nions, J. Anal. At. Spectrom., 1999, 14, G. Shields and J. Veizer, Geochem. Geophys. Geosys., 2002, 3, G. Faure and T. M. Mensing, in Isotopes: Principles and Applications, 2005, John Wiley and Sons, Inc., New Jersey, N. E. Holden and R. L. Martin, Pure Appl. Chem., 1984, 56, T. Ohno and T. Hirata, Anal. Sci., 2007, 23, K. Irisawa and T. Hirata, J. Anal. At. Spectrom., 2006, 21, A. Shinohara, M. Chiba, and Y. Inaba, Anal. Sci., 1998, 14, T. Ohno and T. Hirata, Anal. Sci., 2004, 20, E. R. Monsen, L. Hallberg, M. Layrisse, M. Hegsted, J. D. Cook, W. Mertz, and C. A. Finch, Am. J. Clin. Nutr., 1978, 31, H. Tapiero, L. Gate, and K. D. Tew, Biom. Pharmac., 2001, 55, R. Yip, Present Knowledge in Nutrition, ed. B. A. Bowman and R. M. Russel, 8th ed., 2001, Int. Life Sci. Inst. Press, Washington, H. Haraguchi, J. Anal. At. Spectrom., 2004, 19, 5.