Antibody Labeling and Elemental Mass Spectrometry (Inductively

Jul 29, 2013 - Biochemistry Laboratory, Central University Hospital of Asturias, Celestino Villamil s/n, 33006 Oviedo, Spain. ABSTRACT: Ferritin, an i...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/ac

Antibody Labeling and Elemental Mass Spectrometry (Inductively Coupled Plasma-Mass Spectrometry) Using Isotope Dilution for Highly Sensitive Ferritin Determination and Iron-Ferritin Ratio Measurements Tobias Konz,† Elena Añoń Alvarez,‡ Maria Montes-Bayon,*,† and A. Sanz-Medel*,† †

Department of Physical and Analytical Chemistry, Faculty of Chemistry, University of Oviedo, Julian Clavería 8, 33006 Oviedo, Spain ‡ Biochemistry Laboratory, Central University Hospital of Asturias, Celestino Villamil s/n, 33006 Oviedo, Spain

ABSTRACT: Ferritin, an iron storage protein, is a sensitive clinical biomarker for iron metabolic disorders. It is mainly accumulated in the liver hepatocytes and is present in human plasma at trace levels (picomolar or nanograms per milliliter). Therefore, highly sensitive analytical methods are required to perform ferritin quantification in plasma with high precision and accuracy. For this purpose, we present a mass spectrometry-based analytical strategy (inductively coupled plasma-mass spectrometry, ICP-MS) combined with antibody labeling in a sandwich assay format for ferritin determination. The developed methodology involves two ferritin monoclonal antibodies, one of them biotinylated and the other one labeled with a ruthenium chelate [Ru(bpy)3]2+. The complex formed in solution between ferritin and the two antibodies is then captured using streptavidin-coated magnetic microparticles and directly introduced into ICP-MS for Ru monitoring. Since the Ru complex also allows one to obtain electrogenerated chemiluminescence (ECL), the combination of both sets of data (ICP-MS and ECL) will permit the establishment of the ferritin:Ru stoichiometry. This serves as a basis for further quantification studies using flow injection analysis with isotopically enriched 99Ru as a carrier with ICP-MS detection. Such strategy permits absolute ferritin determination at a picomolar level with good precision (below 5%) and accuracy (85−109% recovery in the existing ferritin reference material, NIBSC code 94/572). Furthermore, the development of a new strategy to address ferritin:iron-ferritin ratios by ICP-MS opens the door also to address the potential of such ratios as a new clinical biomarker for Fe metabolic disorders.

F

(at least 1500 Fe atoms/molecule), whereas H subunit-rich ferritin from the heart and brain have lower average iron content (less than 1000 Fe atoms per molecule).5 Ferritin is also present in human plasma and despite its unclear origin, it is one of the most sensitive biomarkers for Fe metabolic disorders.6 Clinically, serum ferritin is most commonly obtained in combination with other iron parameters (e.g., % transferrin saturation) to gauge the iron status of a specific patient.7 Of the clinical laboratory parameters, the values of serum ferritin are the most useful in the diagnosis and

erritin is a 450 kDa iron-binding protein that exists in both intracellular and extracellular compartments. Apoferritin forms a roughly spherical container of 24 subunits of two types, light (L) and heavy (H), within which ferric iron is stored as a ferrihydrite mineral [Fe2O3·0.5H2O].1,2 The L monomer contains 174 amino acids and has a molecular weight of 19900 Da, while the H monomer has 183 amino acids with a molecular weight of 21200 Da. Ferritin shows ferroxidase activity by converting Fe2+ in Fe3+ that is further internalized and sequestered in the ferritin mineral core.3 The molar ratio Fe:ferritin is variable, depending on the relative abundance of the different subunits. Under “saturation” conditions, ferritin might contain up to 4500 Fe atoms per ferritin molecule.4 For instance, the ferritin molecules from the liver and spleen, which are L subunit-rich, have a relatively high average iron content © 2013 American Chemical Society

Received: June 6, 2013 Accepted: July 29, 2013 Published: July 29, 2013 8334

dx.doi.org/10.1021/ac401692k | Anal. Chem. 2013, 85, 8334−8340

Analytical Chemistry

Article

follow-up of iron deficiency and iron-overload diseases.8 However, the determination of serum ferritin is not trivial since the regular level of this biomarker is in the picomolar (nanogram per milliliter) range. Clinically, a threshold value of 20 ng mL−1 has proved to be useful in the detection of prelatent iron deficiency and provides a reliable indication of exhaustion of the iron reserves available for hemoglobin synthesis.9,10 Latent iron deficiency is defined as a fall below the 12 ng mL−1 ferritin threshold. Therefore, only highly sensitive methods can be used for its direct determination. Among them, methods that combine the specificity of immunoassays (e.g., ELISA) with highly sensitive chemiluminescence detection (e.g., electrochemiluminescence, ECLIA) are preferred to determine ferritin at those low concentration levels.11 The ECLIA method shows high sensitivity (detection limits about 0.5 ng mL−1) and a linear dynamic range up to 2000 ng mL−1. However, the concentration results are calculated based on a calibration curve that needs to be updated daily using a reference ferritin standard of known concentration (third International Recombinant Standard for ferritin, NIBSC code 94/572). Since there is not a reference method for ferritin quantification, the concentration of the reference ferritin standard (used for validation of the ECLIA method) has been obtained, among others, with chemiluminescence methods.12 Therefore, sensitive and absolute methods for ferritin quantitative analysis based on calibration-free methodologies are highly desirable. In this vein, inductively coupled plasma-mass spectrometry (ICP-MS) using isotope dilution analysis (IDA) has been proposed as a quasi-absolute method.13 The concentration results calculated by IDA are based on the measurement of an isotope ratio (obtained with good precision by ICP-MS) between the added tracer also called spike and the analyte present in the sample.14 However, no calibration curve is required, and it could be considered as a definitive method.15 The use of ICP-MS with IDA as a quantification methodology has been successfully applied for the accurate determination of metalloproteins such as transferrin16 (Fe), hemoglobin17 (Fe), or superoxide dismutase18 (Cu) in biological fluids. In all these cases, the concentration calculations are based on the determination of the metal associated to the protein by IDA after chromatographic separation and assessing the stoichiometry metal:protein preservation during sample handling and chromatography.19 Unfortunately, this is not the case of serum ferritin since the Fe:ferritin ratios are not well-established and, what is more important, they can be affected by different clinical pathologies (e.g., anemia of chronic disease, hemochromatosis, etc.). Therefore, here we propose a new alternative for ferritin determination based on the immune complex formation with a monoclonal ferritin antibody that is labeled with a Ru chelate ([Ru(bpy)3]2+). Due to the specificity of the antibody, this strategy does not need chromatographic separation,20,21 but the stoichiometry between Ru and ferritin has to be evaluated.22 This can be accomplished by comparing the ICP-MS Ru results (obtained by isotope dilution analysis) and electrochemically induced chemiluminescence of the Ru complex. The validity of this methodology is evaluated using the reference recombinant ferritin standard by National Institute for Biological Standards and Control (NIBSC). In addition, here we also present the possibility of using the combination of ICP-MS and chemiluminescence measurements to address the Fe:ferritin ratios at very low concentration levels, which should be further

evaluated as a potential clinical biomarker of iron metabolic disorders.



MATERIALS AND METHODS Instrumentation. All ICP-MS experiments during this study were performed using a Thermo Element 2 (Thermo Fisher Scientific, Bremen, Germany) mass spectrometric device, equipped with a double-focusing sector field mass analyzer and applying low resolution (m/Δm = 400) for Ru and medium resolution (m/Δm = 4000) for Fe detection, respectively. The optimized parameters of the Element 2 instrument are summarized in Table 1. The ICP-MS instrument was fitted Table 1. Instrumental Operating Conditions for Ru measurements using DF-ICP-MS with Flow Injection ICP-MS

Thermo Element 2

RF power mass resolution R isotopes monitored nebulizer spray chamber cooling gas auxiliary gas sample gas flow injection (FI)

1290 W 400 (low resolution) 99 Ru, 101Ru, 102Ru, 104Ru concentric Scott double pass (21 °C) 15.5 (L min−1) 0.90 (L min−1) 0.99 (L min−1) peristaltic pump

flow rate mobile phase injection volume mass flow 99Ru spike

5 rpm (0.35 mL min−1) 0.1% HNO3 20 μL 0.36 (g min−1)

with a concentric nebulizer and a Scott double-pass spray chamber. For the flow injection setup, the solvent was pumped using a peristaltic pump (0.35 mL min−1), and the sample injection was conducted with a dual-mode injection valve from Rheodyne, model 9125 (Cotati, California), fitted with a 20 μL PEEK injection loop (Upchurch Scientific, Oak Harbor, Washington). For the evaluation of ferritin purity after its isolation from other serum components, we used size-exclusion chromatography (SEC). The HPLC separation was carried out using a dual-piston liquid chromatographic pump (Shimadzu LC-10AD, Shimadzu Corporation, Kyoto, Japan) equipped with a sample injection valve from Rheodyne, fitted with a 100 μL injection loop and a size-exclusion chromatography column Superdex 200 10/300 GL (300 mm × 10 mm i.d., GE Healthcare Bio-Sciences, Sweden). The electrochemiluminescence system was a Roche E-170 Ferritin “ECLIA” (Roche Diagnostics GmbH, Mannheim, Germany) from the Hospital General de Asturias. Chemicals and Materials. All solutions were prepared using 18 MΩ cm−1 deionized water obtained from a Milli-Q system (Millipore, Bedford, MA). Ferritin standards from equine and human spleens were purchased from Sigma-Aldrich (St. Louis, MO). Isotopically enriched Ru (2.03 μg g−1) with relative abundances of 0.80% 96Ru, 0.29% 98Ru, 85.90% 99Ru, 2.33% 100Ru, 2.79% 101Ru, 4.96% 102Ru, and 2.93% 104Ru was obtained from Cambridge Isotope Laboratories. Dilutions of the concentrated Ru standards to a final concentration of 8.69 pg g−1 were prepared in 0.1% HNO3 (Merck, Darmstadt, Germany) and directly pumped by the peristaltic pump into the DF-ICP-MS for flow injection analysis. Isotopically enriched Fe with relative abundances of 0.03% 54Fe, 4.70% 56Fe, 94.82% 57 Fe, and 0.45% 58Fe were prepared by dilution of the stock 8335

dx.doi.org/10.1021/ac401692k | Anal. Chem. 2013, 85, 8334−8340

Analytical Chemistry

Article

Figure 1. Schematic workflow of the used immunoassay for determination of ferritin in combination with electrochemiluminescence and ICP-MS detection.

solution (22.59 μg g−1) obtained from Spectrascan (Teknolab A.S. Dröbak, Norway). The ECLIA kit (Roche Diagnostics GmbH) contained the reagents for the sandwich assay: biotinylated antiferritin mouse monoclonal antibodies (3 mg L−1), labeled [Ru(bpy)3]2+ antiferritin mouse monoclonal antibodies (6 mg L−1), and streptavidin-covered magnetic microparticles (0.72 mg mL−1). For ferritin-specific isolation, precipitation with MeOH (Merck) was conducted followed by ultrafiltration using 100 kDa Amicon ULTRA centrifugal filters (Millipore, Tullagreen, Ireland). The third international recombinant standard for ferritin, NIBSC code 94/572 was purchased from the National Institute for Biological Standards and Control (Health Protection Agency, Hertfordshire, U.K.). Serum samples from healthy volunteers (n = 30) were kindly provided by the Hospital Central of Asturias, Laboratory for Biochemical Analysis (Oviedo, Spain). Samples were anonymous and collected in accordance with protocols approved by the relevant institutional review boards and with the Declaration of Helsinki. All blood samples were collected using BD Vacutainer-containing gel for serum separation. The blood samples were allowed to coagulate and were centrifuged at 5000 rpm. An aliquot of the clear serum was used for direct ferritin determination by ECLIA and another aliquot for DFICP-MS analysis. Samples were frozen at −20 °C and thawed only once afterward. Sandwich Immunoassay with ECL and DF-ICP-MS Detection. The first incubation uses 10 μL of serum for ECL analysis and 50 μL in the case of the ICP-MS experiments, a biotinylated ferritin-specific antibody, and a Ru-labeled ferritinspecific antibody to form a sandwich complex. The second incubation occurs after the addition of magnetic microparticles that cause the complex to bind to the solid phase by streptavidin−biotin interactions. In the case of the ECLIA, the reaction mixture is then aspirated into the measuring cell where the microparticles are magnetically captured onto the surface of the electrode. Unbound substances are then removed, and the application of a voltage to the electrode

induces a chemiluminescent emission, which is measured by a photomultiplier. Results are determined via a calibration curve. In the case of the DF-ICP-MS measurements, the microparticles are collected by an external magnet placed on the bottom of the vial, and the unbound substances are then removed by aspirating the solution out of the vial. The magnetic particles attached to the bottom of the vial are washed four times with buffer and then resuspended in 200 μL of 0.1% HNO3. Twenty microliters (20 μL) of that solution is directly injected into the flow injection system (FI-DF-ICP-MS). Purification of Serum Ferritin for Fe:Ferritin Ratio Measurements. For isolation of serum ferritin, a multidimensional sample preparation protocol was applied including: (1) protein precipitation with MeOH, (2) heat precipitation, and (3) ultrafiltration (100 kDa). For the first part, MeOH was added to 1.5 mL of serum to give a final concentration of 40% (v/v). After being thoroughly mixed, the samples were centrifuged at 15,000g for 30 min (4 °C). Subsequently, the supernatant was collected, heated to 75 °C, and this temperature was maintained for 10 min, causing the denaturation of a large proportion of serum proteins. However, ferritin is heat stable and remained in the supernatant, which was separated from precipitated proteins by centrifugation. Finally, the last purification step included the ultrafiltration through 100 kDa that served also to eliminate the methanolic solution. The remaining ferritin content was determined by ECLIA, the purity by SEC-DF-ICP-MS and the Fe content by DF-ICP-MS (using isotope dilution of Fe).



RESULTS AND DISCUSSION Ferritin Monitoring by DF-ICP-MS-Linked Sandwich Immunoassay (IA). The commercially available ECLIA method using [Ru(bpy)3]2+ as a chemiluminescent probe has been adapted here to highly sensitive DF-ICP-MS for Ru determination. Ruthenium exhibits excellent properties as an elemental label to conduct ICP-MS-linked immunoassay as it (a) has a highly sensitive response in ICP-MS and is free from 8336

dx.doi.org/10.1021/ac401692k | Anal. Chem. 2013, 85, 8334−8340

Analytical Chemistry

Article

Table 2. Results Obtained for the Analysis of the 3rd International Recombinant Standard for Ferritin, NIBSC Code 94/572 Reference Material with the Proposed ICPMS-Linked Quantitative Immunoassay ferritin (ng/mL)

injected (fmol)

found (fmol)

recovery (%)

50.3 150.2 310.3 513.0 793.9 1031.0

0.50 2.15 4.41 7.53 11.01 14.19

0.43 2.34 4.30 6.58 10.33 15.48

85.0 108.9 97.4 87.4 93.9 109.1

Figure 2. Ruthenium signals obtained by flow-injection (FI)-DF-ICPMS for a ferritin concentration of 150 ng mL−1. The blue trace corresponds to 102Ru and the black one to the isotopically enriched solution of 99Ru at 9 pg mL−1 with a constant flow of 0.36 g min−1.

Figure 4. Chromatograms that correspond to the Fe trace from (A) a ferritin standard from human spleen 400 ng mL−1 and (B) a serum sample containing initially about 530 ng mL−1 of ferritin after protein isolation and preconcentration obtained by SEC-DF-ICP-MS.

is collected on magnetic microparticles coated with streptavidin. The microparticles are then gathered at the bottom of the reaction vial by using an external magnet, and the excess of reagents is removed as shown in Figure 1. The particles are then dissolved in 0.1% HNO3 and introduced into the DF-ICPMS. The immunoassay principle is the same used for ECL measurements, but in this case, there is no need to apply any external current (to induce [Ru(bpy)3]2+ excitation and emission of photons) since Ru can directly be measured by ICP-MS. As shown in Figure 1, the dissociation of Ru from the formed sandwich complexes by acidification (0.1% HNO3) enables long-term storage of the microparticles before analysis, simplifying assay protocols. The dissociated complex is then

Figure 3. Curves representing the relationship between (A) Ru peak area (observed by FI-DF-ICP-MS) and ferritin concentration in ng mL−1 (obtained with the ECLIA measurements) and (B) Ru absolute amount in fmol (obtained by isotope dilution-DF-ICP-MS) and ferritin absolute amount in fmol (ECLIA).

spectral interferences in most of its isotopes, (b) occur at naturally low concentrations in the body and environment, and (c) has good biocompatibility and the ability to easily conjugate to biomolecules.23,24 Here, we couple a sandwich-type ferritin immunoreaction where the bioconjugated complex (biotinylated antibody + ferritin + antibody labeled with [Ru(bpy)3]2+) 8337

dx.doi.org/10.1021/ac401692k | Anal. Chem. 2013, 85, 8334−8340

Analytical Chemistry

Article

Figure 5. Graphical plot of the results obtained for the serum samples of controls (A) representing ferritin−iron versus serum ferritin concentration and (B) Fe:ferritin molar ratio verus ferritin concentration.

injected in a flow injection system to minimize sample consumption and to conduct Ru direct quantification by DFICP-MS. One of the promising fields of application of element-labeled immunoassay combined with ICP-MS detection is the development of SI-traceable primary methods for biocompounds to be used in interlaboratory comparisons and quality assurance by so-called isotope dilution.19 In this line, we have developed a quantitative strategy using isotopically enriched 99 Ru as a carrier (in 0.1% HNO3) in the flow injection system. The type of profiles obtained for the two Ru isotopes (99Ru and 102 Ru) can be observed in Figure 2, using about a 9 pg mL−1 constant 99Ru flow. The use of DF-ICP-MS permits excellent sensitivities for Ru, while a calibration curve-free methodology is used to address Ru concentration in the immune complex by directly calculating the Ru isotopic ratios and applying the isotope dilution equation on each point.13 Evaluation of the Complex Stoichiometry: Complementary ECL and DF-ICP-MS. To translate quantitative elemental Ru results into ferritin concentration it is necessary to address the stoichiometry Ru:ferritin to correlate both sets of data. The number of Ru labels per antibody (or the number of antibodies taken per protein, particularly in the case of ferritin that is formed by 24 subunits) is an important factor. It will affect the sensitivity and final quantification capabilities of the ICP-MS detection (since the signal intensity is linearly proportional to the number of Ru metallic atoms of the

metal-containing tag). For this purpose, a number of serum samples (n = 12) were selected, in which ferritin concentration covered a broad linear range (50 to 1200 ng mL−1) measured by the well-established ECLIA methodology. Because the immunoreaction steps of the two used methods (ECLIA and DF-ICP-MS-linked IA) are identical, their outcome can be directly compared (that is, the electrochemiluminesce produced will be proportional to the [Ru(bpy)3]2+ complex concentration and, subsequently, to the Ru elemental concentration). In other words, we should be able to have a direct correlation between the ferritin concentration (obtained by ECLIA) and the Ru peak area (obtained by DF-ICP-MS). Figure 3A shows the resulting plot obtained with the two sets of experimental data. As can be observed, a good linear correlation (correlation coefficient R2 = 0.98) covering 2 orders of magnitude concentration range of ferritin was obtained. By applying the isotope dilution equation, any Ru peak area can easily be converted into a Ru absolute concentration (fmol) that can also be plotted versus ferritin concentration (fmol). The results observed are shown Figure 3B, where the slope of the obtained curve will directly provide the stoichiometry between Ru and ferritin (necessary, as explained before for further quantitative application of this ICP-MS-based methodology). In this case, the linear regression (y = 23.1x − 1.7, R2 = 0.98) provides a stoichiometry of 1:23 (ferritin:ruthenium). This means that per mol of ferritin, we have 23 mol of ruthenium (this stoichiometry could be ascribed to the 24 ferritin 8338

dx.doi.org/10.1021/ac401692k | Anal. Chem. 2013, 85, 8334−8340

Analytical Chemistry

Article

Measurement of Fe:Ferritin Ratios in Serum Samples. Serum ferritin is a very valuable parameter in the diagnosis of Fe-related disorders, as mentioned before. However, secondary factors such as infection or the presence of certain tumors can also increase serum ferritin values. Some previous experiments have pointed out that the measurement of ferritin-bound iron/ ferritin ratios could offer a higher potential to diagnose iron disorders than pure ferritin measurements.26,27 For this reason, the second goal of this work was to establish an accurate methodology to address ferritin-bound iron/ferritin ratios measurements in serum samples. Unfortunately, the previously developed methodology using magnetic (therefore Fe-containing) microparticles provided too high of soluble iron concentrations, thus hindering the measurement of much smaller physiological variations of iron. This could be improved by modifying the immunoassay using latex microparticles instead of magnetic ones. In that case, simultaneous Ru and Fe measurements could be taken within the same experiments, obtaining absolute ferritin concentration and Fe:ferritin results in a single run. However, in this case additional ferritin purification experiments had to be conducted in order to eliminate free Fe or Fe associated to other plasma proteins (e.g., transferrin) before the final elemental ICP-MS determinations. Three protocols (heat treatment, MeOH precipitation, and ultrafiltration) were independently evaluated (and also combined in sets of two), and the resulting solutions were injected in the SEC column coupled to DF-ICP-MS. In all the cases, we observed some other Fe-containing peaks in the chromatograms impairing the quantification of Fe-ferritin by total Fe determination using ICP-MS. Best results were observed with a sequential treatment that involved first the addition of MeOH, to a final concentration of 40% (v/v), followed by thermal precipitation that removes about 65% of the serum proteins but not ferritin. Finally, the methanolic solution was removed by ultracentrifugation using 100 kDa cutoff filters that served also for preconcentration of ferritin. The samples treated in this way were analyzed by SEC-DFICP-MS and compared with a ferritin commercial standard. The results are shown in Figure 4 (where Figure 4A corresponds to the protein standard and Figure 4B to the purified serum sample). As can be observed, there is a good match between the elution volumes of both species, and no other Fe-containing species could be detected in the sample. Therefore, total Fe determination using isotope dilution analysis in the purified fraction would be sufficient for evaluating the Fe-ferritin content. Simultaneously, an aliquot of the same fraction was used for ferritin quantification, using the previously described ECLIA strategy. Since the initial ferritin concentration results were provided by the University Hospital (using ECLIA), we decided to use the same methodology for the determination of the protein concentration after purification. By comparing serum ferritin concentrations before and after the isolation treatment, the protein recovery turned out to be in the range of 30−40%. This low recovery can be ascribed to ferritin (present at nanograms per milliliter) coprecipitation with high abundant proteins in the serum (e.g., HSA at 30 mg mL−1 or Tf at 3 mg mL−1) during purification. Figure 5A shows the ferritin-bound iron measured in serum that ranged in normal controls (n = 40) from 3 to 25 ng Fe/ mL−1 in agreement with the scarce literature available on this aspect.26,27 Figure 5B also shows the plotted graph of the molar

subunits) that explains the increase of the sensitivity in respect to 1:1 common immunochemical reactions. In fact, the detection limits of immunoassay-linked ICP-MS methods are normally improved by multiple labeling (either by increasing the number of labels attached to an antibody or the number of metal atoms in the labels (e.g., using metallic nanoparticles).25 The first option has some limitations because of the interference of the bulky labels in antibody−antigen binding and the increased fraction of nonspecific binding. Also, in the case of nanoparticles, the determination of the specific stoichiometry (including the number of atoms per particle) further increases the complexity of quantitative calculations. Here, thanks to the use of two complementary (elementary and molecular) methods, the stoichiometry can be directly determined. The analytical features of the developed DF-ICP-MS linked immunoassay for ferritin have been compared with those of the original molecular ECLIA method. In terms of sensitivity, the ICP-MS-based strategy shows comparable limits of detection (0.48 ng mL−1 of ferritin) to those reported by ECLIA (0.5 ng mL−1). Precision among samples, expressed as % CV, turned out to be about 5−7% (n = 5) with ICP-MS and about 2% with the ECLIA. Such difference could be ascribed to the use of the streptavidin-covered magnetic microparticles necessary for the assay that are then introduced into the nebulizer of the ICPMS, increasing the noise. To evaluate if such particles had any effect on the detected Ru signal, the assay was also conducted without particles and just immobilizing the biotinylated antibody on an ELISA well plate. Then, the serum sample was applied and after washing, the Ru-labeled antibody was added. Finally, by adding 0.1% HNO3 into the ELISA well plates, it was possible to disrupt the immune complex and inject the resulting solution into the ICP-MS. The observed plot (data not shown) provided a slope of 22.5 and a correlation coefficient of R2 = 0.98, which are in good agreement with the result obtained using the magnetic microparticles. However, the observed % CV among injections was not significantly improved (probably, in this case, due to differences in the coating of the plate with the antibody). Therefore, further quantitative studies were always done using the microparticles. Validation by Using the Ferritin Reference Material NIBSC 94/572. In order to validate the proposed DF-ICP-MS linked immunoassay for ferritin in serum, the reference material (NIBSC 94/572) was purchased and analyzed. This is a recombinant ferritin standard that contains very little Fe and is commonly used for calibration of the ECLIA systems. The freeze-dried material was reconstituted in water and then treated (as shown in Figure 1) using different dilutions to cover the previously evaluated concentration range (from approximately 50 to 1000 ng mL−1). The obtained results of the ferritin quantification can be observed in Table 2. In summary, the ferritin concentration results are in good agreement with the expected ones, and quantitative recoveries for all the concentrations assayed (down to 0.55 fmol using 50 μL for the assay) can be obtained. Therefore, the proposed methodology using the DF-ICP-MS-linked immunoassay with isotope dilution as a definitive quantification strategy (does not resort to calibration curves) can be used as a reference method at femtomol levels of the protein. It could be used as an independent method to address the concentration of ferritin in standards used for calibration or for analysis of real serum samples. 8339

dx.doi.org/10.1021/ac401692k | Anal. Chem. 2013, 85, 8334−8340

Analytical Chemistry



ratio Fe-ferritin vs ferritin concentrations found for the same samples. As can be observed in this case, there are two different types of correlation when ferritin levels are below and above 100 ng mL−1, respectively. Below 100 ng mL−1 ferritin, the number of Fe atoms per mol of ferritin is quite disperse, reaching values up to 3800, while most individuals are in the range from 1500 to 2500, representing about 50% saturation (if a stoichiometry of 4500 is assumed). Above 100 ng mL−1 ferritin, most analyzed samples (n = 14) show similar saturation levels corresponding to approximately 10%. Such variations have been previously reported for diseases like liver cell damage, where the ferritin iron saturation in serum was significantly higher than that found in groups with iron overload diseases, probably indicating the release of intracellular iron-rich ferritin into the blood.27 Therefore, Fe-ferritin and its relation to circulating ferritin seem to be interesting parameters to evaluate further in patients with iron metabolic disorders. In this work, we provide a useful analytical tool to conduct such studies in large patient cohorts to investigate possible correlations.



CONCLUSIONS



AUTHOR INFORMATION

Article

REFERENCES

(1) Knovich, M. A.; Storey, J. A.; Coffman, L. G.; Torti, S. V.; Torti, F. M. Blood Rev. 2009, 23, 95−104. (2) Theil, E. C. J. Nutr. 2003, 133, 1549S−1553S. (3) Liu, X.; Jin, W.; Theil, E. C. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3653−3658. (4) Theil, E. C. Handbook of Metalloproteins; Messerschmidt, A., Huber, R., Poulos, T., Weighart, K., Eds.; Wiley, Chichester, West Sussex, U.K., 2001; 771−781. (5) Bou-Abdallah, F.; Zhao, G.; Biasiotto, G.; Poli, M.; Arosio, P.; Chasteen, N. D. J. Am. Chem. Soc. 2008, 130, 17801−17811. (6) Muñoz, M.; García-Erce, J. A.; Remacha, A. F. J. Clin. Pathol. 2011, 64, 287−296. (7) Del Castillo Busto, M. E.; Montes-Bayón, M.; Bettmer, J.; SanzMedel, A. Analyst 2008, 133, 379−384. (8) Andrews, N. C. N. Engl. J. Med. 1999, 341, 1986−1994. (9) Finch, C. A.; Bellotti, V.; Stray, S.; Lipschitz, D. A.; Cook, J. D.; Pippard, M. J.; Huebers, H. A. West J. Med. 1986, 145, 657−663. (10) Borch-Iohnson, B. Analyst 1995, 120, 891−903. (11) Blackmore, S.; Hamilton, M.; Lee, A.; Worwood, M.; Brierley, M.; Heath, A.; Thorpe, S. J. Clin. Chem. Lab. Med. 2008, 46, 1450− 1457. (12) Thorpe, S. J.; Walker, D.; Arosio, P.; Heath, A.; Cook, J. D.; Worwood, M. Clin. Chem. 1997, 43, 1582−1587. (13) Heumann, K. G.; Rottmann, L.; Vogl, J. J. Anal. At. Spectrom. 1994, 9, 1351−1355. (14) Rodríguez-González, P.; Marchante-Gayón, J. M.; García Alonso, J. I.; Sanz-Medel, A. Spectrochim. Acta 2005, 60, 151−207. (15) Hulanicki, A. Pure Appl. Chem. 1995, 67, 1906−1911. (16) Del Castillo Busto, M. E.; Montes-Bayon, M.; Sanz-Medel, A. Anal. Chem. 2006, 78, 8218−8226. (17) Del Castillo, M. E.; Montes-Bayón, M.; Añoń , M. E.; SanzMedel, A. J. Proteomics 2011, 74, 35−43. (18) Nuevo-Ordoñez, Y.; Montes-Bayón, M.; Blanco-González, E.; Sanz-Medel, A. Anal. Chem. 2010, 82, 2387−2394. (19) Bettmer, J. Anal. Bioanal. Chem. 2010, 397, 3495−3502. (20) Zhang, S.; Zhang, C.; Xing, Z.; Zhang, X. Clin. Chem. 2004, 50, 1214−1221. (21) Baranov, V. I.; Quinn, Z.; Bandura, D. R.; Tanner, S. D. Anal. Chem. 2002, 74, 1629−1636. (22) Bandura, D. R.; Baranov, V. I; Ornatsky, O. I.; Antonov, A.; Kinach, R.; Lou, X.; Pavlov, S.; Vorobiev, S.; Dick, J. E.; Tanner, S. D. Anal. Chem. 2009, 81, 6813−6822. (23) Zhang, C.; Wu, F. B.; Zhang, X. R. J. Anal. At. Spectrom. 2002, 17, 1304−1307. (24) Hu, S. H.; Zhang, S. C.; Hu, Z. C.; Xing, Z.; Zhang, X. R. Anal. Chem. 2007, 79, 923−929. (25) Bustos, A. R. M.; Trapiella-Alfonso, L.; Encinar, J. R.; CostaFernandez, J. M.; Pereiro, R.; Sanz-Medel, A. Biosens. Bioelectron. 2012, 33, 165−171. (26) Yamanishi, H.; Iyama, S.; Yamaguchi, Y.; Kanakura, Y.; Iwatani, Y. Clin. Biochem. 2002, 35, 523−529. (27) Nielsen, R.; Günther, U.; Dürken, M.; Fischer, R.; Düllmann, J. J. Lab. Clin. Med. 2000, 135, 413−418.

Ferritin absolute determination has been achieved based on a calibration-free methodology and using ICP-MS-linked immunoassay with isotope dilution analysis that could be applied as a reference method to calibrate routinely used ECLIA methods and ferritin reference standards. The combination of ICP-MS-linked immunoassay with isotope dilution has been successfully explored here for the first time, to our knowledge, and shows excellent potential to be extended to the quantitative analysis of many other clinical biomarkers using the same principle. One of the salient features of the developed strategy is that it permits femtomol quantitative analysis of ferritin in serum samples by direct isotope Ru ratio measurements. Furthermore, the comparison of luminescence and MS-based results has permitted the stoichiometry between the antigen (ferritin) and the elemental label (Ru) (1:23) to be established. Such a relationship is the basis for posterior protein quantitative studies. Quantitative ferritin recoveries established in the ferritin NIBSC reference material provided additional validation of the developed strategy. Additionally, the experimental ratio between ferritin concentration and its Fe content could also be evaluated, after careful optimization of the ferritin purification procedure, in a number of serum samples. Ferritin-bound iron levels could serve as a new parameter to address Fe-related metabolic disorders, since it should not be affected by inflammation, etc. However, larger sample cohorts should be evaluated for such purpose with the here-developed tools.

Corresponding Author

*M.M.B.: e-mail, [email protected]. A.S.-M.: e-mail, [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors want to acknowledge the support of the MICINN for T.K. Grant AP2008-04449 and for financial support through the project CTQ2006-02309. 8340

dx.doi.org/10.1021/ac401692k | Anal. Chem. 2013, 85, 8334−8340