Heme oxygenase (HO) is a renoprotective protein in the microsome that degrades heme and produces biliverdin. Biliverdin is then reduced to a potent antioxidant bilirubin by biliverdin reductase in the cytosol. Because HO activity does not necessarily correlate with HO mRNA or protein levels, a reliable assay is needed to determine HO activity. Spectrophotometric measurement is tedious and requires a relatively large amount of kidney samples. Moreover, bilirubin is unstable and spontaneously oxidized to biliverdin in vitro. We developed a novel and sensitive liquid chromatography-tandem mass spectrometry (LC-MS/MS) method to quantify biliverdin to measure HO activity in mice. Biliverdin and its internal standard, a deuterated biliverdin-d4, have MS/MS fragments with m/z transitions of 583 to 297 and 587 to 299, respectively. We prepared lysates of mouse kidneys, and added excess hemin, NADPH, and bilirubin oxidase to convert all bilirubin produced to biliverdin. After 30-min incubation at 37 or 4°C, the samples were analyzed by LC-MS/MS. The difference in the amount of biliverdin between the two temperatures is HO activity. Treating mice with cobalt protoporphyrin, which induces the expression of HO, increased HO activity as determined by biliverdin production. Measuring the production of biliverdin using LC-MS/MS is a more sensitive and specific way to determine HO activity than the spectrophotometric method and allows the detection of subtle changes in renal or other HO activity.
- heme oxygenase
heme oxygenase (HO) catalyzes heme degradation and produces iron, biliverdin, and carbon monoxide (CO), which is an important vasodilator (Fig. 1) (8). Biliverdin is then converted to a potent antioxidant, bilirubin, by biliverdin reductase (1).
There are two isoforms of HO: HO-1 is induced by UVA light, reactive oxygen species, heavy metals, and metalloporphyrins, while HO-2 is a constitutive isoform that is responsible for normal HO activity (8). HO-1 induction is especially important in response to oxidative stress. The nuclear factor-erythroid 2 (Nrf2) pathway is one of the most important systems that enhance cellular protection against oxidative stress. Induction of Nrf2 increases the protein levels of vascular endothelial growth factor (VEGF) via its target protein, HO-1, and its metabolite, CO (7).
The cytoprotective properties of HO-1 were first recognized in heme protein-induced acute kidney injury (AKI) (10), and a recent study reports that the induction of HO-1 by hemin attenuates cisplatin-induced AKI in rats (3). HO-1 responds to oxidative stress in the kidney and plays a pivotal role in the maintenance of renal function (9). In addition, HO-1 may be useful for treating atherosclerosis, kidney disease, and some hypertensive disorders including preeclampsia (4, 5, 14). HO activity is usually estimated by the expression levels of HO mRNA or protein, but HO activity does not necessarily correlate with its protein or mRNA levels (11), at least partly because HO activity measures both HO-1 and HO-2 without discriminating them. Gas chromatography-mass spectrometry (GC-MS) is also used to measure HO activity by measuring CO generation (6). However, CO production in tissues involves both HO-dependent and HO-independent pathways such as lipid peroxidation (15). Spectrophotometry can also be used to measure HO activity; however, mouse HO expression in the kidney is not high enough to be detected using spectrophotometry. In addition, brown bilirubin solution quickly changes its color to green, suggesting that bilirubin is unstable and spontaneously oxidized to biliverdin in vitro, which questions the reproducibility of bilirubin measurement. To circumvent these problems, here we report a novel and highly sensitive method of quantifying biliverdin to measure HO activity using liquid chromatography-tandem mass spectrometry (LC-MS/MS).
MATERIALS AND METHODS
Biliverdin-d4, cobalt protoporphyrin (CoPP), and cisplatin were purchased from Taiyo Nippon Sanso (Tokyo, Japan), Frontier Scientific (Logan, UT), and Wako Pure Chemical Industries (Osaka, Japan), respectively.
Wild-type C57BL/6J mice (CLEA Japan) were treated with an HO-1 inducer, CoPP (5 mg/kg body weight ip), or vehicle (1% DMSO/saline), and euthanized 24 h later for further analysis. In another experiment, C57BL/6J mice were treated with a cisplatin (30 mg/kg body weight ip) to induce acute kidney injury and euthanized 24 h later to obtain kidney samples. All procedures were approved by the Ethics Committee of Animal Experiments, Tohoku University.
Preparation of tissue lysate and microsome.
Tissues were frozen immediately upon collection. Samples were homogenized in 3 ml/g tissue of the homogenizing buffer containing 0.25 M sucrose, 20 mM Tris·HCl, 1 mM EDTA, and 0.1% protease inhibitor cocktail (Sigma). One and one-half murine kidneys, or 100 mg of the liver, were homogenized, centrifuged at 9,000 g for 20 min at 4°C, and the supernatant was used as the tissue lysate.
The tissue lysate was ultracentrifuged at 105,000 g for 60 min at 4°C, and the microsome pellet was washed with the homogenizing buffer. The pellet was homogenized and resuspended in 100 mM KH2PO4 solution containing 20% glycerol for use.
Microsome bilirubin production as an HO activity assay determined by the spectrophotometric method.
An HO activity assay by the spectrophotometric method was performed as previously described (2, 14). An aliquot of microsome (0.5 mg protein) in 947.2 μl was incubated with 252.8 μl of reaction solution [20 mM glucose-6-phosphate Na (120 μl), 2 U glucose-6-phosphate dehydrogenase (0.8 μl), 8 mM NADP (120 μl), and 2 mM hemin (12 μl)] for 60 min at 37 or at 4°C, protected from light. Five hundred microliters of chloroform was added and vortexed. The sample was centrifuged at 15,000 g for 10 min at 4°C. The chloroform layer was removed, evaporated, and resuspended with 70 μl DMSO and 70 μl acetonitrile. Optical density at 460 nm was measured by using a spectraMAX190 (Molecular Devices). Bilirubin production was calculated by subtracting the bilirubin value of the above microsome reaction incubated at 4°C from that incubated at 37°C.
Lysate or microsome biliverdin production for an HO activity assay by LC-MS/MS.
An aliquot of lysate or microsome (0.5 mg protein) in 78 μl was incubated with reaction solution [20 mM NADPH (5 μl), 2.5 mM hemin (5 μl), 2.5 M sucrose (10 μl), 2 M Tris·HCl (1 μl), and 2.5 U/ml bilirubin oxidase (1 μl)] for 30 min at 37 or at 4°C, protected from light. After this incubation, 98 μl of 0.1% formic acid in methanol were added to stop the reaction together with 2 μl of 10 μg/ml biliverdin-d4. The reaction solution was centrifuged at 15,000 g for 20 min at 4°C. The supernatant was filtered through membranes (pore size 0.22 μm, Merck Millipore) and analyzed by LC-MS/MS. Biliverdin production was calculated by the biliverdin values of lysate incubated at 37°C subtracted by those incubated at 4°C.
LC-MS/MS analyses of biliverdin.
The quantitative analysis of biliverdin by LC-MS/MS was performed using a Nanospace Nasca LC System (Shiseido, Tokyo, Japan) coupled to a TSQ Quantiva (Thermo Fisher Scientific) and operated in the positive ion mode. Each sample (10 μl) was injected onto a 150 × 2.0-mm YMC-Pack Pro C18 column (YMC) with a flow rate of 0.2 ml/min. For gradient elution, mobile phase A was H2O/acetonitrile/HCOOH = 98/2/0.1 and mobile phase B was H2O/acetonitrile/HCOOH = 2/98/0.1. The linear and stepwise gradient was programmed as follows: 0–1 min: 10% solvent B; 1–2 min: 10–55% solvent B; 2.1–4 min: 55–90% solvent B; 4.1–6 min: 90–100% solvent B; 6.1–11 min: 100% solvent B; and 11.1–15 min: 10% solvent B. Bilirubin and biliverdin-d4 were detected in SRM mode by monitoring the transitions of the m/z 583 to 297 and 587 to 299, respectively. The spray voltage was 3,500 V, the vaporizer temperature was 275°C, and the ion transfer tube temperature was 350°C.
We validated our LC-MS/MS method in accordance with the Guidance for Industry, Bioanalytical Method Validation protocol from the US Food and Drug Administration (13). Three doses (5, 50, and 500 pg/injection) and six samples on three different days were used for method validation. The accuracy of an analytic method is the closeness of mean test results obtained by the method to the actual value of the analyte. The mean value should be within 15% of the actual value. The precision of an analytic method is the closeness of individual measures of an analyte when the procedure is applied repeatedly to multiple aliquots of a single homogeneous volume of biological matrix. The precision determined at each concentration level should be within 15% of the coefficient of variation. The recovery of an analyte in an assay is the detector response obtained from an amount of the analyte added to and extracted from the biological matrix, compared with the detector response obtained for the true concentration of the analyte in solvent.
mRNA expression was quantified using real-time PCR using the following primers: Hmox-1 forward 5′-CCTCACTGGCAGGAAATCATC-3′ and reverse 5′-CCTCGTGGAGACGCTTTACATA-3′; Hmox-2 forward 5′-TACTTCACATACTCAGCCCT-3′ and reverse 5′-ATGGGCCACCAGCAGCTCTG-3′. β-Actin was used as an internal control.
Values are reported as means ± SD. Data were analyzed by Student's t-test using JMP software version 11 (SAS Institute, Cary, NC).
Biliverdin measurement using LC-MS/MS.
To develop a sensitive method of determining HO activity in the mouse kidney, we decided to develop an LC-MS/MS-based measurement of biliverdin standard and then measured biliverdin production using mouse kidney lysate. For LC-MS/MS analysis, positive ion mode monitoring was selected for biliverdin, and biliverdin-d4 was used as an internal standard. The transitions of the precursor ions into the major product ions were monitored as m/z 583.1 → 297.1 for biliverdin (Fig. 2A) and as m/z 587.1 → 299.1 for biliverdin-d4. Quantification analyses were performed in the selected reaction monitoring (SRM) mode. The retention time of both biliverdin and biliverdin-d4 was 4.8 min (Fig. 2B). When the murine kidney lysate was incubated with NADPH and hemin, both biliverdin and bilirubin were produced and easily detected (Fig. 3A).
The change in the brown color of bilirubin to green suggests bilirubin is quickly and spontaneously oxidized to biliverdin in vitro, and the bilirubin measurement may not be reproducible (Fig. 4). Moreover, bilirubin and biliverdin may not behave exactly the same in LC-MS/MS. Accordingly, we used excess bilirubin oxidase to convert all bilirubin to biliverdin (Fig. 3B) and decided to measure only biliverdin for the HO activity assay.
The biliverdin and biliverdin-d4 peaks were integrated automatically by XCALIBUR 2.2 software (Thermo Fisher Scientific). The relative area ratio (i.e., the area of biliverdin/area of biliverdin-d4) was used to construct a calibration curve, which was linear in the range of 1 to 100 pg/μl of biliverdin/10-μl injection, with a regression coefficient of 0.999 over (Fig. 5A). In contrast, the quantitative limit using spectrophotometry was 500 pg/μl for both bilirubin and biliverdin (Fig. 5B). We conclude that measuring biliverdin using LC-MS/MS is superior to other methods currently used to evaluate HO activity.
In our experimental range, i.e., 5–500 pg, the accuracy was 4.33% (intraday, n = 6) and 1.75% (interday, n = 3) (Table 1). The precision was 7.52% (intraday, n = 6) and 0.023% (interday, n = 3) (Table 2). The recovery of the determination of biliverdin was 101% (intraday, n = 6) and 98.3% (interday, n = 3) (Table 3). We conclude that this LC-MS/MS method has good reliability and reproducibility.
HO activity measured by LC-MS/MS and the mRNA expression of HO.
We next determined the amount of biliverdin production in the mouse kidney lysate. HO activity (pmol·mg protein−1·h−1) was calculated as follows: [biliverdin (pmol) ÷ amount of protein in the lysate (mg) ÷ time of incubation at 37°C (h)] − [biliverdin (pmol) ÷ amount of protein in the lysate (mg) ÷ time of incubation at 4°C (h)]. Biliverdin values in the murine kidney lysate incubated at 37 and at 4°C as described in materials and methods were between 3 and 7 pg/μl and <0.01 pg/μl, respectively. We used the LC-MS/MS method to measure HO activity in the kidneys of untreated mice (control) and mice treated with the HO-1 inducer CoPP. The expression of the HO-1 and HO-2 genes, heme oxygenase-1 (decycling; Hmox-1) and Hmox-2, were both quantified in the kidney using RT-PCR. CoPP treatment significantly increased renal HO activity (8.4 ± 0.9 vs. 5.5 ± 0.9 pmol·mg protein−1·h−1 in the control, P = 0.049) and Homx-1 expression, but did not affect Homx-2 expression (Fig. 6). CoPP dramatically increased HO activity and Hmox-1 expression in the liver (Fig. 7).
Comparison of HO activity measured by LC-MS/MS and by spectrophotometry.
To compare the new method by LC-MS/MS with the spectrophotometric method, we determined the amount of biliverdin or bilirubin production in the kidney microsome obtained from cisplatin-induced AKI mice, a well-known model of HO induction (16). Biliverdin values measured by LC-MS/MS of this microsome preparation incubated at 37 and at 4°C were 33.1 ± 19.1 and 0.4 ± 0.2 pg/μl, respectively, calculated using the calibration curve shown in Fig. 8. HO activity was calculated to be 22.5 ± 13.3 pmol biliverdin·mg microsome protein−1·h−1. In contrast, bilirubin values in the cisplatin-induced AKI model mouse kidney microsome measured by the spectrophotometric method was ∼36.9 ± 9.5 pg/μl using the calibration curve (Fig. 8). However, these amounts of bilirubin were below quantitative limit using spectrophotometry (Fig. 5B).
We developed a sensitive, selective, and reliable method for measuring HO activity using LC-MS/MS. Our method has multiple advantages over other methods. First, this method uses tissue lysate instead of microsomes. Tissue lysate is easy to prepare with minimal loss of sample volume, whereas obtaining a microsome fraction involves multiple steps and sample loss is difficult to avoid, as our preliminary experiment showed that at least one and one-half murine kidneys are necessary. Second, we used excess bilirubin oxidase to convert bilirubin to biliverdin. Because bilirubin is unstable and spontaneously oxidizes to biliverdin, and because bilirubin and biliverdin likely behave differently in LC-MS/MS, quantifying only one substance, i.e., biliverdin, makes the assay more accurate. Third, using biliverdin-d4 as an internal standard enabled accurate measurement of biliverdin, because biliverdin and a stable isotope-labeled biliverdin behave the same way using this analytic method. Biliverdin dimethyl ester was used as an internal standard in a previous report (12), but we found biliverdin dimethyl ester to be unstable and to degrade rapidly.
Our data showed that the changes in HO activity differ from the changes in the mRNA levels (Figs. 6, 7). CoPP treatment increased the expression level of Hmox-1 in the kidney >20-fold, but it increased HO activity only 1.5-fold. This could be because the relative contribution of HO-2 to HO activity in the kidney is greater than that in the liver. These data indicate that this LC-MS/MS-based HO activity assay is superior to measuring HO-1 mRNA expression levels as a measure of total HO activity including both HO-1 and HO-2.
The comparison between this LC-MS/MS method and the spectrophotometric method demonstrated that only the LC-MS/MS method is sensitive enough to measure kidney HO activity of a single cisplatin-induced AKI mouse. The amount of bilirubin measured by spectrophotometry was below the quantitative limit. To measure the amount of bilirubin spectrophotometrically, pooling kidneys from >100 mice are required, indicating that this LC-MS/MS method is superior to the spectrophotometric method.
In conclusion, we report a LC-MS/MS-based method that we developed to evaluate HO activity. Our method is more sensitive and specific than the standard spectrophotometric method and can detect subtle changes in HO activity in small samples of renal or other tissues. This method is expected to be useful for estimating the effect of drugs on HO activity in small-animal models.
Our work was supported by Grants-In-Aid from the Japan Society of Promotion of Science (JSPS, 24659409, 24790832), Translational Research Network Program of Ministry of Education, Culture, Sports, Science and Technology of Japan (J140001192), the Naito Foundation, and Grants-In-Aid for Diabetic Nephropathy Research from the Ministry of Health, Labor and Welfare of Japan.
No conflicts of interest, financial or otherwise, are declared by the authors.
Author contributions: S. Iwamori, E.S., D.S., and K.Y. performed experiments; S. Iwamori, E.S., and N.T. analyzed data; S. Iwamori, E.S., S. Ito, H.S., and N.T. interpreted results of experiments; S. Iwamori, E.S., and N.T. prepared figures; S. Iwamori drafted manuscript; S. Iwamori, E.S., D.S., K.Y., S. Ito, H.S., and N.T. approved final version of manuscript; E.S., D.S., and N.T. provided conception and design of research; E.S., H.S., and N.T. edited and revised manuscript.
The authors are grateful to the members of the Division of Clinical Pharmacology and Therapeutics, Graduate School of Pharmaceutical Sciences, Tohoku University.
- Copyright © 2015 the American Physiological Society