Liquid Chromatography–Coupled Tandem Mass Spectrometry Based

---

ARTICLE pubs.acs.org/ac

Liquid Chromatography
Coupled Tandem Mass Spectrometry Based Assay to Evaluate Inosine-50 -monophosphate Dehydrogenase Activity in Peripheral Blood Mononuclear Cells from Stem Cell Transplant Recipients ric Levesque*,†,§ Isabelle Laverdiere,†,‡ Patrick Caron,†,‡ Felix Couture,§ Chantal Guillemette,*,†,‡ and E Pharmacogenomics Laboratory, †Centre Hospitalier de l’Universite Laval Research Center, ‡Faculty of Pharmacy, and § Faculty of Medicine, Laval University, Quebec, Canada ABSTRACT:

Combinations of immunosuppressive drugs are routinely used post-transplantation to prevent rejection and/or other complications and optimize outcomes. The prodrug ester mycophenolate mofetil (MMF) is frequently used in solid-organ and stem cell transplantation settings. A growing body of evidence supports therapeutic monitoring of this immunosuppressant to optimize its efficacy and reduce toxicity. Thus, pharmacodynamic monitoring of MMF is a strategy that could potentially improve patient outcomes. Pharmacodynamic measurements require evaluation of inosine-50 -monophosphate dehydrogenase (IMPDH) activity, the target enzyme of the active moiety mycophenolic acid. Various nonradioactive methods using chromatographic separations have been used to quantify xanthosine monophosphate, the catalytic product of the enzyme, to indirectly evaluate IMPDH activity. However, no methods have used mass spectrometry based detection, which provides more specificity and sensitivity. Here, we describe a liquid chromatography
coupled tandem mass spectrometry (LC
MS/MS) method for the quantification of xanthosine monophosphate and adenosine monophosphate (for normalization) in lysates of peripheral blood mononuclear cells (PBMCs) from hematopoietic stem cell transplant (HSCT) recipients. Linearity, precision, and accuracy were validated over a large range of concentrations for each compound. The method could measure analytes with high sensitivity, accuracy (93
116%), and reproducibility (CV < 7.5%). Its clinical application was validated in PBMC lysates obtained from healthy individuals (n = 43) and HSCT recipients (n = 19). This reliable and validated LC
MS/MS method could be a useful tool for pharmacodynamic monitoring of patients treated with MMF.

reaction.6 Given that this pathway is essential for the mitogenic function of T cells, MPA leads to cessation of their proliferation and thereby to immunosuppression.6,7 Much evidence supports the pharmacokinetic monitoring of MPA because this drug is characterized by a narrow therapeutic index and wide interindividual variability that can result in serious complications including rejection and severe toxicity. Therefore, individualization of MMF dose might help improve patient outcomes.8
11 Therapeutic drug monitoring based on a

C

ombinations of immunosuppressive drugs are routinely used in post-transplantation to prevent rejection and engraftment failure.1
3 Mycophenolate mofetil (MMF) is a standard immunosuppressive therapy used worldwide in solid-organ transplants and in hematopoietic stem cell transplant (HSCT) recipients.4,5 The therapeutic effect of the MMF ester prodrug is mediated by its active moiety, mycophenolic acid (MPA), and is based on the potent, selective, and reversible inhibition of inosine-50 -monophosphate dehydrogenase (IMPDH). IMPDH is the rate-limiting enzyme involved in de novo synthesis of guanosine nucleotides; IMPDH catalyzes the oxidation of inosine 50 -monophosphate (IMP) to xanthosine 50 -monophosphate (XMP) by a nicotinamide adenine dinucleotide (NAD)+-dependent r 2011 American Chemical Society

Received: September 10, 2011 Accepted: November 17, 2011 Published: November 17, 2011 216

dx.doi.org/10.1021/ac202404y | Anal. Chem. 2012, 84, 216–223

Analytical Chemistry

ARTICLE

Table 1. Chromatographic Conditionsa

pharmacokinetic
pharmacodynamic approach is one of diverse strategies that have been proposed to optimize MMF therapy.12
14 This dual monitoring evaluates a patient’s drug exposure and the efficacy of MPA. Because MPA targets IMPDH, this enzyme is a promising biomarker to assess the efficacy and toxicity of MMF.15
17 Various methods using radiometric detection or chromatographic separation methods coupled with UV detection have been described to measure IMPDH activity indirectly.18
23 Those methods are based on in vitro quantification of XMP, a catalytic product of IMPDH, in tissues, whole blood, or extracts of erythrocytes or peripheral blood mononuclear cells (PBMCs). The innovative aspect of this approach is that the concentration of intracellular adenosine monophosphate (AMP), rather than cell count or protein concentration, is used to normalize results.15,23 The normalization of IMPDH activity using AMP levels has several advantages, namely, that this approach better reflects the intracellular level of the enzyme released by disrupted cells and is less susceptible to interference by extracellular proteins or erythrocyte contamination. However, none of these methods have used mass spectrometry (MS) to detect XMP formation. Indeed, tandem MS is highly sensitive and selective compared to UV-based quantification. Here we developed a new liquid chromatography
coupled tandem MS (LC
MS/MS) method to measure IMPDH activity based on the quantification of XMP formation normalized by intracellular AMP level. We evaluated and validated this method in PBMC lysates from healthy volunteers and HSCT recipients.

mobile phase composition (%) elution time (min)

solvent A

solvent B

solvent C

0

85

15

0

2.00

30

15

55

3.50

30

15

55

3.51

85

15

0

7.50

85

15

0

a

Content of each solvent A
C: A, 3 mM ammonium formate in water; B, 3 mM ammonium formate in methanol; C, 100% methanol.

and 100 mM potassium chloride containing 0.25 mM of each of IMP and NAD+. The reaction was stopped by the addition of 120 μL of methanol, and then the mixture was centrifuged at 12 470g for 5 min at 4 °C. Samples were stored at
20 °C until LC
MS/MS. The effect of MPA on IMPDH activity was addressed by incubating lysates from healthy subjects and HSCT recipients with 0, 10, or 100 nM MPA. MPA-mediated inhibition of IMPDH activity was assessed in pooled cell lysates from healthy volunteers by adding MPA with final concentrations ranging from 0.01 to 1000 nM, similar to what was previously described.24 The potential influence of other often coadministered immunosuppressants was also tested by adding 100
800 ng/mL CsA, 5
50 ng/mL TCL, 0.02
20 μg/mL PSL, or 300
600 ng/mL Mtx. High-Performance Liquid Chromatography
Tandem Mass Spectrometry. Stock Solutions, Working Solutions, Calibration Standards, and Quality Control Samples. Stock solutions were prepared by dissolving XMP (10 mM) and Br-AMP (0.1 mM) in water and AMP (1 mM) in 0.5% NH4OH. All solutions were stored at
80 °C. Working solutions for each compound were prepared to yield target concentrations ranging from 0.25 to 35 μM. The internal standard Br-AMP working solution was prepared from stock solution at the time of assay by dilution in 95% water and 5% methanol. Calibration standards were prepared by diluting 50 μL of each working solution with 130 μL of solution consisting of 40 mM sodium dihydrogen phosphate (pH 7.4) and 100 mM potassium chloride, to which 50 μL of bovine serum albumin and 120 μL of methanol were added. Quality control (QC) samples of XMP/AMP were prepared in the absence of analytes (0 μM) or at low (2 μM each), medium (10 μM), and high (18 μM) concentrations and kept at
20 °C. Sample Preparation. Various compounds sharing structural similarities with AMP and XMP were tested to identify the most appropriate internal standard. Br-AMP has chromatographic properties similar to those of AMP and XMP (data not shown) and was thus chosen as the internal standard. Prior to chromatographic separation, QC samples, standard samples for the calibration curve, and clinical samples from individuals were thawed at room temperature; for analysis, 50 μL of sample was mixed with 150 μL of the Br-AMP internal standard. LC
MS/MS Conditions. High-performance liquid chromatography (HPLC) was performed with a Prominence system (Shimadzu Scientific Instruments Inc., Columbia, MD, U.S.A.) coupled to an API4000 mass spectrometer (Applied Biosystems, ON, Canada). Chromatographic separation was achieved with a Gemini C18 column (100 mm 4.6 mm, particle size 3 μm; Phenomenex). The column was eluted with various combinations of three solvents, namely, 3 mM ammonium formate in

’ EXPERIMENTAL SECTION Chemicals and Reagents. IMP, XMP, AMP, 8-bromoadenosine 50 -monophosphate (Br-AMP), MPA, cyclosporine A (CsA), tacrolimus (TCL), bovine serum albumin, sodium dihydrogen phosphate, and potassium chloride were from SigmaAldrich Canada (ON, Canada). All chemicals required for phosphate-buffered saline (PBS, pH 7.4) were also obtained from Sigma-Aldrich Canada except for NaCl and NAD+, which were from EMD Chemicals Inc. (QC, Canada). Methotrexate (Mtx) and prednisolone (PSL) were from Toronto Research Chemical Inc. (ON, Canada) and Steraloids (Newport, RI, U.S. A.), respectively. Ammonium formate was from Laboratoire Mat (QC, Canada). Methanol and Leucocep tubes were from VWR (QC, Canada). Ficoll-Paque solution was from GE Healthcare (QC, Canada). Gemini C18 columns were obtained from Phenomenex (Torrance, CA, U.S.A.). All reagents were of the highest grade commercially available. Sample Collection and Preparation. Venous peripheral blood samples from HSCT patients (n = 19) and healthy volunteers (n = 43) were collected and processed for PBMC preparation. Each participant provided written informed consent, and the institutional review board approved the research protocol. The PBMCs were isolated by density gradient centrifugation using Ficoll-Paque Plus and prepared within 4 h of blood collection. Cells, 10 million/mL, were then frozen at
80 °C until analysis. Assay Conditions and the Influence of Immunosuppressant Drugs. IMPDH activity was evaluated in PBMC lysates as described (Glander et al.)23 with slight modifications. Briefly, IMPDH activity was calculated based on enzymatic production of XMP normalized by intracellular AMP level. The assay was performed at 37 °C for 150 min. The reaction was initiated by adding 50 μL of thawed lysate to 180 mL of incubation mixture consisting of 40 mM sodium dihydrogen phosphate (pH 7.4) 217

dx.doi.org/10.1021/ac202404y |Anal. Chem. 2012, 84, 216–223

Analytical Chemistry water, 3 mM ammonium formate in methanol, and methanol. The flow rate was 0.9 mL/min. The chromatographic conditions are presented in Table 1. The retention times were as follows: XMP, 1.47 min; AMP, 2.7 min; Br-AMP, 2.81 min. The MS was operated in multiple reactions monitoring mode and equipped with a turbo-spay source. Electrospray ionization was performed in positive-ion mode with an ionization voltage of 5500 V, a declustering potential of 20 V, collision energy of 20 (XMP), 32 (AMP), or 27 V (Br-AMP), and source temperature of 450 °C. The analytes were detected according to the following mass of transitions: 365.1 f 97.1 (XMP), 348.1 f 136.1 (AMP), and 428.0 f 216.0 (Br-AMP). Method Validation. The intraday precision and interday precision were assessed based on the coefficient of variation (CV, %), whereas the accuracy (%) was determined as follows: [measured QC concentration/reference QC concentration]  100. The intra- and interday assessments were validated by analyzing five replicates of QC samples on three different days. The recovery of AMP and XMP was estimated by analyzing cell lysates containing a specific concentration (range, 2
18 μM) of each exogenously added analyte. The recovery was defined as [(obtained value/theoretical value)  100%] and expressed as a percentage. The theoretical value for each of XMP and AMP was expressed as [endogenous level + exogenous input]. The endogenous level refers to the XMP or AMP concentration measured in the incubated cell lysate samples. The exogenous input corresponds to the amount of XMP and/or AMP added at the various QC concentrations plus the XMP and AMP levels present in the commercial sources (see the Results) as determined in blank buffer solution. The long-term stability of analytes was determined in triplicate samples at all QC concentrations by analyzing samples stored at
20 °C for 4 months. To evaluate freeze/thaw stability, QC samples were also subjected to three cycles of freezing at
20 °C and thawing at room temperature (over 1 h). Stability of XMP and AMP was assessed by keeping QC samples at room temperature for 24 h and subsequently comparing the measured concentration to that of freshly extracted samples. Stock solution stability was investigated for analyte solutions stored at
20 °C. A nine-point calibration curve was prepared by spiking blank matrix with a specific amount of each analyte. The linear regression of XMP/Br-AMP and AMP/Br-AMP peak area ratios was weighted by 1/x2. The coefficient of determination (R) was used to evaluate the linearity of the calibration curve. The lower limit of quantification was defined as the minimum value at which the ratio of signal-to-noise was g5:1. Data Analysis. IMPDH activity was calculated according to XMP formation and normalized by AMP.23 The following equation was used (XMP and AMP concentrations are expressed in micromoles per liter and incubation time in seconds): IMPDH activity (μmol 3 s
1 3 mol
1 AMP) = [produced XMP  106/(incubation time  measured AMP)]. MPA-mediated inhibition of IMPDH activity was described by the residual IMPDH activity after MPA addition and was calculated as the percentage of enzymatic activity remaining in lysate containing MPA compared to the activity measured in the absence of MPA. Statistical analysis was performed using Sigma Plot 11 (San Jose, CA, U.S.A.). All statistical tests were two-tailed with a statistical significance level of 0.05. The comparisons between healthy subjects and HSCT recipients were performed with a Student’s t test. The between-group differences in residual IMPDH activity were evaluated using a Kruskal
Wallis nonparametric analysis of

ARTICLE

Figure 1. Typical chromatograms of bovine serum albumin matrix spiked with 35 μM XMP, 35 μM AMP, and 25 μM Br-AMP.

variance on ranks, with the assumption of unequal variance between groups. Results are expressed as percentages, median, and ranges.

’ RESULTS IMPDH activity was based on the formation of XMP from the substrate IMP and was normalized by the amount of AMP in cell lysates. We initially confirmed the selectivity of the chromatographic method for the targeted analytes, XMP and AMP, and the internal standard, Br-AMP, by evaluating the signal of all analytes in bovine serum albumin matrix. Figure 1 shows a representative chromatogram of the analytes in bovine serum albumin matrix. No additional peak was observed. We further tested for potential contamination by XMP and/or AMP in commercial sources of the substrates, IMP and NAD+; namely, we tested an IMP sample from Saccharomyces cerevisiae, IMP sample from muscle, NAD+ and NAD+ hydrate from SigmaAldrich, disodium IMP from Alfa Aesar, and lithium NAD+ from EMD Chemicals Inc. Interestingly, XMP and AMP were detected in all sources. As a consequence, when using the substrate with the smallest amount of contaminating XMP and AMP (see the Chemicals and Reagents section), levels of XMP and AMP were corrected for the background signal present in the commercial source of the substrates. The lowest effective concentration of reaction components was used to perform assays to minimize the AMP and XMP contaminants. A concentration of 0.25 mM IMP and 0.25 mM NAD+ led to linear formation of XMP and AMP (data not shown) under the assay conditions 218

dx.doi.org/10.1021/ac202404y |Anal. Chem. 2012, 84, 216–223

Analytical Chemistry

ARTICLE

Figure 2. Chromatographic separation of XMP and AMP in samples incubated in a buffer solution containing IMP and NAD+ before (A) and after (B) adding pooled cell lysates and in a sodium dihydrogen phosphate
potassium chloride solution free of IMP and NAD+ before (C) and after (D) adding of pooled cell lysates. In panel A, baseline resolution was achieved for the contaminants XMP and AMP with retention times of 1.47 and 2.7 min, respectively. In the absence of the substrate IMP and cofactor NAD+, a sample incubated without cell lysate did not show an XMP or AMP peak (C). In panel D, the peak seen at 2.7 min reflects the release of intracellular AMP by disrupted cells. The MRM transitions were 365.1 f 97.1 (XMP), 348.1 f 136.1 (AMP), and 428.0 f 216.0 (Br-AMP).

described above.20,22 Figure 2 presents chromatograms showing XMP and AMP in buffer solution and cell lysates supplemented (or not) with IMP and NAD+ substrates. The quantification assay was linear in the range of 0.25
35 μM of XMP and AMP (R: 0.998) with a lower limit of quantification of 0.25 μM for both analytes. Table 2 presents the inter- and intraday precision (CV, %) and accuracy (bias, %). All CVs were less than 7.5%, and accuracy ranged from 93.5% to 115.4%. The recovery of XMP ranged from 92.1% to 109.5%, and that of AMP was 98.7% to 105.4%. Table 3 presents values for long-term stability (4 months at
20 °C), freeze/thaw, and wet stability (24 h at room temperature). The stock solutions were stable for at least 4 months. The values are expressed as accuracy (%) compared with the initial concentration. Notably, IMPDH activity is stable in heparinized blood stored up to 24 h at room temperature prior to PBMC isolation.23

IMPDH Activity in PBMC Lysates from Healthy Individuals and HSCT Recipients. IMPDH activity was evaluated in PBMC

lysates from healthy volunteers (n = 43) and HSCT recipients (n = 19) (Figure 3). We defined the basal condition as the IMPDH activity determined at a random time for healthy individuals and immediately prior to intake of immunosuppressant drugs (trough level) for HSCT patients. The median basal IMPDH activity was higher in healthy individuals (74.8 μmol 3 s
1 3 mol
1 AMP) than in HSCT recipients (45.2 μmol 3 s
1 3 mol
1 AMP; p < 0.01). IMPDH activity was quite variable between individuals in the HSCT group (5.3-fold) compared to the healthy volunteers (4.1fold). Furthermore, the addition of MPA (10 or 100 nM) to lysates resulted in similar inhibition between healthy subjects (n = 12) and HSCT recipients (n = 6); as expected, greater inhibition was obtained using 100 nM MPA (p < 0.05) (Figure 3). The addition of other immunosuppressive drugs 219

dx.doi.org/10.1021/ac202404y |Anal. Chem. 2012, 84, 216–223

Analytical Chemistry

ARTICLE

Table 2. Precision of the Assaya AMP QC endo intraday

interday

recovery

QC low

XMP

QC med

QC high

QC endo

QC low

QC med

QC high

AMP
XMP-spiked (μmol/L)

0

2

10

18

0

2

10

18

CV (%)

6.65

2.92

5.83

4.00

7.29

4.13

4.46

4.24

accuracy (%)

103.34

106.92

106.93

105.74

99.83

93.54

103.00

115.39

AMP
XMP-spiked (μmol/L)

0

2

10

18

0

2

10

18

CV (%)

6.37

4.75

5.17

4.90

6.26

5.14

4.11

4.62

accuracy (%)

103.35

106.24

107.24

103.03

100.46

94.74

104.47

113.05

(%)

N/A

102.0

105.4

98.7

N/A

92.1

104.5

109.5

The intra- and interday validation was performed by analyzing five replicates of QC samples on three different days at different concentrations. The QC endo level was determined by the concentration of XMP and AMP measured in the incubated cell lysates alone. The QC low, medium (med), and high were obtained by spiking QC endo samples with XMP and AMP at 2, 10, and 18 μM, respectively.

a

Table 3. Stability of the Analytesa AMP QC endo long-term

a

QC low

XMP QC med

QC high

QC endo

QC low

QC med

QC high

CV (%)

2.9

0.5

1.6

1.5

3.9

2

2.14

accuracy (%)

113.3

100.2

98.2

86.8

98.3

83.6

109.2

2.5 111.5

freeze/thaw

CV (%)

2.2

2.2

2.18

1.4

3.8

1.4

2.9

1.7

accuracy (%)

114.6

104.8

104.2

95.1

95.4

85.2

108.6

107.6

wet

CV (%)

4.96

2.26

5.38

1.52

5.77

1.34

4.10

1.04

accuracy (%)

106.80

107.39

108.29

107.08

103.51

94.53

104.14

116.15

See the footnotes of Table 2 for details.

Figure 3. (A) Example of the variability of IMPDH activity for healthy individuals (n = 43; samples collected randomly) and HSCT recipients (n = 19; samples collected at the immunosuppressant drug trough levels) determined under basal conditions (see the Experimental Section). Each box plot shows the median, interquartile range, and outliers for IMPDH activity for each group. The median basal IMPDH activity in healthy individuals was 74.8 μmol 3 s
1 3 mol
1 AMP (range 32.8
134.5; 95% CI 71.1
84.6 μmol 3 s
1 3 mol
1 AMP) compared to 45.2 μmol 3 s
1 3 mol
1 AMP (range 14.7
79.0; 95% CI 33.4
52.9 μmol 3 s
1 3 mol
1 AMP) in HSCT recipients, p < 0.01. (B) Example of the variability of residual IMPDH activity for healthy volunteers (n = 12) and HSCT patients (n = 6) in lysates supplemented with 10 or 100 nM MPA. At 10 nM MPA, the IMPDH activity in healthy volunteers and HSCT patients was, respectively, 73.5 μmol 3 s
1 3 mol
1 AMP (range 67.0
84.0) vs 78.0 μmol 3 s
1 3 mol
1 AMP (range 72.0
88.0), p = 0.11. At 100 nM MPA, healthy volunteers and HSCT patients display, respectively, an IMPDH activity at 40.0 μmol 3 s
1 3 mol
1 AMP (range 29.0
55.0) and 40.0 μmol 3 s
1 3 mol
1 AMP (range 9.0
53.0), p = 0.61.

dependent manner, with an IC50 of 6.5 nM. Similarly, the IMPDH activity determined in PBMCs from an HSCT recipient receiving MMF-based immunosuppressive therapy was inhibited

(CsA, Mtx, PSL, TCL) to incubating pooled lysates did not interfere with IMPDH activity, as expected (Figure 4). In fact, IMPDH activity was reduced by MPA in a concentration220

dx.doi.org/10.1021/ac202404y |Anal. Chem. 2012, 84, 216–223

Analytical Chemistry

ARTICLE

samples from transplant patients undergoing immunosuppressive therapy that includes MMF. Of significance, while validating our method we found XMP and AMP contamination in all commercial sources. This did not affect the ability of our method to measure XMP in those samples, however, because we corrected the data based on a blank sample. Previously described methods did not provide such information, and therefore UV detection-based methods may potentially overestimate the XMP concentration. Our observations also probably reflect the higher sensitivity of MS. The potential clinical application of our method was confirmed using cell lysates from healthy volunteers and HSCT recipients. Previous studies performed with samples from renal transplant patients supported an inverse relationship between systemic MPA concentration and IMPDH residual activity, with the lowest enzymatic levels at the maximal plasma concentration of MPA.12,16,26 We therefore tested the ability of our method to quantify XMP and AMP in PBMC extracts from HSCT recipients receiving various immunosuppressive drugs and supplemented in vitro with MPA (0.01
1000 nM), which reflects the range of MPA concentration in white blood cells in samples for which the concentration of MPA in plasma ranged from 0.032 to 3200 mg/L.27,28 This large range of concentrations included MPA plasma levels that exceed those reported in kidney organ transplants, which usually range from ∼0.01 to 5 mg/L (at trough level) but may vary up to 30 mg/L (at peak level following a dose of