Development and validation of a LC-MS/MS method for in vivo quantification of meta-iodobenzylguanidine (mIBG)

Abstract

meta-iodobenzylguanidine (mIBG) is a radiopharmaceutical used for the diagnosis and treatment of neuroendocrine cancers. Previous quantification of mIBG in biodistribution and pharmacokinetic studies mainly relied on the use of radiolabeled mIBG, which involves the handling of highly radioactive materials. The goal of this study was to develop a nonradioactive analytical method for quantifying mIBG in mouse plasma and tissue homogenates using high performance liquid chromatography-tandem mass spectrometry (LC-MS/MS). Samples were prepared for analysis using a protein precipitation method. Mass spectrometry analysis was performed using 4-hydroxyphenformin as the internal standard, and the mass-to-charge transitions were 276.1 → 217.0 for mIBG and 222.1 → 121.0 for 4-hydroxyphenformin. The quantification limit of mIBG was 0.98 ng/mL, and the method was linear up to 500 ng/mL. The accuracy, inter-day and intra-day precision were 96–112%, 5.5–14.4%, and 3.7–14.1%, respectively, suggesting that the method was accurate and precise in quantifying mIBG at multiple concentrations in mouse plasma and liver homogenates. The extraction recovery was 96–106% and the matrix effect was 95–110%, indicating that the method was reproducible in quantifying mIBG with minimal impact from the biological matrices. In summary, we have developed and validated a fast, high-throughput quantification method of non-radiolabeled mIBG using LC-MS/MS. This method is reproducible, accurate, and precise, and can be used to quantify mIBG in plasma and tissue matrices to determine the pharmacokinetics and biodistribution of mIBG in preclinical animal models.

1. INTRODUCTION

Two isotopes of radioiodine-labeled meta-iodobenzylguanidine (mIBG) are used for the diagnosis and targeted radiotherapy of neuroendocrine tumors. ¹²³I-labeled mIBG is used for whole body imaging of neuroblastoma to detect primary tumors and identify metastatic sites [1]. ¹³¹I-labeled mIBG is approved by the US Food and Drug Administration (FDA) for the treatment of advanced pheochromocytoma and paraganglioma [2]. ¹³¹I-mIBG is also under investigation in multiple clinical trials for the treatment of high-risk neuroblastoma [3, 4]. These cancers originate from the neural crest and exhibit high expression of the norepinephrine transporter (NET), a transmembrane protein involved in the transport of norepinephrine across cell membranes [5]. As a structural analog of norepinephrine, ¹³¹I-mIBG is selectively transported into the tumor cells through the NET pathway, where radioactive decay of 131I produces DNA damage, cell death, and tumor necrosis [2, 5].

Clinically, mIBG is administered to patients by intravenous infusion. The drug does not undergo significant hepatic metabolism in vivo, and the majority of the administered dose is excreted into the urine unchanged [3, 6]. Besides tumor cells, mIBG accumulates in several normal tissues (e.g., liver, salivary glands, and heart), which can lead to tissue toxicities [3, 7, 8]. Recently, we and others have shown that mIBG is a substrate of the polyspecific organic cation transporters and that these transporters could play important roles in the pharmacokinetics and tissue-specific distribution of mIBG [9–12]. To further understand the roles of these transporters in the disposition and tissue-specific uptake of mIBG in vivo, it is necessary to perform biodistribution and pharmacokinetic studies of mIBG in appropriate animal models. Previous analytical methods have primarily relied on mIBG labeled with radioactive iodine (¹²³I or ¹³¹I) to quantify its levels in plasma and tissue samples [11, 13]. However, the use of radioactive mIBG in the research setting is challenging and requires measures to protect personnel from radiation exposure. Besides the safety concerns of handling radioactive iodine, the radiolabeled form of mIBG is not readily available due to the limited number of facilities that can synthesize the compound. Further, the short half-lives of ¹²³I and ¹³¹I (13 hrs and 8 days, respectively) make the use of radiolabeled mIBG challenging due to its short shelf life. Last, these methods detect the radiation emitted by the iodine radionuclide, and they typically do not differentiate between intact mIBG and the free iodide that can dissociate from mIBG in the body [14]. To support preclinical pharmacokinetic and biodistribution studies of mIBG, a bioanalytical method that measures concentrations of non-radiolabeled mIBG in biological matrices is highly desirable.

Liquid chromatography-tandem mass spectrometry (LC-MS/MS) has been routinely used to quantify drugs in pre-clinical and clinical studies [15]. In particular, highly sensitive and specific LC-MS/MS instrumentation has been widely utilized for rapid and accurate determination of drug concentrations in biological matrices to determine their pharmacokinetic profiles [15–18]. Additionally, LC-MS/MS methods developed to quantify tissue concentrations of drugs in transporter-knockout animals have provided novel insights into the mechanisms involved in their disposition [18–21]. To further investigate the molecular mechanisms involved in mIBG pharmacokinetics and tissue uptake, we developed and validated a reliable and simple LC-MS/MS method for the in vivo quantification of non-radiolabeled mIBG in mouse plasma and liver homogenates.

2. MATERIALS AND METHODS

2.1. Chemicals and animals

mIBG was purchased from Sigma Aldrich (St. Louis, MO). 4-hydroxyphenformin (4OHP) was purchased from USV (Mumbai, India). High-grade acetonitrile, methanol, and formic acid were obtained from Thermo Fisher (Rockford, IL). Plasma and liver tissues were collected from adult FVB mice. The mice were housed in a specific pathogen-free facility and fed a standard diet. The mice were euthanized by CO₂ overdose followed by cardiac puncture to collect whole blood using a heparinized needle and syringe. The blood was centrifuged at 2,000 x g for 10 minutes and the plasma supernatant was collected. The liver was subsequently collected, snap-frozen, and placed in −80°C for storage. The liver samples were homogenized at a 1:5 ratio of liver weight to phosphate buffered saline volume with an Omni Bead Ruptor from Omni International (Kennesaw, GA). All animal studies were approved by the Institutional Animal Care and Use Committee of the University of Washington.

2.2. Instrument and LC-MS/MS conditions

An Agilent 1290 series HPLC system from Agilent Technologies (Santa Clara, CA) was used to separate mIBG and 4OHP with a Zorbax Eclipse Plus C18 column (2.1 × 50mm, 1.8 µm) from Agilent Technologies (Santa Clara, CA) accompanied with a Security Guard column from Phenomenex (Torrance, CA). The column was maintained at room temperature and the autosampler at 4°C. The mobile phases consisted of 0.1% (vol/vol) formic acid in water (A) and acetonitrile (B) with gradient elution at 0.4 mL/min flow as follows: 3% B until 0.1 min, increased to 90% B by 2 min and maintained at 90% B for 1 min. The composition then returned to 3% B over 0.1 min and holding for 1.9 min for a total run time of 5 min.

Mass spectrometry analysis was performed on an Agilent 6410 Triple Quad mass spectrometer from Agilent Technologies (Santa Clara, CA) operated in the positive electrospray ionization mode. The following parameters were set for the mass spectrometer: capillary voltage at +5 kV, gas flow at 11 L/min, and gas temperature at 300°C. The multiple reaction monitoring (MRM) mode was used to detect and quantify mIBG and 4OHP with mass-to-charge (m/z) transitions at 276.1 → 217.0 and 222.1 → 121.0, respectively. For mIBG, the fragmentor voltage was 129 V, the collision energy was 21 V, and the cell accelerator voltage was 4 V. For 4OHP, the fragmentor voltage was 98 V, the collision energy was 25 V, and the cell accelerator voltage was 4 V. Instrument control and data processing were performed using Agilent MassHunter software from Agilent Technologies (Santa Clara, CA).

2.3. Stock and working solutions

Stock standard solutions were prepared in methanol at 5 mg/mL for mIBG and 1 mg/mL for 4OHP. Diluted “working” standard solutions were prepared in methanol with concentrations of 500 µg/mL for mIBG and 20 µg/mL for 4OHP. A protein-precipitation solution was prepared by a 2000-fold dilution of the working solution of 4OHP in acetonitrile.

2.4. Calibration standards, linearity, and quality control samples

The calibration standards were prepared from the working solutions of mIBG. The highest calibration standard solution was prepared such that a 2 µL addition to 20 µL of blank plasma or liver homogenate matches a concentration of 500 ng/mL of mIBG. The remaining calibration standard solutions were prepared via 2-fold serial dilution. The lower limit of quantification (LLOQ) was defined as the lowest sample with S/N > 5 and accuracy within 15% of the nominal concentration. The standard curve ranged from 0.98 to 500 ng/mL mIBG in plasma and liver, and the low quality control (QC_low), medium quality control (QC_med), and high quality control (QC_high) samples contained 2, 25, and 200 ng/mL of mIBG, respectively. A zero standard sample with only 4OHP and a double blank sample with neither mIBG nor 4OHP were also prepared.

2.5. Sample preparation

Samples were prepared by a protein precipitation method. For each sample, an aliquot of 20 µL of blank plasma or liver homogenate was transferred to a 1.5 mL Eppendorf tube, followed by addition of 178 µL of the protein-precipitation solution. The samples were spiked with 2 µL of acetonitrile with or without mIBG. The samples were vortexed for approximately 5 seconds and centrifuged at 18,000 x g for 10 min at 4°C. A 100 µL aliquot of the organic supernatant was transferred to a 96-well plate, followed by evaporation under nitrogen gas at 40°C for approximately 20 mins. The samples were then re-constituted with 200 µL of 10% acetonitrile in water and analyzed by LC-MS/MS.

2.6. Method validation

Method validation was performed based on the FDA bioanalytical method validation guidance for industry.[22]

2.6.1. Specificity & selectivity

Specificity was evaluated by analyzing the blank mouse plasma or liver homogenate for interfering peaks. Selectivity was determined by comparing the signal in the blank mouse plasma or liver homogenate samples to the samples spiked with mIBG at the LLOQ concentration.

2.6.2. Calibration Model and Linearity

The calibration standard curves described in Section 2.4 were run independently four times with duplicate samples for each run. A linear equation with 1/x² weighting was fit to the data using least-squares linear regression (y = b₀ + b₁x), where y is the mIBG to 4OHP peak area ratio (PAR), b₀ is the intercept, b₁ is the slope, and x is the nominal mIBG concentration in ng/mL. The runs were considered acceptable when 80% of the calibration standards were within ± 15% of the nominal concentrations.

2.6.3. Accuracy and precision

Accuracy was expressed as the deviation of the measured mean concentration from the nominal concentrations. The inter- and intra-day precision were calculated using a root mean squares approach based on the following analysis of variance (ANOVA) equations:

$$Intra-day\hspace{0.17em}precision=\frac{\sqrt{M{S}_{intra}}}{mea{n}_{all}}x\hspace{0.17em}100\%$$ (1)

$$Inter-day\hspace{0.17em}precision=\frac{\sqrt{\left(M{S}_{inter}-M{S}_{intra}\right)/n}}{mea{n}_{all}}x\hspace{0.17em}100\%$$ (2)

where MS_intra and MS_inter are the mean squares within a group and between a group, respectively. mean_all is the arithmetic mean of all the measured concentrations and n is the number of samples used for the tested concentration. Accuracy and precision were determined at each QC level and were considered acceptable if the values were within ± 15% of the nominal concentrations.

2.6.4. Matrix effect and recovery

Three types of samples were used to determine the matrix effect and recovery from mouse plasma and liver homogenates (prepared as described in Section 2.5). “Extracted” samples were defined as samples that were spiked with mIBG before adding the protein precipitation solution to the blank plasma or liver homogenates. Similarly, “non-extracted” samples were samples that were spiked with mIBG after the protein precipitation solution was added. Last, “neat” samples were prepared using 20 µL of acetonitrile instead of mouse plasma or liver homogenates. The matrix effect was determined by the ratio of the non-extracted samples to that of the neat samples. The extraction recovery was determined by the ratio of the extracted samples to that of the non-extracted samples. Three independent experiments were run with three to six replicates for QC_low, QC_med, and QC_high to capture the matrix effect and recovery.

3. Results and Discussion

3.1. Method development

The in vivo method presented in this manuscript was adapted from an in vitro method developed to quantify mIBG from cell lysates [10]. In the in vitro quantification method, glyburide was used as an internal standard. The structure of glyburide is considerably different from mIBG, which may prove problematic in complex matrices such as plasma and tissue homogenates. Thus, we developed an in vivo quantification method using 4OHP as an internal standard, which is structurally more similar to mIBG than glyburide (Fig. 1). The chromatographic conditions were optimized to achieve distinguishable and sharp peaks for mIBG and 4OHP. The retention times were 1.9 and 1.6 mins for mIBG and 4OHP, respectively, in all matrices with a total run time of 5 min (Fig. 2), allowing for fast and efficient analysis of multiple samples. The mass spectrometry conditions were optimized for mIBG and 4OHP. The MS/MS product ion spectra for mIBG and 4OHP and their fragmentation patterns are shown in Figure 1. The most abundant ions in the product-ion mass spectra were formed by the loss of the guanidine groups with a m/z of 217.0 for mIBG and a m/z of 121.0 for 4OHP. Therefore, the m/z transitions used for quantification of mIBG and 4OHP were 276.1 → 217.0 and 222.1 → 121.0, respectively.

Figure 1.

MS/MS product ion spectra from monitoring a precursor ion of 276.1 m/z for mIBG (A) and a precursor ion of 222.1 m/z for 4OHP (B).

Figure 2.

LC-MS/MS chromatograms of mIBG and 4OHP. The MRM transition monitored for 0.98 ng/mL of mIBG were 276.1 → 217.0 m/z in plasma (A) and liver homogenate (B). The MRM transition monitored for 20 ng/mL of 4OHP were 222.1 → 121.0 m/z in plasma (C) and liver homogenate (D).

The proposed LC-MS/MS method provides multiple benefits compared to the use of radioactive mIBG for quantification. First, it completely eradicates the hazards of handling radioactive iodine. Second, it substantially reduces the cost and time constraints of using a radioactive isotope with a short half-life. Third, it eliminates the confounding effects of the radioactive decay half-life when analyzing pharmacokinetic parameters. Last, it directly quantifies intact mIBG rather than detecting the radiation emitted by the iodine. The proposed LC-MS/MS method required a relatively small volume of mouse plasma and liver homogenates for a one-step protein precipitation assay that is simple, fast, and inexpensive.

3.2. Method validation

3.2.1. Specificity and selectivity

Representative chromatograms of blank and LLOQ-spiked plasma and liver homogenates are shown in Figure 2. The LLOQ was determined in both plasma and liver homogenates to be 0.98 ng/mL. A small and inconsistent interfering peak of mIBG was observed in some blank samples but was deemed negligible as the mean peak area was less than 20% of the LLOQ. The chromatograms showed that the method was selective in both matrices with no expected interference to the quantification of mIBG.

3.2.2. Calibration model and linearity

A linear equation was fit to the data using a weighting factor of 1/x² to describe the concentration and PAR relationship. The calibration curves were linear within the concentration range of 0.98–500 ng/mL of mIBG with correlation coefficients ranging from 0.9881–0.9978 and 0.9893–0.9972 for mouse plasma and liver homogenates, respectively (Fig. 3). The mean (± S.D.) regression parameters were b₀ = 0.009 ± 0.003 and b₁ = 0.062 ± 0.018 for plasma and b₀ = 0.0188 ± 0.0063 and b₁ = 0.0584 ± 0.0201 for liver homogenates. The back-calculated concentrations for both matrices were accurate to within 5% and had coefficients of variation (% CV) less than 10% for each of the standard concentrations, indicating a good fit to the data with the weighted linear regression (Table 1).

Figure 3.

Calibration curves for mIBG in mouse plasma and liver homogenates. Data represent the mean ± S.D. of three (liver) and four (plasma) independent experiments conducted in duplicates.

Table 1.

Back-calculated concentrations from calibrators run in duplicates on three (liver) and four (plasma) independent experiments.

Nominal Concentrations (ng/mL)	Plasma		Liver
	Mean ± S.D. (%CV) (ng/mL)	Accuracy (%)	Mean ± S.D. (%CV) (ng/mL)	Accuracy (%)
0.98	0.98 ± 0.04 (4%)	101	0.97 ± 0.04 (4%)	100
1.95	1.97 ± 0.09 (5%)	101	2.0 ± 0.1 (6%)	102
3.91	3.9 ± 0.2 (4%)	99	3.9 ± 0.2 (4%)	99
7.81	7.6 ± 0.2 (3%)	97	7.8 ± 0.4 (6%)	100
15.6	16 ± 1 (6%)	102	16.3 ± 0.7 (4%)	105
31.3	30.8 ± 0.9 (3%)	99	31 ± 2 (5%)	99
62.5	62 ± 3 (5%)	99	62 ± 2 (3%)	98
125	128 ± 6 (5%)	102	125 ± 4 (3%)	100
250	256 ± 7 (3%)	103	250 ± 20 (7%)	99
500	500 ± 40 (8%)	101	490 ± 40 (9%)	98

3.2.3. Accuracy and precision

Accuracy and inter-day and intra-day precision were determined using the extracted samples from plasma and liver homogenates at three concentration levels. The method quantified the mIBG concentrations with accuracies ranging from 96–112% in both matrices (Table 2). Furthermore, the inter-day and intra-day precision values were within 15% for each mIBG concentration in both matrices, indicating that the method was accurate and precise in quantifying mIBG concentrations in both mouse plasma and liver homogenates.

Table 2.

Intra- and inter-day precision and accuracy for mIBG in mouse plasma and liver homogenates conducted in three to six replicates on four independent experiments.

Matrix	Nominal Concentrations (ng/mL)	Intra-day Precision (%)	Inter-day Precision (%)	Accuracy (%)
Plasma	2	14.4	10.0	112
	25	8.4	6.4	105
	200	5.3	10.2	99
Liver	2	11.3	3.7	96
	25	7.2	9.8	99
	200	5.5	14.1	101

3.2.3. Matrix effect and recovery

The matrix effect and extraction recovery were determined in plasma and liver homogenate at three concentration levels. The recovery for the extraction in all concentration levels ranged between 98–106% for plasma and 96–102% for liver homogenates (Table 3). The matrix effect at each tested concentration ranged between 96–100% for plasma and 95–110% for liver homogenates. These data suggest that the extraction method is reproducible, and the matrix from the plasma and liver homogenate has negligible effects on quantification of mIBG.

Table 3.

Recovery and matrix effect in mouse plasma and liver homogenates from four independent experiments.

Matrix	Nominal Concentrations (ng/mL)	Recovery (mean ± SD, (%CV) %)	Matrix Effect (mean ± SD, (%CV) %)
Plasma	2	106 ± 7 (6%)	100 ± 10 (13%)
	25	104 ± 8 (8%)	96 ± 6 (6%)
	200	98 ± 3 (3%)	100 ± 4 (4%)
Liver	2	96 ± 7 (7%)	110 ± 10 (11%)
	25	100 ± 6 (6%)	95 ± 7 (7%)
	200	102 ± 2 (2%)	100 ± 10 (11%)

4. Conclusion

We report the development and validation of a LC-MS/MS method for the quantification of mIBG in plasma and tissue samples. Previously published methods for quantification of mIBG in preclinical species relied on counting the gamma radiation emitted by the radioactive iodine in mIBG. The proposed method for quantification of non-radiolabeled mIBG allows for a safe, fast, and high-throughput analysis of mIBG concentrations in preclinical species. The validation of this method indicates that this method is reproducible, accurate, precise, and is applicable to both plasma and tissue matrices.

Highlights:

• A novel LC-MS/MS method was developed for quantification of meta-iodobenzylguanidine (mIBG)

• The method consists of a reproducible and simple extraction with a 5 min run time

• Method validation was performed with mouse plasma and liver homogenate

• The method quantifies mIBG selectively, accurately, and precisely in plasma and tissue matrices

Abbreviations:

4OHP

4-hydroxyphenformin

CV

coefficient of variation

FDA

Food and Drug Administration

LC-MS/MS

Liquid chromatography-tandem mass spectrometry

LLOQ

lower limit of quantification

mIBG

meta-iodobenzylguanidine

NET

norepinephrine transporter

PAR

peak area ratio

QC

quality control

SD

standard deviation

S/N

signal-to-noise ratio