A validated liquid chromatography‐tandem mass spectroscopy method for the quantification of tolinapant in human plasma

Abstract

Tolinapant (ASTX660), a pan‐selective inhibitor of apoptosis protein antagonist with dual cIAP/XIAP activity, was identified as a clinical candidate in preclinical efficacy, pharmacokinetic and safety studies. In order to assess tolinapant in first‐in‐human Phase I/II clinical trials, a validated bioanalytical method was required to determine plasma pharmacokinetics. Tolinapant and d₄‐tolinapant were extracted from human plasma using liquid‐liquid extraction. Separation chromatography was performed on a Acquity BEH C18 1.7 µM, 50 mm × 2.1 mm i.d. column, using a mobile phase of 0.1% formic acid in water and 0.1% formic acid in acetonitrile. Mass spectrometry detection was performed by positive turbo ion spray ionisation, in multiple reaction monitoring mode. The method was validated according to the US Food and Drug Administration (FDA) guidelines. The method has a quantifiable linear range of 1–500 ng/mL (r ² = 0.999). The intra‐ and inter‐day coefficients of variation were < 11.4%. Dilution QC samples agreed with prepared concentrations, with a precision of 1.5% and accuracy of 101%. Tolinapant mean recoveries ranged from 85.1–94.4 % with negligible matrix effects. A highly sensitive and selective LC‐MS/MS bioanalytical method was developed and validated. The method was successfully applied in Phase 1/2 clinical trials to determine the human pharmacokinetic profile of tolinapant.

Abbreviations

AUC

area under curve

cIAP

cellular IAP

CV

coefficient of variation

FDA

US Food and Drug Administration

HPLC

high performance liquid chromatography

HQC

high quality control

IAP

inhibitor of apoptosis

LC‐MS/MS

liquid chromatography‐tandem mass spectrometry

LLOQ

lower limit of quantification

LQC

low quality control

MRM

multiple reaction monitoring

MS

mass spectrometry

MQC

middle quality control

QC

quality control

SMAC

second mitochondria derived activator of caspases

SPE

solid phase extraction

XIAP

X‐linked IAP

1. INTRODUCTION

A distinctive hallmark of human cancer in many tumour types is the evasion of apoptosis, ⁵ with this dysregulation of programmed cell death considered a contributing cause of resistance to available therapies. ³ The inhibitors of apoptosis (IAP) family of proteins are important regulators of cell survival in cancer, ² regulating various cellular processes such as cell death, cell proliferation and cell migration, which contribute to tumour development and growth. Examples of such IAP proteins are cellular IAP (cIAP) 1 and 2, and X‐linked IAP (XIAP). Studies have demonstrated that pan‐selective IAP antagonists of both cIAP and XIAP are required for efficient cell death induction, in comparison to c‐IAP selective antagonists, ⁸ such as SMAC (second mitochondria‐derived activator of caspases) peptidomimetics. Potent orally bioavailable non‐ peptidomimetic antagonists with dual cIAP/XIAP activity were discovered using fragment‐based drug design. ¹ Further medicinal chemistry optimisation resulted in the development of the clinical candidate tolinapant. ⁶ Preclinical pharmacokinetic studies demonstrated that tolinapant was bioavailable and efficacious in vivo xenograft models. ⁹ We present a selective and robust validated liquid chromatography‐tandem mass spectrometry (LC‐MS/MS) method for the quantification of tolinapant in human plasma, which is supporting Phase I/II clinical trials in adults with advanced solid tumours or lymphomas. ⁷

2. EXPERIMENTAL SECTION

2.1. Reference standards

Tolinapant reference standard, (batch A0478‐050‐03, chemical purity 98.7% w/w) was supplied by Aptuit Limited and was stored at ‐20^○C with desiccant (Figure 1). The internal standard, d₄‐tolinapant (batch 93203‐01, isotopic purity 88.4%), was supplied by Aptuit and stored at 4 ^○C in a closed container with desiccant (Figure 1).

FIGURE 1.

Chemical structure of tolinapant (A) and d₄‐tolinapant (B)

2.2. Chemicals and reagents

HPLC grade acetonitrile was purchased from Fisher Chemicals. Methyl tert‐butyl ether and certified A.C.S formic acid were obtained from Sigma. Deionized distilled water was generated using an in‐house Milli‐Q system. Drug‐free (blank) human plasma and lipemic plasma were obtained from BioreclamationIVT, Inc. (anticoagulant K₂ EDTA) and stored at ‐20^○C. Human whole blood was obtained from BioreclamationIVT, Inc. and stored at 4^○C.

2.3. Mass spectrometry instrumentation

A SIL‐30AC HT autosampler and LC‐30AD high‐performance LC pump (Shimadzu), coupled with an API 5500 mass spectrometer (Sciex) was used. ABSciex Analyst software (version 1.6) controlled the LC‐MS/MS system.

2.4. Chromatography

The chromatographic separation of analytes was performed on an Acquity BEH C18 (1.7 µM, 50 mm x 2.1 mm i.d.) analytical column, protected by a 0.2 µM pre‐filter (Waters), at ambient temperature. The mobile phase consisted of 0.1% formic acid in water and 0.1% formic acid in acetonitrile. The elution starts with isocratic elution at 15% organic for 0.2 min, followed by a step gradient from 15% organic, increasing to 50% organic from 0.2 to 1.0 min, then increasing from 50% to 95% organic from 1.0 to 1.5 min, remaining at 95% organic for 0.7 min and then linearly decreasing to 15% over 0.25 min. An equilibration time of 0.95 min at 1% organic was maintained post‐injection. The mobile phase was delivered at a rate of 0.6 ml/min. The injection volume for each sample was typically 5 µl.

2.5. Mass spectrometry

Tolinapant and d₄‐tolinapant were ionised using turbo ion spray ionisation in positive ion mode. Final conditions of Turbolon spray voltage (5000 V), Turbolon spray temperature (500^○C), curtain gas (30 psi), collision gas (8 psi), nebulizing gas (50 psi) and auxiliary gas (50 psi), were optimal for the detection for parent and product ions.

2.6. Data acquisition and quantification

Analytes were detected using multiple reaction monitoring. The characteristic ion dissociation transition for tolinapant was 540.5 → 439.3 with a declustering potential of 110 V and collision energy of 38 V. The internal standard d₄‐tolinapant ion dissociation transition was 554.5 → 443.3 with a declustering potential of 120 V and collision energy of 40 V. The MS response (peak area ratio of standard to internal standard) of each calibration standard was plotted against the prepared concentration and subjected to a weighted (1/x ²) linear regression analysis to generate a calibration curve.

2.7. Stock solutions

Separate calibration and quality control (QC) stock solutions of tolinapant were prepared at 1 mg/ml in acetonitrile: water (50:50 v/v) and stored at 4^○C for up to 6 months. Dilutions of the stock were made in acetonitrile: water (50:50 v/v), as appropriate, to give working solutions with concentrations of 0.02–10 µg/ml. Working solutions were stored at 4^○C for 1 month. A d₄‐tolinapant stock solution was prepared at 1 mg/ml in acetonitrile:water (50:50 v/v) and stored at 4^○C for up to 6 months. Dilutions of the d₄‐tolinapant stock solution were made in acetonitrile:water (50:50 v/v) to give working solutions with concentrations of 10000, 100 and 300 ng/ml and were stored at 4^○C for up to 2 months.

2.8. Preparation of calibration and QC samples

A calibration curve in the final concentration range of 1–500 ng/ml was prepared fresh daily by serial dilution of the standard calibration stock in blank plasma. QC samples were prepared in human plasma following serial dilution of the QC stock to give high (HQC: 360 ng/ml), medium (MQC: 40 ng/ml), low (LQC: 3 ng/ml) and lower limit of quantification (LLOQ: 1 ng/ml) final concentrations and stored at ‐70^○C.

2.9. Sample extraction

Plasma samples (50 µl) were thawed at room temperature and vortex mixed. Internal standard working solution (d₄‐tolinapant, 30–100 ng/ml) was added to all samples, except the blanks. Blanks were spiked with 50 µl acetonitrile:water (50:50 v/v). Water (100 µl) and methyl tert‐butyl ether (700 µl) were added to all samples, followed by vortexing for 3 min, starting at a vortex speed of 1500 rpm and increasing to 1800 rpm. Samples were centrifuged at 2500 rpm for 5 min. The supernatant (100 µl) was evaporated to dryness under nitrogen using an SPE concentrator at 35^○C and a flow rate of 40 L/min, followed by reconstitution in acetonitrile: water (10:90, v/v) containing 0.2% formic acid.

2.10. Method validation

The method was validated according to the US Food and Drug Administration guidelines for bioanalytical method validation. ⁴ Validation consisted of three consecutive precision batches, each batch containing two human plasma calibration curves, blank human plasma, and replicate (n = 6) LQC, MQC and HQC samples, in order to assess precision and accuracy. System suitability samples were included prior to batch analysis.

2.10.1. Selectivity and matrix effect

Six independent batches of blank human plasma were assessed for endogenous interferences, as blanks and spiked with tolinapant at the LLOQ concentration. The matrix effect was assessed by six batches of an extracted blank matrix, spiked post‐extraction at the LLOQ concentration, against standard working solutions at the same concentration. The matrix effect was determined using the normalised matrix factor, as calculated with the following equation:

$$\begin{array}{ccc}\hfill \text{Matrix}\text{Factor}& =& \text{Post}\text{extracted}\text{spiked}\text{peak}\text{area}\text{ratio}/\hfill \\ & & \text{Working}\text{solution}\text{mean}\text{peak}\text{area}\text{ratio}\hfill \end{array}$$

2.10.2. Sensitivity

Sensitivity was determined by the assessment of the accuracy and precision at the lower limit of quantification, based on a replicate (n = 6) analysis of the LLOQ (1 ng/ml) over three assay batches.

2.10.3. Accuracy and precision

The intra‐ and inter‐assay batch accuracy and precision of the method were assessed and calculated from replicate (n = 6) LQC, MQC and HQC samples, in three individual batches. Accuracy, as assessed by comparing the observed QC concentrations with the theoretical concentrations, was calculated with the following equation:

$$\%\mathrm{Accuracy}=\mathrm{Measured}\mathrm{concentration}/\mathrm{nominal}\mathrm{concentration}\times 100$$

The precision of the assay was measured as coefficient of variations (CV) and calculated as follows:

$$\%\mathrm{CV}=\mathrm{Standard}\mathrm{deviation}/\mathrm{mean}\times 100$$

2.10.4. Inter‐assay variation of standards

The inter‐assay variation of calibration standards was assessed for accuracy and precision across five validation batches.

2.10.5. Recovery

To assess extraction efficiency, tolinapant and d₄‐tolinapant were extracted from plasma or spiked into blank plasma extract at LQC, MQC and HQC. The results were expressed as recovery (%) = (area of extracted analyte/unextracted analyte) × 100.

2.10.6. Carryover

Carryover was assessed by injections of matrix or solvent blanks immediately after the highest calibration standard.

2.10.7. Dilution

Dilution QCS were prepared in human plasma at 2500 ng/ml and a dilution factor of 10 was evaluated for dilution linearity.

2.10.8. Haemolysis and lipemic plasma effect

The effect of haemolysis and lipemic plasma on the quantitation of precision and accuracy was evaluated using replicate sets (n = 6) of extracted LQC samples containing either 2% haemolysed blood or lipemic plasma, quantified against a freshly prepared calibration curve.

2.10.9. Stability

The stability of tolinapant in whole blood sample processing was assessed using replicate (n = 3) sets of LQC and HQC at 0 h and after storage for 2 h in wet ice and room temperature. Benchtop stability at room temperature for 20 h, processed sample stability after storage at 2–8^○C for 71 h, long term storage (33, 187 and 419 days) at ‐20 and ‐70^○C, and after 4 freeze/thaw cycles at ‐70^○C was assessed using replicates (n = 6) of LQC and/or MQC and HQC, quantified against a freshly prepared calibration curve. To assess re‐injection reproducibility, extracted tolinapant calibration standards and replicates of LQC, MQC and HQC (n = 6) were stored at 2–8^○C after the original analysis for at least 48 h, followed by re‐injection. The stability of tolinapant standard stock was assessed at room temperature for 6 h and after storage at 4^○C for 30 days. Working standard solution stability was assessed at 4^○C for 30 days.

3. RESULTS

3.1. Selectivity

Six independent batches of blank human plasma were free from significant chromatographic interference as defined by ≤20% of tolinapant LLOQ response or ≤5% of mean d₄‐tolinapant response of the calibration standards. Measured concentrations of tolinapant LLOQ samples had a mean %CV of 11.5% and a mean accuracy of 96.7%.

3.2. Sensitivity

At the LLOQ (1 ng/ml) the intra‐batch precision expressed as %CV, ranged from 6.1%–11.4% with an accuracy range of 94.6%–99%. Inter‐batch precision and accuracy were 2.3% and 96.7%, respectively (Table 1). Representative chromatograms of blank plasma and analytes are shown in Figure 2.

TABLE 1.

Intra and inter‐assay quality control samples and lower limit of quantification (LLOQ) for tolinapant

	LLOQ	LQC	MQC	HQC
Intra‐assay	1 ng/ml	3 ng/ml	40 ng/ml	360 ng/ml
Mean (ng/ml) ± SD	0.965 ± 0.11	3.13 ± 0.164	40.7 ± 1.09	360 ± 8.02
CV (%)	11.4	5.2	2.7	2.2
Accuracy (%)	96.5	104	102	99.9
Mean (ng/ml) ± SD	0.990 ± 0.060	2.91 ± 0.162	39.6 ± 1.46	346 ± 12.9
CV (%)	6.1	5.6	3.7	3.7
Accuracy (%)	99	97	99	96.2
Mean (ng/ml) ± SD	0.946 ± 0.066	2.99 ± 0.063	40.4 ± 2.11	355 ± 7.32
CV (%)	7.0	2.1	5.2	2.1
Accuracy (%)	94.6	99.6	101	98.5
Inter‐assay
Mean (ng/ml) ± SD	0.967 ± 0.022	3.01 ± 0.112	40.2 ± 0.557	354 ± 6.83
CV (%)	2.3	3.7	1.4	1.9
Accuracy (%)	96.7	100	101	98.2

FIGURE 2.

Representative chromatograms of blank plasma and d₄‐tolinapant (A), extracted lower limit of quantification (LLOQ) (1 ng/ml) tolinapant and d₄‐tolinapant calibration standard (B) and clinical tolinapant plasma sample and d₄‐tolinapant (C)

3.3. Human plasma calibration curve

The calibration curve was linear over the range of 1–500 ng/ml, with a mean correlation coefficient determination of 0.999 (n = 5). Weighted least squares regression (1/x ²) was used. Calibration standards at each concentration level demonstrated acceptable precision (1.9%–7.1%) and accuracy (98.5%–102%) across five analytical batches (data not shown).

3.4. Accuracy and precision

The intra‐batch accuracy ranged between 96.2% and 104% across all QC levels, with inter‐assay accuracy in the range of 98.2%–101%. The intra‐assay precision of QC samples (n = 6) ranged from 2.1% to 5.6% at 3 ng/ml, 2.7% to 5.2% at 40 ng/ml and 2.1% to 3.7% at 360 ng/ml. The inter‐assay precision (n = 18) ranged from 1.4% to 3.7% across all QC levels (Table 1).

3.5. Recovery

Recoveries of tolinapant from human plasma by liquid‐liquid extraction were 94.4%, 93% and 85.1% at LQC, MQC and HQC, respectively. Adequate recoveries of d₄‐tolinapant were observed in the range of 83.1 ‐ 85.8%. A negligible matrix effect was observed for tolinapant, with a normalized matrix factor of 0.931 (data not shown).

3.6. Carryover

Tolinapant carryover was observed at approximately 40%–60% of the LLOQ response and less than 5% for the internal standard, d₄‐tolinapant (data not shown).

3.7. Dilution

The observed concentrations for the dilution QC samples (n = 6) were in good agreement with the prepared concentrations, with a precision of 1.5% and an accuracy of 101% (Table 2).

TABLE 2.

Precision and accuracy of tolinapant in dilution quality control samples

	Concentration (ng/ml)
	2500 (10 x dilution)
n	6
Mean ± SD	2530 ± 37
CV (%)	1.5
Accuracy (%)	101

3.8. Haemolysis and lipemic effect

The presence of 2% haemolysed human whole blood or lipemic plasma did not affect assay performance at the LQC level, with a precision of 5.9% or 6.6% and accuracy of 92.3% or 97%, respectively.

3.9. Stability

3.9.1. Sample stability

Tolinapant was stable in human blood for at least 2 h at room temperature and on wet ice at LQC and HQC concentrations, under sample collection conditions, with a difference of < 9.3%. Tolinapant was stable in processed samples stored at 6^○C for 71 h, with a precision range of 3.1%–5% and accuracy of 94.8%–100% across QC levels (data not shown).

3.9.2. Stability in human plasma

Tolinapant was stable in human plasma at room temperature for 20 h and after storage for up to 419 days at –20 and –70^○C. In addition, tolinapant was stable after four free‐thaw cycles at ‐70^○C (Table 3).

TABLE 3.

Stability of tolinapant in human plasma

Treatment	LQC	MQC	HQC
	3 ng/ml	40 ng/ml	360 ng/ml
Room temperature (20 h)
Mean (ng/ml) ± SD	2.86 ± 0.131		350 ± 8.89
CV (%)	4.6		2.5
Accuracy (%)	95.3		97.4
Freeze‐thaw stability (‐70○C)
Mean (ng/ml) ± SD	2.85 ± 0.216		354 ± 6.75
CV (%)	7.6		1.9
Accuracy (%)	95.1		98.3
419 days (–20○C)
Mean (ng/ml) ± SD	3.27 ± 0.193	49.9 ± 1.22	379 ± 7.73
CV (%)	5.9	2.8	2.0
Accuracy (%)	109	107	105
419 days (–70○C)
Mean (ng/ml) ± SD	3.38 ± 0.211	41.3 ± 1.69	369 ± 9.22
CV (%)	6.2	4.1	2.5
Accuracy (%)	113	103	103

3.9.3. Re‐injection reproducibility

Re‐injection reproducibility was acceptable for up to 90 h at 6^○C, with an accuracy range of 95.5%–97.9% across QC levels (data not shown).

3.9.4. Stability in stock solutions

Tolinapant was stable in standard stock solutions for 6 h at room temperature and up to 30 days at 4^○C in both stock and working solutions (data not shown).

4. DISCUSSION AND CONCLUSION

An LC‐MS/MS method is described in this paper, for the quantification of tolinapant in human plasma, in order to evaluate pharmacokinetics in clinical studies. Sample preparation employed a simple liquid‐liquid extraction, with acceptable and reproducible analyte recoveries and negligible matrix effects. The specificity and selectivity were demonstrated by using different sources of plasma including haemolyzed and lipemic plasma. A deuterated internal standard of tolinapant improved the accuracy and precision of quantitation, as well as the robustness of the method. Observed tolinapant carry over, at approximately 40%–60% of the LLOQ, exceeded acceptable validation criteria. Specific measures, such as avoiding randomization of study samples during analysis and injecting blank samples after the top calibration standard, HQC and samples with expected high concentrations, were undertaken in clinical studies, to minimise any impact on assay performance.

In Phase 1 dose‐escalation and expansion studies, tolinapant was administered in the range of 15–270 mg. ⁷ As a dynamic range of 1–500 ng/ml was achieved and validated, clinical samples with analyte concentrations up to 10 times above the limit of this calibration range could be accurately and precisely quantified following dilution into range. In addition, acceptable stability of tolinapant was demonstrated in biological matrixes and processed samples, under various storage conditions, enabling efficient handling, storage, and analysis of clinical samples.

The development of a robust bioanalytical method enabled human pharmacokinetic assessment of tolinapant in Phase 1 or 2 studies. Based on preclinical mouse xenograft data, tolinapant was dosed orally on an intermittent schedule. ⁷ Tolinapant was rapidly absorbed within 0.5 and 1 h and showed dose‐proportional exposure in the range of 15–180 mg. Plasma exposure was higher on day 7 (mean AUC_0‐24 2960 h ng/ml) compared to day 1 (mean AUC_0‐24 1690 h ng/ml) due to mild accumulation, at the recommended Phase 2 dose. ⁷

In Phase 2 studies with 28‐day dosing cycles, biphasic profiles were exhibited with comparable accumulation on day 7, as observed in Phase 1 studies. ¹⁰ , ¹¹ No continuous accumulation in exposure was observed cycle over the cycle. ⁷ , ¹⁰ Clinical exposures reached the target therapeutic efficacious range observed in nonclinical models and there was evidence of pharmacodynamic and preliminary clinical activity. ⁷ , ⁹ , ¹⁰ , ¹¹

In conclusion, the data presented demonstrate that the developed LC‐MS/MS method for the quantitative analysis of tolinapant in human plasma is sensitive, accurate, reproducible and meets the required criteria as set by the FDA for bioanalytical method validation. ⁴ This method is currently supporting Phase 2 clinical trials in adults with advanced solid tumours and lymphomas.

CONFLICT OF INTEREST

The authors have declared no conflict of interest.

Martins V, Wilsher N, Lin S, Oganesian A. A validated liquid chromatography‐tandem mass spectroscopy method for the quantification of tolinapant in human plasma. Anal Sci Adv. 2022;3:198‐204. 10.1002/ansa.202200009

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.