Bioanalytical Methods for Poly(ADP-Ribose) Polymerase Inhibitor Quantification: A Review for Therapeutic Drug Monitoring

Abstract

Background: 

Therapeutic drug monitoring (TDM) of poly(ADP-ribose) polymerase inhibitors (PARPis) is an exploratory practice aimed at improving the quality of treatment through personalized therapy. Currently, there are 4 European Medicines Agency-approved and US Food and Drug Administration–approved PARPis available clinically whose quantification requires validated analytical methods: olaparib, niraparib, rucaparib, and talazoparib. The purpose of this literature review was to highlight the pharmacological features of PARPis that could support their TDM practice and provide a detailed discussion of the available liquid chromatography coupled with tandem mass spectrometry methods for their quantification.

Methods: 

Using several Medical Subject Heading terms, the literature was searched using several research engines, including SciFinder, Web of Science, Google Scholar, and PubMed, to find articles published before August 2022.

Results: 

Exposure-efficacy and exposure-safety profiles, drug–drug interactions, and hepatic/renal impairment of PARPis provide the potential rationale to monitor their concentrations through TDM. Several bioanalytical methods for their quantification have been reported and compared, and a great deal of heterogeneity has been found among methods, regarding both their analytical and regulatory aspects.

Conclusions: 

In addition to reducing toxicity and increasing the efficacy of PARPis therapy, TDM could be beneficial to thoroughly investigate the exposure–response relationships of PARPis and to establish pharmacokinetic thresholds for clinical decisions. Based on the comparison of published bioanalytical methods, their transferability and validation both play a key role in method selection. For future use in clinical TDM, we anticipate that bioanalytical methods should address every analytical need more thoroughly and should be validated with standardized guidelines.

INTRODUCTION

Poly(ADP-ribose) polymerase inhibitors (PARPis) are a rapidly growing class of targeted anticancer drugs that have been developed over the past 2 decades. PARP belongs to a family of enzymes that play a key role in DNA damage repair.^1,2 In normal cells, the repair of DNA double-strand breaks is regulated by proteins encoded by breast cancer (BRCA) 1 and BRCA2 genes through the homologous recombination mechanism of repair (HRR). Homologous recombination mechanism of repair is an essentially error-free mechanism; however, when altered, as in BRCA-mutated tumors, other repair mechanisms take over to minimize the damage. One of these alternate repair mechanisms requires proper functioning of the PARP enzyme. However, when an inhibitor impairs PARP function, permanent damage accumulates in DNA, leading to a cytotoxic effect.^2–4 The process outlined is based on the synthetic lethality theory that initiated the development of PARPis.

Since their first clinical trial in 2003,^2,4 the following 4 PARPis have been approved by the US Food and Drug Administration (FDA) and European Medicines Agency (EMA) for cancer therapy: olaparib (Lynparza), rucaparib (Rubraca), niraparib (Zejula), and talazoparib (Talzenna). These drugs are primarily used as monotherapies for the maintenance treatment of platinum-sensitive advanced or recurrent ovarian cancer and human epidermal growth factor receptor (HER) 2–negative locally advanced or metastatic breast cancer, which are most commonly associated with BRCA mutations.^5–8 In addition, olaparib has been approved for the treatment of prostate and pancreatic cancer, further stimulating the interest and investigation of PARPis for their therapeutic potential in other tumor types, regardless of the BRCA mutation status of the patient.^3,4

Poly(ADP-ribose) polymerase inhibitors are orally administered at a fixed dose over a prolonged period and are substrates of various drug-metabolizing enzymes and transporters.^5–8 These factors, along with the pharmacogenetic background, varying concomitant medications, and dietary regimens of the patient, often result in wide interindividual variability in drug exposure. Consequently, some patients may experience an insufficient response because of suboptimal drug concentrations, while others may experience severe side effects when the drug exceeds optimal levels.⁹ To optimize response and personalize treatment, therapeutic drug monitoring (TDM) may be a useful strategy. Specifically, this practice measures drug concentrations in biological fluids of the patient and helps support clinical decisions with the goal of maximizing treatment efficacy and minimizing the occurrence of toxicity.

The aim of this review was to highlight the relevant pharmacological features of PARPis that suggest their use in TDM practice and to discuss the available analytical methods for PARPis quantification in human plasma using liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS).

METHODS

This review considered published methods for quantifying PARPis in human plasma, which is the standard method of quantifying pharmaceutical levels in clinical trials and practice. In addition, the search was refined to include only those methods that use LC-MS/MS with a triple quadrupole because this is the most accurate technique for drug quantification in human plasma.¹⁰ Finally, only methods that quantify FDA-approved and EMA-approved drugs and provide sufficient information to be reproduced by analysts were considered. The search was performed using SciFinder, Google Scholar, PubMed, and Web of Science with the keywords “PARP inhibitors,” “bioanalytical methods,” “LC-MS/MS,” “human plasma,” “TDM,” and a combination thereof. Medical subject heading terms such as “Poly(ADP-ribose) Polymerase Inhibitors,” “Chromatography, High Pressure Liquid,” “Tandem Mass Spectrometry,” “Drug Monitoring,” and “Analytic Sample Preparation Methods” were considered during the PubMed search. Thirteen manuscripts,^11–23 published before August 2022, met the inclusion criteria described above. Moreover, the investigation of PARPis pharmacology and the clinical utility of TDM was based on the official documents of the regulatory agencies, namely the FDA multidisciplinary reviews,^5–8,24 EMA public assessment reports,^25–29 and other relevant papers found in the literature.^30–40 In the following discussion, information on the pharmacology, physicochemical features of the drugs, methods used for biological sample preparation, hardware used for the analysis (LC and MS), and validation issues are provided.

TDM AND PARP INHIBITORS

Therapeutic drug monitoring is a branch of clinical chemistry and pharmacology that is applied to several classes of drugs, such as antibiotics, immunosuppressants, antipsychotics, and anticonvulsants. Clinical studies have also demonstrated a correlation between drug exposure [eg, area under the curve (AUC) and trough concentration (C_trough)] and treatment response (efficacy and/or toxicity) in some cancers.³⁴ However, a strong correlation has not been found for many anticancer drugs, and their potential for TDM is still being investigated. In the case of PARPis, several authors point out that they should be further explored as candidates for TDM.^30,31,35 Recently, a review by Groenland et al³⁰ gathered together 7 criteria known to be valid indications of the suitability of a drug for TDM, and PARPis satisfy most of these criteria. Specifically, they are prescribed as long-term therapy, taken daily for at least 2 years or until disease progression or unacceptable toxicity.^41–44 Furthermore, their dose can be easily adapted to meet individual patient needs because PARPis are formulated as fixed dose capsules or tablets.^25–29 Moreover, no measurable biomarkers to assess drug effects are available,^5–8,24 and validated bioanalytical methods for their quantification have been published.^11–23 Finally, PARPis display significant variability in pharmacokinetic (PK) exposure, as highlighted by relevant studies on olaparib,^35,45 niraparib,⁴⁶ rucaparib,^6,47 and talazoparib.⁴⁸ By contrast, the therapeutic window of PARPis is wide,³⁷ and not all PARPis have clear exposure–response relationships.³⁵ In the subsequent sections, the controversial exposure–response relationships, drug–drug interactions, and renal and hepatic impairments have been addressed thoroughly. The last 2 circumstances are included in the current discussion because a change in dosage might be involved, leading to nonoptimal plasma concentrations, which could be monitored by TDM.

Pharmacology of PARP Inhibitors

Olaparib was first formulated as capsule with a standard dose of 400 mg twice daily (BID) and then as tablet with a standard dose of 300 mg BID. Both are addressed in the subsequent discussion.

Exposure Safety

For talazoparib⁸ and olaparib,⁵ an exposure–toxicity relationship for anemia has been reported, that is, a higher risk of anemia at higher steady-state concentrations. For olaparib specifically, the likelihood of anemia has been correlated with its maximum concentration (C_max). In addition, olaparib tablets are associated with a higher risk of anemia than the capsules because the drug exposure of the tablets exceeds that of the capsules.⁵ Moreover, a C_trough of 2500 ng/mL for olaparib was identified in a recent retrospective study³⁶ as the threshold for early adverse events. In addition, higher exposure to talazoparib was associated with a higher risk of thrombocytopenia.³⁹ For niraparib, a correlation between AUC and thrombocytopenia has been observed⁷; however, its tolerability could be managed through laboratory monitoring. For rucaparib, a correlation between exposure and some hematological parameters,⁴⁰ including grade 2+ creatinine, has been reported.⁶

Exposure-Efficacy

The regulatory paper by FDA reports that higher exposure to talazoparib leads to longer progression-free survival (PFS).⁸ For rucaparib, the situation is unclear and requires further investigation. Based on previous studies on olaparib and niraparib, the FDA and EMA report a desired target C_trough of 650 ng/mL for rucaparib⁶ and that all patients treated with the recommended dose (600 mg BID) had a plasma rucaparib concentration well above this threshold.^6,27 By contrast, a recent review³⁵ reported the absence of a correlation between exposure and efficacy end points. Notably, the exposure–efficacy relationship for olaparib and niraparib was not significant. However, this relationship should be further studied because frequent dose modification and/or treatment discontinuation occurred in past studies.^7,24

Drug–Drug Interactions

Olaparib is primarily metabolized by cytochrome P450 (CYP) 3A4; therefore, the concomitant administration of potent CYP3A4 inhibitors or inducers should be avoided.³² When strong and moderate CYP3A4 inhibitors are present, a dose reduction to 100 and 150 mg BID, respectively, is recommended for tablet administration.²⁴ When talazoparib is coadministered with potent P-glycoprotein (P-gp) inhibitors, dose reduction from the standard dose of 1 mg–0.75 mg is suggested and patients should be monitored for adverse effects so the dose can be adjusted according to tolerability.³³ Similarly, breast cancer resistance protein (BCRP) inhibitors may increase talazoparib exposure.⁸ Niraparib has a low risk of drug–drug interactions because it is primarily metabolized by carboxylesterases, rather than cytochromes.⁷ Finally, rucaparib has been reported to interact with many CYPs and transporters, resulting in a potential effect on other drugs.³⁵

Renal and Hepatic Impairment

According to the latest guidelines,⁴⁹ renal impairment is considered either mild or moderate when creatinine clearance is between 60 and 89 mL/min or 30 and 59 mL/min, respectively. For olaparib, the creatinine clearance values used to assess renal impairment refer to older guidelines.⁵⁰ Hepatic impairment is considered mild and moderate when the Child–Pugh score is classified as A and B, respectively.⁵¹

For olaparib, no dose adjustment is required for patients with mild renal impairment²⁴; however, patients should be monitored for toxicity.⁵ By contrast, a tablet dose reduction to 200 mg BID is suggested for patients with moderate renal impairment.²⁴ Furthermore, no dose adjustment is required in cases of mild or moderate hepatic impairment.³⁸ Similarly, for talazoparib, dose reduction is required if the renal impairment experienced by the patient is moderate (0.75 mg),^8,33 whereas no adjustments are required if mild renal or hepatic impairment is present.²⁸ No dose adjustments are required for niraparib if the renal impairment experienced by the patient is mild or moderate⁷; however, a recent postmarketing study⁵² suggested a dose reduction to 200 mg for cases with moderate hepatic impairment. In the case of rucaparib, the situation requires further investigation because mild or moderate renal impairment is known to increase the AUC of rucaparib, although no dose adjustment is recommended.⁶ However, in cases of severe renal impairment, some authors³⁵ indicate that TDM of rucaparib might be beneficial.

BIOANALYTICAL METHODS FOR PARP INHIBITOR QUANTIFICATION

General Considerations

The compounds analyzed in this literature review and their most relevant physicochemical features are listed in Table 1. The information presented in Table 1 is useful for the development of a robust bioanalytical method and is directed to the analyst required to prepare the analyte solutions and set up the sample preparation protocol, as well as the LC and MS methods and conditions. In addition, a discussion of other analytical aspects (ie, single versus multianalyte methods and analyte calibration range) relevant to TDM is provided.

TABLE 1. - Main Physicochemical Features of EMA-Approved and FDA-Approved Poly (ADP-Ribose) Polymerase Inhibitors (PARPis)

	Olaparib	Rucaparib	Niraparib	Talazoparib
MW (g/mol)	434.46	555.67	510.61	552.56
Free base MW (g/mol)	N/A	323.36	320.4	380.35
Counterion	N/A	Camsylate	Tosylate	Tosylate
pKa	N/A	∼9.5	9.95	N/A
Solubility in water (mg/mL)	0.1 at 20°C	1.7 at 37°C	0.7–1.1 at 37°C	Low
Solubility in organic solvents (mg/mL)	ACN = 3.1; EtOH = 5.5; MeOH = 10.6	N/A	N/A	N/A

ACN, acetonitrile; EtOH, ethanol; MeOH, methanol; N/A, not applicable or reported.

Physicochemical Properties of the Analytes

Olaparib is a neutral compound with a molecular weight (MW) of 434.46 g/mol that is sold as a free base species. It has a basic pKa of −1.25 and an acidic pKa of 12.07, which means that it is unionized at physiological pH.⁵ It is only slightly soluble in aqueous media (0.1 mg/mL at 20°C), but it is soluble in organic solvents (Table 1).⁵

Niraparib, rucaparib, and talazoparib are weak bases that are commercialized as salts with their specific counterions in their pharmacological formulations (Table 1). Niraparib is found in capsules as a tosylate salt (monohydrate). It has a MW of 510.61 g/mol and a free base MW of 320.40 g/mol. It has a pKa of 9.95, and its aqueous solubility at 37°C ranges between 0.7 and 1.1 mg/mL below its pKa (ionized form).⁷ Rucaparib is found as tablets in the form of a camsylate salt with a MW of 555.67 g/mol and a free base MW of 323.36 g/mol. Based on a report by EMA,²⁷ its solubility decreases above pH 9.5, likely because of deprotonation, and its structure is that of a secondary amine, suggesting its pKa to be approximately 9.5. It has an aqueous solubility of 1.4 mg/mL at 25°C and 1.7 mg/mL at 37°C, which is pH independent between pH 3 and 7 (under its pKa value).⁶ Talazoparib is commercialized in capsules in the form of a tosylate salt with a MW of 552.56 g/mol and a free base of 380.35 g/mol. Information on its solubility and pKa was not clearly indicated in the official EMA²⁸ and FDA⁸ documents, which state that the solubility of talazoparib is “low across the physiological pH range” (at 37°C), but they report that 1 mg of talazoparib is soluble in less than 250 mL of aqueous medium (pH not specified).

Other Analytical Considerations

For TDM purposes, either single-analyte or multianalyte methods can be found in the literature. Of the 13 methods considered in this review, more than half are single-analyte methods. Specifically, 4 methods for the quantification of olaparib,^12,13,15,16 3 for that of rucaparib,^14,18,22 and one for that of niraparib¹⁷ were considered. The major advantage of single-analyte methods is the simplicity of their development as both sample preparation and detection are optimized for this analyte, and chromatography does not require the separation of multiple species. However, multianalyte methods are more convenient because they save time and money by analyzing different molecules simultaneously. Indeed, several authors use multianalyte methods for PARPi quantification. For example, Krens et al²¹ investigated olaparib and niraparib together with other anticancer drugs, and similarly, Jolibois et al,²⁰ Pressiat et al,¹⁹ and Sparidans et al¹¹ investigated olaparib together with other anticancer drugs. These methods could be useful if the analyst is interested in quantifying all the drugs included in the assay; otherwise, a single-analyte method might be more convenient and easier to transfer from one laboratory to another. Finally, the method proposed by Bruin et al²³ allows the use of a single assay to simultaneously quantify all the FDA-approved and EMA-approved PARPis, along with veliparib, which is yet to be approved by the regulatory agencies. Reportedly, this is the first published method to quantify talazoparib in human plasma.²³Table 2 summarizes the details of all bioanalytical methods.

TABLE 2. - Summary of General Features, Sample Preparation, and LC Conditions of Bioanalytical Methods for PARPi Quantification

Author (yr)	Analytes	IS	Calibration Range	Matrix Volume	Extraction Procedure	Addition of IS	Injection Solvent	Mobile Phases	Instrument Specification and Set Up	Column	Elution Type	Runtime	Other Chromatographic Parameters (TF, Oven T, AS T, Vinj)	Reference Number
Bruin et al (2020)	Olaparib, rucaparib, niraparib, talazoparib, veliparib	2H8-olaparib,13C,2H3-rucaparib, 13C6-niraparib as hydrochloride salt, 13C,2H4-Talazoparib, 13C,2H3,15N-Veliparib as dihydrochloride salt	100–10000 for olaparib, 30–3000 for niraparib, 50–5000 for rucaparib, and veliparib; 0.5–50 for talazoparib (ng/mL for all)	50 µL	PP (1 + 2 ACN)	Before precipitation	Dilution 1 + 1 with MPA	0.1% HCOOH in water (MP A) and 0.1% HCOOH in MeOH (MP B)	Shimadzu nexera 2 with binary pump, degasser, autosampler, valco diverter valve and oven	Waters acquity UPLC BEH C18 (100 × 2.1 mm, 1.7 µm) + its vanguard pre-column	Gradient	6.2 min	TF: 0.3; oven T: 40; AS T: 2–8	23
Krens et al (2020)	Olaparib, niraparib, regorafenib, regorafenib M2, vemurafenib, cobimetinib, cabozantinib and dabrafenib	2H8-olaparib,13C6-niraparib	200–20000 for olaparib and 200–2000 for niraparib (ng/mL for all)	50 µL	PP (1 + 4 ACN)	With precipitant	Dilution 1 + 9 with water	0.1% HCOOH in water (MP A) and 0.1% HCOOH in ACN (MP B)	Waters acquity H-class UPLC	Waters Cortex UPLC C18 (50 × 2.1 mm, 1.6 µm)	Gradient	7 min	TF: 0.8; oven T: 50; AS T: RT; Vinj: 10	21
Shapiro et al (2019)	Rucaparib	d7-rucaparib	5–10000 ng/mL	20 or 50 µL	PP with 96-well plate, solvent not reported	N/A	N/A	20% ACN and 0.1% HCOOH in H2O	N/A	Agilent Polaris C18-A (50 × 2.1 mm, 3 µm)	Isocratic	N/A	TF: 0.5	22
Jolibois et al (2019)	Olaparib, cabozantinib, palbociclib HCl, pazopanib HCl, sorafenib, Sunitinib l-malate, N-desethyl-sunitinib HCl	2H8-olaparib	170–20000 ng/mL	150 µL	LLE with EtOAc	Before extraction	Reconstitution with 55% MeOH/45% water-based solvent	92% ACN (vol/vol) and 8% of H2O in 0.1% HCOOH	Dionex HPLC	Agilent zorbax Bonus-RP (150 × 2.1 mm, 1.8 µm)	Isocratic	2.5 min	TF: 0.5, then 0.3; oven T: 50; AS T: 4; Vinj: 2	20
Pressiat et al (2018)	Olaparib, ruxolitinib, vismodegib, pazopanib	2H8-olaparib	100–100000 ng/mL	50 µL	PP (1 + 2,5 ACN)	With precipitant	Organic supernatant injected	10 mM ammonium formate buffer containing 0.1% HCOOH (MP A) and ACN with 0.1% HCOOH (MP B)	Waters acquity I-class UPLC	Waters acquity UPLC BEH C18 (50 × 2.1 mm, 1.7 µm)	Gradient	5 min	TF: 0.3; oven T: 40; AS T: RT; Vinj: 0.15	19
Gorijavolu et al (2018)	Rucaparib	13C,2H3-rucaparib	5–5000 pg/mL	400 µL	LLE with EtOAc and DCM	Before extraction	Reconstitution with mobile phase	10 mM ammonium formate and MeOH 20:80 vol/vol	Agilent 1200 series HPLC	Waters Xbridge C18 (50 × 4.6 mm, 5 µm)	Isocratic	2.5 min	TF: 0.7; oven T: 40; Vinj: 5	18
van Andel et al (2017)	Niraparib, Niraparib-M1	M002151 (niraparib deuterated IS), D5-M1	1–500 ng/mL for niraparib and Niraparib–M1	100 µL	PP (1 + 2 ACN/MeOH) then dryness	Before precipitation	Reconstitution with initial mobile phase	20 mM ammonium acetate in water (MP A) and 0.1% HCOOH in ACN/MeOH 50:50 (vol/vol) (MP B)	Waters acquity I-class with diverter valve	Waters SunFire C18 (50 × 2.1 mm, 5 µm)	Gradient	7 min	TF: 0.7; oven T: 40; AS T: 8; Vinj: 3	17
Rolfo et al (2015)	Olaparib	2H8-olaparib	0.5–500 and 0.02–20 µg/mL	100 µL	SPE elution with ACN	Before extraction	Reconstitution with mobile phase	1 mM pH 3 ammonium formate buffer/ACN (73/27 vol/vol)	N/A	Waters xterra phenyl (50 × 2.1 mm, 3.5 µm)	Isocratic	N/A	TF: 0.2	16
Roth et al (2014)	Olaparib	2H8-olaparib	0.5–5000 ng/mL	100 µL	LLE with EtOAc containing IS	With extraction solvent	Reconstitution with 50% ACN in water	0.1% HCOOH in H2O (MP A) and 0.1% HCOOH in ACN (MP B)	Waters acquity UPLC	Waters acquity UPLC BEH C18 (50 × 2.1 mm, 1.7 µm)	Gradient	5 min	TF: 0.25; oven T: 30; AS T: 4; Vinj: 10	15
Sparidans et al (2014)	Rucaparib	Gefitinib	1.25–2000 ng/mL	50 µL	PP (1 + 1,5 ACN)	With precipitant	Dilution 1 + 1 with water	0.02% HCOOH in H2O (MP A) and MeOH (MP B)	Shimadzu LC10 with diverter valve	Varian polaris 3 C18-A (50 × 2 mm, 3 µm)	Gradient	3 min	TF: 0.5; oven T: 40; AS T: 4; Vinj: 2	14
Nijenhuis et al (2013)	Olaparib	2H8-olaparib	10–5000 ng/mL	100 µL	LLE with MTBE	Before extraction	Reconstitution with MeOH	10 mM ammonium acetate in water (MP A) and MeOH (MP B)	Shimadzu LC20AD prominence with diverter valve	Phenomenex gemini C18 (50 × 2.0 mm, 5 µm) + its guard column	Gradient	6 min	TF: 0.25; oven T: 40; AS T: 4; Vinj: 1	13
Rajan et al (2012)	Olaparib	2H8-olaparib	2–2000 ng/mL	100 µL	LLE with EtOAc	N/A	Reconstitution with mobile phase	0.1% HCOOH in water (MP A) and ACN (MP B)	N/A	Phenomenex luna C18(2) (50 × 2 mm, 5 µm)	Gradient	7.5 min	Vinj: 10	12
Sparidans et al (2011)	Olaparib, melphalan	Erlotinib	10–5000 ng/mL	100 µL	PP (1 + 1,5 ACN)	With precipitant	Dilution 1 + 1 with water	0.1% HCOOH in water (10% vol/vol) + water (47% vol/vol) + MeOH (43% vol/vol)	Accela system with diverter valve	Waters acquity UPLC BEH C18 (30 × 2.1 mm, 1.7 µm) + its VanGuard pre-column	Isocratic	1.2 min	TF: 0.6; oven T: 40; AS T: 4; Vinj: 1	11

ACN, acetonitrile; AS T, autosampler temperature (°C); DCM, dichloromethane; EtOAc, ethyl acetate; HCOOH, formic acid; MeOH, methanol; MP A, mobile phase A; MP B, mobile phase B; MTBE, methyl tert-butyl ether; N/A, not applicable or not reported; oven T, oven temperature (°C); RT, room temperature; TF, total flow (mL/min); V_inj, volume of injection (µL).

One of the most important aspects of a bioanalytical method is the calibration range of the analytes that can be reliably and reproducibly quantified. This range is established by considering several factors, including whether the purpose of the analysis is TDM or PK, the expected concentration of the analyte in the matrix, and the need for dilution.^53,54 In TDM, the range of analytes is generally limited because the main objective is to determine whether the analyte reaches a target drug value, often referred to as C_trough.^23,31 By contrast, the purpose of PK studies was to examine the entire concentration range from intake to elimination. Only 4 recent methods claim that their goal is TDM^19–21,23 and they use the amount of data for C_trough accumulated by earlier studies, whose goal was a more comprehensive PK analysis. It is also possible to use other methods for TDM, but some of them^12,14,17,18 may require a sample dilution step because they do not properly include C_trough values in their quantification range (Table 2).

Sample Preparation: Extraction from Human Plasma

Among the many extraction techniques available,⁵⁵ the reviewed papers used protein precipitation (PP),^{11,14,17,19,21–23} liquid–liquid extraction (LLE),^{12,13,15,18,20} and solid-phase extraction (SPE) (Table 2).¹⁶ Sample preparation is necessary to avoid any ion suppression or enhancement phenomena that may occur because of matrix effects (MEs).⁵⁶ As reported by Nijenhuis et al¹³ and Gorijavolu et al,¹⁸ when problems similar to these occur, the method is flawed and the sample preparation must be revised. In both studies, the preliminary PP had to be changed to LLE^13,18 and the analytes were extracted using ethyl acetate/dichloromethane and tert-butyl methyl ether, respectively. Although LLE proved to be the superior extraction technique and was necessary in the aforementioned cases, PP was used in most of the methods^{11,14,17,19,21–23} because it is faster and more simple. Most of the studies that used PP extracted the analytes using various amounts of acetonitrile, starting from 1.5 volumes,^11,14 all the way up to 4 volumes.^19,21,23 Alternatively, one of these studies used a mixture of acetonitrile/methanol as the extraction solvent.¹⁷ In a clinical TDM environment where high throughput is essential, a fast and easy PP might be the best choice for extraction, especially because it is easily amenable to automation with 96-well plates, as reported by Shapiro et al.²² However, special care should be taken to not contaminate the electrospray ionization (ESI) source. Some authors managed to avoid this using a divert valve^{11,13,14,17,23} that directed the flow to the mass spectrometer only when analytes were eluting from the column.

Another aspect to consider is the matrix volume needed for the analysis as almost all authors using PP kept the volume of human plasma down to 50 µL.^{14,19,21–23} This is small compared with the average amount required by LLE and SPE, which ranges between 100 and 400 µL.^{12,13,15,16,18,20} Therefore, when the analysis requires a limited matrix volume, the use of PP is a good solution.

All methods reviewed here added an internal standard (IS) during sample preparation to compensate for losses that might occur during sample handling and injection. Specifically, almost every study presented in Table 2 used the following stable isotope-labelled analogs of the analytes as the IS (SIL IS): ²H₈-olaparib, ¹³C,²H₃-rucaparib, ¹³C₆-niraparib, ¹³C,²H₄-talazoparib, and ¹³C,²H₃,¹⁵N-veliparib. These molecules have the same chemical behavior as the analytes; therefore, they elute together from the column, minimizing the potential MEs⁵⁷ at their retention time. This aspect was highlighted in the work of Nijenhuis et al,¹³ where the use of dasatinib, erlotinib, or imatinib as an IS for olaparib led to an unacceptable lack of accuracy and precision. However, when ²H₈-olaparib became commercially available and could be used as the IS, these problems were resolved. Importantly, van Andel et al¹⁷ demonstrated the use of one SIL IS for niraparib and another SIL IS for its metabolite niraparib-M1, highlighting this method as the best choice for reliable metabolite quantification. However, when an SIL IS is neither commercially available nor accessible because of its high cost, a generic organic compound could be used as the IS. This was reported by Sparidans et al,^11,14 who used erlotinib and gefitinib, respectively. Regardless of the type of IS used, IS addition during the sample preparation was performed in 2 main ways in the papers reviewed. In 6 papers,^{13,16–18,20,23} the addition was performed by dispensing a small amount of the IS before adding the extraction solvent. In another 5 papers,^{11,14,15,19,21} the IS was dissolved in the extraction solvent and added to the matrix in one step. The necessity of either approach depends on the solubility of the IS; however, the dissolution of the IS in the extraction solvent can shorten sample preparation and reduce sources of error.

Some studies, such as those by Nijenhuis et al¹³ and Pressiat et al,¹⁹ ended the sample preparation after the extraction of the analyte, whereas others required further steps. Although direct injection of the organic extract seems to be the simplest and the fastest route, it can lead to solvent effects that interfere with chromatography in the form of fronting peaks, as reported by Bruin et al.²³ In this case, the organic extract was diluted with an equivalent amount of an aqueous mobile phase to improve the peak shape of the analytes. Similarly, Sparidans et al^11,14 diluted the organic extract with an equivalent amount of water, whereas Krens et al²¹ chose to inject a solution with an even higher (>90%) percentage of water. Other studies included an evaporation step with a subsequent reconstitution. Four of them used the mobile phase as the reconstituting agent,^12,16–18 while 2 of them used a 1/1 aqueous/organic solution.^15,20 Ideally, the composition of the final extract should match the initial composition of the mobile phase, which can be easily obtained through evaporation and reconstitution steps. Moreover, this procedure is often advantageous when low limits of quantification must be reached but is subject to more errors because of extensive manipulation of the sample.

Liquid Chromatography Separation

In previously published methods, chromatographic separation of PARPis was performed exclusively on reverse phase columns. Almost every author used columns functionalized with C18 chains, except Rolfo et al, who took advantage of the phenyl functionalization to retain olaparib. Regarding C18 columns, most studies selected columns based on hybrid particles built with resistance to high pH,^{11,13,15,16,18,19,23} although they were used at low to mid pH. In 2 studies, the researchers selected ordinary silica-based C18 columns suitable for work at low to mid pH,^12,17 and one study used a core–shell column,²¹ which should improve both the resolution and efficiency. Finally, only a few studies used silica columns functionalized with embedded polar groups that provide good peak shapes for mixtures of polar and nonpolar compounds, especially for bases at low pH.^14,20,22 This last column choice might be the most appropriate for the group of analytes and conditions used in the methods reviewed, where the elution of rucaparib, niraparib, talazoparib (weak bases), and/or olaparib (neutral) was performed under acidic conditions.^{11,12,14–17,19–23} In fact, protonated bases (at a pH below their pKa) are known to produce tailing, which should be minimized to avoid inaccuracies during peak integration. An FDA guideline⁵⁸ recommends the tailing factor (TF) to be less than 2. However, although the TF was clearly visible in some papers, it was neither calculated nor reported by any of them.

In addition to the chemistry of the columns, their physical properties varied substantially among the methods reviewed. In about half of them, the columns were packed with a stationary phase consisting of sub 2 μm particles,^{11,15,19–21,23} whereas the others used columns with particle sizes in the range of 3–5 μm.^{12–14,16–18,22} Smaller particles exhibit superior performance than larger ones for resolution but are associated with higher back pressures and typically require specific equipment. This is referred to as ultra-high-performance liquid chromatography (UHPLC) or ultra-performance liquid chromatography (UPLC). Although the use of such a system allows for faster methods with lower solvent consumption and better resolution, it requires cleaner samples and more frequent maintenance. For laboratories that are not equipped with this instrumentation or cannot afford its cost, a few of the methods presented in Table 2 explicitly use the conventional high-performance liquid chromatography (HPLC) system.^13,14,18 However, it should be noted that a similar run time^13,15,17 and solvent flow^{13,14,18,19,21} has been reported in comparisons among the HPLC with the UHPLC methods. Therefore, the use of UHPLC in the papers reviewed here might be appropriate only when many species are to be separated and/or when a higher resolution is needed.

Most of the methods listed in Table 2 used a gradient elution program,^{12–15,17,19,21,23} but isocratic separation was still widely used.^{11,16,18,20,22} With a duration of less than 3 minutes, a major advantage of isocratic methods is their rapidity,^11,18,20 although some papers in the literature debate whether isocratic separation is always faster than a gradient elution program.⁵⁹ Indeed, the method developed by Sparidans et al¹⁴ used a gradient with a total run time of only 3 minutes, which is very close to the mean isocratic performance. Such a short run time translates into higher throughput and lower solvent consumption, which are necessary features in contexts where many samples are analyzed each day, such as in a hospital. Another advantage of isocratic methods is their ease of transferability from one system to another, a property of paramount importance when running an existing method from the literature in one's laboratory. Although methods using gradients are transferable, some fundamental features of the original equipment should be known (eg, the dwell volume) to avoid gradient delays and distortions after the transfer.^60–62 However, none of the methods reviewed reported these features, and in addition, some did not report any hardware specifications.^{11,12,16,18,20,22}

Mass Spectrometry Conditions

The ESI source was used in all the reviewed methods. Because it generally allows for high sensitivity with polar analytes, it is the most common ionization source chosen in clinical laboratories. Only Gorijavolu et al¹⁸ examined the atmospheric pressure chemical ionization (APCI) source as a viable alternative to ESI because APCI is generally less susceptible to MEs. However, they reported better results with ESI for the detection of rucaparib.⁵⁷

Almost every reviewed paper reported the type of mass spectrometer used for this study. The most common instrument brand was AB Sciex, followed by ThermoFisher Scientific and Waters. Moreover, all the authors accurately listed the fundamental parameters of the mass spectrometer source for temperature, voltage, and gases. This information is useful when transferring a method from the literature to a laboratory using the same type of mass spectrometer.

All authors ionized PARPis in positive mode and detected them using selected reaction monitoring (SRM) or multiple reaction monitoring (MRM) MS programs. Detailed information on the fragmentation of each analyte was provided in each paper, except for that by Rajan et al.¹² For olaparib, a precursor ion of 435 m/z was reported in all methods,^{11,13,15,16,19–21} except in a study by Bruin et al²³ wherein the [M + H]⁺ + 1 isotopologue at 436 m/z was used instead. For rucaparib, a precursor ion of 324 m/z was described in all methods,^14,22,23 except in the case of Gorijavolu et al¹⁸ wherein a molecule (the nature of which was not described) at 323 m/z was used. Furthermore, the niraparib precursor ion at 321 m/z was used by all authors,^17,21,23 and the study by van Andel et al¹⁷ additionally analyzed the niraparib M1 metabolite at 322 m/z. Finally, talazoparib was reported by Bruin et al,²³ with a precursor ion at 381 m/z. Regardless of the analyte involved, only a few studies^11,14,19,20 have used 2 different product ions for the analyte quantification and confirmation, although the importance of this is emphasized in official guidelines.^63,64 Moreover, most studies used a product ion of the analyte which had the same m/z as the product ion of its relative SIL IS.^{15–18,20,21,23} This choice could be problematic for quantification, particularly when the 2 product ions are collected closely—one immediately after the other—in the MS acquisition method. Instead, the product ion of the SIL IS should still contain stable isotopes for more robust quantification.

Finally, 5 studies explored the fragmentation pattern of the analytes in detail and reported the representative figures. Three of the studies examined olaparib,^11,13,21 2 of them examined niraparib,^17,21 and 1 of them examined rucaparib.¹⁴ No information was found for the fragmentation pattern of talazoparib. Moreover, the fragmentation pattern was not reported for any IS. This information is superfluous in studies where the SIL IS has the same product ion as the relative analyte; however, it might be useful where other compounds were used as the IS in place of the SIL IS.^11,14

Validation Issues

Regulatory agencies, such as the FDA⁵³ and EMA,⁵⁴ define the key parameters required to validate a bioanalytical method. Among the methods reviewed (Table 3), 3 of them^12,16,22 did not specify how validation was performed, 4 of them^13,17,20,21 followed the EMA guidelines, and 7 of them^{11,13–15,17–19} followed the older version of the FDA guidelines from 2001.⁶⁵ In addition, Jolibois et al²⁰ also used the French Accreditation Committee (COFRAC) guidelines for biological methods, and Bruin et al²³ used another validation protocol.⁶⁶ Most of the studies evaluated linearity, accuracy, precision, sensitivity, and stability using the rationale proposed by the respective regulatory agencies with minimal deviations; therefore, they do not require further comments. Some studies also evaluated carryover,^13,17,20,23 recovery,^{11,14,15,17–21} and dilution,^12,17,21 and Pressiat et al¹⁹ did a thorough investigation of the possible interference of other drugs (selectivity). However, the evaluation of MEs was performed using several approaches, some of which were more exhaustive than other experimental protocols. The reason for this heterogeneity might be the complexity of ME evaluation, as already pointed out in a recent perspective article.⁶⁷ Most of the reviewed studies^{13–15,17,19–21} used the postextraction method,⁶⁸ which requires evaluation from 6 different matrices at high and low concentrations. Some of these studies used more levels of concentration¹⁹ or less than 6 matrices¹⁵ or did not specify these details,^14,19 but their rationales remained the same. One study¹⁸ used an alternative method that was unclear to us. Finally, Sparidans et al¹¹ evaluated MEs with the postcolumn infusion method,⁶⁹ a qualitative procedure often used during method development to evaluate ion enhancement/suppression phenomena at the retention time of analytes.⁵⁶ This method is not currently sufficient for the regulatory agencies, which require a quantitative determination of MEs. As demonstrated, the evaluation of MEs is one of the most challenging investigations required by regulatory agencies and is susceptible to misinterpretation.⁷⁰

TABLE 3. - Summary of Mass Spectrometry Conditions and Validation Issues of Bioanalytical Methods for PARPi Quantification

Author (yr)	Mass Spectrometer Set up	Detection Mode	MS Transitions (m/z)	Type of Validation	Validation Experiments	Evaluation of ME	Reference Number
Bruin et al (2020)	Triple quadrupole 6500+ (AB Sciex)	ESI, positive ion mode, MRM	Olaparib 436 > 253, niraparib 321 > 205, rucaparib 324 > 236, talazoparib 381 > 298, veliparib 245 > 90, olaparib IS 443 > 253, niraparib IS 327 > 211, rucaparib IS 328 > 236, talazoparib IS 386 > 303, veliparib IS 250 > 90.	Simplified validation procedure for TDM purposes	Linearity, A&P, selectivity, sensitivity, carryover, stability	Not evaluated	23
Krens et al (2020)	Xevo RQ-S micro tandem MS (Waters)	ESI, positive ion mode, MRM	Olaparib 435 > 281, niraparib 321 > 205, olaparib IS 433 > 281, niraparib IS, 327 > 211	EMA 2011	Linearity, A&P, selectivity, sensitivity, recovery (not done according to guidelines), matrix effect, dilution, stability	Determined with postextraction spike from 6 plasma batches	21
Shapiro et al (2019)	API4000 system (AB Sciex)	ESI, positive ion mode, SRM	Rucaparib 324.1 > 293.1, rucaparib IS 331.2 > 300.2	N/A	N/A	N/A	22
Jolibois et al (2019)	Endura (Thermo Fisher Scientific) triple quadrupole MS	ESI, positive ion mode, MRM	Olaparib 435.03 > 367.11 and 435.03 > 281.00; olaparib IS 443.26 > 375.16 and 443.26 > 281.07	EMA 2011 + French committee of accreditation's guidelines for biological method (2018), when appropriate.	Linearity, A&P (not clear), selectivity, sensitivity, carryover, recovery, matrix effect, stability	Determined with postextraction spike from 6 plasma batches	20
Pressiat et al (2018)	Triple quadrupole (TQ-S) Xevo MS (Waters)	ESI, positive ion mode, MRM	Olaparib 435.21 > 367.18 and 435.21 > 133.12; olaparib IS 443.29 > 375.25 and 443.29 > 280.03	FDA 2001 + others	Linearity, A&P, selectivity, sensitivity, recovery, matrix effect, stability	Determined with postextraction spike at 3 concentration levels (not specified how many different plasma batches)	19
Gorijavolu et al (2018)	API4000 triple quadrupole instrument (AB Sciex)	ESI, positive ion mode, MRM	Rucaparib 323.4 > 170.1, rucaparib IS 328.4 > 170.1	FDA 2001	Linearity, A&P, selectivity, sensitivity, recovery, matrix effect, stability	Not evaluated with standard procedure according to EMA guidelines	18
van andel et al (2017)	API5500 tandem mass spectrometer (AB Sciex)	ESI, positive ion mode, MRM	Niraparib 321 > 304, niraparib-M1 322 > 304, niraparib IS 328 > 304, niraparib-M1 IS 329 > 304	FDA 2001 and EMA 2011	Linearity, A&P, selectivity, sensitivity, carryover, recovery, matrix effect, dilution, stability	Determined with postextraction spike from 6 plasma batches, plus one hemolyzed sample in triplicate	17
Rolfo et al (2015)	N/A	N/A	Olaparib 435.2 > 281.1, olaparib IS 443.0 > 281.1	N/A	N/A	N/A	16
Roth et al (2014)	Quattro micro triple quadrupole MS (Waters)	ESI, positive ion mode, MRM	Olaparib 435.4 > 281.1, olaparib IS 443.2 > 281.1	FDA 2001	Linearity, A&P, selectivity, recovery, matrix effect, stability	Determined with postextraction spike from 5 plasma batches	15
Sparidans et al (2014)	Finnigan TSQ Quantum Discovery Max triple quadrupole (Thermo Electron)	ESI, positive ion mode, SRM	Rucaparib 324.1 > 293 and 324.1 > 236; gefitinib 447.1 > 128	FDA 2001	Linearity, A&P, selectivity, sensitivity, recovery, matrix effect, stability	Determined with postextraction spike (not specified how many plasma batches)	14
Nijenhuis et al (2013)	Finnigan TSQ quantum ultra triple quadrupole (Thermo Fisher Scientific)	ESI, positive ion mode, MRM	Olaparib 435.3 > 367.1; olaparib IS 443.1 > 375.2	FDA 2001 and EMA 2011 in GLP environment	Linearity, A&P, selectivity, sensitivity, carryover, matrix effect, stability	Determined with postextraction spike at high and low levels of QC from 6 plasma batches	13
Rajan et al (2012)	5500 Qtrap (AB Sciex)	MRM	N/A	N/A	Linearity, A&P, dilution	N/A	12
Sparidans et al (2011)	Finnigan TSQ quantum ultra triple quadrupole (Thermo Fisher Scientific)	ESI, positive ion mode, SRM	Olaparib 435.2 > 281.1 and 435.2 > 367.2; erlotinib 394.1 > 278.1	FDA 2001 plus others	Linearity, A&P, selectivity, recovery, stability	Determined with postcolumn infusion-no matrix effect at the retention time of the analytes	11

A&P, accuracy and precision; GLP, good laboratory practice; N/A, not applicable or not reported; QC, quality control.

DISCUSSION, GAP ANALYSIS, AND OUTLOOK

Our literature search identified 13 validated bioanalytical methods for the quantification of PARPis, suggesting a high level of interest in this subject. This could be attributed to the fact that PARPis are good candidates for TDM because they satisfy most of the criteria reported in the literature.³⁰ Moreover, interactions with other drugs, as well as renal and hepatic impairments, are situations in which clinical monitoring may be especially useful. In addition, the relationship between exposure and response to PARPis needs to be better defined. Although some clear relationships were found for both safety and efficacy, PK thresholds were not established for all the reviewed drugs. This ambiguity arises from various reasons, including frequent dose modification and treatment discontinuation during clinical trials. Accordingly, further studies should be conducted to mitigate this. TDM might be a useful tool to accomplish this task with the additional aims of reducing toxicity and increasing treatment efficacy.

This comparison of PARPi quantification methods highlighted the most relevant aspects to consider when choosing the appropriate method from the literature to reproduce. The experimental procedures of a method and its ease of transfer from one laboratory to another are crucial in the choice. Detailed information regarding all these aspects is necessary if the method has to be reproduced and used by the scientific community. In addition, the validation of these bioanalytical methods is fundamental, and several guidelines exist for different purposes. The lack of information provided by some authors and the heterogeneity of validation protocols are significant issues for the analyst that adopts a published method. In the future, newer and standardized guidelines focusing specifically on the clinical practice of TDM are required to help the scientific community develop robust and reliable bioanalytical methods.