Optimized LC-MS/MS methods for quantifying antibody–drug conjugate payloads in cell culture media containing phenol red

Abstract

Aim: Phenol red is commonly used in cell culture media, but can be detrimental to bioanalysis of in vitro samples as it may impact instrument reliability. Many researchers do their final stage of culture in ‘phenol red free’ media, but in collaborative work this is not always feasible.

Materials & methods: A comparison was made between typical extraction methods to reduce phenol red matrix interferences, including organic solvent precipitation and solid phase extraction.

Results: The final method was demonstrated to be precise and accurate for the measurement of a target analyte by LC-MS/MS, and was applied to an in vitro ADC deconjugation study.

Conclusion: This method allows for for continued bioanalytical support of in vitro models used in drug development.

Graphical Abstract

Plain language summary

Summary points.

• In vitro models are being widely used in oncology drug discovery, but from an analytical perspective the cell culture media components may cause significant issues.

• Before LC-MS/MS analysis, analytes of interest are traditionally extracted from biological matrices, and the same application is required for in vitro study samples.

• This article describes the application of three approaches for the removal of phenol red during the extraction of an oncology treatment drug, and compares the relative matrix effects and recovery.

• Liquid–liquid extraction was found to have the best analyte recovery (67%) whereas HLB solid phase extraction was found to show the least impactful matrix effect (11%).

• Adjustment of the solid phase extraction material to include ion exchange interactions showed a marked improvement in phenol red filtration, and reduced matrix effect (1%).

• Some fit for purpose criteria were assessed for the target analyte, and found to be precise and accurate over a suitable range.

• The final method was applied to in vitro study samples and generated useful data for the application of a 3D bone marrow model for oncology drug development.

• Future applications of this extraction method to other target analytes in cell culture media will allow for accurate drug analysis while reducing the impact of phenol red on the analytical system and accuracy of the data.

1. Background

Antibody–drug conjugates (ADCs) are widely used for cancer treatment and are considered as a type of targeted therapy, designed to selectively deliver cytotoxic drugs (payload) to cancer cells while minimizing damage to normal healthy cells. The antibody component is designed to specifically recognize and bind to antigens expressed on the surface of cancer cells and selectively deliver a potent drug that will induce apoptosis. The bioanalytical targets therefore include both the ADC and small molecule payload, and sensitive robust assays are required for each. During the drug development process the toxicity, viability and stability (deconjugation) of ADC candidates must be assessed. Recent advancements in 3D in vitro systems suggest promising and ethical alternatives for drug development [1,2]. Scaffold-based systems and organoids [3], are useful for measuring cell responses as they have metabolic competency [4] and depending on the target tissue, the models can be adapted to have the most suitable conditions and culture additives for progenitor maintenance, lineages differentiation and cell proliferation [5].

Specifically for drug safety evaluations in bone marrow, a 3D humanized in vitro model creates a physiologically relevant system that closely replicates the intricate cellular interactions and microenvironments[6]. This model enhances translational value in evaluating drug hematotoxicity such as the effects of drug treatment on cell lineage and differentiation [7], providing more precise predictions of human responses, and mitigating the risk of species-specific variations in drug effects. The general layout of a 3D bone marrow scaffold involves an assay plate insert, which mimics human BM pore size and structure, and a layer of mesenchymal (MSC) and hematopoietic (HSC) stem cells, differentiated to maintain a similar environment to in vivo [4,8]. It is key that the scaffold cells are cultured for a few days before experimentation, to allow for target cells to have a suitable differentiation before drug application.

An essential component of successful in vitro systems is the cell culture medium, which has specific additives (such as amino acids, vitamins and minerals and growth factors) to promote efficient cell growth and differentiation [9,10]. In addition to nutrients, the media may also contain physiochemical factors, such as pH indicator buffers [11], with phenol red (phenolsulfonphthalein) being the most widely used. Phenol red is a weak acid (pKa ≈ 8) and has a zwitterion structure in acidic conditions (Figure 1A) [12]. An obvious colour change occurs in solutions containing phenol red, over the pH range of 6–8, which makes it a suitable monitor for increased carbon dioxide levels in cell culture platforms [11].

Figure 1.

Phenol red and bioanalytical extraction techniques. (A) Chemical structures and colours of phenol red (phenolsulfonphthalein) when in different pH conditions, making it a suitable indicator of pH change in cell culture media. (B) Three bioanalytical extraction methods which could remove phenol red from cell media prior to analysis (image not to scale, created in Biorender.com). (C) The process of matrix effect and recovery assessment, performed for each extraction method.

Image not to scale, created in Biorender.com.

While a useful indicator of cell stress and oxidative activity, phenol red may sometimes be a hindrance to in vitro experiments – such as impacting redox reactions [13] or acting as a hormonal agonist [14] [15] [16] [17]. Interactions of phenol red with target analytes can also be detrimental to assay results and may lead to the formation of cytotoxic artifacts under certain circumstances [18] [19]. The presence of phenol red in a cell culture sample can also interfere with the implementation of chromogenic assays such as ELISA or imaging, specifically across the yellow and red absorbance ranges (420–430 and 500–520 nm). In chromatographic assays, phenol red could severely interfere with the separation on HPLC columns [20]. As shown in Figure 1, the structure of phenol red presents functional groups that can form both polar and ionic interactions with the analytical column, suggesting it is a highly retentive material. It may also coelute with target analytes, causing difficulties for HPLC-UV detection and purification assays.

To ensure the supernatants are suitable for analysis, the majority of in vitro bioanalysis studies have the final few days of cell culture or secretion experiments in phenol red free media[21] [22] [23] [24], at the risk of not monitoring oxidative stress in the system. But in some instances of specific cell culture or collaborative research, the samples may already be collected in phenol red media and additional sample clean up procedures are required.

Some simple entrapment or filtration assays have been shown to be suitable at removing phenol red [25] but these may not be compatible with LC-MS/MS analysis. Alternatively, assuming chromatographic separation is possible, a divert valve could be used to prevent phenol red being introduced into the mass spectrometer. However, the development of a suitable extraction technique allows for reduced pressures and longevity of the LC column. Extraction assays often focus on manipulating physiochemical properties such as hydrophilicity, to remove target analytes from complex biological samples [26]. There are a number of well used techniques which can do this, for example organic solvent precipitation and immiscible organic and aqueous solvent interactions. By comparing the logD values of phenol red (approximately 3 at low pH) and common cytotoxic drugs (i.e., Exatecan, logD of -2 at low pH) it may be possible to transfer the target analyte to a suitable supernatant while retaining the phenol red in aqueous buffer.

During a recent development of a novel ADC compound containing a topoisomerase I inhibitor payload, some experiments required the quantitation of payload only following ADC incubation in a bone marrow in vitro model. The key aim of the methods detailed in this manuscript was to reduce the amount of phenol red in the final extracts, while trying to optimize the recovery of the target analyte. For improved extraction of the small molecule payload, three techniques were compared (Figure 1B) and investigations into assay sensitivity, matrix effect and recovery were concluded. In addition, the amount of phenol red in each final sample extract was monitored, with the aim of finding a specific, selective and suitable assay for significant reduction of the matrix interference in cell media samples.

2. Materials & methods

Unless stated otherwise, all reagents were commercially sourced and used as supplied. HPLC grade methanol, acetonitrile, dimethyl sulfoxide (DMSO), ethyl acetate and water (Fisher Scientific, Loughborough, UK) were used for both extraction and LC-MS/MS solvents. Reagent grade formic acid and acetic acid (Sigma-Aldrich, Poole, UK) were used for solvent preparation.

Intermediate solutions of Drug A and its stable isotopically labelled form (known as Drug A-IS) were prepared in DMSO and stored at -20°C. These were used to prepare calibration standards and quality controls (QC), in blank StemSpan™ SFEM II media (StemCell Technologies, France), fortified with specific cytokines (Peprotech, London, U.K.) for hematopoietic stem cell differentiation.

All experimental procedures were performed on an H-Class Acquity (Waters, MA, USA) LC system coupled to an AB SCIEX 6500+ triple quadrupole mass spectrometer (Sciex, MA, USA). The target analytes and internal standard were detected using specific Q1 and Q3 ions and quantified by multiple reaction monitoring. Chromatographic data were processed using Sciex Analyst (version 1.7.2) and calculations and data analysis were performed in Microsoft Excel.

3. Extraction methods

The following extraction methods were compared for the analysis of calibration standards (0.05–50 ng/ml) and QCs (0.15, 4, 40 ng/ml) prepared in cell culture media, containing phenol red. In addition, matrix effect and recovery samples were prepared using blank water and media, respectively, which were spiked post extraction with Drug A at the expected final extracted concentrations for a 4 ng/ml QC sample (350 pg/ml QC for the WAX SPE method) (Figure 1C).

3.1. Protein precipitation

25 μl calibration standards and QCs were aliquoted into a LoBind 96 well plate, 25 μl internal standard working solution (2.5 ng/ml Drug A-IS in 20% acetonitrile (aq)) was added and the plate was mixed for 1 min at 1000 rpm. The samples were precipitated with 250 μl acetonitrile, mixed for 2 min at 1000 rpm and centrifuged at 3500 ×g for 5 min. A 200 μl aliquot of supernatant was transferred to a clean plate and evaporated under nitrogen at 40°C. The samples were reconstituted in 75 μl 0.2% formic acid (20% acetonitrile (aq)), mixed for 3 min at 1200 rpm and centrifuged at 3000 ×g for 3 min, prior to injection on the LC-MS/MS.

3.2. Liquid–liquid extraction

25 μl calibration standards and QCs were aliquoted into a LoBind 96-well plate, 25 μl internal standard working solution (2.5 ng/ml Drug A-IS in 20% acetonitrile [aq] was added and the plate was mixed for 1 min at 1000 rpm. To the samples, 250 μl of ethyl acetate was added before being mixed for 2 min at 1000 rpm, and centrifuged at 3500 ×g for 5 min. A 200 μl aliquot of supernatant was transferred to a clean plate and evaporated under nitrogen at 40°C. The samples were reconstituted in 75 μl 0.2% formic acid (20% acetonitrile [aq]), mixed for 3 min at 1200 rpm and centrifuged at 3000 ×g for 3 min, prior to injection on the LC-MS/MS.

3.4. Solid phase extraction

25 μl calibration standards and QCs were aliquoted into a LoBind 96 well plate, 25 μl internal standard working solution (2.5 ng/ml Drug A-IS in 1% formic acid in 20% acetonitrile [aq]) was added and the plate was mixed for 1 min at 1000 rpm. All of the diluted sample was loaded onto an Oasis μElution HLB solid phase extraction plate (Waters) and held under positive pressure on a SPE manifold. The SPE plate was washed sequentially with 150 μl 0.1% formic acid (aq) and 150 μl 1% acetic acid in 10% methanol (aq), and samples eluted into a clean plate with 75 μl 5% acetic acid in 60% methanol (aq). The sample eluents were evaporated under nitrogen at 40°C and reconstituted in 75 μl 0.2% formic acid in 20% acetonitrile (aq). The plate was mixed for 3 min at 1200 rpm and centrifuged at 3000 ×g for 3 min, prior to injection on the LC-MS/MS.

For assay comparison, an Oasis μElution WAX solid phase extraction plate (Waters) was preconditioned with 150 μl 0.2% formic acid (aq) and then the premixed 25 μl sample and 25 μl internal standard working solution (2.5 ng/ml Drug A-IS in 1% formic acid in 20% acetonitrile [aq]) was added. The SPE plate was washed sequentially with 150 μl 0.1% formic acid (aq) and 150 μl 1% acetic acid in 10% methanol (aq), and samples eluted into a clean plate with 75 μl 5% acetic acid in 60% methanol (aq). The samples were evaporated under nitrogen at 40°C and reconstituted in 75 μl 0.2% formic acid in 20% acetonitrile (aq). The plate was mixed for 3 min at 1200 rpm and centrifuged at 3000 ×g for 3 min, prior to injection on the LC-MS/MS.

4. LC-MS/MS methods

The extracted sample (10 μl) was injected onto an Acquity BEH C18 Column (2.1 mm x 50 mm, 1.7 μm, Waters) set at 60°C, with starting mobile phases set to 90% A (0.2% formic acid [aq]) and 10% B (0.2% formic acid in acetonitrile) and a flow rate of 0.5 ml/min. The analytes were separated over a 2 minute gradient from 10 to 40% B, followed by a flush of 98% B for 1 min before returning to initial conditions, resulting in an overall runtime of 4 min. The sample manager wash solvent was ethanol, and the sample purge and seal wash solvent were 10% acetonitrile (aq).

Turbo ion spray ionization was applied in positive mode, within the mass spectrometer source (500°C), with an ion spray voltage of 3 kV, curtain gas of 30 psi and a high collision gas setting. The analytes were monitored by unit/unit resolution, and the specific transitions and voltages for Drug A, internal standard (Drug A-IS) and phenol red are described in Table 1.

Table 1.

Specific transitions and voltages for the LC-MS/MS method.

Analyte	Q1 (m/z)	Q3 (m/z)	Dwell time (ms)	DP	CE	CXP	Approx. retention time (min)
Drug A	404.2	360.0	45	130	45	20	0.9
Drug A-IS	409.0	365.3	45	130	45	20	0.9
Phenol red	355.3	261.1	45	130	28	20	0.8

Masses are nominal, and parameters are instrument specific so may vary between instruments.

CE: Collision energy; CXP: Cell exit potential; DP: Declustering potential.

5. Results & discussion

For all extraction methods, the transitions for the target analyte (Drug A), the internal standard (Drug A-IS) and phenol red were monitored and example chromatograms are shown in Figure 2. For all experiments a precision and accuracy assessment was performed, with an acceptance criteria of ± 20% (Table 2). The sensitivity of the drug was similar between methods, and all lower limit of quantitation (LLOQ) samples (0.05 ng/ml) show good chromatography with sufficient signal to noise (greater than 5:1). But there were some differences in phenol red response between each extraction technique, and the recovery and matrix effect results for each assay were compared (Table 3).

Figure 2.

Chromatograms of Drug A and phenol red from LLOQ (0.05 ng/ml) samples, extracted by either (A) protein precipitation, (B) liquid–liquid, or (C) HLB solid phase extraction.

Table 2.

Precision and accuracy results for media QC samples (n = 4 at each level) when extracted by the each assay.

Protein precipitation	QC low (0.15 ng/ml)	QC med (4 ng/ml)	QC high (40 ng/ml)
Mean concentration	0.144	3.84	33.2
Standard deviation	0.007	0.738	0.873
%CV (Precision)	5.4	19.2	2.6
%RE (Accuracy)	−4.0	−4.0	−17.0

Liquid–liquid	QC low (0.15 ng/ml)	QC med (4 ng/ml)	QC high (40 ng/ml)
Mean concentration	0.130	3.52	33.4
Standard deviation	0.004	0.571	0.416
%CV (Precision)	3.1	16.2	1.3
%RE (Accuracy)	−13.3	−12.0	−16.5

Solid-phase extraction (HLB)	QC low (0.15 ng/ml)	QC med (4 ng/ml)	QC high (40 ng/ml)
Mean concentration	0.163	3.95	37.4
Standard deviation	0.017	0.039	0.661
%CV (Precision)	10.4	1.0	1.8
%RE (Accuracy)	8.7	−1.3	−6.5

Table 3.

Recovery and matrix effect values for all of the extraction methods; where %recovery = (extracted sample response/post spiked sample response)*100 and %matrix effect = (post spiked media sample/post spiked water sample)*100

Extraction technique	%CV (Extracted QC)	%CV (post spiked media)	%CV (post spiked water)	%Recovery	%Matrix effect
Protein precipitation	19.2	4.0	0.3	57.2	35.1
Liquid–liquid extraction	16.2	4.1	0.5	66.5	−11.6
Solid-phase extraction (HLB)	1.0	1.5	2.2	58.9	11.2
Solid-phase extraction (WAX)	2.2	2.4	1.8	70.9	0.7

5.1. Protein precipitation

Following precipitation with acetonitrile some phenol red was visible in the final sample as the extracts were slightly pink in colour, even after centrifugation/reconstitution. This was also confirmed from the chromatograms, which demonstrated significant peaks for phenol red. Even though this was the simplest and crudest extraction technique, the chromatography for Drug A had relatively low background noise and suitable linearity (Table 2).

The introduction of a divert valve would allow for the separation of phenol red to elute to waste, but care must be taken not to lose the target analyte also (in this instance the similarities in retention time mean this may not be possible). Also, as shown in Figure 2, phenol red does not have the best chromatographic peak shape, and overall a high background noise was seen during all test LC combinations. As the detrimental effects of long term exposure of phenol red on the analytical column and LC-MS system are currently unknown, if alternative approaches give cleaner extracts these may be more favourable.

5.2. Liquid–liquid extraction

Following ethyl acetate addition there were clearly defined layers of immiscible organic and aqueous solvents, with phenol red visibly remaining in the aqueous layer. This allowed for an easier transfer of supernatant for analysis, and the final extracts were visibly clear in colour. This was confirmed by the chromatography, as no phenol red peaks were detected. The LLE precision and accuracy performed very well (Table 2); the sensitivity even appeared suitable for a lower analytical range than protein precipitation. Stability QCs and re-extraction results were inconsistent, and lead to some concern over Drug A polarity and isomer structure in neutral extraction conditions. However, the LLE extraction may prove useful in future for separating alternative analytes from phenol red media.

5.3. Solid phase extraction

Based on the polarity of Drug A, the internal standard working solution was adjusted to contain acid, which had a visual change on the samples (pink to yellow). The SPE wash and elution solvents were also acidic, and for the initial experiments with HLB plate the final eluent still had a slight yellow colour. Phenol red was detected by LC-MS/MS, but at a slightly lower response than the protein precipitation method. The sensitivity of Drug A was comparable to the other extraction methods, and QCs were found to be precise and accurate (Table 2).

To understand more about each assay, a solution at the mid QC level (adjusted for 100% recovery) was added to extracted blank media and water samples (n = 6). The peak area of these were compared with the extracted QC medium results and used to calculate the percentage recovery and matrix effect (Table 3). This experiment is performed during bioanalytical assay development and validations, to understand whether the extraction conditions and matrix have an impact on analytical sensitivity and accuracy. The recovery values for all three extraction methods were similar, but the matrix effects ranged from -12 to + 35%. The correlation in matrix effect with phenol red peak area was positive, suggesting that the presence of matrix in the final sample may be increasing the chromatography background noise and therefore affecting quantitative results. Based on these experiments it would seem the liquid–liquid extraction was the most efficient for recovering the analyte (and with the lowest background noise), however as described earlier it was noted that the physiochemical properties of Drug A were not suited to this assay.

Based on the chromatography, QC accuracy, recovery and matrix effect results, it was decided that the solid phase extraction assay would be taken forward, with some additional attempts to optimize the extraction. The SPE plate was changed to utilize ionic interactions, to improve isolation of phenol red from the samples. A weak anionic exchange (WAX) SPE plate has a reversible positively charged piperazine group, which was hypothesized to interacted with phenol red when under acidic conditions (Figure 3). Other ion-exchange sorbents are available (MAX, MCX, WCX) which could theoretically be used in a similar way by changing pH based charge states, creating a versatile toolkit to filter target analytes from phenol red media interference.

Figure 3.

Solid-phase extraction optimization. (A) Differences between weak anion exchange (WAX) and hydrophilic–lipophilic (HLB) chemistries in solid phase extraction plates. Inset a visual example of phenol red retention on a WAX SPE plate, showing pink coloration to the sorbent bed. (B) Peak area responses for phenol red in calibration samples extracted via either WAX or HLB solid-phase extractions, with the mean response tabulated for comparison. (C) Chromatograms for Drug A in QC Low (0.15 ng/ml) samples extracted by WAX and HLB solid-phase extraction.

By changing to a more selective SPE chemistry, an obvious improvement in sample extraction was made with visible confirmation that phenol red was being retained on the SPE plate. Compared with the HLB results, there was a significant decrease of matrix effect, confirmed by the reduced phenol red response. In addition, the ion exchange SPE assay resulted in a slight increase in analyte sensitivity both by a reduction in background noise and an increase in recovery.

This assay was updated to extract 100 μl of sample, to allow for a more sensitive LLOQ. The updated analytical range (5–5000 pg/ml) was assessed for precision and accuracy in both fresh QCs and those stored for four hours storage at two possible extraction conditions – room temperature and on wet ice (Table 4). The only results outside of typical bioanalytical acceptance criteria (± 20%) were the 4 h room temperature QCs (at 4 ng/ml), which confirmed the importance of performing the extraction on wet ice. The assay was also found to be suitable for analysis of stored calibration standards, QCs and study samples (after short- and long-term storage at -70°C, Figure 4).

Table 4.

Precision and accuracy results for media QC samples (n = 6 at each level) when extracted by the WAX solid phase extraction assay, immediately after preparation and after four hours storage at two extraction temperatures (RT; room temperature (approx. 22°C) and on wet ice (approx. 4°C)).

WAX solid phase extraction	QC LLOQ (5 pg/ml)	QC low (15 pg/ml)	QC med (350 pg/ml)	QC high (4000 pg/ml)
Mean concentration	5.68	14.7	349	4002
Standard deviation	0.593	1.50	7.65	84.8
%CV (precision)	10.4	10.2	2.2	2.1
%RE (accuracy)	13.7	−1.9	−0.3	0.1

Extraction stability assessment	QC low 4 h RT	QC low 4 h wet ice	QC high 4 h RT	QC high 4 h wet ice
Mean concentration	12.8	16.1	3028	3640
Standard deviation	0.869	1.02	67.0	75.3
%CV (precision)	6.8	6.3	2.2	2.1
%RE (accuracy)	−14.5	7.0	−24.3	−9.0

Figure 4.

Fit-for-purpose application of the final extraction method. (A) Chromatograms for Drug A, Drug A-IS and phenol red in a LLOQ sample (5 pg/ml) showing acceptable signal to noise, and calibration curve over the final analytical range (linear, 1/x² regression). (B) Comparison of data from all repeat analysis of calibration and study samples after short (7 days) and long (6 months) term storage at -70°C, where %Diff = ((repeat concentration – original concentration)/original concentration)*100. Note for incurred sample reanalysis a %difference result < ± 20% is deemed acceptable by bioanalytical clinical regulations. (C) Concentration data for study samples, from an assessment of Drug A deconjugation in a 3D bone marrow in vitro system, at three concentration levels.

The final method was applied to a small in vitro study, where a relevant ADC was incubated in a 3D bone marrow model and the release of Drug A was quantified over 14 days at three different concentrations (Figure 4C). The data showed a consistent and dose dependant profile of Drug A deconjugation from the ADC, with sensitive quantitation and no matrix interference from phenol red. This initial study confirmed the WAX SPE assay was fit for purpose and can be applied to wide-scale studies involving ADCs with Drug A payload.

6. Application of the assay to other drug analytes

Drug A is not the only candidate for oncology treatment, and as ADC modalities are being developed the solid phase extraction assay may be suitable to a wider market of payload compounds. As with most bioanalytical methods a new target analyte would require at least a precision and accuracy assessment, but as a proof of application the final WAX SPE method was used to extracted two similar payloads from cell media containing phenol red (described as Drug B and C). Both analytes were detected in the extracted samples and were found to be free from phenol red, but some variability in recovery was observed (Supplementary Figure S1). This is due to difference in analyte chemical properties, and some specific changes to elution solvent composition could improve the recovery if required.

7. Conclusion

In vitro bone marrow systems are becoming more widely used to understand the toxicity and interactions of ADCs. The application of regenerative stem cells for continuous measurement of cytotoxic drug activity in human relevant systems[27] [28] reduce the requirement for multiple preclinical rodent studies, and could be considered a more ethical approach for oncology drug development.

Phenol red is a common and useful additive for cell culture assays and in vitro systems which are important for oncology drug optimization, but due to chemical interactions it is preferable that it is removed before LC-MS/MS analysis. As described, three bioanalytical methods were developed and compared for the extraction and clean-up of cell media samples. All assays showed suitable sensitivity of Drug A, but the chromatography results highlighted the advantage of liquid–liquid extraction and ion exchange solid phase extraction for the removal of phenol red. The LLE method using ethyl acetate is a clean and simple technique and, while not compatible with Drug A, may be an efficient assay for other target analytes which do not need pH adjustment/acidification for stability or suitable polarity. The solid phase extraction method was found to be reliable and robust at measuring Drug A over a suitable analytical range and generated some beneficial study sample results for understanding ADC deconjugation in a bone marrow in vitro model.

LLE and SPE methodologies have been presented here that demonstrate the ability to extract analytes from cell media, and remove phenol red in the process. For Drug A, SPE was more appropriate and conditions were optimized to create a robust method to provide concentrations for in vitro assessments to support early stage ADC drug discovery. Overall this manuscript describes solid phase extraction as a viable approach to filter interferences from in vitro matrices, but for future applications and alternative ADC payloads the assay can be adjusted to suit the physiochemical properties of the analyte and the context of use.

Supplementary material

Supplementary data for this article can be accessed at https://doi.org/10.1080/17576180.2024.2349422

Author contributions

RE Foreman, R Lucey and AR Leaney contributed to the methodology and concept of the study, with RE Foreman and R Lucey performing the experiments. H Naseer and A Wilson supervised and managed the project. RE Foreman prepared the original manuscript and all authors contributed to the final version.

Financial disclosure

The authors have no financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Competing interests disclosure

At the time of writing all authors are employees of AstraZeneca. The authors have no other competing interests or relevant affiliations with any organization or entity with the subject matter or materials discussed in the manuscript apart from those disclosed.

Writing disclosure

No writing assistance was utilized in the production of this manuscript.