An automated mass spectrometric blood test for therapeutic drug monitoring of infliximab

Graphical abstract

Highlights

• •LC–MS/MS method for quantification of infliximab suitable for routine clinical use.
• •Streamlined sample preparation to reduce assay time, cost, and complexity.
• •Method accurate to drug manufacturer target concentrations.
• •No interference in the presence of endogenous antibodies to infliximab.

Abstract

Infliximab is a monoclonal antibody therapy used to treat several chronic immune-mediated diseases, including Crohn’s disease, ulcerative colitis, and rheumatoid arthritis. Infliximab acts by binding to tumor necrosis factor and, thus, inhibiting the inflammatory cascade. While it is a highly effective therapy, a subset of patients on infliximab will develop a loss of response to therapy. In these circumstances, therapeutic drug monitoring of infliximab offers a rational approach to clinical decision making and is associated with improved outcomes. While infliximab has most commonly been measured by immunometric approaches, mass spectrometric approaches offer the opportunity to improve test accuracy and reduce test costs. Herein, we describe a simple, bottom-up high performance liquid chromatography tandem mass spectrometry (LC–MS/MS) approach for quantitation of infliximab in serum. Method development included pre-digestion and digestion experiments to determine critical sample preparation steps, optimization of the workflow and selection of rapidly produced proteolytic peptide(s) for quantitation. The workflow was further improved by automating all sample preparation steps on a robotic liquid handler, facilitating implementation in routine clinical use. A method comparison was performed against a Health Canada and US Food and Drug Administration licensed enzyme-linked immunosorbent assay. Our LC–MS/MS assay accurately reported concentrations based on drug manufacturer targets and demonstrated no interference from endogenous antibodies to infliximab; immunoassay methods did not share these performance characteristics. This LC–MS/MS method provides a workflow amenable to implementation in a clinical laboratory and desired performance characteristics for guiding clinical decision making.

1. Introduction

Infliximab (IFX) is a monoclonal antibody therapeutic used to treat chronic inflammatory disorders, including Crohn’s disease, ulcerative colitis, and rheumatoid arthritis. IFX acts by binding to tumor necrosis factor (TNF) and, thus, inhibiting the inflammatory cascade. While it is a highly effective therapy, a subset of individuals will develop loss of response to IFX. In these circumstances, therapeutic drug monitoring (TDM) of IFX offers a rational approach to clinical decision making [1], [2], [3] and is associated with improved outcomes [4], [5]. In the IFX TDM workflow, samples with low serum or plasma IFX are reflexed for testing for the presence of endogenous antibodies to infliximab (ATI), for which the IFX cut-off for reflexive testing is assay dependent (reviewed in [6]).

TDM for IFX in clinical practice has relied almost exclusively on the use of ligand binding assays including sandwich immunoassays (e.g., enzyme-linked immunosorbent assays [ELISA]) and the homogenous mobility shift assay. The benefits of immunometric approaches include relatively high analytical sensitivity and the availability of commercial assay kits. For clinical laboratories, the requirement of antibodies in the reagent kit typically results in costly and restrictive batch formatting (e.g., 96-well plates or multiples of 8-well strips). Moreover, as sandwich immunoassays and the mobility shift assay indirectly detect IFX via molecular binding interactions, they lack selectivity [6], [7]. For mobility shift assays, this is particularly troubling as it can result in the reporting of the erroneous therapeutic molecule [8].

Immunometric IFX methods have also demonstrated discordance and several lack evidence of accuracy based on gravimetric targets [9], [10]. Immunoassays are also well known to be susceptible interferences from endogenous (e.g., heterophile antibodies, analyte-binding proteins) and exogenous sources, resulting in falsely increased or decreased results [7]. Specific to IFX, is the documented analytical error in IFX measurement associated with the presence of ATI and its ligand TNF [10]. A comparison of four different immunoassays revealed inaccurate measurement of IFX in the presence of ATI, with reduced recoveries in the range of 6–75% of the actual IFX concentration [10]. These types of interferences are concerning with ATI’s developing in up to 61% of individuals receiving IFX [11] and significantly higher concentrations of plasma/serum TNF found in inflammatory bowel disease [12], [13].

Amid these concerns, an alternate method is needed that is: 1) selective and directly measures IFX, 2) unaffected by interferences common to immunometric assays, 3) benchmarked against gravimetric targets, and 4) suitable for clinical lab implementation. Mass spectrometry, specifically high performance liquid chromatography tandem mass spectrometry (LC–MS/MS), provides such a solution. LC–MS/MS instruments are ubiquitous in clinical laboratories, having been used for decades for quantification of small molecules based on their ability to selectivity distinguish between closely related molecules and avoid common immunoassay interferences [7]. For proteins, LC–MS/MS provides similar advantages enabling selective detection of the analyte of interest.

In a bottom-up LC–MS/MS method, the target molecule is enzymatically digested and one or several unique proteolytic peptides are selected for monitoring. The identity (sequence) of the peptides is confirmed via further fragmentation in the mass spectrometer. Known transitions from the intact peptide to its fragments are monitored via a process known as multiple reaction monitoring (MRM). As this LC–MS/MS approach requires no ligand or antibody binding processes for analyte capture or detection, it avoids interferences common to immunometric assays. LC–MS/MS approaches for quantitation of infliximab have been previously described [14], [15], [16], [17]; however, the lengthy (e.g., overnight tryptic digest) and complicated (e.g., analyte purification) workflows proposed warrant reconsideration, as we describe herein, to improve the feasibility for use by clinical laboratories routinely performing TDM for IFX.

2. Material and methods

2.1. Material sources and instrumentation

Ammonium bicarbonate [A6141], ammonium sulfate [A4418] 2,2,2-trifluoroethanol (TFE) [T63002] and formic acid (FA) [F0507] were obtained from Sigma-Aldrich (Canada). Acetonitrile [A955] and isopropyl alcohol [A461] were obtained from Fisher Scientific (Canada). Tosyl phenylalanyl chloromethyl ketone-treated trypsin [LS003744] was obtained from Worthington (USA). Lyophilized IFX (Remicade®, Janssen Biotech, Inc. Canada) was sourced from the drug manufacturer. For internal standards (IS), stable-isotope-labeled peptides were synthesized by New England Peptide (USA). Eppendorf 2.0 mL Protein LoBind tubes [13-698-795], Eppendorf 1.5 mL tubes [0540734], NUNC 2 mL 96-deep well plates (DWP) [12-565-606] and capmats [12-565-570] were purchased from Fisher Scientific (Canada), and 12 × 75 mm polystyrene tubes [CA60830-021] from VWR (Canada). The ELISA method used was by Immunodiagnostik (30-INFHU-E01; ALPCO, USA).

LC–MS/MS sample preparation was automated on a STARlet robotic liquid handler (Hamilton, USA). Peptides were resolved using Kinetex 2.6 μm C18, 100 Å, 50 × 3.0 mm column (Phenomenex, USA) on a Shimadzu 20AD LC system and analyzed on a SCIEX QTRAP 5500 triple quadrupole mass spectrometer.

2.2. Specimens

All specimens included in this study were received at St. Paul’s Hospital for TDM of IFX as part of routine care. Specimens were aliquoted and stored at −70 °C prior to analysis. For clinical testing, one aliquot was referred to the University of Alberta for measurement of free IFX using an ELISA from Immunodiagnostik. A second aliquot was used for in-house LC–MS/MS analyses; additional in-house analyses (e.g. interference studies) were performed on the ELISA from Immunodiagnostik. For LC–MS/MS assay development (e.g., calibrators and controls), we prepared both IFX-naïve and IFX-treated patient serum pools from discarded patient specimens. This study was undertaken with Providence Health Care Research Institute ethics approval.

2.3. LC–MS/MS calibrator preparation

Deionized water was added to a lyophilized vial of Remicade® to obtain a 10 mg/mL stock solution. This stock solution was serially diluted with deionized water to 1 mg/mL and 100 μg/mL working solutions. Stock and working solutions were aliquoted in 2 mL tubes and stored at −70 °C. Calibrators were prepared by spiking exact volumes of the stock and working solutions into pooled IFX-naïve serum to make calibrators at 0.5, 1, 2.5, 5, 7.5, 10, 20 and 40 μg/mL of IFX. Controls for validation experiments were prepared similarly, to make 4 controls at 1, 6, 12 and 25 μg/mL of IFX. Calibrators were aliquoted into 1.5 mL tubes and controls into 12 × 75 mm tubes and stored at −70 °C.

2.4. Proteotypic peptide selection and MRM design

We considered the following unique IFX tryptic peptides, i.e. not found in the human proteome, for quantitation based in part on previous analyses [15]: YASESMSGIPSR (light chain), DILLTQSPAILSVSPGER (light chain), SINSATHYAESVK (heavy chain) and GLEWVAEIR (heavy chain). We evaluated their suitability for MRM design, surrogates for measurement of the intact molecule, and ability to produce rapid and stable digestion kinetics.

2.5. Streamlining sample preparation to create a simple and rapid digestion workflow

Using existing sample preparation workflows as a guide [14], [15], we explored means to develop simpler and faster workflows. Steps in the sample preparation process (i.e., denaturation, reduction, alkylation, digestion) were studied to determine if they resulted in improvements in the method as judged by overall ease of the workflow, generation of rapidly released proteotytic peptides from IFX, and improvement in signal intensity and/or reproducibility. Any unnecessary reagents and/or steps were omitted or replaced with alternate methods.

As part of this analysis, we performed digestion time course experiments analyzing 0, 5, 15, and 30-minute, and 1, 2, 5, 10 and 24-hour time-points.

Using parameters previously developed by our group [18], we explored the use of endogenous human IgG1 and IgG2 (i.e., proteotypic peptide GPSVFPLAPSSK and GLPAPIEK, respectively) as internal digestion controls. As calibrators and controls are prepared in human serum pools, the signal intensity (peak area) for total IgG1 or IgG2 in the calibrators was used as a reference (mean IgG1 or IgG2 concentration), to which individual patient specimens were compared.

2.6. LC–MS/MS optimization

MS parameters were optimized via direct injection of trypic digests of IFX in buffer. Precursor and product ion scans were performed to determine potential MRMs. Source parameters (TEM, ISVoltage, GS1, GS2, CAD and CUR) and compound dependent parameters (CE, CXP, DP, EP) were subsequently optimized for the selected MRMs.

LC parameters were developed to resolve interferences in the sample extracts while maintaining a reasonable run time per injection. For LC, mobile phases utilized included (A) 0.1% formic acid in water and (B) 0.1% formic acid in 25:75 isopropyl alcohol:acetonitrile, which were run using a 0.28 mL/min flow rate with a column temperature of 45 °C. Twenty-five microliters of sample extract were injected and peptides separated with the following gradient of mobile phase B: 5% from 0 to 1.5 min, 5–25% from 1.5 to 7.5 min, 25–65% 7.5–9.5 min, 65% for 9.5–10.5 min, followed by re-equilibration for 2 min to initial conditions. Total run time was 12.5 min.

2.7. Method validation and comparisons

Method validation followed the CLSI C62-A guideline and related recommendations [19], [20]. Precision studies were performed using a modification of the CLSI EP05-A3 protocol by doing quintuplicate measurements over five days with spiked human pools over a period of 5 weeks [21]. Imprecision was expressed as the % CV with acceptance criteria set at ≤20% CV at the lower limit of the measuring interval (LLMI) and ≤15% at all other levels. The LLMI was defined as the analyte peak having a signal-to-noise (S/N) > 20 and was determined by diluting 15 individual patient specimens into blank serum to ∼0.5 μg/mL.

Linearity and the analytical measuring range were determined by spiking a blank serum pool at 100 μg/mL, and diluting at 1:2, 1:5, 1:10, 1:20, 1:50, 1:100 and 1:200 into blank serum (samples were prepared in triplicate).

Carryover for light and heavy chain peptides was assessed by injecting sample extracts with high IFX concentrations followed by blank samples. Significant carryover was defined by detection of an IFX peak (peak area > 20% of LLMI) in the blank samples following the high IFX samples.

Ion suppression was assessed by post-column infusion, whereby 1 μg/mL mixture of the IS peptides was infused into the MS, post-column, at 7 μL/min. Ten unique patient samples were extracted without the addition of IS and injected through the LC to the MS.

IFX sample stability was assessed at room temperature (1, 3 and 7 days) and 4 freeze-thaw cycles by stressing aliquots of the control pools at various conditions and comparing them to non-stressed controls. Sample extract stability was determined by re-injecting two separate batches 72 h after original injection and storage at 2–8 °C. Peptide stability was determined by spiking the IS peptides before and after tryptic digest.

A method comparison was performed between our LC–MS/MS assay and the ELISA, with data analyzed using Passing-Bablok regression. Potential interferences, including hemolysis, lipemia and icterus were evaluated by spiking hemolysate, intralipid and bilirubin into IFX-containing serum pools. ATI interference was tested by assessing IFX recovery in the presence of increasing amounts of ATI, using (1) a recombinant, neutralizing anti-idiotypic ATI (BioRad, HCA213) and, (2) mixture of ATI-negative (ATI⁻) and ATI-positive (ATI⁺) samples. For the recombinant ATI+ samples (rec-ATI⁺), rec-ATI⁺ was added to a serum pool containing 5 μg/mL IFX. For the ATI⁻ sample, IFX-naïve serum samples were pooled and IFX was spiked in to a concentration within the therapeutic range (12 μg/mL). For the ATI⁺ sample, serum was used from a patient who had received IFX therapy and had an ATI concentration above the upper limit of the measuring interval (>20 μg/mL) and undetectable IFX by both ELISA and LC–MS/MS. Mixtures of the rec-ATI⁺, ATI^-, and ATI⁺ samples were incubated for 18 h at room temperature prior to analysis in triplicate.

2.8. Data analysis

Data was analyzed using Analyst® software (SCIEX v.1.6), cp-R [22], and Excel (Microsoft). To normalize and visualize the internal digestion control data, raw immunoglobulin sub-type peak areas for calibrators, controls and samples were divided by the mean calibrator peak area for that sub-type.

3. Results

3.1. Peptide selection and digestion kinetics

A proteolytic digest MRM method was developed for IFX (Fig. 1). Of the four IFX proteolytic peptides initially considered for quantitation, only YASESMSGIPSR and GLEWVAEIR were studied further after the initial LC–MS/MS analysis based on their higher signal intensity relative to the other two peptides. The light chain peptide YASESMSGIPSR demonstrated rapid and stable production, reaching a digestion asymptote by 30 min, and was therefore selected for quantitation (Fig. 2).

Fig. 1.

Schematic of the automated IFX LC–MS/MS assay.

Fig. 2.

A brief 1 h trypsin digestion step was required for the IFX light chain peptide (YASESMSGIPSR, solid line) based on digestion time course experiments; the IFX heavy chain peptide (GLEWVAEIR, dashed line) was also monitored but display slower digestion kinetics. Displayed as the mean peak area of three replicates ± SD.

3.2. Streamlined sample preparation workflow

Via analysis of various steps in the sample preparation workflow, it was determined that inclusion of DTT (reduction) and IAA (alkylation) yielded no benefit (Fig. 3). The final sample preparation procedure consisted of protein precipitation of 100 μL of patient sample (or calibrator or control) using a 50% (w/v) saturated ammonium sulfate solution, followed by reconstitution of the protein pellet with 100 μL of 50 mM ammonium bicarbonate. Samples were incubated at 55 °C for 30 min after addition of 75 μL of TFE, after which 100 μL of denatured sample was transferred to a DWP, followed by 400 μL of deionized water, 100 μL of 50 mM ammonium bicarbonate, and 50 μL of 2 mg/mL trypsin prepared in deionized water. After a 1 h incubation at 37 °C, 20 μL of concentrated formic acid was added to arrest digestion. Ten microliters of a 10 μg/mL solution of IS peptides was added. Following centrifugation for 10 min, the plate was loaded on the LC–MS/MS for analysis of the light chain peptide. All pipetting was automated on a robotic liquid handler, with a total sample preparation time of 3.5 h (including the 1 h digestion step).

Fig. 3.

Effect of changing sample preparation and digestion conditions including omitting the use of DTT, IAA, trypsin and TFE, and adding twice the amount of trypsin.

3.3. Internal digestion control

As an internal control for trypsin digestion, we found total endogenous IgG1 and IgG2 to be useful controls, as demonstrated by the analysis of both calibrators and patient samples (Fig. 4).

Fig. 4.

IgG1 (black) and IgG2 (gray) metric plots of calibrators (non-shaded regions), and quality control and patient specimens (shaded region); as an internal digestion control, samples with peak areas 3 SD below the calibrator mean (black and gray dashed-lines, respectively) trigger review.

3.4. LC–MS/MS optimization

Optimal source parameters were as follows: curtain gas, 40; collisionally activated dissociation gas, Med; ionspray voltage, 5500 V; source temperature, 700 °C and gas 1 and 2, 70. MRM parameters can be found in Table 1.

Table 1.

Mass spectrometer parameters for tryptic peptides specific to IFX and to human IgG subtype peptides.

Tryptic peptide	Description	Q1 (m/z)	Q3 (m/z)	DP (V)	CE (V)	EP (V)	CXP (V)
IFX light chain (YAS-)	Quantifier	632.8	834.4	80	30	10	15
	Qualifier	632.8	1050.4	80	30	10	15
	IS	649.8	365.3	80	30	10	15
IFX heavy chain (GLE-)	MRM-1	536.8	488.1	80	25	10	15
	MRM-2	536.8	587.4	80	25	10	15
	IS	543.8	495.4	80	30	10	15
Human IgG1	Digestion control	593.8	418.2	80	9	10	15
Human IgG2	Digestion control	412.7	486.3	80	5	10	15

3.5. Quantitation

Quantitation of IFX was performed using the light chain peptide YASESMSGIPSR with linear, 1/x-weighted regression (Fig. 5, Fig. 6). The heavy chain peptide GLEWVAEIR was also acquired, but not used for quantification of IFX as there was unacceptable carryover.

Fig. 5.

Representative chromatogram demonstrating relevant tryptic peptide MRMs from a human serum specimen containing 5 µg/mL of IFX. An unknown peak (*) with the same MRM as the IFX light chain quantifier is resolved chromatographically.

Fig. 6.

External calibration curve using the IFX peptide light chain peptide YASESMSGIPSR and stable isotope labeled YASESMSGIPSR as the IS.

3.6. Figures of merit from method validation

Within-run and between-run imprecision was 4.0–7.9% and 6.5–12.2%, respectively (Table 2). The LLMI was 0.5 μg/mL with a S/N > 70. Post-column infusion experiments demonstrate no significant ion suppression at the retention time of the IFX light chain peptide MRMs. IFX was stable up to 7 days at room temperature and for 4 freeze–thaw cycles (Table 3). IFX sample extracts were stable for up to 72 h of refrigerated storage with a median difference between the original and re-injection of −3.00% and a Passing-Bablok R² of 0.995. The LC–MS/MS was linear over the range of 0.5–100 μg/mL based on polynomial least squares regression analysis of the linearity dilution series results (Fig. 7).

Table 2.

Precision metrics of the IFX LC–MS/MS method.

Serum pool	Target IFX (μg/mL)	IFX grand mean (μg/mL)	Within-run CV (%)	Between-run CV (%)	Total CV (%)
∼LLMI	1.0	1.03	7.9	12.2	14.5
Low	N/A	6.38	4.9	8.5	9.8
Medium	12.0	11.74	4.1	6.5	7.7
High	25.0	25.21	4.0	7.1	8.1

Table 3.

Room temperature (RT) and freeze/thaw cycle stability.

Serum pool	Target IFX (μg/mL)	IFX mean untreated (μg/mL)	Difference after RT storage (%)				IFX mean untreated (μg/mL)	Difference after freeze/thaw cycles (%)
			1-day	3-day	7-day			4 cycles
∼LLMI	1.0	0.85	2.34	−3.17	6.10		1.03	4.87
Low	N/A	5.55	−1.59	−7.46	6.74		6.39	−2.13
Medium	12.0	10.71	−1.92	−1.81	1.63		11.66	0.51
High	25.0	22.16	−0.54	−6.23	3.43		25.08	−1.36

Fig. 7.

The IFX LC–MS/MS assay was linear over the measurement range from 0.5 to 100 μg/mL.

Differences between replicate analysis of samples spiked with hemolysate, bilirubin or intralipid at increasing amounts were within 15% of the mean concentration of untreated samples (data not shown). There was no effect of increasing amounts of human or recombinant ATI on the accuracy of the measurement of IFX by LC–MS/MS (Table 4, Table 5).

Table 4.

Accuracy of LC–MS/MS measurement of a serum pool containing 5 μg/mL IFX in the presence of recombinant, neutralizing anti-idiotypic ATI (measurements performed in triplicate).

rec-ATI+ (μg/mL)	Mean IFX (μg/mL)	Recovery (%)
0	5.06	101
0.1	5.44	109
0.5	5.27	105
1.0	5.05	101
2.0	5.29	106
5.0	5.34	107
10	5.36	107
25	5.58	112

Table 5.

Accuracy of LC–MS/MS measurement of IFX in the presence of ATI (measurements performed in triplicate).

Mixture (v/v ratio of ATI+:ATI−)a	Expected IFX (μg/mL)	Measured IFX (μg/mL)	Recovery (%)
100:0		0
71:29	8.41	8.18	97.3
83:17	9.81	9.50	96.9
90:10	10.70	10.52	98.4
95:5	11.21	11.07	98.8
98:2	11.54	10.97	95.1
0:100		11.77	100

^aATI⁺: A serum sample with a high ATI concentration with undetectable IFX; ATI⁻: Serum sample from IFX-naïve individual, with a known concentration of IFX added.

The method comparison with a US Food and Drug Administration and Health Canada licensed ELISA kit revealed the methods were positively correlated, but demonstrated significant bias (Fig. 8). To explore the observed method bias, both methods were compared against the IFX drug concentration assigned by the manufacturer demonstrating the accuracy of the LC–MS/MS assay, whereas the ELISA demonstrated under-recovery of IFX in the range of −18 to −55% (Table 6).

Fig. 8.

Method comparison (n = 112) between the LC–MS/MS and ELISA assays. A) Linear regression: 95% confidence interval of the slope (shaded region) [2.087, 2.385], 95% confidence interval of the intercept [0.409, 1.073], line of identity (dashed line). B) Difference of measures: median difference [87.4%], 95% confidence interval (shaded region).

Table 6.

Accuracy of the LC–MS/MS and ELISA methods compared to manufacturer-assigned IFX mass.

Target	LC–MS/MS			ELISA
IFX (μg/mL)	Measured IFX (μg/mL)	Absolute bias (μg/mL)	Bias (%)	Measured IFX (μg/mL)	Absolute bias (μg/mL)	Bias (%)
0.50	0.59	0.09	18	0.4	−0.1	−20
2.5	2.6	0.1	5.6	1.12	−1.38	−55
5.0	5.3	0.3	5.0	2.55	−2.45	−49
10.0	10.6	0.6	6.0	8.12	−1.88	−18
20.0	18.3	−1.7	−8.5	12.07	−7.93	−39

4. Discussion

We have developed a streamlined sample preparation workflow for the quantitation of IFX by LC–MS/MS. We determined that commonly used reduction and alkylation steps were unnecessary when other aspects of sample preparation were optimized. In addition to reducing time, eliminating reduction and alkylation steps reduced the complexity of the sample preparation. The workflow was further streamlined by using an internal digestion control (i.e., endogenous IgG1 or IgG2), thus avoiding the cost and additional steps associated with the addition of an exogenous digestion control. Specimens with internal digestion controls <3 SD below the calibrator mean trigger medical review and, potentially, repeat analysis if deemed warranted. Utilizing digestion kinetics to help select an appropriate peptide for quantitation, we reduced the proteolytic digestion step to 1 h. With the simplification of sample preparation, all steps were automated on a liquid handler with a total sample preparation time of 3.5 h and implemented in our clinical laboratory for IFX quantitation.

A driving motivation for the development of a quantitative IFX LC–MS/MS method was concern related to the accuracy of available immunoassays, and lack of standardization (i.e., comparability of IFX results) between manufacturers and laboratories. Sandwich immunoassays for IFX are subdivided into “free” and “total” assays, the latter involving an extra step in the procedure (e.g., acid-dissociation) to disrupt binding of IFX to potential ATIs. We found that a free-IFX ELISA variably under-recovered IFX in the range of −18 to −55% of the manufacturer-assigned concentration and previous studies have demonstrated lack of concordance between various immunoassay methods [9], [10]. The IFX LC–MS/MS assay was able to reproduce manufacturer-assigned drug concentrations with minimal bias.

Accuracy concerns extended to measurement of IFX in the presence of clinically relevant molecules including ATI. Previous studies of IFX immunoassays have demonstrated variable and unpredictable interference in the presence of ATI, with increasingly negative interference at higher ATI concentrations. The proteolytic digest LC–MS/MS method, however, was not affected by the presence of ATI, even at high ATI titers. The LC–MS/MS method does not use molecular interactions for quantitation, such as antibody-analyte binding, and, therefore, is not affected by common immunoassay interferences that can lead to false negative and false positive results including ATI binding to IFX, TNF binding to IFX, or endogenous polyreactive antibodies binding to the immunoassay substrates (e.g., capture and detection antibodies) [6], [7]. Identification of analytical interferences commonly occurs when the test result is incongruous with the clinical picture (e.g. elevated troponin result with no other evidence of a myocardial infarction). In the case of IFX TDM such “incongruous” results are unlikely as testing is commonly ordered when there is loss of response to therapy and in this setting IFX results in the range of undetectable to ‘high’ are plausible and clinically actionable [6]. Due to difficulties in identifying analytical interferences in this clinical setting, it is desirable to have a method, like LC–MS/MS, that is not susceptible to these particular analytical interferences.

Ideally a clinical test for TDM of an exogenous analyte would be standardized. Assay standardization would enable comparability between testing performed by different labs/methods; for individuals on IFX therapy, this would mean that they do not have to have their testing performed by the same laboratory to enable meaningful comparison to their previous result(s). While immunoassays cannot achieve standardization due to reliance on different antibodies and epitopes, harmonization is possible. The LC–MS/MS approach to IFX quantitation in serum, calibrated to gravimetric targets, demonstrates the future potential for standardization of IFX LC–MS/MS TDM assays.

5. Conclusion

Via optimization of sample pretreatment steps, including elimination of unnecessary reduction and alkylation steps, and consideration of rapidly produced tryptic peptides, we created a simple and rapid workflow for quantitation of IFX by proteolytic digestion LC–MS/MS. The LC–MS/MS method had minimal bias to manufacturer-assigned IFX gravimetric targets and no analytical interference in the presence of ATIs, demonstrating its suitability for TDM in this clinical population.

Conflict of interest statement

J.G.V. and M.L.D. have nothing to disclose. B.B. reports financial support from Janssen Biotech, Merck and Pfizer, outside the submitted work.