Simultaneous Quantification of Vedolizumab, Ustekinumab, and Infliximab in Human Serum by LC-MS/MS: Application to Therapeutic Drug Monitoring in Inflammatory Bowel Disease

Abstract

Purpose

With the expanding use of therapeutic monoclonal antibodies (mAbs) in inflammatory bowel disease (IBD), there is a critical need for robust, multiplex analytical tools to support therapeutic drug monitoring (TDM) and pharmacokinetic studies. This study aimed to develop and validate a rapid and sensitive liquid chromatography-tandem mass spectrometry (LC-MS/MS) method for the simultaneous quantification of infliximab (IFX), vedolizumab (VDZ), and ustekinumab (UST) in human serum.

Patients and Methods

A universal immunocapture strategy utilizing Protein G magnetic beads was used to enrich total IgG from human serum. Samples were subjected to tryptic digestion and analyzed in multiple reaction monitoring (MRM) mode, with rituximab as a full-process internal standard. Method validation was performed in accordance with regulatory guidelines, evaluating linearity, precision, accuracy, selectivity, recovery, matrix effect, and stability. The validated method was subsequently applied to residual serum samples from IBD patients undergoing treatment.

Results

The method demonstrated excellent linearity (R²>0.99) over the concentration ranges of 0.5–100 μg/mL for IFX and UST, and 1–200 μg/mL for VDZ. Total precision ranged from 1.1% to 14.9%, and accuracy ranged from −12.1% to 14.5%. Validation parameters, including selectivity, carryover, recovery, matrix effects, dilution integrity, and stability, were all within acceptable criteria. Analysis of the 37 clinical samples revealed marked inter-individual variability in drug concentrations. However, the absence of key covariates (eg, body weight and dosing history) precluded a formal pharmacokinetic analysis or attribution of this variability to specific clinical factors.

Conclusion

The developed LC-MS/MS method provides a practical, specific, and reproducible tool for the multiplex quantification of therapeutic mAbs. Based on a preliminary exploratory study of 37 patient samples, this method demonstrates analytical feasibility and serves as a promising tool for pharmacokinetic research; however, its definitive clinical applicability and readiness for routine TDM require further confirmation through cross-platform comparison, incurred sample reanalysis, and external validation in larger cohorts.

Introduction

Inflammatory bowel disease (IBD) is a chronic, relapsing inflammatory disorder of the gastrointestinal tract with a rapidly global disease burden, particularly in newly industrialized regions such as China.1–6 In recent years, biologic therapies, especially therapeutic monoclonal antibodies (mAbs), have become central to the treatment of moderate-to-severe IBD. However, substantial interindividual variability in treatment response remains a major clinical challenge.7–9 Primary non-response and secondary loss of response are common, and are often associated with insufficient drug exposure caused by anti-drug antibody formation, altered pharmacokinetics, or increased drug clearance.4,10–16

Therapeutic drug monitoring (TDM) has become a pivotal strategy for optimizing biologic therapy in IBD. Current guidelines, including those from the American Gastroenterological Association (AGA) and Australian consensus groups, primarily endorse reactive TDM—targeted measurements in patients with primary non-response or secondary loss of response—rather than advocating for universal proactive monitoring in all patients.17,18 The 2018 Expert Consensus of the Chinese Society of Gastroenterology similarly endorsed TDM-guided decision-making for infliximab (IFX) and adalimumab, emphasizing concentration-based optimization during both induction and maintenance phases. These recommendations are underpinned by extensive evidence linking trough concentrations of IFX, vedolizumab (VDZ), and ustekinumab (UST) to key therapeutic outcomes.19–24 However, the successful implementation of TDM relies not only on clinical algorithms but also on robust bioanalytical methods capable of accurately quantifying mAbs in complex biological matrices.

Currently, two major analytical platforms are used for mAb quantification: enzyme-linked immunosorbent assay (ELISA) and liquid chromatography-tandem mass spectrometry (LC-MS/MS). Although ELISA is widely adopted, it faces limitations including potential epitope masking by anti-drug antibodies (ADAs), restricted multiplexing capability, and relatively long processing times.25–29 In contrast, LC-MS/MS offers superior accuracy, specificity, and multiplexing capability, making it a promising alternative for high-throughput TDM of biologics.26,30–35 Crucially, a fundamental analytical distinction exists between these platforms: while most commercial ELISAs preferentially measure free, functionally active drug, Protein G-based LC-MS/MS workflows quantify total circulating antibody, including drug sequestered within immune complexes. This difference in measured molecular species implies that therapeutic thresholds historically derived from ELISA cannot be automatically extrapolated to mass spectrometry platforms without formal cross-validation and assay harmonization.36,37 Furthermore, while multiplex LC-MS/MS excels at quantifying drug levels, it often does not incorporate ADA detection, which remains a critical component of reactive TDM for biologics like IFX. Addressing these analytical and clinical gaps is essential for the successful translation of LC-MS/MS into routine practice.

Despite these advantages, existing LC-MS/MS methods for mAb quantification still face several challenges in clinical implementation. For IFX, previously reported methods often require large amounts of immunoaffinity reagents or prolonged enzymatic digestion procedures.38,39 For UST, currently available LC-MS/MS methods generally have relatively narrow linear ranges, which may require additional dilution steps for high-concentration samples.40 For VDZ, some high-throughput approaches have insufficient sensitivity for low-concentration samples.41 Moreover, although multiplexed detection has been attempted with techniques like nano-surface and molecular-orientation limited (nSMOL) proteolysis, the associated sample preparation remains relatively complex.42 Most importantly, while several multiplex LC-MS/MS assays for therapeutic antibodies have been reported, a streamlined method capable of simultaneously quantifying the three most relevant biologics for IBD (IFX, VDZ, and UST) in a single analytical run remains lacking.

To address these limitations, we developed and validated a rapid and multiplex LC-MS/MS method for the simultaneous quantification of IFX, VDZ, and UST in human serum. Total immunoglobulin G (IgG) was enriched using a simplified Protein G-based immunocapture workflow, followed by tryptic digestion and multiple reaction monitoring (MRM)-based quantification of drug-specific signature peptides. Compared with previously reported methods, the present workflow reduces sample preparation complexity and analytical turnaround time while maintaining robust analytical performance across clinically relevant concentration ranges. The developed method was successfully applied to serum samples from IBD patients, demonstrating its applicability for streamlined, multiplex therapeutic drug monitoring in clinical practice.

Material and Methods

Reagents and Materials

Vedolizumab (Batch No.: 232459, purity ≥98%) and ustekinumab (Batch No.: 146121, purity ≥98%) were purchased from TargetMol (Boston, MA, USA). Infliximab (CAS: 170277–31-3, purity ≥95%) was obtained from Shanghai Acmec Biochemical Co., Ltd. (Shanghai, China). Rituximab (Batch No.: J03GS150610, purity ≥97%), used as the internal standard (IS), was sourced from Yuanye Biological Technology Co., Ltd. (Shanghai, China). Recombinant trypsin (Batch No.: 2220329001) was acquired from Solarbio Science &and Technology Co., Ltd. (Beijing, China). Methanol (MeOH), acetonitrile (ACN) and isopropanol (IPA) (HPLC grade) were purchased from Honeywell Burdick and Jackson™ (Morris Plains, NJ, USA). Formic acid (LC-MS grade) was obtained from Sigma-Aldrich (St. Louis, MO, USA). Phosphate-buffered saline (PBS, Batch No.: AK30795319) was procured from HyClone (Logan, UT, USA). Bovine serum albumin (BSA, Batch No.: WXBD7762V) was procured from Sigma-Aldrich (St. Louis, MO, USA). Pure water used in this study was purified by the Milli-Q™ system from Millipore (Bedford, MA, USA). Protein low-binding tubes were supplied by Eppendorf (Hamburg, Germany). Protein G Mag Sepharose Xtra magnetic beads were purchased from Cytiva (Uppsala, Sweden). Blank serum (free of analytes and IS) was collected from healthy volunteers in Peking Union Medical College Hospital and was stored in a - 30°C refrigerator until use (Ethics Approval No.: KST2019001).

Instruments

The analysis was conducted on the Citrine^TM Triple Quad^TM system (Sciex, Framingham, MA, USA) and data acquisition was performed using Analyst MD 1.6.3 software (Sciex, USA) for peak extraction and integration, generating a data matrix containing multiple reaction monitoring (MRM) transitions, retention times, and peak areas for quantification.

Lc-Ms/Ms

Samples were separated on a Jasper^TM HPLC system using an Aeris 2.6 µm peptide XB-C₁₈ 100 Å column (2.1 × 150 mm, Phenomenex, USA) at 40°C with mobile phases: A (H₂O + 0.1% formic acid) and B (acetonitrile + 0.1% formic acid) under a gradient. The flow rate was maintained at 0.3 mL/min with the following gradient elution program: The gradient distribution was 0–1.0 min, 10% B; 1.0–7.5 min, 10–35% B; 7.5–8.0 min, 35–90% B; 8–10 min, 90% B; 10–10.1 min, 90–10% B; 10.1–12.0 min, 10% B.

A 5-µL injection was performed with MeOH:ACN:IPA:H₂O (1:1:1:1, v/v/v/v) needle wash. Detection was carried out using the positive electrospray ionization-MRM mode of the AB Sciex Citrine 6500 mass spectrometer. The source and gas parameters were set as follows: ion source gas 1, 55 psi; ion source gas 2, 55 psi; curtain gas, 40 psi; collision-associated dissociation gas, 8; source temperature, 550°C; spray voltage, 4500 V. The mass spectrometry parameters of the monoclonal antibody are shown in Table 1.

Table 1.

Mass Spectrometry Parameters of the Components to Be Tested

Compound	Transitions (m/z)	Retention Time (Minutes)	Declustering Potential (V)	Collision Energy (V)	Peptide Selection
Ustekinumab	887.4/1019.5	7.72	50	42.5	GLDWIGIMSPVDSDIR
	887.4/888.4	7.72	50	44.5
Vedolizumab	1119.0/956.4	6.01	50	45.0	LEWIGEIDPSESNTNYNQK
	1119.0/429.2	6.01	50	35.0
Infliximab	536.8/587.4	6.20	50	25.3	GLEWVAEIR
	536.8/773.4	6.20	50	25.3
Rituximab	803.9/926.5	4.54	50	44.4	FSGSGSGTSYSLTISR
	803.9/839.5	4.54	50	41.4

Preparation of Calibration Standards and Quality Control Samples

VDZ (4.89 mg/mL), UST (5 mg/mL), and IFX (10 mg/mL) stock solutions were used for the preparation of calibration standards and quality control (QC) samples. Calibration standards were prepared in blank human serum at eight concentration levels: 1/0.5/0.5, 2.5/1.25/1.25, 5/2.5/2.5, 10/5/5, 25/12.5/12.5, 50/25/25, 100/50/50, and 200/100/100 μg/mL for VDZ/UST/IFX, respectively. QC samples were prepared at concentrations of 3/1.5/1.5, 15/7.5/7.5, and 160/80/80 μg/mL. Rituximab (RTX), used as the internal standard (IS), was prepared at 0.1 mg/mL in PBS. Recombinant trypsin was prepared as a 10 mg/mL stock solution and diluted to a working concentration of 0.5 mg/mL before use. All stock and working solutions were stored at −80 °C until analysis.

Sample Preparation

The sample preparation workflow is illustrated in Figure 1. Briefly, 40 μL of Protein G magnetic beads were washed twice with PBS and resuspended in PBS containing 0.05% Tween-20. Subsequently, 20 μL of sample (standard, QC, or clinical sample) and 10 μL of RTX internal standard solution (0.1 mg/mL) were added and incubated for 1 h with gentle shaking. After sequential washing with PBS containing 0.1% BSA and deionized water, the captured antibodies were eluted with 100 μL of 0.25% formic acid for 10 min.

Figure 1.

Schematic workflow for the multiplex quantification of IFX, VDZ, and UST in human serum. The procedure involves (a) capture of therapeutic antibodies using Protein G magnetic beads (1 h incubation), (b) sequential washing steps to remove matrix interferences, (c) acid elution of bound antibodies, (d) denaturation at 95 °C, and (e) enzymatic digestion with trypsin (2 h incubation). The resulting signature peptides are quantified by LC-MS/MS using a 12-min chromatographic method. Although the total batch preparation time is 3.5–4 h, the workflow is designed for parallel processing, enabling the analysis of over 100 samples daily and a turnaround time of 4–5 h per batch.

For denaturation, 90 μL of eluate was mixed with 17 μL of 1 mol/L ammonium bicarbonate and heated at 95 °C for 10 min. After cooling to room temperature, 15 μL of trypsin solution (0.5 mg/mL) was added for digestion at 37 °C for 2 h. Digestion was terminated by adding 23 μL of 50% acetonitrile containing 10% formic acid. Following centrifugation, the supernatant was transferred to glass inserts for LC-MS/MS analysis.

Method Validation

Method validation was performed in accordance with the International Council for Harmonisation (ICH) Guideline M10 (May 2022) and the US Food and Drug Administration (FDA) Bioanalytical Method Validation Guidance for Industry, including assessment of selectivity, linearity, accuracy, precision, lower limit of quantitation (LLOQ), carryover, recovery, matrix effects, dilution integrity, and stability.20,21,43,44

Selectivity, Specificity and Carry-Over

Selectivity was evaluated using blank serum obtained from six independent sources. Double blank (DB), blank with internal standard (BK), and lower limit of quantification (LLOQ) samples were prepared and analyzed for each matrix source. Interference at the retention times of VDZ, IFX, UST, and the internal standard was considered acceptable when the response in blank matrices was below 20% of the analyte response at the LLOQ and below 5% of the internal standard response.

Specificity was assessed using QC samples spiked at the upper limit of quantification (ULOQ) for each analyte (VDZ: 200 μg/mL; UST: 100 μg/mL; IFX: 100 μg/mL). No significant cross-interference was observed among analytes.

Carryover was evaluated by injecting blank serum immediately after the highest calibration standard. Responses in blank samples remained below 20% of the analyte response at the LLOQ and below 5% of the internal standard response.

Linearity

The calibration curve was constructed using eight concentration points. A weighted least squares regression (weighting factor = 1/x²) was performed with the analyte-to-internal standard peak area ratio (Y-axis) versus nominal concentration (X-axis). The back-calculated concentrations of all standards should fall within 85–115% of their nominal values, except for LLOQ which should be within 80–120%. At least 75% of the calibration standards with a minimum of 6 calibration standard levels should meet the above criteria. Two calibration curves are required per batch, to be placed at the beginning and the end of the analytical run, respectively. These two curves should be prepared independently. For each concentration level, at least 50% of the calibration standards must meet the acceptance criteria.

Accuracy, Precision and LLOQ

Method validation was performed at LLOQ and low, medium, high concentration levels with six replicates each to evaluate intra-day accuracy (percentage deviation from nominal concentration) and precision (coefficient of variation(CV)). Inter-day variability was assessed over three consecutive days with six replicates of QC and LLOQ samples per day (n=18/level). Acceptance criteria required accuracy and precision within ±15% for all levels except LLOQ, and within ±20% for LLOQ, for both intra- and inter-day evaluations.

Recovery and Matrix Effect

Recovery and matrix effect were evaluated at low, medium, and high QC levels. Recovery was determined by comparing the measured concentrations of serum samples spiked before immune-enrichment and digestion with those of samples spiked after immune-enrichment. Matrix effects were assessed using six independent serum matrices by comparing the peak areas of post-extraction spiked samples with those of corresponding solutions prepared in H₂O/PBS. The reproducibility of matrix effects across different matrices was evaluated using the coefficient of variation (CV), with values within ±20% considered acceptable.

Dilution Factor and Stability

The concentrations of dilution quality control (DQC) samples (VDZ/UST/IFX: 800/400/400 μg/mL) should exceed the ULOQ. The samples are to be diluted 10-fold with blank matrix, with a minimum of 5 replicates, and analyzed within the same batch to ensure that the measured concentrations fall within the calibration curve range and are accurately and precisely quantified. To evaluate the stability of VDZ, IFX and UST, QC samples were assessed under following conditions: short-term room temperature (RT), three freeze-thaw cycles (- 80 °C to RT), and long-term storage at - 80 °C for 30 days. The repeatability of replicate injections and the stability after sample preparation were also evaluated. Accuracy and precision were to be within ±15% of nominal value.

Clinical Sample Collection

Serum samples were obtained from residual clinical specimens of 37 IBD patients receiving VDZ, UST, or IFX treatment at the Department of Gastroenterology, Peking Union Medical College Hospital (PUMCH). Written informed consent was obtained from all participants for the use of residual clinical specimens, and this consent procedure was approved by the Scientific Research Ethics Committee of PUMCH (Approval Nos.: I-23PJ2147, I-22PJ1092). The study was conducted in accordance with the Declaration of Helsinki. Blood samples were collected using vacuum blood collection tubes and allowed to clot at room temperature for 30 min before centrifugation at 3000 × g for 10 min. Serum samples were aliquoted and stored at −80 °C until analysis.

Results

Method Development and Optimization

The rational selection of signature peptides was performed using a combined in silico and experimental strategy. Amino acid sequences of the mAbs were retrieved from UniProt, and Skyline software was used to generate theoretical tryptic peptides. Subsequent high-resolution peptide mapping identified multiple candidates, which were evaluated based on sequence uniqueness, signal intensity, and matrix interference. Four surrogate peptides with optimal performance were selected for quantification: LEWIGEIDPSESNTNYNQK for VDZ, GLDWIGIMSPVDSDIR for UST, GLEWVAEIR for IFX, and FSGSGSGTSYSLTISR for RTX (internal standard) (Figure 2). The optimized multiple reaction monitoring (MRM) transitions were m/z 1119.0→956.4, 887.4→1019.5, 536.8→587.4, and 803.9→926.5, respectively (Figure 3).

Figure 2.

Peptide scan of 4 mAbs. (a) GLEWVAEIR(IFX); (b) GLDWIGIMSPVDSDIR(UST); (c) LEWIGEIDPSESNTNYNQK(VDZ); (d) FSGSGSGTSYSLTISR(RTX).

Figure 3.

Representative MRM chromatograms of vedolizumab, ustekinumab, infliximab and internal standard.

Rituximab (RTX) was employed as a full-process internal standard (IS). Its characteristic ion transition demonstrated coefficients of variation below 5% across all validation experiments.

Regarding sample pretreatment, Protein G magnetic bead-based immunocapture yielded superior specificity compared to direct denaturation-digestion. The use of protein low-binding tubes and Tween further minimized nonspecific adsorption. Comparative evaluation of proteolytic enzymes revealed that trypsin outperformed chymotrypsin in terms of digestion efficiency and signal response (Figure 4). Consequently, trypsin was selected for the final workflow.

Figure 4.

Comparative signature peptide profiles of four mAbs generated by trypsin and chymotrypsin digestion. (a) Ustekinumab (UST), (b) Vedolizumab (VDZ), (c) Infliximab (IFX), and (d) Rituximab (RTX). Relative intensity (%) of signature peptides is shown across three conditions: Serum (blank), Serum+mab (serum with antibody), and H₂O+mab (water with antibody). Trypsin (blue bars) consistently yielded higher intensity and more specific peptides compared to chymotrypsin (red bars), particularly in the complex serum matrix, supporting the selection of trypsin for the final assay.

Method Validation

Selectivity, Specificity and Carry-Over

Representative chromatograms of blank serum, VDZ, UST, IFX, and IS are shown in Figure 5. In blank human serum, no significant interfering peaks were detected at the retention times corresponding to the signature peptides of the four mAbs. The selectivity was acceptable, and the accuracy at the LLOQ ranged from 82.8% to 113%. For specificity assessment, the results demonstrated that the interference peak areas produced by these drugs at the retention times of the target mAbs were all ≤20% of the LLOQ value, confirming no significant interference. Carry-over was not observed for UST, VDZ, IFX, and the IS (RTX).

Figure 5.

Validation chromatograms for (a) ustekinumab, (b) vedolizumab, and (c) infliximab, each shown in double blank (DB), blank + IS (BK), and LLOQ samples. (d) Representative clinical sample overlay.

Linearity

Linear calibration curves (VDZ/UST/IFX: 1/0.5/0.5–200/100/100 µg/mL; r²>0.99) were established with acceptable accuracy and precision across all analytical batches. The data demonstrated good linearity over the tested concentration ranges. The typical linear regression equations were as follows: VDZ: y = 0.00962x - 0.002 (r = 0.9970); UST: y = 0.0622x + 0.00628 (r = 0.9973); IFX: y = 0.053x + 0.00535 (r = 0.9989). (x: analyte concentration; y: peak area ratio of analyte to IS; (Figure 6).

Figure 6.

Representative calibration curves for (a) ustekinumab, (b) vedolizumab, and (c) infliximab.

Accuracy, Precision and LLOQ

We evaluated the quantification accuracy and precision of all three mAb drugs at four concentration levels. The RE% and RSD% were all less than 15% at four levels of QC samples across the tested concentration range. Both repeatability (intra-day precision) and intermediate precision (inter-day precision) demonstrated good reproducibility, with CV values up to 14.9% and 12.6%, respectively. These accuracy and precision results are comprehensively presented in Table 2.

Table 2.

Inter-Run and Intra-Run Accuracy and Precision of All Analytes

Run Batch	Analytes	LLOQ		LOQ		MOQ		HOQ
		RSD	RE	RSD	RE	RSD	RE	RSD	RE
		(%)	(%)	(%)	(%)	(%)	(%)	(%)	(%)
Intra-day 1 (n=6)	vedolizumab	11.4	−13.0	3.2	14.5	1.8	6.0	2.3	−11.8
	ustekinumab	10.9	−3.0	5.7	−1.1	2.5	0.6	3.7	−8.5
	infliximab	5.0	−0.6	1.8	3.6	3.2	4.0	2.6	−2.2
Intra-day 2 (n=6)	vedolizumab	12.3	−7.8	10.4	1.6	5.7	2.3	4.7	−12.1
	ustekinumab	14.9	−7.9	3.6	−3.8	1.1	−4.5	2.0	−8.5
	infliximab	7.2	−11.1	2.2	0.8	1.6	2.8	1.4	1.5
Intra-day 3 (n=6)	vedolizumab	12.9	−4.4	7.5	−10.9	2.3	0.4	6.0	3.2
	ustekinumab	8.1	0.3	2.5	2.1	3.5	−4.0	3.7	−8.5
	infliximab	12.7	−2.7	3.9	10.9	1.7	7.0	3.3	5.0
Inter-day (n=18)	vedolizumab	11.9	−8.9	12.6	1.7	4.1	2.9	9.1	−6.9
	ustekinumab	11.3	−3.5	4.6	−0.9	3.4	−2.7	3.0	−8.5
	infliximab	9.8	−4.8	5.0	5.1	2.8	4.6	3.9	1.4

Abbreviations: LLOQ, Lower Limit of Quantification; LOQ, Limit of Quantification; MOQ, Minimum Operating Quantity; HOQ, Highest Observable Quantity; RSD, Relative Standard Deviation; RE, Relative Error.

Recovery, Matrix Effect

The recovery and matrix effect profiles of VDZ, UST, and IFX were evaluated at three QC levels (Table 3). The mean recoveries were 65.9%, 69.7%, and 53.3%, respectively. The corresponding coefficients of variation (CVs) were 2.70%, 6.08%, and 10.9%.

Table 3.

Matrix Effects and Extraction Recovery of All Analytes

Analytes	Concentration	Matrix Effect		Extraction Recovery
	(μg/mL)	Mean (%)	RSD (%)	Mean (%)	RSD (%)
Vedolizumab	3.0	61.9	4.3	70.4	4.4
	15.0	53.7	9.5	65.2	3.6
	160	53.0	3.8	73.6	6.6
Ustekinumab	1.5	91.2	5.4	67.8	3.9
	7.5	101.3	9.0	64.3	2.6
	80	83.2	2.1	65.6	6.3
Infliximab	1.5	67.1	7.0	59.4	3.8
	7.5	70.1	4.4	47.9	3.1
	80	63.9	2.7	52.4	7.0
Rituximab	100	NA	NA	59.5	6.3

Matrix effects ranged from 53.0% to 61.9% for VDZ, 83.2% to 101.3% for UST, and 63.9% to 70.1% for IFX. All matrix effect measurements showed good reproducibility, with CVs consistently below 15%.

Dilution Factor and Stability

Ten-fold diluted QC samples showed acceptable accuracy, with recoveries of 96.6% to 99.8%. The stability of three mAb-spiked serum samples was investigated under various conditions. The results indicated that all surrogate peptides for each mAb remained stable during 48 h storage at 4 °C, with accuracy ranging from 85.7% to 112.1%. The accuracy for injection reproducibility ranged from 86.7% to 111.6%. In addition, as shown in Table 4, the three mAbs in spiked serum samples were stable under the following conditions: at room temperature for 13 h, at −80°C for 30 days, and through three freeze-thaw cycles, with the bias ranging from −12.8% to 12.7%.

Table 4.

Stability of All Analytes Under Different Conditions in Serum

Conditions	Analytes	LQC		HQC
		RE	RSD	RE	RSD
		(%)	(%)	(%)	(%)
Autosampler (4°C for 48 h)	Vedolizumab	12.1	4.0	−14.3	3.2
	Ustekinumab	2.6	7.2	−12.7	1.0
	Infliximab	3.9	2.7	0.9	1.0
Freeze-thaw (3 cycles)	Vedolizumab	−3.8	6.4	−9.0	1.7
	Ustekinumab	1.2	3.5	6.9	1.8
	Infliximab	−8.9	3.9	5.7	1.6
Short-term (room temperature for 13 h)	Vedolizumab	12.7	4.9	−10.0	2.5
	Ustekinumab	−3.3	3.2	−5.6	3.9
	Infliximab	−12.0	6.8	−12.8	3.1
Long-term(−80°C for 30 days)	Vedolizumab	6.7	6.8	−3.8	2.1
	Ustekinumab	2.4	5.8	−1.4	2.2
	Infliximab	−9.7	3.7	−9.4	2.2

Clinical Application

The multiplex LC-MS/MS method was applied to analyze serum samples from 37 IBD patients receiving VDZ (n=15), UST (n=8), or IFX (n=14). Samples were collected at trough levels from residual clinical serum at Peking Union Medical College Hospital, and all data were anonymized to protect patient privacy. Quantitative results and clinical characteristics are summarized in Table 5.

Table 5.

Demographic and Clinical Characteristics of the 37 Patients and Corresponding Drug Concentrations

Subject Number	Drug	Age, y	Sex	Diagnosis	Dose, mg	Dosing Interval	Last Dose	Sampling Date	Concentration (μg/mL)
1	VDZ	30	M	CD	300	q8w	2024-07-09	2024-07-08	3.55
2	VDZ	54	M	UC	300	q8w	2024-07-09	2024-07-08	9.96
3	VDZ	47	M	UC	300	q8w	2024-07-09	2024-07-08	11.4
4	VDZ	36	F	UC	300	q8w	2024-07-10	2024-07-09	4.73
5	VDZ	35	M	UC	300	q8w	2024-07-10	2024-07-09	6.30
6	VDZ	31	M	UC	300	q8w	2024-07-12	2024-07-10	3.07
7	VDZ	33	M	UC	300	q6w	2024-07-12	2024-07-11	10.1
8	VDZ	61	M	UC	300	q8w	2024-07-15	2024-07-12	7.15
9	VDZ	31	F	UC	300	q8w	2024-07-16	2024-07-15	5.19
10	VDZ	41	F	UC	300	q4w	2024-07-16	2024-07-15	26.5
11	VDZ	64	F	UC	300	q8w	2024-07-17	2024-07-16	21.6
12	VDZ	53	F	UC	300	q4w	2024-07-17	2024-07-16	18.2
13	VDZ	41	F	UC	300	q8w	2024-07-18	2024-07-17	6.46
14	VDZ	53	M	UC	300	q4w	2024-07-18	2024-07-18	11.2
15	VDZ	42	F	UC	300	q4w	2024-07-19	2024-07-18	11.7
16	IFX	21	M	CD	400	q8w	2024-07-12	2024-07-10	13.4
17	IFX	21	F	CD	400	q8w	2024-07-11	2024-07-11	BLQ
18	IFX	28	M	CD	400	q8w	2024-07-10	2024-07-08	2.60
19	IFX	28	F	CD	300	q6w	2024-07-11	2024-07-09	3.42
20	IFX	36	M	CD	500	q8w	2024-07-10	2024-07-09	0.930
21	IFX	22	M	CD	400	q8w	2024-07-10	2024-07-09	4.06
22	IFX	23	F	CD	400	q6w	2024-07-11	2024-07-09	14.8
23	IFX	16	F	CD	400	q8w	2024-07-15	2024-07-12	BLQ
24	IFX	34	F	CD	300	q8w	2024-07-17	2024-07-15	7.00
25	IFX	45	M	CD	600	q6w	2024-07-16	2024-07-15	5.66
26	IFX	28	F	CD	300	q6w	2024-07-17	2024-07-16	7.09
27	IFX	46	M	CD	400	q6w	2024-07-19	2024-07-18	4.98
28	IFX	18	M	CD	600	q6w	2024-07-19	2024-07-18	5.67
29	IFX	50	M	CD	300	q8w	2024-07-19	2024-07-18	2.72
30	UST	31	M	CD	90	q8w	2024-07-12	2024-07-10	1.02
31	UST	31	M	CD	90	q6w	2024-07-16	2024-07-15	3.87
32	UST	28	M	CD	90	q8w	2024-07-16	2024-07-15	1.98
33	UST	21	M	CD	90	q8w	2024-07-17	2024-07-15	0.782
34	UST	60	F	CD	90	q4w	2024-07-17	2024-07-16	4.07
35	UST	53	F	CD	90	q8w	2024-07-18	2024-07-16	1.47
36	UST	58	F	CD	90	q6w	2024-07-18	2024-07-17	0.935
37	UST	32	M	CD	260	q4w	2024-07-11	2024-07-10	3.41

Notes: :BLQ, below the lower limit of quantification (0.5 μg/mL for IFX, 0.5 μg/mL for VDZ, and 1.0 μg/mL for UST). Values reported as BLQ indicate no detectable signal and are clinically meaningful, potentially reflecting immunogenic clearance, secondary non-response, etc. Last Dose Date refers to the date of the most recent monoclonal antibody administration prior to sampling. Time from Last Dose is calculated as the interval (days) between the last administration and blood collection. Due to the retrospective nature of the study and standard clinical documentation practices, body weight was not recorded systematically at every visit and is thus not included in this table.

Abbreviations: M, male; F, female; CD, Crohn’s disease; UC, ulcerative colitis; IFX, infliximab; UST, ustekinumab; VDZ, vedolizumab.

Considerable inter-individual variability in trough concentrations was observed despite standardized dosing and sampling. The concentration ranges were 3.07–26.5 μg/mL for VDZ, 0.930–14.8 μg/mL for IFX, and 0.782–4.07 μg/mL for UST, corresponding to 8.6-, 15.9-, and 5.2-fold differences, respectively. These results demonstrate that the method covers the therapeutic concentration ranges reported in clinical practice.

Discussion

Method Development and Optimization

The accurate quantification of highly humanized mAbs remains challenging due to the high sequence homology between therapeutic antibodies and endogenous IgG.45 To address this, our signature peptide selection was not solely based on empirical screening but was underpinned by rigorous in silicovalidation. Candidate peptides were evaluated by BLAST searches against the UniProt human proteome database to ensure sequence uniqueness and avoid homology with common endogenous immunoglobulin variants or co-administered therapeutic antibodies. This in silicostrategy, combined with experimental verification in six blank matrices, provided robust evidence for the specificity of our targets: LEWIGEIDPSESNTNYNQK (VDZ), GLDWIGIMSPVDSDIR (UST), and GLEWVAEIR (IFX). Notably, the IFX peptide (GLEWVAEIR) is highly conserved across all currently approved infliximab biosimilars. Consequently, our assay is designed to quantify total infliximab-related drug products (reference biologic plus biosimilars), a deliberate and clinically relevant feature for real-world TDM where multiple biosimilars are in use. This systematic approach ensured the selection of peptides with optimal specificity, effectively distinguishing the target drugs from the complex serum background.

A limitation of the current workflow is the use of rituximab (RTX) as a universal full-process internal standard. We acknowledge that RTX cannot perfectly replicate the analyte-specific Protein G binding affinity or proteolytic digestion kinetics of IFX, VDZ, and UST. However, this compromise was deliberate and pragmatic. While stable isotope-labeled (SIL) peptides are considered the gold standard for internal standards due to their identical physicochemical behavior, their prohibitive cost and lengthy synthesis times pose significant barriers to high-throughput clinical applications.45,46 In contrast, RTX served as a surrogate protein IS capable of compensating for variability introduced during the entire pre-analytical process, including antibody capture, elution, denaturation, and enzymatic digestion—a feat that SIL peptides cannot achieve as they are added only post-digestion. Crucially, the observed CV < 5% for the RTX signal across validation runs confirms its effectiveness in correcting for analytical variability and instrument drift, supporting the robustness of our quantification within the evaluated concentration ranges. Future method refinements may consider stable isotope-labeled full-length antibodies to achieve superior analyte-specific correction while retaining the benefits of full-process monitoring.

Sample preparation is often the bottleneck for mAb quantification in biological matrices. While direct digestion is simpler, it suffers from high background noise. The use of Protein G magnetic beads provides a robust solution for IgG enrichment, significantly reducing matrix complexity and improving the signal-to-noise ratio.38 Combined with the use of low-binding consumables, this workflow enhances both sensitivity and reproducibility. Furthermore, the choice of trypsin over chymotrypsin was justified by its more efficient and specific cleavage pattern, resulting in more robust peptide markers for quantification.

Recovery, Matrix Effect

Although the absolute recovery of IFX was relatively moderate (~53%), its high reproducibility (CV < 11%) demonstrates that the sample processing workflow is robust and stable.

It is noteworthy that matrix effects varied among the three mAbs; however, the consistent CVs (< 15%) across different batches and concentrations suggest that these effects, while present, are predictable and manageable within our laboratory setting. Crucially, the use of a full-process protein internal standard effectively compensated for variability arising from immunocapture, enzymatic digestion, and LC-MS/MS analysis. Nevertheless, we acknowledge that moderate recovery and substantial ion suppression may pose challenges to inter-laboratory robustness, particularly across different reagent lots, disease states, or immunoglobulin concentrations. Therefore, while the method demonstrates reliable quantitative accuracy within this validation framework, future multi-center studies are warranted to fully establish its transferability and robustness in diverse clinical environments.

Clinical Application and Inter-Individual Variability

The observed inter-individual variability is consistent with previously reported therapeutic ranges: VDZ (~27.8 μg/mL),47 IFX (5–25 μg/mL),18,48 and UST (0.9–14.7 μg/mL).40 Rather than validating therapeutic thresholds, this alignment merely demonstrates the technical commutability and plausibility of our method compared to existing mass spectrometry platforms. While these findings confirm the assay’s capability to accurately quantify drugs within clinically relevant concentration windows, we acknowledge that establishing definitive therapeutic targets requires prospective correlation with clinical, biochemical, and endoscopic outcomes, which was beyond the scope of this proof-of-concept study.

Nevertheless, the current sample size (n = 37) precludes robust pharmacokinetic modeling and correlations with clinical endpoints. Consequently, this study represents a proof-of-concept rather than a definitive validation of therapeutic targets or remission associations. Larger multicenter studies are warranted to further elucidate the clinical utility of this assay.

Comparison with Previous Multiplex mAb Assays

Although several LC-MS/MS-based multiplex strategies for therapeutic mAb quantification have been reported, important differences remain in assay design, sample preparation complexity, and clinical applicability (Table 6).

Table 6.

Comparison of Representative LC-MS/MS Based Multiplex Methods for Therapeutic mAb Quantification

Method	No. of mAbs	Sample Pretreatment	Key Advantage	Main Limitation
Willeman et al39	7	Peptide-level workflow	Applied to adalimumab TDM in Crohn’s disease	Extensive peptide optimization required
Hallin et al41	11	Caprylic acid precipitation	High throughput, low cost	Limited sensitivity for low-concentration mAbs
Iwamoto et al42	7	nSMOL proteolysis	Reduces endogenous IgG interference	Technically demanding, limited scalability
Amrani et al48	7	Generic sample preparation	Broad mAb coverage	Labeled IS and peptide optimization required; LLOQ ~1 mg/L;Long turnaround time.
Present method	3	Universal Protein G capture	No mAb-specific reagent; single-run IBD TDM	Limited to IgG-type therapeutics

Willeman et al39 first implemented a multiplex LC-MS/MS-based workflow for adalimumab TDM in patients with Crohn’s disease, demonstrating the clinical feasibility of quantifying multiple therapeutic mAbs by LC-MS/MS. Other high throughput approaches, such as caprylic acid precipitation for up to 11 therapeutic antibodies,41 provide cost effective sample preparation but may lack the sensitivity needed for low concentration mAbs such as vedolizumab. nSMOL based strategies allow Fab selective digestion to reduce interference from endogenous IgG;42 however, the associated sample preparation remains relatively complex. Amrani et al48 subsequently reported a generic sample preparation method for the simultaneous quantification of seven therapeutic mAbs in human serum or plasma; although effective, this approach requires extensive selection of signature peptides, synthesis of stable isotope labeled internal standards, and peptide specific optimization, which may limit scalability and routine implementation in clinical TDM.

In contrast, the present method employs a universal Protein G based immunocapture step that is applicable to all IgG type therapeutic antibodies. This eliminates the need for mAb-specific capture reagents; however, the identification of a unique signature peptide and validation of its analytical performance remain requisite for each new therapeutic antibody. By enabling the simultaneous quantification of IFX, VDZ, and UST within a single LC-MS/MS run, this workflow offers a simple and reproducible approach to multiplex therapeutic drug monitoring in patients with IBD.

Comparison with ELISA and Clinical Implications

While ELISA remains the reference standard for biologic TDM, our LC-MS/MS assay offers distinct advantages. The present assay demonstrated excellent linearity (R² > 0.99) and precision (CV <15%), which are comparable to the analytical performance reported for commercially available ELISA assays for IFX, VDZ, and UST.36,49 Mass spectrometry is also less prone to ADA interference, a limitation of immunoassays that may underestimate measured drug concentrations.36,50 Previous studies have demonstrated good agreement between LC-MS/MS and ELISA for therapeutic monoclonal antibodies, suggesting that the two analytical platforms can provide comparable quantitative results under appropriate validation conditions.36,49,51 However, such cross-platform validation has not yet been performed in the present study.

The key methodological strength is the proposed assay is its multiplexing capability, enabling the simultaneous quantification of three widely used biologics for IBD within a single analytical run.36,52 In contrast, ELISA typically requires separate assays for each therapeutic antibody. Furthermore, our simplified Protein G immunocapture workflow eliminates the need for drug-specific capture reagents, thereby reducing workflow complexity while maintaining a universal sample preparation strategy.38

Operationally, the workflow achieves a favorable balance between preparation time and throughput. Although batch sample preparation requires approximately 3.5–4 h due to the immunocapture and digestion steps, these procedures can be performed in parallel for multiple samples. Compared with conventional ELISA workflows, which typically require separate assays for each therapeutic antibody, the multiplex LC-MS/MS approach simplifies laboratory workflow and may reduce the overall time required for multi-analyte analysis.36,49

From an economic perspective, the assay has the potential to offer a favorable cost-benefit profile. The estimated reagent cost of the proposed assay (excluding labor, instrument depreciation, maintenance, and other operational expenses) is substantially lower than the cumulative reagent cost associated with performing three separate ELISA assays.36,38 While LC-MS/MS requires a higher initial investment in instrumentation, its multiplex capability and cost-per-sample reagent costs may provide operational advantages in laboratories processing multiple therapeutic antibodies. Further head-to-head comparative studies with ELISA will be necessary to further evaluate both the clinical and economic utility of the proposed workflow.

Nevertheless, it is important to acknowledge that this study represents a preliminary exploratory analysis based on only 37 residual patient samples. A formal incurred sample reanalysis (ISR) study was not performed because the limited availability of residual clinical specimens precluded repeat analysis. Consequently, the definitive clinical applicability of the assay cannot be established without further cross-platform validation, and therapeutic thresholds established using ELISA or other ligand-binding assays should be interpreted with caution when applied to this LC-MS/MS platform. Prospective studies incorporating paired ELISA measurements, ISR, and external laboratory validation will be required before routine clinical implementation.

Conclusion

This study developed and validated a rapid, cost-effective, and multiplex LC-MS/MS method for simultaneous quantification of three key monoclonal antibodies—infliximab, ustekinumab, and vedolizumab—in human serum of Chinese IBD patients. By integrating Protein G–based immunocapture with tryptic digestion and MRM analysis, the method achieved robust analytical performance across clinically relevant concentration ranges with acceptable accuracy, precision, and reproducibility. The assay was successfully applied to serum samples from IBD patients, confirming its analytical feasibility for measuring these biologics. However, definitive clinical applicability and interchangeability with existing assays require further confirmation. Overall, this work provides a promising and technically robust tool for pharmacokinetic research and exploratory TDM studies, pending future cross-platform comparison, incurred sample reanalysis, external laboratory validation, and integration with anti-drug antibody testing.

Funding Statement

This work was supported by the National Natural Science Foundation of China (Grant No. 82574508) and the National High Level Hospital Clinical Research Funding (Grant No. 2025-PUMCH-A-097).

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

Ruichen Liu and Xianglong Zhao are employed by SCIEX. Their contributions to this study were strictly limited to technical support for instrument operation, consultation during assay optimization, assistance with data processing, and manuscript review. SCIEX did not provide reagents or financial support for this work. All clinical data collection, data interpretation, and the decision to submit this manuscript were conducted independently by the academic authors. The remaining authors declare no competing interests.