Hello

2.1. Chemical substances and laboratory materials

Reference baricitinib (lot 22Z102-M1, purity: 98.68%) and internal standard baricitinib-d5 (lot 22T083-L1, purity: 98%) (Figure 1) were procured from Shanghai Zzbio Co., Ltd. (Shanghai, China). HPLC-grade methanol, acetonitrile, dimethyl sulfoxide (DMSO), ammonium acetate, and ACS-grade formic acid were obtained from Aladdin Company (Shanghai, China). The ultrapure water used in this study was produced using an Elga Purelab Flex 3 water system (ELGA, High Wycombe, UK). Blank plasma and whole blood from healthy volunteers were donated by the Phase I Clinical Laboratory of Tongji Hospital (ethical approval number: TJ-IRB20220737). Blank aqueous humor was provided by patients who underwent cataract surgery in our hospital. Written informed consent was obtained from all the patients.

AABB

2.2. Preparation of Calibration Standards and Quality Controls

Accurately weighed amounts of baricitinib and baricitinib-d5 (IS) were dissolved in methanol:DMSO (40:60, v/v) to yield standard stock solutions of 1.0 mg/mL and 40 μg/mL, respectively. The baricitinib stock solution was diluted with acetonitrile to obtain a series of standard (4–4,000 ng/mL) and quality control (QC) working solutions (4, 12, 150 ng/mL, 3,000 ng/mL). Similarly, a 40 ng/mL IS solution was prepared by diluting the stock IS solution.

Calibration standards for baricitinib in human plasma or aqueous humor were prepared at concentrations of 0.2, 0.5, 2, 5, 10, 40, 100, and 200 ng/mL. This was achieved by mixing 20 μL of standard working solutions with 180 μL of blank plasma or aqueous humor. QC samples were independently prepared using the same method. These QC samples included 0.20 ng/mL (lower limit of quantification, LLOQ), 0.6 ng/mL (low-quality control, LQC), 7.5 ng/mL (medium-quality control, MQC), and 150 ng/mL (high-quality control, HQC). Furthermore, a dilution quality control (DQC) sample was prepared at a concentration of 400 ng/mL. Subsequently, the DQC sample was diluted with blank plasma or aqueous humor (dilution factor: 5) to produce an 80 ng/mL dilution test sample.

2.3. Sample Preparation

An aliquot of 20 μL IS working solution (40 ng/mL) was added to 50 μL of plasma or aqueous humor sample. The mixture was vortex-mixed for 10 s. Subsequently, an aliquot of 100 μL acetonitrile was added to facilitate protein precipitation. The mixture was then thoroughly extracted on a vortex mixer for 5 min and centrifuged at 12,000 rpm for 10 min at 4 °C. Next, 100 μL of the protein-free supernatant was carefully transferred to a new 1.5 mL plastic vial, followed by 100 μL of an acetonitrile–water solution (50:50, v/v). Finally, a 6 μL sample of the resulting mixture was injected for detection and analysis.

2.4. Instrumental Settings

LC-MS/MS analysis was performed using a Shimadzu LC-20A liquid chromatographic system (Shimadzu Corp., Kyoto, Japan) coupled to a QTRAP 5500 mass spectrometer (AB Sciex, MA, USA). The mass spectrometer was equipped with an electrospray ionization (ESI) source. Positive multiple reaction monitoring (MRM) mode was used to analyze the samples. The quantitative and confirming MRM mass transitions (m/z) were 372.3 > 186.1 and 372.2 > 251.1 for baricitinib, and 377.2 > 186.1 and 377.2 > 251.1 for IS, respectively. A simple gradient elution method was employed using an Ultimate XB-C₁₈ column (2.1 mm × 50.0 mm, 5 μm) (Welch, Shanghai, China). The mobile phase consisted of 10 mM ammonium formate in water (phase A) and 0.2% formic acid in a mixture of methanol, acetonitrile, and isopropanol (75:15:15, v/v/v; phase B). The column oven and autosampler were set at 40 °C and 15 °C, respectively.

The total analysis time for each sample was 2.5 min, with a constant flow rate of 0.72 mL/min. The gradient elution program was as follows: 0 – 0.2 min: 20% B; 0.2 – 1.3 min: 20 – 88% B; 1.3 – 1.8 min: 88% B; 1.8 – 2.0 min: 88 – 20% B; 2.0 – 2.5 min: 20% B. This gradient elution program allowed for the effective separation and analysis of the compounds of interest.

The specific source and gas parameters used were as follows: collision gas set to medium; curtain gas maintained at 40 psi; ion source gas 1 set to 50 psi; ion source gas 2 adjusted to 55 psi; ion spray voltage set at 5500 V; and ion spray temperature held at 600 °C. The transitions and optimized settings are listed in Table 1. Data acquisition and analysis were performed using the Analyst software package (Applied Biosystems, Version 1.6.3).

MRM, multiple reaction monitoring

2.5. Method Validation

All method validation procedures adhered to the bioanalytical method validation guidance provided by the United States Food and Drug Administration and the European Medicines Agency (ref26 ; ref27 ).

2.6. Applicability of the Method for TDM

The utility of the validated method was assessed by analyzing plasma or aqueous humor samples from 10 patients with NIU. Inclusion criteria were as follows: (1) age between 18 and 65 years; (2) meeting the diagnostic criteria for NIU, including intermediate uveitis, posterior uveitis, or panuveitis; (3) active uveitis at day 1/baseline examination with at least one of the following: (1) active inflammatory chorioretinal lesions and/or inflammatory retinal vasculopathy; (2) ≥2+ anterior chamber cells (according to the SUN criteria); (3) ≥2+ inflammatory vitreous opacity (according to the SUN criteria). Patients with the following conditions were excluded: (1) infectious uveitis or uncontrolled tuberculosis or hepatitis B; (2) severe glaucoma or cataract unrelated to uveitis; (3) elderly or obese individuals or those with a history of deep vein thrombosis or pulmonary thrombosis, surgery or other bedridden conditions, severe cardiovascular and cerebrovascular diseases, or uncontrolled systemic diseases such as hypertension and diabetes. All participants provided informed consent before inclusion in the study.

All patients were orally administered 4 mg baricitinib (OLUMIANT^®) once daily at 10:00 AM. Whole blood was drawn into EDTA-K2 vacuum tubes immediately pre-dose (0 h), and 0.5, 1, and 1.5 h post-dose after administration of the second dose. Aqueous humor samples were drawn into sterile plastic tubes immediately before administration of the second dose (0 h). Whole blood was centrifuged at 1800 × g for 10 min, and the plasma and aqueous humor were stored at -70 °C until analysis. The trapezoid rule was used to calculate the area under the blood concentration–time curve over 1.5 h (AUC₀–_1.5), and then the median, first quartile, and third quartile of the AUC₀–_1.5 adjusted by the daily dose were calculated. Pearson correlation coefficients were used to assess the association between the concentration in aqueous humor and the concentration or AUC₀–_1.5 of plasma using R software (version 4.2.1; http://www.r-project.org; accessed on January 30, 2023). Statistical analysis and figure generation were performed using the tidyverse and ggplot2 packages.

3.1. Optimization of Chromatography and Mass Spectrometry

A 40 ng/mL solution of baricitinib and the IS was used to determine the mass spectrometric parameters. Consistent with previous reports (Kaneko et al. 2021), a more stable and higher signal-to-noise intensity ratio was observed in the ESI positive-ion mode. This generated the most abundant prominent protonated molecular ions [M+H]⁺ at 372 and 377 for baricitinib and its isotopic IS, respectively. Both analytes produced predominant fragment ions at m/z 186.1 and 251.1. Finally, we adopted 372.2 > 186.1 and 377.2 > 186.1 as the quantified mass transitions for baricitinib and IS, respectively, while the other two pairs of mass transitions were used for compound confirmation.

Because a single-step protein precipitation was sufficient to cover the sensitivity and concentration range, acetonitrile was chosen as the extracting solution due to its superior ability to yield cleaner and more efficient results than methanol. Following this step, the supernatants were reconstituted with an equivalent volume of 50% methanol containing 0.1% formic acid to further minimize the matrix effect and maximize ionization intensity.

To achieve optimal chromatographic separation, the composition of the mobile phase was systematically optimized. The aqueous phase consisted of various concentrations of ammonium formate or ammonium acetate solution, while the organic phase was evaluated in different stages. Initially, a binary mixture of methanol and acetonitrile was employed to balance the insufficient ionization efficiency associated with acetonitrile and the relatively high column pressure induced by methanol. Subsequently, the addition of a small proportion of isopropanol to this binary mixture improved peak shape by reducing tailing, yielding symmetric and sharp peaks. The final mobile phase comprised 10 mM ammonium formate and 0.2% formic acid in water as the aqueous component, and a ternary mixture of methanol, acetonitrile, and isopropanol (75:15:15, v/v/v) as the organic component.

Multiple commercially available C₁₈ columns were evaluated to enhance chromatographic peak shape, peak width, and separation. The Ultimate XB-C₁₈ column (2.1 mm × 50.0 mm, 5 μm) was selected due to its favorable peak shape, narrow peak width, and cost-effectiveness. Under the aforementioned chromatographic conditions, baricitinib and the IS were effectively separated with a retention time of 1.27 min within a total running time of 2.5 min, resulting in a shorter analysis time than prior methods (Veeraraghavan et al. 2016; Ezzeldin et al. 2020; Koller et al. 2020; Zhao et al. 2020; Wickremsinhe et al. 2023).

3.2. Method Validation

3.3. Clinical Application

The plasma concentration–time profiles are shown in Figure 4, and the demographic and laboratory characteristics are summarized in Table 5. The plasma concentrations of baricitinib at 0 h (before the second dose) and at 0.5, 1, and 1.5 h after the second dose were 2.25 (1.79, 3.31), 30.18 (11.27, 44.02), 44.64 (30.46, 60.02), and 43.53 (36.87, 54.34) ng/mL, with %CV values of 49.6%, 81.0%, 48.6%, and 30.4%, respectively. The calculated AUC₀–_1.5 and daily dose-normalized AUC₀–_1.5 (AUC0–1.5/D) were 53.30 (40.62, 87.0) and 13.30 (10.15, 18.57) ng·h/mL/mg, respectively, with both %CV values being 51.4%. Furthermore, half of the patients did not exhibit a delayed absorption peak within 1.5 h, indicating that not all patients achieved the peak concentration within 1.0 h. This finding is inconsistent with results from earlier phase I and II clinical trials (Ito et al. 2024). One plausible explanation for this discrepancy is that our study, conducted in a real-world clinical setting, did not strictly require patients to take the medication on an empty stomach, and some may have taken it after a meal, despite the fact that oral bioavailability is not affected by food.

^a Calculated according to the Chronic Kidney Disease Epidemiology Collaboration equation and converted to L/h;

^b Represented as mean ± SD.

The trough concentration (C_min) of baricitinib after the first dose in the anterior chamber was 0.94 (0.89, 1.16) ng/mL, with a %CV of 20.2%, which was significantly lower than the plasma C_min (49.6%). Furthermore, significant correlations between the concentrations in the aqueous humor at 0 h and the plasma concentrations at 0, 0.5, 1, and 1.5 h, as well as the corresponding plasma AUC₀–_1.5, were not observed (Figure 5). These findings indicate that drug concentrations in the anterior chamber are less susceptible to fluctuations in plasma concentrations and that it may not be feasible to predict the therapeutic efficacy of ocular diseases by monitoring drug concentrations in plasma.

3.4. Comparison with Previous Studies

Several methods have been established to quantify baricitinib in the plasma of rats (Veeraraghavan et al. 2016; Ezzeldin et al. 2020) and humans (Koller et al. 2020; Zhao et al. 2020; Wickremsinhe et al. 2023). To date, these quantification methods have been validated by LC-MS/MS in MRM-positive mode, with three reports (Zhao et al. 2020; Wickremsinhe et al. 2023; Cafaro et al. 2024) using isotopic IS and analysis times exceeding 2.8 min. In addition, the application of microsampling technology in one pharmacokinetic study (Wickremsinhe et al. 2023) conducted in pediatrics, although innovative, presented high technical barriers with no superior cost-effectiveness. The latest published method by Cafaro et al. (Cafaro et al. 2024), which is similar to ours, requires only a small sample volume, a straightforward sample preparation protocol, and a low detection limit. However, their method did not cover both peak and trough points to verify its applicability, nor did it explore drug concentrations in aqueous humor. Furthermore, all other studies based on human plasma used liquid–liquid extraction for sample pretreatment, which is complicated and time-consuming. In summary, there are currently no reports on the determination of baricitinib in aqueous humor. This study provides improvements by using LC-MS/MS; for instance, the sensitivity was adequate without the need for complicated and time-consuming preprocessing, the analytical time was shorter, and the matrix effect was eliminated.

An important factor contributing to the success of ocular therapy is the ability of drug molecules to penetrate targeted ocular tissues at effective concentrations. Moreover, aqueous humor is an accessible site for the systemic administration of baricitinib. Given the presence of the blood-ocular barrier, baricitinib in the aqueous humor is closer to the ciliary body or choroid than that in plasma, and monitoring its PK in the aqueous humor is relatively more suitable than plasma for evaluating the exposure-efficacy relationship in patients with NIU. Nevertheless, it must be clearly recognized that aqueous humor remains a surrogate matrix. Researchers have advocated the development of baricitinib ophthalmic liposomes, as direct ocular administration may achieve better penetration into target tissues than oral administration, potentially enhancing therapeutic efficacy (Garros et al. 2022).

This study also presents some limitations. Firstly, plasma concentrations at the terminal elimination phase were not monitored during the second dosage interval. Therefore, AUC_0-1.5 could not fully substitute for the AUC₂₄ of baricitinib in plasma. Secondly, the small sample size in our study may limit the generalizability of the pharmacokinetic variability observed. In addition, variability in drug administration (fasted vs fed states) may have contributed to the observed PK variability, but was not controlled in this real-world study. Lastly, the use of Pearson correlation provides only a preliminary assessment of the relationship between plasma and aqueous humor concentrations. More robust analyses (e.g., population pharmacokinetics) would better characterize the compartmental dynamics and exposure–effect relationship in NIU cohorts but would require larger sample sizes. These limitations highlight opportunities for future research to optimize baricitinib TDM in ocular diseases.