Pharmacokinetic profile of N-acetylcysteine amide and its main metabolite in mice using new analytical method

Journal Pre-proof

Pharmacokinetic profile of N-acetylcysteine amide and its main metabolite in mice using new analytical method

Rui He conceived the study , Wenyi Zheng performed experiments and analyzed data , Tobias Ginman purified NACA-MPOZNAC-MPOZ , Hakan Ottosson performed 1H-NMR and made data interpretation , Svante Norgren , Ying Zhao performed the animal experiment , Moustapha Hassan supervised the study and acquired fundingAll authors participated in the preparation of the man PII: DOI: Reference:

S0928-0987(19)30431-2 https://doi.org/10.1016/j.ejps.2019.105158 PHASCI 105158

To appear in:

European Journal of Pharmaceutical Sciences

Received date: Revised date: Accepted date:

6 August 2019 3 October 2019 15 November 2019

Please cite this article as: Rui He conceived the study , Wenyi Zheng performed experiments and analyzed data , Tobias Ginman purified NACA-MPOZNAC-MPOZ , Hakan Ottosson performed 1H-NMR and made data interpreta Svante Norgren , Ying Zhao performed the animal experiment , Moustapha Hassan supervised the study and acq Pharmacokinetic profile of N-acetylcysteine amide and its main metabolite in mice using new analytical method, European Journal of Pharmaceutical Sciences (2019), doi: https://doi.org/10.1016/j.ejps.2019.105158

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier B.V.

Pharmacokinetic profile of N-acetylcysteine amide and its main metabolite in mice using new analytical method

Rui He1, Wenyi Zheng1, Tobias Ginman2, Håkan Ottosson3, Svante Norgren4, Ying Zhao1,5, Moustapha Hassan1,5*

1. Experimental Cancer Medicine, Clinical Research Center, Department of Laboratory Medicine, Karolinska Institutet, Huddinge, 141 86 Stockholm, Sweden 2. Sprint Bioscience, Huddinge, 141 86 Stockholm, Sweden 3. Department of Biosciences and Nutrition, Karolinska Institutet, Huddinge, 14186 Stockholm, Sweden 4. Children´s and Women´s Health Theme, Karolinska University Hospital and Karolinska Institutet, Stockholm, Sweden 5. Clinical Research Center, Karolinska University Hospital, Huddinge, 141 86 Stockholm, Sweden *: Corresponding author: [email protected] Postal address: Experimental Cancer Medicine (ECM); Novum, Plan 6, Hiss F, Lab 601; Hälsovägen 7-9, 14157 Huddinge, Sweden

1

Abstract N-acetylcysteine amide (NACA) is the amide derivative of N-acetylcysteine (NAC) that is rapidly converted to NAC after systemic administration. It has emerged as a promising thiol antioxidant for multiple indications; however, the pharmacokinetic property is yet unclear due to lack of an accurate quantification method. The present investigation aimed to develop an analytical method for simultaneous quantification of NACA and NAC in plasma. A new reagent (2-(methylsulfonyl)-5-phenyl-1,3,4-oxadiazole, MPOZ) was introduced for thiol stabilization during sample processing and storage. Further, we utilized tris(2-carboxyethyl) phosphine (TCEP) to reduce the oxidized forms of NACA and NAC. After derivatization, NACA-MPOZ and NAC-MPOZ were quantified using liquid chromatography–mass spectrometry (LC-MS). The new method was validated and found to have high specificity, linearity, accuracy, precision, and recovery for the quantification of NACA and NAC in plasma. Furthermore, the formed derivatives of NACA and NAC were stable for 48 h under different conditions. The method was utilized in pharmacokinetic study which showed that the bioavailability of NACA is significantly higher than NAC (67% and 15%, respectively). The pharmacokinetic of NACA obeyed a two-compartment open model. The glutathione (GSH)-replenishing capacity was found to be three to four-fold higher after the administration of NACA compared to that observed after the administration of NAC. In conclusion, the present method is simple, robust and reproducible, and can be utilized in both experimental and clinical studies. NACA might be considered as a prodrug for NAC. Furthermore, this is the first report describing the pharmacokinetics and bioavailability of NACA in mouse.

Keywords N-acetylcysteine amide; N-acetylcysteine; 2-(methylsulfonyl)-5-phenyl-1,3,4-oxadiazole; liquid chromatography–mass spectrometry; Pharmacokinetics; GSH; Bioavailability 1 Introduction Elevated levels of reactive oxygen species (ROS) can lead to oxidative damage in cells, which has been implicated in development and progression of a broad range of diseases e.g. neurodegeneration, cardiovascular disorder, cancer and inflammations (Ebrahimi et al., 2018; Ilkan and Akar, 2018; Sugumar et al., 2018; Tabriziani et al., 2018).To counteract oxidative damage, strategies have been adopted to either prevent ROS generation or neutralize excess ROS (Poljsak, 2011). These strategies target reinforcing enzymes (like superoxide dismutase, catalase, glutathione peroxidase, and glutathione reductase) in the endogenous antioxidative defense system and reducing equivalents (like glutathione (GSH), nicotinamide adenine dinucleotide phosphate, and nicotinamide adenine dinucleotide) (Matés et al., 1999; Mejia et al., 2018). In addition, supplementation of exogenous antioxidants, including vitamin C, vitamin E, ebselen, and isoflavone, are commonly used for prevention and treatment of oxidative damage-induced events (Jaturakan et al., 2017; Umeno et al., 2016; Yu et al., 2017).

2

N-acetylcysteine (NAC) is well known antidote for paracetamol overdose and as a mucolytic therapy (Pettie et al., 2019; Sadowska et al., 2006). NAC is also a broadly used thiol antioxidant in both experimental and clinical settings. Its mechanism involves replenishing GSH, the major antioxidant, and direct ROS neutralization. Due to the low bioavailability of NAC, it has been reported that a dose of 600 mg per day was found to be inadequate to normalize GSH content under some conditions with critical oxidative stress in human beings (Borgström et al., 1986; Cazzola et al., 2015). Moreover, probably due to the low bioavailability, the in vivo effect of NAC was found to be inconclusive and in several cases contradictory (Aparicio-Trejo et al., 2018; Karlsson et al., 2011). To improve the antioxidative function of NAC, its amide derivative, NACA, was synthesized in 1967 (structure depicted in Figure 1) (Martin et al., 1967). NACA features higher lipophilicity and better membrane permeability compared to NAC which is due to the amide group. Several in vitro and in vivo investigations on NACA have revealed better anti-oxidative effect as well as tolerability compared to that observed when NAC was used (Goyal et al., 2016; Kawoos et al., 2017; Maddirala et al., 2017).

Figure 1. Chemical structures of NAC and NACA.

To unleash the full therapeutic potentials of NACA, it is crucial to understand both the pharmacokinetic properties and the metabolic pathway. One related study was published a decade ago (Wu et al., 2006), in which the authors reported that NACA could be converted to NAC after oral administration. The method used was, however, not selective enough due to the overlapping peaks of NACA and NAC in the chromatogram. Moreover, it has been reported that NAC is readily oxidized as a result of disulfide formation, an event that is plausibly applicable to NACA, which also has one thiol residue. In order to achieve full recovery and accurate quantification of NACA, it is imperative to stabilize the thiol residue during sample storage and processing. In two separate studies, N-(1-pyrenyl) maleimide (NPM) (Wu et al., 2006) and recently, 2-chloro-1-methylpyridinium iodide were investigated (King et al., 2019) as stabilizing/derivatizing agents. In both investigations, stabilization maximum and reaction kinetics were unclear. The quantification of NACA was performed on rat and patient samples obtained after the administration of NACA. However, the pharmacokinetic property of NACA remained uninvestigated. In order to overcome the limitation of existing quantification methods, our strategy was to find a new thiol-reactive reagent, which could achieve rapid, stable and selective thiol labeling. The derivatives should be separated by conventional reverse phase chromatography 3

using electrospray ionization-mass spectrometry (ESI-MS) for detection. NACA is known to be oxidized during sample preparation process as well as in vivo, forming dimer and conjugates with endogenous thiols like cysteine. To account for the oxidized NACA in in vivo, tris(2-carboxyethyl) phosphine (TCEP) could be applied for reductive cleavage of disulfide. With the newly developed method, our main aim was to investigate the stability of NAC and NACA in biological samples as well as in aqueous solutions and to study the stability of the formed derivatives for preclinical and clinical use. Moreover, we aimed to investigate the pharmacokinetic differences between NAC and NACA, to compare the bioavailability of both compounds, and to detect GSH replenishment after the administration of NACA or NAC. 2 Materials and Methods 2.1 Materials NACA was provided by Dr Glenn Goldstein, Sentient Life Sciences Inc, NY, New York, USA. NAC, GSH, TCEP, N-Ethylmaleimide (NEM) and ammonium bicarbonate (NH4HCO3) were purchased from Sigma-Aldrich (St. Louis, Missouri, USA). 2-(methylsulfonyl)-5phenyl-1,3,4-oxadiazole (MPOZ) was purchased from Aurum Pharmatech (New Jersey, USA). Acetonitrile (ACN) and formic acid were of LC-MS grade and obtained from Merck (Darmstadt, Germany), while dimethyl acetamide (DMA) was from Fluka (Seelze, Germany). H2O (resistance ≥ 18.2 MΩ) was prepared by a special purification system from ELGA (High Wycombe, United Kingdom). Pooled blank human plasma from healthy individuals was provided by the blood and transfusion center at Karolinska University Hospital, Huddinge, while blank murine plasma was collected from control BALB/c mice (Charles River). 2.2 Synthesis of NACA-MPOZ and NAC-MPOZ NACA-MPOZ and NAC-MPOZ were synthesized as following: 20 mg NACA or NAC was dissolved in 1 ml NH4HCO3 (10 mM in H2O), to which 58 mg MPOZ (in 2 ml ACN) was added. After overnight reaction, the raw product was purified through preparative liquid chromatography (Gilson HPLC system with a UV detector) using Symmetry C18 columns. The mobile phase consisted of ACN/H2O, and the flow rate was 35 ml/min. The derivative was then characterized by 1H-NMR (nuclear magnetic resonance, Bruker DRX-400, 400 MHz) and ESI-MS (Thermo TSQ Quantum Ultra). 2.3 Mixed Disulfide formation (Oxidation) of NACA and NAC in plasma Freshly prepared NACA and NAC solutions (5 µl, 100 µg/ml) were mixed with 10 µl blank human plasma. The mixture was either subjected to direct extraction or left overnight at room temperature prior to extraction. For extraction of NACA and NAC, 20 µl MPOZ (0.045 M in DMA) was added to the spiked samples, followed by 5 µl H2O or TCEP (0.2 M in H2O). After reaction for 30 min, deproteinization was achieved by addition of 100 µl ACN and centrifugation (30,000 g, 10 min). The supernatant was collected for LC-MS measurement. 2.4 Sample preparation using TCEP/MPOZ derivatization method

4

NACA and NAC were freshly dissolved in NH4HCO3 (10 mM in H2O) at the same concentration, and mixed at the ratio of 1/1 (v/v). Ten µl of plasma was spiked with 5 µl mixture of NACA and NAC, followed by addition of 40 µl MPOZ (0.045 M in DMA) and 5 µl TCEP (0.2 M in H2O) successively. The mixture was vortexed and left at room temperature for 30 min before deproteinization with 100 µl ACN. The supernatant after centrifugation (30,000g, 10 min) was aspirated for LC-MS measurement. 2.5 LC-MS parameters NACA-MPOZ and NAC-MPOZ were dissolved together in ACN, and continuously infused into ESI-MS together with mobile phase using a T-piece. MS parameters tuning was performed with the Quantum Tune software. For measurement, 2 µl supernatant was injected via an Agilent 1100 HPLC system and passed through a C18 column. TSQ Quantum Mass Spectrometer (Thermo Fisher Scientific) was operated in positive mode. The optimal LC and MS parameters were summarized in Table 1.

Table 1 Optimized LC-MS conditions for NACA-MPOZ, NAC-MPOZ and GSH-MPOZ analysis. Item Column LC

Mobile phase

NACA-MPOZ NAC-MPOZ GSH-MPOZ YMC AQ12S05-1546WT ACN/H2O with 0.1% formic acid

Flow rate Isocratic flow

0.2 ml/min 40% ACN

40% ACN

Spray voltage

5000 V

Aux gas pressure

25 Arb*

Shealth gas pressure

30 Arb

MS Ion sweep gas pressure

0 Arb

Tube lens offset

232 V

Capillary temperature

200 °C

Skimmer offset

-18 V

Monitored m/z value

50% ACN

307

308

452

*Arbitrary unit

2.6 Method validation and stability assessment

5

The bioanalytical validation procedure and stability evaluation were carried out according to the FDA guideline and the one recommended by Shah et al (FDA et al., 2013; Shah et al., 2000). Apart from short-term (up to 48 h) stability of both processed and spiked samples, long-term (2 months) stability of NACA in plasma, stock solution as well as other solvents related to medical use, including saline and 5% glucose, was also studied. 2.7 Application for pharmacokinetics study All the animal experiments described were approved by the Stockholm Southern Ethical Committee (Ethic permit No. S1034-17) and performed in accordance with Swedish Animal Welfare Law. NACA or NAC was dissolved in sterile saline (pH adjusted to neutral) and administered orally (p.o.) or intravenously (i.v.) to female BALB/c mice at the dose of 300 mg/kg. Around 30 µl blood was taken through facial vein at 5, 20, 30, 60, 90, 120, and 180 min post administration, and placed in EDTA-coated tubes. Each mouse provided three samples, and was sacrificed right after the last sample collection. After centrifugation (3,500 g; 15 min), 10 µl plasma was aspirated, added into 5 µl NH4 HCO3 (10 mM in H2O), and processed as mentioned in section 2.4. The mean concentrations obtained at each time point were utilized to calculate the PK parameters. 2.8 Measurement of GSH In order to confirm availability of the derivatization route for GSH, GSH-MPOZ was synthesized similarly as previously described for NACA-MPOZ. After initially verified by HPLC-UV system (consisted of a LKB 2150 pump, a Gilson 234 autosampler and a Spectro Monitor 3100 detector), its structure was characterized by 1H-NMR and ESI-MS as well (Section 2.2). Processed samples from Section 2.7 were also measured using the parameters listed in Table 1 for GSH-MPOZ. 2.9 Pharmacokinetic analysis and statistics Chromatograms and the quantitative results were measured using Xcalibur software (Thermo Fisher Scientific), while PK parameters were calculated by WinNonLin software (standard edition, version 2.0). The concentrations of NACA were fitted to a two compartment open model and weighted using 1/Y*Y, both as oral administration and bolus injection. However, the obtained concentrations of metabolite NAC were fitted to a one compartment open model with lag time, and weight of 1/Y. NAC concentrations were fitted to a one-compartment open model with 1/Y weighting for both the administration routes (p.o. and i.v.) as well. The statistical analyses were all performed by Microsoft Office Excel 2010 and Graphpad Prism (version 4). All data are expressed as means ± SD (standard deviation) unless otherwise described. 3 Results and Discussion 3.1 Selection of thiol-reactive reagent

6

N-Ethylmaleimide (NEM), an analogue of N-(1-pyrenyl) maleimide (NPM), is commonly used for thiol labeling and was therefore assessed regarding its ability to derivatize NAC (Baldwin and Kiick, 2011; Ercal et al., 1996). Nevertheless, the derivatization procedure processed at a slow rate when NEM was dissolved in organic solvent (Figure 2A), and the NAC-NEM adduct was susceptible to hydrolysis in aqueous phase (Figure 2B). These phenomena were in agreement with a previous reported study in which NPM was used for thiol derivatization (Wu et al., 2006), implying that maleimide alkylation of thiol might not be suitable for robust quantification of NACA and NAC. In search of other thiol-reactive reagents, a Julia-Kocienski-like reagent, named MPOZ (2(methylsulfonyl)-5-phenyl-1,3,4-oxadiazole), was introduced by Toda et. al. in 2013 (Toda et al., 2013). As seen in 1H-NMR spectra (Figure S1A and S1B), the presence of protons assigned to phenyl, methyl and methylene indicated that NACA and NAC were successfully conjugated to MPOZ. This was further validated by ESI-MS (Figure S1C and S1D), where the fragment ions with m/z of 307 and 308 ([M+H]+) representing NACA-MPOZ and NACMPOZ, respectively, were the most abundant. Moreover, the reagent MPOZ proved to be highly reactive with both NACA and NAC, with the yields reaching maximum within 5 min and stayed at steady state level (stable) thereafter (Figure 2C). Thus, MPOZ was chosen to stabilize NACA and NAC (Figure 2D) in order to avoid thiol oxidation during sample storage and processing for the present study.

Figure 2. Comparison between NEM and MPOZ as thiol-reactive reagents. (A–B) Yields of NAC-NEM at different time points. The amounts of NAC and NEM were 34 µmol and 68 µmol, respectively. NAC was dissolved in H2O, with pH adjusted to around 7, while NEM was dissolved in DMA (A) or H2O (B). (C) Yields of NACA-MPOZ and NACMPOZ at different time points. The readout of point with the highest peak area was considered as 100%. Three technical replicates at each time point. (D) Reaction scheme for NACA or NAC with MPOZ. 3.2 Mixed Disulfide formation (Oxidation) of NACA and NAC in plasma

7

It has been reported that NACA and NAC could be oxidized through dimer formation or covalent conjugate with plasma proteins (particularly albumin) and other small thiol molecules via the formation of disulfide bonds (Harada et al., 2002; King et al., 2019). In order to investigate the conception, we mixed NACA or NAC with human plasma and measured the percentage of the free (intact) forms after different extraction procedures (Figure S2). Remarkably, almost no free form of NACA or NAC was detected after 24 h incubation at room temperature; in contrast, both were clearly detected in freshly prepared samples. To address whether the low recovery after 24 h incubation was due to thiol oxidation, we exploited TCEP to reductively cleave the disulfide prior to MPOZ derivatization. Following TCEP treatment, a clear increase in the peak area of NACA or NAC was observed. Noteworthy, the response increased by 3-fold approximately after disulfide cleavage for the freshly prepared samples. This indicated the presence of a spanking oxidation event, which necessitates consideration while quantifying NACA and NAC. Moreover, the response was rather similar after TCEP treatment irrespective of freshly prepared or incubated sample, thereby highlighting disulfide cleavage as a determining step in abrogating the divergence originating from sample incubation time. 3.3 Optimization of the TCEP/MPOZ derivatization step Inspired by the advantages of combining TCEP and MPOZ in extraction of NACA and NAC, we further optimized the procedures of TCEP/MPOZ derivatization method. The original publication about MPOZ has reported an important role of aqueous phase in MPOZ reactivity (Toda et al., 2013). We found that increasing organic/aqueous phase ratio beyond 2.5 inversely correlated with derivative yield for both compounds (Figure 3A). Considering MPOZ solubility, an organic/aqueous ratio of 2 was selected for later experiment. In terms of the effect of TCEP or MPOZ amounts on the reaction yield, we observed similar yield curves which could be divided into three phases upon increasing the amounts: an increment phase, intermediate steady state level where the reaction reached its maximum followed by a decrease yield when the reagent increased. An explanation is that the reactant amounts in the first phase were inadequate to convert all NACA or NAC to their final derivatives. While the decrement observed in yield upon increasing TCEP beyond 2 µmol (the third phase; Figure 3B) might be ascribed to changes in pH, since TCEP was supplied as its hydrochloride salt form. The detrimental effect observed using excess of MPOZ (≥ 2 µmol; Figure 3C) most probably is due to its solubility issue, yet the underlining reason is unclear. Eventually, 5 µl of 0.2 M TCEP and 40 µl of 0.045 M MPOZ were applied thereafter. Regarding the addition sequence of TCEP and MPOZ (Zheng et al., 2017), higher yields were found when MPOZ was added prior to TCEP (Figure 3D) which most probably is due to the pre-acidic environment provided by TCEP that is not suitable to initiate the derivatization process.

8

Figure 3. Optimization of the TCEP/MPOZ derivatization method. (A) Yields of NACA-MPOZ and NAC-MPOZ at different organic/aqueous (v/v) phase ratios. The reaction time was 30 min. The amounts of NACA, NAC and MPOZ were the same in each group. (B) Yields of NACA-MPOZ and NAC-MPOZ at different TCEP amounts. The amount of MPOZ was 2 µmol, and the organic/aqueous ratio was 2. (C) Yields of NACAMPOZ and NAC-MPOZ at different MPOZ amounts. The addition of TCEP was 1 µmol, and the organic/aqueous ratio was 2. (D) Yields of NACA-MPOZ and NAC-MPOZ at different addition sequences of MPOZ (1.8 µmol) and TCEP (1 µmol). The readout of point with the highest peak area was considered as 100%. Results were shown as the mean ± SD of three technical replicates.

3.4 Validation of the TCEP/MPOZ derivatization method Using the optimal derivatization parameters and the above described LC-MS conditions, we validated the newly developed method in terms of specificity, linearity, accuracy, precision and recovery. Representative chromatograms of blank and spiked human plasma are shown in Figure 4. No interference peaks from plasma were observed as can be seen in Figure 4A. The calibration curves were linear covering a wide concentration range (0.5-500 µg/ml) for both NACA and NAC, with the regression coefficient (R2) equal to 0.9997 and 0.9994, respectively. The limits of detection (signal/noise= 3) were approximately 10 and 25 ng/ml for NACA and NAC, respectively. Accuracy and precision were calculated using five determinations per concentration, and results are presented in Table 2. In spite of a two-step derivatization procedure, this method showed a high accuracy (96.5-101.2%) and precision (up to 8.8%). The recovery of NACA and NAC from human plasma was compared to that from buffer solution (150 mM NH4HCO3), which held equivalent buffering capability to plasma. Results showed that the recovery ranged between 92.5-102.8% for NACA and 88.8102.9% for NAC.

9

Figure 4. Representative LC-MS chromatograms of NACA and NAC in plasma using TCEP/MPOZ derivatization method. (A) Blank plasma and (B) spiked human plasma containing NACA and NAC. Both samples were processed according the TCEP/MPOZ derivatization method and subjected to LC-MS analysis.

Table 2. Precision, accuracy and recovery of TCEP/MPOZ method in analysis of NACA and NAC in human plasma. Conc. (µg/ml)

NACA

NAC

0.5 20 500 0.5 20 500

Precision (%) Accuracy (%) Recovery (%) Intra-day Inter-day 7.7 5.5 96.5 ± 7.0 92.5 ± 5.4 7.2 3.8 100.1 ± 8.0 102.8 ± 4.9 5.8 0.8 99.5 ± 6.5 100.6 ± 2.3 8.8 8.1 98.8 ± 8.2 88.8 ± 2.9 5.4 2.0 101.1 ± 7.1 102.9 ± 4.8 3.6 0.4 101.2 ± 5.6 100.2 ± 2.7

Precision, accuracy and recovery of the method were investigated at low, middle, and high concentrations. Precision was shown as the relative SD of individual measures of the analyte in multiple aliquots of the same stock solution in one day (intra-day) or in five successive days (interday). Accuracy was the percentage of concentration calculated with calibration curve to the actual value. Recovery from human plasma was compared to that from 150 mM NH4HCO3. Results were shown as the mean ± SD of five technical replicates.

10

3.5 Stability assessment In order to reveal factors that might affect quantification, we monitored the short- and longterm stability of processed samples, spiked plasma, and stock solution under different storage conditions. The processed samples were measured after being stored at RT (bench-top), 4°C or -20°C for 48 h. As it is shown in Table 3 and S1, the obtained concentrations were within the range of 89.0-107.3% for NACA-MPOZ and 99.8-112.0% for NAC-MPOZ, respectively, showing good stability required for sample analysis. The spiked samples with NACA or NAC were kept for only 24 h at RT and 4 °C, and recovery values of 91.9-98.0% for NACA and 89.1-99.3% for NAC were observed. Moreover, spiked samples were less stable after 5 cycles of freeze-thaw procedure, with approximatively 18% loss at low concentration for both NACA and NAC. Three cycles of freeze-thaw procedure had less impact on the stability, with a recovery of 87.1-100.8% for NACA and 90.6-105.0% for NAC, respectively (Table S2). Those findings reflected that freshly spiked plasma should be preferably processed within 24 h and analyzed within 48 h, and frequent freeze-thaw procedures should be avoided.

Table 3. Short-term stability of processed and spiked sample at low concentration. Treatment

Compound Freeze-thaw RT 4°C -20°C NACA 103.6 ± 6.0 95.7 ± 6.5 89.0 ± 4.1 96.5 ± 6.1 Processed sample NAC 109.4 ± 9.1 112.0 ± 8.2 104.6 ± 6.7 106.8 ± 7.2 NACA 81.7 ± 1.9 98.0 ± 9.0 97.5 ± 4.9 95.8 ± 3.2 Spiked plasma NAC 82.9 ± 5.8 89.1 ± 6.8 92.0± 2.5 96.7 ± 7.1 Both processed and spiked samples at the concentration of 0.5 µg/ml underwent 5 cycles of freezethaw procedure. In the meanwhile, the stability of processed or spiked sample was evaluated at 24 h or 48 h after being stored under different conditions, respectively. The stability was presented as the relative percentage of the measured concentration in the indicated sample compared to its actual concentration. Results were shown as the mean ± SD of three technical replicates.

In the long-term (2 months) stability study, NACA and NAC spiked into plasma were stable at -20°C, with a recovery ranging from 103.5% to 111.5% for NACA and from 99.7% to 105.4% for NAC (Table S3), showing that storage at -20°C is an alternative for long term studies when plasma cannot be analyzed immediately. In aqueous solutions (10 mM NH4HCO3, pH 7.4), a recovery rates of 91.8 to102.1% (Table S4) were obtained for NACA and NAC at RT, 4°C or -20°C, indicating that aqueous/stock solutions are stable for long term studies. Table 4 shows that NACA was also stable in physiological saline at RT and 4°C (91.0-116.1%), while less stability was observed in 5% glucose at high concentration at RT (86.6%), suggesting that NACA should be preferably stored at 4°C when 5% glucose is used in future clinical settings.

11

Table 4. Long-term stability of NACA in saline and 5% glucose. Saline

Conc. (mg/ml) RT

5% Glucose 4°C

RT

4°C

1

116.1 ± 3.1 103.7 ± 3.6 108.5 ± 2.6 109.0 ± 8.2

5

97.0 ± 5.3 101.3 ± 4.7 92.3 ± 4.1

10

91.0 ± 5.1

97.3 ± 4.6

93.2 ± 9.2

86.6 ± 5.1 104.2 ± 4.8

NACA was dissolved in saline and 5% glucose at 1, 5 and 10 mg/ml and stored at RT or 4°C for 2 months. The concentration of each sample was measured after being diluted to 100 µg/ml. The recovery was presented as the relative percentage (%) of the measured concentration in the indicated sample compared to its actual concentration. Results were shown as the mean ± SD of three technical replicates.

3.6 Application in pharmacokinetics study To verify the applicability of the current method for experimental settings, we compared the recovery from murine and human plasma. Compared to human plasma, the extraction/reaction yield from murine plasma were 108.9% and 106.8% for NACA and NAC, respectively (Figure S3A). These findings indicate the flexibility and robustness of the method to be utilized in different biological matrices. The PK curves of NACA in murine plasma are depicted in Figure 5. As it can be seen, after a dose of 300 mg/kg of NACA, a Cmax of 726 µg/ml was observed in plasma post intravenous administration, while the Cmax reached only 138 µg/ml after the administration of an equal dose orally. Moreover, the levels of metabolite NAC reached Cmax of 241 and 296 µg/ml within 10 min post intravenous and oral administration, respectively. The acquired data clearly showed a rapid metabolism of NACA to NAC in vivo, which is in agreement with previous published results (King et al., 2019; Wu et al., 2006). Remarkably, the amount of NAC generated from oral NACA within 30 min after administration almost doubled compared to that of oral administered NAC at the same dose (Figure 5B and S3B). This finding prompts a superior potential of NACA compared to NAC, which can be exploited in clinical setting.

12

Figure 5. Concentration-time curves of NACA after systemic administration. Plasma concentrations of NACA (A) and its metabolite NAC (B) were monitored after intravenous and oral administration of NACA at a dose of 300 mg/kg. () Observed concentration after intravenous administration; (◊ ) Observed concentration after oral administration; (______) Predicted concentration using WinNonLin software after intravenous administration; (------) Predicted concentration using WinNonLin software after oral administration. Concentrations are shown as the mean ± SD (n=3).

The pharmacokinetic parameters are displayed in Table 5. NACA exhibited higher bioavailability compared to that found for NAC (66.8% vs 14.6%), suggesting that a lower dose of NACA is probably needed to achieve the same effect obtained after the administration of higher dose of NAC. Furthermore, we have observed NACA is eliminated in biphasic manner after both administrations. The distribution phase is rapid with a half-life of 5.7 or 5.0 min for i.v. or p.o., respectively, however, the elimination half-lives were relatively longer (72.8 and 122.7 min for i.v. and p.o., respectively), indicating that NACA is rapidly distributed into the tissues, which probably reflect the lipophilic character of NACA. These results may explain the local ROS scavenging capability that is reported for NACA in different organs (Goyal et al., 2016; Kawoos et al., 2017). Interestingly, after the administration of an equal dose of NACA, the AUC of the metabolite NAC after oral dose was higher (1.7 fold) than that after intravenous dose (Table S5). This is certainly due to the first pass metabolism, where the majority of NACA is metabolized to NAC in the liver. However, the AUC observed after intravenous administration of NAC was relatively higher than the total AUCs (i.e. AUCNACA + AUCmetabolite NAC) after NACA administration, implying that other metabolic pathways (like deamination) of NACA may exist, except for converting the amine group to hydroxyl group.

13

Table 5. Pharmacokinetic parameters for NACA and NAC after systemic administration. Administered NAC i.v. p.o. i.v. p.o. AUC (h•µg/ml) 130.4 ± 16.1 87.1 ± 12.4 388.4 ± 38.4 56.8 ± 8.6 Vd (ml) 7.4 ± 1.7 --3.7 ± 0.5 --HL10 (min) ----11.2 ± 1.3 17.8 ± 4.8 α-HL (min) 5.7 ± 0.7 5.0 ± 29.9 ----β-HL (min) 72.8 ± 11.1 122.7 ± 29.8 ----Cl (ml/h) 41.4 ± 5.1 62.0 ± 8.5 13.9 ± 1.4 98.6 ± 15.0 Cmax (µg/ml) 726.5 ± 170.2 138.0 ± 42.8 1440.0 ± 190.8 84.0 ± 13.7 Bioavailability (%) 66.8 14.6 Parameters

NACA

All the parameters were calculated using WinNonLin software. NACA and NAC were administered at a dose of 300mg/kg. Results were shown as the mean ± SD (n=3). AUC: Area Under the Curve; Vd: Volume of Distribution; HL10: Elimination half-life in 1-compartment model; α-HL and β-HL: Distribution and elimination half-life in 2-compartment model, respectively; Cl: Clearance; Cmax: The maximum reached plasma concentration.

3.7 Application in GSH measurement GSH is one of the most important molecules for detoxifying several xenobiotics (DeLeve and Wang, 2000; Supratim et al., 2011). NAC is well known for replenishing GSH to avoid liver toxicity when is used as antidote for paracetamol overdose (Pettie et al., 2019). NACA is the amide derivative of NAC that supposed to be an alternative to NAC. Due to the fact that all the three molecules have a free thiol group, we intended to examine the possibility of utilizing the current method to quantify GSH. We firstly synthesized GSH-MPOZ and confirmed the derivatization route and product (Figure 6A). As shown in Figure S4A, the appearance of protons assigned to methine, phenyl and imino group revealed the conjugation between GSH and MPOZ. These results were in agreement with that obtained by ESI-MS (Figure S4B), where the most abundant fragment ion was with m/z of 452 ([M+H]+) for GSH-MPOZ. The derivative was clearly seen in both HPLC-UV and LC-MS chromatograms (Figure S4C and 6B). To minimize the interference from plasma content, the percentage of ACN in the mobile phase was raised to 50%. The relative GSH increment in plasma after NACA and NAC administration were monitored (Figure 6C and 6D). Following i.v. NACA administration, total increment in GSH (as expressed as relative GSH increment vs time) was 4 fold higher compared to that seen after i.v. administration of NAC at the same dose. After oral administration the expression of GSH was 3 fold higher following NACA administration compared to NAC. Moreover, in both administration routes, GSH levels were still 2 times higher at 3 h after NACA administration compared to that observed after NAC. This is the first proof-of-concept kinetic study to illustrate that NACA is superior to NAC regarding GSH replenishment.

14

Figure 6. Detection of GSH using the TCEP/MPOZ method. (A) Reaction scheme between GSH and MPOZ. (B) LC-MS chromatogram of GSH after TCEP-MPOZ derivatization. (C-D) Relative GSH increment after intravenous (C) and oral (D) administration of NACA and NAC. Y-axis is presented as the percentage of the increment at indicated time points in relative to the basal level. () Observed relative increment after intravenous administration; (◊ ) Observed relative increment after oral administration; (______) Predicted relative increment using WinNonLin software after intravenous administration; (-----) Predicted relative increment using WinNonLin software after oral administration. Concentrations are shown as the mean ± SD (n=3).

15

4 Conclusions NACA is a promising antioxidant and has been investigated for multiple indications. However, the available knowledge can neither clearly profile the pharmacokinetic property, nor profoundly present the ability to reinforce the GSH pool. To fill the gap, the present investigation has provided a powerful method to quantify NACA as well its metabolite (NAC) beside the determination GSH in plasma with the same procedure. We have rationally established a two-step, robust, reproducible, and selective derivatization method. We also demonstrated that NACA is stable in the working solution and saline, but relatively less stable in glucose at high concentrations at RT. The results are of high impact for future experimental as well as for clinical studies. The current investigation is the first pharmacokinetic study of NACA showing that NACA is eliminated according to two compartments open model, the bioavailability of NACA is superior compared to that observed for NAC, and GSH increment is several fold higher after NACA administration compared to NAC. The present results indicate that NACA may act as a prodrug for NAC. Given the favorable pharmacokinetic properties of NACA, further endeavors are warranted to explore the pharmacodynamics profiles and the clinical application of NACA. Acknowledgement This study was supported by grants from the Swedish Research Council (2017-00741), Swedish Children cancer foundation (Barncancerfonden; PR2017-0083), KI funds (201802377), and Cancer Research Funds of Radiumhemmet project (161082) to M.H. Rui He and Wenyi Zheng receive the PhD student scholarship from China Scholarship Council. Contributions R.H. W.Z., Y.Z., and M.H. conceived the study. R.H. and W.Z. performed experiments and analyzed data. T.G. purified NACA-MPOZ, NAC-MPOZ and GSH-MPOZ. H.O. performed 1H-NMR and made data interpretation. R.H., W.Z. and Y.Z. performed the animal experiment. M.H. supervised the study and acquired funding. All authors participated in the preparation of the manuscript. Conflict of Interests The authors declare that they have no conflict of interest. Reference Aparicio-Trejo, O.E., Reyes-Fermin, L.M., Briones-Herrera, A., Tapia, E., Leon-Contreras, J.C., Hernandez-Pando, R., Sanchez-Lozada, L.G., Pedraza-Chaverri, J., 2018. Protective effects of N-acetylcysteine in mitochondria bioenergetics, oxidative stress, dynamics and S-glutathionylation alterations in acute kidney damage induced by folic acid. Free radical biology & medicine 130, 379-396. Baldwin, A.D., Kiick, K.L., 2011. Tunable degradation of maleimide-thiol adducts in reducing environments. Bioconjugate chemistry 22, 1946-1953. Borgström, L., Kågedal, B., Paulsen, O., 1986. Pharmacokinetics of N-acetylcysteine in man. European Journal of Clinical Pharmacology 31, 217-222.

16

Cazzola, M., Calzetta, L., Page, C., Jardim, J., Chuchalin, A.G., Rogliani, P., Matera, M.G., 2015. Influence of N-acetylcysteine on chronic bronchitis or COPD exacerbations: a meta-analysis. European respiratory review : an official journal of the European Respiratory Society 24, 451-461. DeLeve, L., Wang, X., 2000. Role of Oxidative Stress and Glutathione in Busulfan Toxicity in Cultured Murine Hepatocytes. Pharmacology 60, 143-154. Ebrahimi, S., Soltani, A., Hashemy, S.I., 2018. Oxidative stress in cervical cancer pathogenesis and resistance to therapy. Journal of cellular biochemistry. Ercal, N., Oztezcan, S., Hammond, T.C., Matthews, R.H., Spitz, D.R., 1996. High-performance liquid chromatography assay for N-acetylcysteine in biological samples following derivatization with N-( 1pyrenyl)maleimide. Journal of Chromatography B 685, 329-334. FDA, CDER, CVM, 2013. Guidance for Industry bioanalytical method validation-US. Goyal, V., Bews, H., Cheung, D., Premecz, S., Mandal, S., Shaikh, B., Best, R., Bhindi, R., Chaudhary, R., Ravandi, A., Thliveris, J., Singal, P.K., Niraula, S., Jassal, D.S., 2016. The Cardioprotective Role of NAcetyl Cysteine Amide in the Prevention of Doxorubicin and Trastuzumab-Mediated Cardiac Dysfunction. The Canadian journal of cardiology 32, 1513-1519. Harada, D., Naito, S., Hiraoka, I., Otagiri, M., 2002. In Vivo Kinetic Analysis of Covalent Binding Between N-Acetyl-L-Cysteine and Plasma Protein Through the Formation of Mixed Disulfide in Rats. Pharmaceutical Research 19, 615-620. Ilkan, Z., Akar, F.G., 2018. The Mitochondrial Translocator Protein and the Emerging Link Between Oxidative Stress and Arrhythmias in the Diabetic Heart. Frontiers in physiology 9, 1518. Jaturakan, O., Dissayabutra, T., Chaiyabutr, N., Kijtawornrat, A., Tosukhowong, P., Rungsipipat, A., Nhujak, T., Buranakarl, C., 2017. Combination of vitamin E and vitamin C alleviates renal function in hyperoxaluric rats via antioxidant activity. The Journal of veterinary medical science 79, 896-903. Karlsson, H., Nava, S., Remberger, M., Hassan, Z., Hassan, M., Ringden, O., 2011. N-acetyl-L-cysteine increases acute graft-versus-host disease and promotes T-cell-mediated immunity in vitro. European journal of immunology 41, 1143-1153. Kawoos, U., McCarron, R.M., Chavko, M., 2017. Protective Effect of N-Acetylcysteine Amide on BlastInduced Increase in Intracranial Pressure in Rats. Frontiers in neurology 8, 219. King, B., Vance, J., Wall, G.M., Shoup, R., 2019. Quantitation of free and total N-acetylcysteine amide and its metabolite N-acetylcysteine in human plasma using derivatization and electrospray LCMS/MS. J Chromatogr B Analyt Technol Biomed Life Sci 1109, 25-36. Maddirala, Y., Tobwala, S., Karacal, H., Ercal, N., 2017. Prevention and reversal of selenite-induced cataracts by N-acetylcysteine amide in Wistar rats. BMC ophthalmology 17, 54. Martin, T.A., Causey, D.H., Sheffner, A.L., Wheeler, A.G., Corrigan, J.R., 1967. Amides of NAcylcysteines as Mucolytic Agents. Journal of Medicinal Chemistry 10, 1172-1176. Matés, J.M., Pérez-Gómez, C., Núñez de Castro, I., 1999. Antioxidant enzymes and human diseases. Clinical Biochemistry 32, 595-603. Mejia, S.A., Gutman, L.A.B., Camarillo, C.O., Navarro, R.M., Becerra, M.C.S., Santana, L.D., Cruz, M., Perez, E.H., Flores, M.D., 2018. Nicotinamide prevents sweet beverage-induced hepatic steatosis in rats by regulating the G6PD, NADPH/NADP(+) and GSH/GSSG ratios and reducing oxidative and inflammatory stress. European journal of pharmacology 818, 499-507. Pettie, J.M., Caparrotta, T.M., Hunter, R.W., Morrison, E.E., Wood, D.M., Dargan, P.I., Thanacoody, R.H., Thomas, S.H.L., Elamin, M., Francis, B., Webb, D.J., Sandilands, E.A., Eddleston, M., Dear, J.W., 2019. Safety and Efficacy of the SNAP 12-hour Acetylcysteine Regimen for the Treatment of Paracetamol Overdose. EClinicalMedicine 11, 11-17. Poljsak, B., 2011. Strategies for reducing or preventing the generation of oxidative stress. Oxidative medicine and cellular longevity 2011, 194586. Sadowska, A., Verbraecken, J., Darquennes, K., Backer, W., 2006. Role of N-acetylcysteine in the management of COPD. International Journal of Chronic Obstructive Pulmonary Disease 1, 425-434. Shah, V.P., Midha, K.K., Findlay, J.W., Hill, H.M., Hulse, J.D., McGilveray, I.J., McKay, G., Miller, K.J., Patnaik, R.N., Powell, M.L., Tonelli, A., Viswanathan, C.T., Yacobi, A., 2000. Bioanalytical Method Validation—A Revisit with a Decade of Progress. Pharmaceutical Research 17, 1551-1557.

17

Sugumar, M., Sevanan, M., Sekar, S., 2018. Neuro protective effect of Naringenin against MPTP induced oxidative stress. The International journal of neuroscience, 1-7. Supratim, R., Bibhas, P., Sumit, D., Subrata, C., 2011. Cyclophosphamide-Induced Lipid Peroxidation and Changes in Cholesterol Content Protective Role of Reduced Glutathione. Iranian Journal of Pharmaceutical Sciences 7, 255-267. Tabriziani, H., Lipkowitz, M.S., Vuong, N., 2018. Chronic kidney disease, kidney transplantation and oxidative stress: a new look to successful kidney transplantation. Clinical kidney journal 11, 130-135. Toda, N., Asano, S., Barbas, C.F., 2013. Rapid, stable, chemoselective labeling of thiols with JuliaKocienski-like reagents: a serum-stable alternative to maleimide-based protein conjugation. Angewandte Chemie 52, 12592-12596. Umeno, A., Horie, M., Murotomi, K., Nakajima, Y., Yoshida, Y., 2016. Antioxidative and Antidiabetic Effects of Natural Polyphenols and Isoflavones. Molecules 21. Wu, W., Goldstein, G., Adams, C., Matthews, R.H., Ercal, N., 2006. Separation and quantification of Nacetyl-l-cysteine and N-acetyl-cysteine-amide by HPLC with fluorescence detection. Biomedical chromatography : BMC 20, 415-422. Yu, Y., Jin, Y., Zhou, J., Ruan, H., Zhao, H., Lu, S., Zhang, Y., Li, D., Ji, X., Ruan, B.H., 2017. Ebselen: Mechanisms of Glutamate Dehydrogenase and Glutaminase Enzyme Inhibition. ACS chemical biology 12, 3003-3011. Zheng, W., Benkessou, F., Twelkmeyer, B., Wang, S., Ginman, T., Ottosson, H., Abedi-Valugerdi, M., Subirana, M.A., Zhao, Y., Hassan, M., 2017. Rapid and Robust Quantification of pXyleneselenocyanate in Plasma via Derivatization. Analytical chemistry 89, 7586-7592.

GRAPHICAL ABSTRACT

18