Biological fate and interaction with cytochromes P450 of the nanocarrier material, d-α-tocopheryl polyethylene glycol 1000 succinate

2.1. Chemicals and materials

Chemicals and materials (suppliers) were as follows: TPGS (Shanghai Ponsure Biotech Inc., Shanghai, China); PEG1000 (Changchun Institute of Applied Chemistry, Changchun, China); diazepam for use as internal standard (I.S., National Institute for the Control of Pharmaceutical and Biological Products, Beijing, China); HPLC-grade distilled water (Watson's Food & Beverage, Guangzhou, China); acetonitrile, isopropanol and formic acid (all HPLC-grade) (Fisher Scientific, Toronto, Canada); O,O-dimethyl-O-2,2-dichlorovinyl phosphate (DDVP) (Sigma–Aldrich, St. Louis, MO, USA); human liver microsomes, NADPH, sulfapyrazole, α-naphthoflavone, quercetin, ticlopidine, ketoconazole and quinidine (Research Institute for Liver Diseases. Shanghai, China); phenacetin, dextromethorphan hydrobromide, tolbutamide, testosterone, mephenytoin, midazolam, bupropion, acetaminophen, hydroxybupropion, dextrorphan, 1-hydroxymidazolam, 4-hydroxytoluamide, 6β-hydroxytestosterone and 4-hydroxymephenytoin (iPhase Pharma Services, Beijing, China); amodiaquine, N-desethylamodiaquine and cyTepa (Toronto Research Chemicals, Toronto, Canada).

2.2. LC‒MS/MS assay

Chromatography was performed on a Shimadzu LC-20ADXR HPLC system incorporating a CTO-20AC column oven maintained at 55 °C, an LC-20ADXR binary pump, SIL-30AC autosampler and a DGU-20A degasser. Separation was carried out on a reversed phase PLRP-S 1000 Å polymer column (50 mm × 4.6 mm, 8 μm) using a mobile phase consisting of 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile-isopropanol (50:50, v/v; solvent B) delivered at a flow rate of 1 mL/min. The gradient elution program was as follows: 0–0.5 min, 20% B; 0.5–1.0 min, 20%–80% B; 1.0–2.0 min, 80% B; 2.0–3.0 min, 80%–90% B; 3.0–6.0 min, 90% B; 6.0–6.1 min, 90%–20% B; 6.1–7.0 min, 20% B.

Analytes and diazepam (I.S.) were detected using a SCIEX Triple quad 6500 mass spectrometer equipped with an ESI source operated in the positive ion mode. Optimized ESI parameters were as follows: Source temperature 400 °C; ion spray voltage 5500 V; curtain gas, heater gas and nebulizer gas (N₂) 25, 30 and 30 psi, respectively. Multiple reaction monitoring (MRM) coupled with in-source CID involved transitions (collision energies eV, declustering potentials V) for TPGS at m/z 557.4 → 99.0 (35, 250), for PEG1000 at 221.3 → 89.1 (15, 180) and for I.S. at 285.2 → 193.1 (40, 100).

2.3. Preparation of calibration standards and quality control (QC) samples

Mixed stock solutions containing TPGS (1 mg/mL) and PEG1000 (1 mg/mL) were prepared in acetonitrile‒water (50:50, v/v) and diluted with the same solvent to produce mixed standard solutions containing 5, 10, 30, 50, 100, 300 and 500 μg/mL of each analyte. Mixed QC solutions containing 15, 80 and 400 μg/mL of each analyte were prepared in a similar way. Calibration standards in rat matrices were prepared by spiking samples of blank plasma, urine, feces homogenate and pooled tissue homogenate containing 0.05% DDVP with mixed standard solutions to produce concentrations of each analyte of 50, 100, 300, 500, 1000, 3000 and 5000 ng/mL for plasma and pooled tissue homogenate and 250, 500, 1500, 2500, 5000, 15,000 and 25,000 ng/mL for urine and feces homogenate. Corresponding QC samples were prepared at concentrations of 150, 800 and 4000 ng/mL and 750, 4000 and 20,000 ng/mL, respectively, by spiking matrices with mixed QC solutions.

2.4. Sample preparation

After thawing at room temperature, samples of rat matrices (50 μL) were mixed with 50 μL I.S. working solution (1.5 ng/mL) and 150 μL acetonitrile before being vortexed for 1 min and centrifuged at 15,000×g for 10 min. Aliquots of supernatants (100 μL) were collected, vortex mixed with 100 μL acetonitrile‒water (30:70, v/v) and 20 μL portions transferred to the LC‒MS/MS system for analysis of TPGS and PEG.

2.5. Assay validation

Calibration curves for quantitation of TPGS and PEG1000 in rat matrices were based on peak area ratios and subjected to weighted linear regression (1/x²) to assess linearity. Other validation parameters were evaluated (shown in Supporting Information Figs. S1‒S9 and Tables S1‒S11) to confirm the assay was suitable for simultaneous determination of the target analytes.

2.6. Animal experiments

Wistar rats (weight 200 ± 20 g) were obtained from the Animal Research Institute of Jilin University. All procedures involving rats were carried out in accordance with the Guidance for the Care and Use of Laboratory Animals of the National Research Council of USA, 1996 and related ethical regulations of Jilin University.

2.7. Interaction of TPGS with CYP450 enzymes

Details of microsomal incubations and assay of probe metabolites (sample preparation and LC‒MS/MS parameters) for CYP450 enzymes are given in Supporting Information Notes S-1 and S-2.

3.1. Bioassay of TPGS and PEG1000

To our knowledge, a method for the direct quantitation of TPGS in biological matrices has not been reported. Radiolabeling has been commonly used to analyze PEG and PEG-related materials²⁶ but it does not differentiate PEG-related materials from their degradation products. Nuclear magnetic resonance (NMR) has been applied to quantitate PEGylated protein²⁷ but its low sensitivity and lack of separation capability make it unsuitable for PK studies in complex biological matrices. Although ELISA using a PEG-antibody conjugate has been shown to provide good sensitivity for PEG and PEG-related materials²⁸, it displayed insufficient selectivity to distinguish between analytes, degradation products and endogenous compounds.

LC‒MS/MS is widely used in the bioanalysis of pharmaceuticals due to its matchless selectivity and sensitivity^29,30. Multiple reaction monitoring (MRM) is the principal method applied to PK and toxicokinetic studies. However, both TPGS and the PEG1000 produced by its hydrolysis are polydisperse molecules which acquire multiple charges and form different adduct combinations in the ESI source leading to complex mass spectra (Fig. 2). Thus, it is a significant challenge to determine all the polymeric constituents using LC‒MS/MS and traditional MRM. Fortunately, the numerous precursor ions formed by polydisperse materials can be further fragmented by in-source CID at a higher declustering potential (DP) to a limited number of TPGS/PEG1000 specific fragment ions which can be tested as surrogate ions of TPGS/PEG1000 to allow quantitation of all polymers in the biological sample (Fig. 3). MRM transitions (m/z 557.4–99.0 and 221.3–89.1) produced by the surrogate ions m/z 557.4 and 221.3 show the best signal-to-noise ratio for quantitation of TPGS and PEG1000, respectively (Supporting Information Figs. S10 and S11). Typical optimized chromatograms of TPGS and PEG1000 are shown in Fig. 4.

3.2. Chromatography development

In order to carry out the simultaneous quantitation of TPGS and PEG1000 by LC‒MS/MS, it is important to obtain adequate chromatographic retention and good peak shape for both analytes. However, the fact that the two analytes exhibit a large difference in polarity makes simultaneous quantitation challenging. During the early phase of method development, we found PEG1000 showed moderate retention on C8 and C18 columns whereas TPGS was strongly bound resulting in an excessively long retention time. In order to reduce the run time, a PLRP-S polymer column (50 mm × 4.6 mm, 8 μm) was evaluated and found to provide appropriate retention for both analytes. Moreover, the large pore size (1000 Å) reduced the risk of the pores becoming blocked by macromolecules thereby improving column efficiency and extending column life.

Both TPGS and PEG1000 are mixtures of polymers of different chain lengths which give rise to broad and asymmetric chromatographic peaks when employing a shallow gradient elution. We found peak shape was improved and the two peaks baseline separated by using a steep gradient of a relatively polar organic phase of isopropanol:acetonitrile (50:50, v/v) (Fig. 4). This separation proved necessary to avoid interference because TPGS produced the same PEG related ions as PEG1000 (e.g., m/z 89.2, 133.2, 177.2, 221.2 for 2, 3, 4 and 5 repeating ethylene oxide subunits.

3.3. Sample preparation

Biological matrices like plasma and tissue homogenate contain large amounts of protein and phospholipid which can give rise to serious matrix effects in the mass spectrometric analysis of TPGS and PEG1000. Therefore, sample preparation with high recovery of analytes was necessary to eliminate matrix interferences and provide high sensitivity for quantitative analysis.

During method development, we found that solid-phase extraction (SPE) and liquid–liquid extraction (LLE) gave low recovery for both TPGS and PEG1000. In contrast, due to the high solubility of PEG in aqueous and organic solvents, simple protein precipitation provided satisfactory recovery from rat plasma and tissue homogenate. Organic solvents including acetonitrile, methanol and isopropanol were tested as precipitation solvents and a 3:1 ratio of acetonitrile:biological sample gave the highest recovery (>90%).

3.4. Plasma PK study

After oral administration of TPGS to rats, neither TPGS nor PEG1000 could be detected in plasma indicating that TPGS has extremely low bioavailability. After i.v. injection of TPGS, Q1 scans of chromatographic peaks in plasma samples at 1.65 and 2.37 min showed they corresponded to PEG1000 and TPGS, respectively (Fig. 5). Moreover, comparison of their MW distributions with those in Q1 scans of reference samples showed there was very little if any difference between them (Fig. 5).

Plasma concentration‒time profiles and PK parameters of TPGS after i.v. administration (Table 1 and Fig. 6) show TPGS is rapidly distributed to tissues and organs with no significant difference between male and female rats. The apparent distribution volume (V_d) of TPGS is much larger than the total volume of body fluid indicating it has a wide distribution in vivo and is likely bound to certain tissues. This together with the long elimination half-life (t_1/2) of TPGS indicates it has the potential to accumulate in tissues on long-term administration. TPGS is rapidly hydrolyzed to free PEG1000 which reaches a peak concentration (C_max) at ∼10 min. Compared with TPGS, PEG1000 is rapidly eliminated making it less likely to accumulate in tissues.

3.5. Tissue distribution

The tissue distribution of TPGS shown in Fig. 7 (Supporting Information Tables S12 and S14) shows it reaches a high concentration in spleen, liver and lungs which correspond to tissues with high blood perfusion rates and high expression of the reticuloendothelial system (RES). The latter is mainly present as macrophages which have been shown to take up polymer nanocarrier materials and contribute significantly to their elimination^31,32. A low expression of the RES in heart and kidney probably explains why the concentrations of TPGS in these tissues are significantly less since heart and kidney both have high blood perfusion rates. There is very little TPGS in brain tissue presumably due to its inability to cross the blood‒brain barrier. Tissue levels are also low in muscle and fat reflecting their low blood perfusion rates. The fact that the concentration in fat is higher than in muscle can be attributed to the presence of the lipophilic vitamin E moiety in TPGS. Interestingly, the concentration of TPGS in testes is below the lower limit of quantitation (50 ng/mL) in contrast to the concentration in ovaries which is even higher than in heart and kidney.

The tissue distribution of PEG1000 shown in Fig. 8 (Supporting Information Tables S13 and S15) is somewhat similar to that of TPGS except that the concentration in kidney is significantly higher than that of TPGS reflecting the fact that the kidney is the main organ of excretion of PEG^26,33.

Tissue concentration‒time profiles of TPGS and PEG1000 after single i,v. injections of TPGS to rats are shown in Fig. 9. After 10 h, the concentration of both analytes in most tissues was less than 10% of their corresponding peak concentrations. The relatively high concentration of PEG in the large intestine at 10 h presumably results from it being the route of elimination. It is worth noting that both TPGS and PEG1000 maintain high concentrations in spleen over time suggesting that long-term administration could lead to accumulation in and potential toxicity to this organ.

3.6. Metabolism and excretion

TPGS was undetectable in urine after i.v. injection indicating it is not cleared by renal excretion. This result is consistent with the fact that, in aqueous solution, TPGS undergoes self-assembly to form nanoscale micelles that are inflexible and too big (particle size >20 nm) to undergo glomerular filtration (size limit 10 nm)²². The fact that TPGS micelles are in equilibrium with free TPGS then raises the question of why free TPGS is unable to undergo renal clearance. Presumably this is because of the lipophilicity of the vitamin E component of amphiphilic TPGS.

The cumulative amount of TPGS excreted in feces (Fig. 10A) reaches a plateau of 85 ± 30.4 μg after 60 h and accounts for 8.72% of the dose of TPGS (Supporting Information Table S16). It has been reported that TPGS is not prone to hydrolysis by pancreatic lipases in the gastrointestinal tract^34,35 but could undergo hydrolysis to PEG1000 by the carboxyl esterase 1 enzyme in cells, a suggestion supported by a molecular docking model of enzyme binding³⁶. Since previous studies have indicated that TPGS may be excreted in feces and urine as PEG1000, we also determined PEG1000 in urine and feces after i.v injection of TPGS.

The cumulative amounts of PEG1000 excreted into urine and feces over 120 h (Fig. 10B and C) of 210 ± 67 and 212 ± 66 μg, respectively, accounting for 32.4% and 32.6% of the dose of TPGS, respectively (Supporting Information Tables S17 and S18). This indicates that the total PEG1000 excreted in urine and feces after an i.v. dose accounts for 65% of the dose of TPGS and that the cumulative amount of TPGS and PEG1000 excreted (Fig. 10D) accounts for approximately 73% of the dose.

3.7. Effect of TPGS on CYP450 enzymes

The activity–concentration profiles of TPGS on various CYP450 enzymes in human liver microsomes are shown in Fig. 11 (Supporting Information Table S19). TPGS showed no inhibition of CYP1A2 and 2B6 but varying degrees of concentration-dependent inhibition of the other 5 enzymes. The IC₅₀ values for inhibition of these 5 enzymes are given in Table 2.

Previous studies have shown that drugs with IC₅₀ values in the range 10–50 μmol/L are generally considered to be weak inhibitors of CYP450 enzymes and those with IC₅₀ > 50 μmol/L to have negligible inhibitory effects^37,38. On this basis, our results indicate that inhibition of CYP450 enzymes by TPGS is negligible except for CYP3A4 where a weak inhibitory effect may be observed. Whether this is of clinical concern depends on the plasma concentration of TPGS attained during clinical use of a TPGS-based NDDS. Since such NDDS normally contain a significant proportion of TPGS, it would appear at least possible that a concentration of TPGS sufficient to inhibit CYP3A4 could be reached. This points to the possibility that monitoring the plasma concentration of TPGS during clinical use of a TPGS-based NDDS is an appropriate strategy to avoid this type of drug interaction. Of course the main concern must apply to CYP3A4 in the wall of the small intestine since it is rich in CYP3A4 and plays an important role in first-pass metabolism of orally administered drugs³⁹. Inhibition of CYP3A4 by TPGS at this site could potentially lead to an increased absorption of a co-administered drug or, in the worst case scenario, an overdose.