Biotransformation of Penindolone, an Influenza A Virus Inhibitor

Penindolone (PND) is a novel broad-spectrum anti-Influenza A Virus (IAV) agent blocking hemagglutinin-mediated adsorption and membrane fusion. The goal of this work was to reveal the metabolic route of PND in rats. Ultra-high-performance liquid chromatography tandem high-resolution mass spectrometry (UHPLC–HRMS) was used for metabolite identification in rat bile, feces and urine after administration of PND. A total of 25 metabolites, including 9 phase I metabolites and 16 phase II metabolites, were characterized. The metabolic pathways were proposed, and metabolites were visualized via Global Natural Product Social Molecular Networking (GNPS). It was found that 65.24–80.44% of the PND presented in the formation of glucuronide conjugate products in bile, and more than 51% of prototype was excreted through feces. In in vitro metabolism of PND by rat, mouse and human liver microsomes (LMs) system, PND was discovered to be eliminated in LMs to different extents with significant species differences. The effects of chemical inhibitors of isozymes on the metabolism of PND in vitro indicated that CYP2E1/2C9/3A4 and UGT1A1/1A6/1A9 were the metabolic enzymes responsible for PND metabolism. PND metabolism in vivo could be blocked by UGTs inhibitor (ibrutinib) to a certain extent. These findings provided a basis for further research and development of PND.


Introduction
Influenza A virus (IAV) circulated in the human population as an annually recurring epidemic disease with significant impact on human health as well as on economy. Fewer types of drugs available and resistance of antiviral drugs make its treatment remain a challenge [1][2][3]. Furthermore, the pandemic of COVID-19 and influenza highlights the importance of antiviral drug development. Thus, it is necessary and desperate to discover and develop novel anti-influenza drugs.
Penindolone (PND), a new diclavatol indole adduct produced by an Antarctic deepsea derived fungus (Penicillium crustosum PRB-2), was discovered to be a broad-spectrum anti-IAV agent with low risk of inducing drug resistance. It is the first dual inhibitor to block hemagglutinin-mediated adsorption and membrane fusion. Even though its bioavailability was low, PND can protect mice against IAV-induced death and weight loss after oral dosing, superior to the effects of the clinical drug oseltamivir [4].
Although the anti-IAV activities and mechanism of PND, as well as its plasma pharmacokinetics in vivo were explored [4,5], the disposition fate of PND remains unclear. Drug metabolism is an integral part of pharmaceutical development process, facilitating the identification of metabolic "soft spots" and providing information on the fate of candidate drugs to aid lead optimization with improved pharmacokinetics, toxicology and efficacy. Regulatory agencies recommend that metabolites should be given full consideration in Table 1. UHPLC-HRMS-based metabolites identifications of PND in rat urine, feces, bile and plasma. Phase I reactions of oxidation, methylation, demethylation and dehydrogenation and Phase II reactions of glucuronidation and sulfation were the observed metabolic reactions of PND in rats. Noteworthy, glucuronide conjugates were only detected in plasma and bile and were readily hydrolyzed by β-glucuronidase ( Figure 2). These results supported the structural assignment for glucuronide conjugates. Whereas corresponding phase I metabolites but not the glucuronide conjugates were only detected in rat feces, indicating the hydrolysis reaction of glucuronide conjugates in intestinal canal. The proposed metabolism pathway was summarized in Figure 3.
In addition, as shown in Table 2, the metabolites of PND varied in different administration ways and in different biological samples, suggesting that disposition processes affected the fate of PND.  Figure S2b-2,c-2,d-2), m/z 474.19 yielded by neutral loss of glucuronyl moiety was observed and owned similar fragmentation behavior with M0 in positive mode. However, the two symmetrical clavatol moieties result in multiple metabolites with the same MS 2 spectra information. Therefore, they were tentatively identified as glucuronidation metabolites of PND, which was consistent with our previous study of rat plasma [5].          [5]. Both negative and positive m/z of precursor ions were 16 Da higher than that of PND. It was preliminarily speculated that it was produced by adding an oxygen atom to the PND molecule. Based on the obtained MS 2 data ( Figure S3 17. Apart from the above information, the other MS 2 also revealed similar fragmentation regularity of PND in positive ion mode. However, those binding sites could not be determined based on the current information. Therefore, M14 and M15 were preliminarily identified as oxidation, followed by glucuronidation metabolites of PND.  6 , error = 1.23 ppm). It was 14 Da higher than that of PND in negative ion mode, indicating that it probably was produced by adding a methylene to the PND molecule. As shown in Figure S4 6 , error < 2.5 ppm) and were eluted at 25.69 and 26.17 min, respectively. They were 14 Da lower than that of PND, which were consistent with the loss of methylene group in the molecular structure of PND. The MS 2 spectrum of M9 and M10 ( Figure S5a,b) showed the informative ion at m/z 151.04, which suggested that acetophenone moiety at the indole ring was incomplete. The fragment ions at m/z 280.0982 and m/z 165.05 indicated that demethylation occurred at C-2 or C-3 acetophenone moiety. Hence, M9 and M10 were assigned as demethylation metabolites of PND.
M16, M17 and M18 were eluted at 21.56, 21.78 and 22.89 min, respectively. Three metabolites were detected with the same theoretical [M−H] − ion at m/z 634.19 (C 33 H 34 NO 11 , 1.42 ≤ error ≤ 1.57 ppm) in negative ion mode. All of them were 176 Da higher than that of M7 and M8. They might be glucuronidation products of them. MS 2 spectra showed that m/z 458. 16 12 , error = 1.08 ppm). It was 176 Da higher than that of M11, which was consistent with the type of glucuronidation metabolite. Also, the fragment-ions of M19 in positive mode owned similar fragmentation behavior with M11. Given to above information, M19 was tentatively identified as dehydrogenation (−4H), followed by glucuronidation metabolite of PND. . They were 2 Da lower than that of PND in positive ion mode, which was consistent with dehydrogenation (−2H) products. The MS 2 spectrum ( Figure S7a,b) showed characteristic fragment ions at m/z 292.10 and m/z 294.11, which were respectively produced by losing C 10 H 12 O 3 and C 10 H 10 O 3 moiety, suggesting that dehydrogenation occurred at one of the acetophenone moiety. In a word, M12 and M13 were produced by dehydrogenation of PND.
Metabolites M20, M21 and M22, which were eluted at 21.96, 22.91 and 23.38 min, respectively, were detected with the same theoretical [M+H] + ion at m/z 648.21 (C 34 H 33 NO 12 , 0.46 ≤ error ≤ 0.62 ppm). They were 176 Da higher than that of M12 or M13 in positive ion mode, indicating that they were glucuronide conjugate of M12 or M13. For MS 2 spectra, the ion at m/z 472.17 (neutral loss 176 Da) and similar fragmentation behavior with M12 or M13 in positive mode was observed. Therefore, they were tentatively identified as glucuronide conjugate of M12 or M13. All of them were 80 Da higher than that of PND, suggesting that they might be sulfation products of PND. MS 2 spectra ( Figure S8) showed that m/z 470.16 yielded by neutral loss of sulfate moiety was observed and owned similar fragmentation behavior with M0 in negative mode. Therefore, they were identified as the sulfation metabolites.

Molecular Networking
The molecular networks in both positive and negative ion modes were generated ( Figure 4) to understand the metabolism of PND in rats and validated the relation of PND and its metabolites. According to UHPLC-HRMS information, PND is usually metabolized in the liver by losing hydrogen and methyl, producing hydroxyl and methyl metabolites and/or resulting glucuronide and sulfate conjugates. In order to highlight the relation networks between metabolites, cluster A

Metabolites Quantification
According to the results of metabolite profiling above, four analytes (parent of PND and its oxidation, methylation and glucuronidation metabolites) were further quantified for the recovery study. The concentrations of parent and glucuronidation metabolites of PND in bile were determined. The accumulation-time curves of determined compounds are shown in Figure 5a. PND rapidly accumulated in bile at 0-12 h in the formation of glucuronide conjugate products. After 72 h of administration via gavage and injection, the accumulation of PND and glucuronidation metabolites corresponded to 2.68-3.35% and 65.92-83.79% of dosing. Fecal and urinary cumulative excretion-time curves of PND and its determined metabolites (oxidation and methylation of PND) are shown in Figure 5b,c. The majority of PND and its determined metabolites were excreted in feces at 12-24 h. Moreover, all these compounds were hardly found in feces after 48 h post dose. PND and its metabolites were barely detected in urine. The overall excretion of PND and its determined metabolites corresponded more than 51.06% of intake dose in feces, whereas it was less than 0.5% in urine.

In Vitro Metabolism of PND by Rat, Mouse and Human LMs
The probe substrates metabolism of various enzymes (CYPs and UGTs) in LMs indicates the availability of metabolic system ( Figure S9). PND was discovered to be eliminated in rat, mouse and human LMs to different extents in the presence of NADPH and UDPGA. The disappearance of PND at various time points is presented in Figure 6. In the presence of NADPH in LMs, the significant species differences were also indicated by the different metabolic rate of PND. Approximately 96% and 89% of PND was metabolized in rat and mouse LMs within 1 h, respectively. Less than 16% of PND was diminished in human LM system within 60 min. Furthermore, we also investigated the stability of PND in LMs based on UGTs enzyme. We found that PND was unstable in human, rat and mouse LMs with more than 65% of prototype drug eliminated in the presence of UDPGA. In addition, PND was stable in all species LMs without co-factors (NADPH or UDPGA). . The color of the node is referring to the presence characteristics of metabolites in various samples. The edges (lines) are connecting the nodes in accordance with the "cosine score" (fragment match/similarity score ranging 0.7-1), and the thickness of the edges is related to the relative abundance of metabolites within a network. Bile, feces, plasma and urine samples are labeled as G1, G2, G3 and G4, respectively. Three clusters (A-C) were selected and discussed in this article.

Metabolites Quantification
According to the results of metabolite profiling above, four analytes (parent of PND and its oxidation, methylation and glucuronidation metabolites) were further quantified for the recovery study. The concentrations of parent and glucuronidation metabolites of PND in bile were determined. The accumulation-time curves of determined compounds are shown in Figure 5a. PND rapidly accumulated in bile at 0-12 h in the formation of glucuronide conjugate products. After 72 h of administration via gavage and injection, the accumulation of PND and glucuronidation metabolites corresponded to 2.68-3.35% and 65.92-83.79% of dosing. Fecal and urinary cumulative excretion-time curves of PND and its determined metabolites (oxidation and methylation of PND) are shown in Figure 5b, c. The majority of PND and its determined metabolites were excreted in feces at 12-24 h. Moreover, all these compounds were hardly found in feces after 48 h post dose. PND and its metabolites were barely detected in urine. The overall excretion of PND and its determined metabolites corresponded more than 51.06% of intake dose in feces, whereas it was less than 0.5% in urine. . The color of the node is referring to the presence characteristics of metabolites in various samples. The edges (lines) are connecting the nodes in accordance with the "cosine score" (fragment match/similarity score ranging 0.7-1), and the thickness of the edges is related to the relative abundance of metabolites within a network. Bile, feces, plasma and urine samples are labeled as G1, G2, G3 and G4, respectively. Three clusters (A-C) were selected and discussed in this article.
The metabolism kinetics of PND was further explored for rat, mouse and human LMs. The kinetic plots of PND were shown in Figure S10, and the parameters are listed in Table 3. The transformation reactions of PND in LMs conformed to Michaelis-Menten kinetics. For CYPs reaction system, The K m value of PND in HLM (97.70 µM) was over six times that in RLM (15.24 µM) and MLM (16.11 µM), but the V max was quite similar in all species (1.1-1.3 nmol/(min × mg protein)). Additionally, the CL int value of HLM (0.013 µL/min) was far smaller than that of RLM and MLM, suggesting a lower metabolism in HLM. Besides, the significant species differences were found in UGTs reaction system. PND was found to have K m and V max values of 31.61 µM and 6 nmol/(min × mg protein) in MLM, respectively, resulting in the highest CL int (0.198 µL/min) among the three species. Moreover, it was shown that the clearance of PND by UGTs in MLM and HLMs was faster than that by CYPs based on the higher CL int of UGTs.

In Vitro Metabolism of PND by Rat, Mouse and Human LMs
The probe substrates metabolism of various enzymes (CYPs and UGTs) in LMs indicates the availability of metabolic system ( Figure S9). PND was discovered to be eliminated in rat, mouse and human LMs to different extents in the presence of NADPH and UDPGA. The disappearance of PND at various time points is presented in Figure 6. In the presence of NADPH in LMs, the significant species differences were also indicated by the different metabolic rate of PND. Approximately 96% and 89% of PND was metabolized in rat and mouse LMs within 1 h, respectively. Less than 16% of PND was diminished in human LM system within 60 min. Furthermore, we also investigated the stability of PND in LMs based on UGTs enzyme. We found that PND was unstable in human, rat and mouse LMs with more than 65% of prototype drug eliminated in the presence of UDPGA. In addition, PND was stable in all species LMs without co-factors (NADPH or UDPGA). The metabolism kinetics of PND was further explored for rat, mouse and human LMs. The kinetic plots of PND were shown in Figure S10, and the parameters are listed in Table 3. The transformation reactions of PND in LMs conformed to Michaelis-Menten kinetics. For CYPs reaction system, The Km value of PND in HLM (97.70 μM) was over six times that in RLM (15.24 μM) and MLM (16.11 μM), but the Vmax was quite similar in all species (1.1-1.3 nmol/(min × mg protein)). Additionally, the CLint value of HLM (0.013 μL/min) was far smaller than that of RLM and MLM, suggesting a lower metabolism in HLM. Besides, the significant species differences were found in UGTs reaction system. PND was found to have Km and Vmax values of 31.61 μM and 6 nmol/(min × mg protein) in MLM, respectively, resulting in the highest CLint (0.198 μL/min) among the three species. Moreover, it was shown that the clearance of PND by UGTs in MLM and HLMs was faster than that by CYPs based on the higher CLint of UGTs.

Identification of Enzymes Responsible for the Metabolism of PND in LMs
Chemical inhibition assay was carried out in the LMs reaction system for the metabolism enzyme types identification (CYPs: 1A2, 2C19/C11, 2C9/6, 2D6/D2, 3A4 and 2E1) and UGTs: 1A1, 1A3, 1A4, 1A6, 1A9 and 2B7), which mediated PND biotransformation. The effects of the CYPs and UGTs inhibitors on PND metabolism in LMs were shown in Figure 7. Reaction at 0 and 30 min was defined as control and normal metabolism samples, respectively. Metabolism of PND was decreased significantly in the presence of CYP2E1 inhibitor (disulfiram). Inhibitors of CYP2C9 (sulphafenazole) and CYP3A4 (ketoconazole) inhibited PND metabolism by 19-48% in different species LMs. Moreover, the overall

Identification of Enzymes Responsible for the Metabolism of PND in LMs
Chemical inhibition assay was carried out in the LMs reaction system for the metabolism enzyme types identification (CYPs: 1A2, 2C19/C11, 2C9/6, 2D6/D2, 3A4 and 2E1) and UGTs: 1A1, 1A3, 1A4, 1A6, 1A9 and 2B7), which mediated PND biotransformation. The effects of the CYPs and UGTs inhibitors on PND metabolism in LMs were shown in Figure 7. Reaction at 0 and 30 min was defined as control and normal metabolism samples, respectively. Metabolism of PND was decreased significantly in the presence of CYP2E1 inhibitor (disulfiram). Inhibitors of CYP2C9 (sulphafenazole) and CYP3A4 (ketoconazole) inhibited PND metabolism by 19-48% in different species LMs. Moreover, the overall metabolism of PND was moderate inhibited by troglitazone (UGT1A6 in human LMs, 17.32%) and significantly inhibited by troglitazone (UGT1A6 in RLM and MLM, 68.62 and 61.80%), bilirubin (UGT1A1 in RLM, 64.32%) and niflumic acid (UGT1A9 in HLM, 36.69%).  Since the significant species differences in metabolic kinetics and isozymes, the detectable metabolites so far described in rat plasma, bile, urine and feces may not be the only ones in the downstream metabolic pathway.

The Effect of Ibrutinib on PND Pharmacokinetics
As shown in Figure 8, the pharmacokinetic behavior of PND was significantly altered by ibrutinib. The pharmacokinetic parameters are presented in Table 4. After pretreatment Since the significant species differences in metabolic kinetics and isozymes, the detectable metabolites so far described in rat plasma, bile, urine and feces may not be the only ones in the downstream metabolic pathway.

The Effect of Ibrutinib on PND Pharmacokinetics
As shown in Figure 8, the pharmacokinetic behavior of PND was significantly altered by ibrutinib. The pharmacokinetic parameters are presented in Table 4. After pretreatment with ibrutinib through i.p. injection, the maximum plasma concentration (C max ) of PND was elevated by 18.78%. More importantly, twin peaks in the plasma concentration-time curve disappeared. The area under the plasma concentration-time curve (AUC 0-∞ ) value after combination with ibrutinib increased 30.56%. In addition, PND showed higher t 1/2 and Vd values under the effect of ibrutinib (p < 0.05). Under such conditions, liver had some contributions to the low bioavailability of PND. After the majority of the UGT enzymes were inhibited, the elimination rate of PND decreased. The finding of altered pharmacokinetics by UGT inhibitor provided further evidence that PND could be metabolized by UGTs. The potential DDI with other drugs will be evaluated in vitro and in vivo.

Animals
Male Sprague-Dawley rats (weight: 220 ± 20 g; age: 6-8 weeks) from Beijing Vital River Laboratory Animal Technology Co., Ltd. (NO. SCXK-2019-0009) were used. Rats were maintained under 12 h light-dark cycle, 20-25 • C and 50-60% humidity. They were fed with free access to water and a standard diet. All animal care and experimental procedures were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Animal Ethics Committee of School of Medicine and Pharmacy, Ocean University of China (Qingdao, China) (Approval code: OUC-SMP-2021-03-03).

Identification of the Metabolites of PND by UHPLC-HRMS
Identification analysis of metabolites in bio-samples was performed as described previously [5] using a Dionex Ultimate 3000 UHPLC system combined with a Q-Exactive Focus Orbitrap mass spectrometer (Thermo Fischer Scientific, MA, USA) fitted with a CORTECS T3 column (2.1 × 150 mm, 2.7 µm; Waters, MA, USA). An electrospray ionization (ESI) source in both negative-and positive-ionization mode was used. Data recording and processing were carried out using the Xcalibur software (version 2.2, Thermo Fischer Scientific, MA, USA) and Compound Discover software (version 3.2, Thermo Fischer Scientific, MA, USA).
A total of 100 µL plasma, bile, urine and feces homogenate samples were deproteinized with 500 µL, 500 µL, 500 µL and 900 µL of acetonitrile, respectively. After vortex-mixing and centrifugation at 14,000 rpm for 10 min twice, 5 µL of supernatant was injected into the UHPLC-HRMS system for qualitative analysis.

Molecular Networking
The raw UHPLC-HRMS data were converted into mzXML files by MSConvert (ProteoWizard) and uploaded to the web-based platform, Global Natural Product Social Molecular Networking (GNPS), for generating MS based molecular networks [18]. The default GNPS networking parameters were adjusted by keeping the mass tolerance at ± 0.02 Da (for precursor ion), and its threshold was tolerance at ± 0.02 Da for the fragment ions to create the simplified resultant networks. To generate a consensus spectrum from the identical MS/MS spectra and to reduce more less-related spectra of the network, the minimum matched fragment ion was set at 6 (MS fragments) where minimum cosine score for connecting the nodes was set as 0.7. Cystoscope software (version 3.9.0) was used to visualize GNPS data and to generate large as well as subnetwork portions. The cystoscope data was supplemented with the node color referring to the individual source file (different samples), whereas the thickness of the network edges (lines) reflected the cosine similarity score indicating the high (thick line) and low (thin line) MS/MS spectral match. The result link and dataset were as follows: https://gnps.ucsd.edu/ProteoSAFe/status.jsp?task=b4eaee952aeb432aa55508a98e1 7a3b6# (accessed on 27 November 2021) and https://gnps.ucsd.edu/ProteoSAFe/status. jsp?task=2bf77b60877746859c7dba84c5248fed (accessed on 27 November 2021).

Determination of PND and Its Metabolites by UPLC-MS/MS
The quantitative determination of PND and its metabolites in bile, urine and feces samples was carried out on an ultra-performance liquid chromatography tandem mass spectrometry (UPLC-MS/MS) instrument consisting of a UPLC H-Class PLUS system and a Xevo TQ-XS triple quadrupole mass spectrometer (Waters, MA, USA) with Masslynx software (version 4.2). The transitions and corresponding cone voltages and collision voltages for PND and its metabolite were optimized as follows: m/z 472.16 > 306.03 (48 V, 18 V) for PND, m/z 486.14 > 266.01 (50 V, 46 V) for methylated PND, m/z 488.29 > 164.97 (30 V, 36 V) for oxidation PND, and m/z 500.29 > 179.02 (28 V, 38 V) for internal standard (IS, HDYL-GQQ-2399). Other chromatography and MS parameters followed our previous study [5].
Bile, urine and feces homogenate samples (100 µL) were taken and added 500 µL, 500 µL and 900 µL acetonitrile containing IS (20 ng/mL), respectively. In addition, 50 µL of each bile sample was mixed with 30 µL of β-glucuronidase-solution and 420 µL Tris-HCl buffer and incubated at 37 • C for 2 h. Following hydrolysis, 100 µL incubation samples were mixed with 500 µL acetonitrile (containing IS, 20 ng/mL). The mixture was vortexed and centrifuged at 14,000 rpm for 10 min twice. Finally, 2 µL of the supernatant was injected into the analysis system.

Data Analysis
Given corresponding authentic standards of metabolites were unavailable, two derivatives, 1-methlated PND (QL-Vir-09) and 5-hydroxylated PND (lbh3-78-2) were used to prepare the calibration curve to tentatively quantify the methylated and oxidation metabolites.
Figures and statistics were plotted and calculated using GraphPad Prism (version 8.4.3) and Microsoft Excel (version 2019). Metabolic pathways were illustrated with Chem-BioDraw Ultra (version 13.0). Pharmacokinetic parameters were estimated using noncompartmental methods (WinNonlin version 2.0, Pharsight Co., Certara, NJ, USA). All pharmacokinetic parameters were given as the mean ± standard deviation. All statistical analyses were performed using SAS (version 9.4; SAS Institute Inc., Cary, NC, USA). A p value < 0.05 was considered statistically significant.

Conclusions
A total of 25 metabolites including 9 phase I metabolites and 16 phase II metabolites were characterized and identified according to accurate mass, elemental composition and MS 2 spectra. The metabolism of PND followed several known biotransformation pathways including oxidization, demethylation, dehydrogenation, methylation, sulfation and glucuronidation. The overall recovery of PND in rat's excreta was determined as more than 51%. Moreover, 65.24-80.44% of the PND rapidly accumulated in bile in formation of glucuronide conjugate metabolites. CYPs (CYP2E1, CYP2C9 and CYP3A4) and UGTs (UGT1A1, UGT1A6 and UGT1A9) isoenzymes mediated the metabolism of PND, and the metabolic showed species difference. PND metabolism in vivo could be blocked by ibrutinib, a UGTs inhibitor, to a certain extent. These findings provided a basis for further research and development of PND.