Sorafenib increases cytochrome P450 lipid metabolites in patient with hepatocellular carcinoma

Hepatocellular carcinoma (HCC) is a leading cause of cancer death, and medical treatment options are limited. The multikinase inhibitor sorafenib was the first approved drug widely used for systemic therapy in advanced HCC. Sorafenib might affect polyunsaturated fatty acids (PUFA)-derived epoxygenated metabolite levels, as it is also a potent inhibitor of the soluble epoxide hydrolase (sEH), which catalyzes the conversion of cytochrome-P450 (CYP)-derived epoxide metabolites derived from PUFA, such as omega-6 arachidonic acid (AA) and omega-3 docosahexaenoic acid (DHA), into their corresponding dihydroxy metabolites. Experimental studies with AA-derived epoxyeicosatrienoic acids (EETs) have shown that they can promote tumor growth and metastasis, while DHA-derived 19,20-epoxydocosapentaenoic acid (19,20-EDP) was shown to have anti-tumor activity in mice. In this study, we found a significant increase in EET levels in 43 HCC patients treated with sorafenib and a trend towards increased levels of DHA-derived 19,20-EDP. We demonstrate that the effect of sorafenib on CYP- metabolites led to an increase of 19,20-EDP and its dihydroxy metabolite, whereas DHA plasma levels decreased under sorafenib treatment. These data indicate that specific supplementation with DHA could be used to increase levels of the epoxy compound 19,20-EDP with potential anti-tumor activity in HCC patients receiving sorafenib therapy.


Introduction
Liver cancer is a global issue, being the most common cancer and the leading cause of cancer death in transition countries.In 2020, almost 9,06,000 patients were diagnosed with liver cancer and over 8,30,000 deaths were documented worldwide.Hepatocellular carcinoma (HCC) has the highest prevalence among the different subtypes of liver cancer (Sung et al., 2021).Viral infections, more specifically hepatitis B and C virus (Datfar et al., 2021), lifestyle factors such as alcohol intake (Matsushita and Takaki, 2019), as well as type 2 diabetes and non-alcoholic fatty liver disease (NAFLD) (Estes et al., 2018) remain the leading risk factors, depending on the region considered.Hepatocellular carcinoma remains one of the most common causes of cancer death, especially in men, and has one of the lowest 5-year survival rates of all different cancer types (Siegel et al., 2022).
In addition to chronic inflammation, tissue remodeling and changes in cellular signaling are pathogenetic factors of carcinogenesis (Refolo et al., 2020).Interestingly, patients with a non-cirrhotic HCC, mostly caused by NAFLD, seem to show a more severe HCC histopathology on one hand but better overall survival on the other hand (Gawrieh et al., 2019).
The first targeted and currently a widely used systemic therapy for HCC is the oral multikinase inhibitor sorafenib binding in an ATP-binding pocket to inhibit kinase function (Hwang et al., 2013), which predominantly inhibits angiogenesis via binding the vascular endothelial growth receptor (VEGFR).Furthermore, it targets the cell proliferation and differentiation via the rapidly accelerated fibrosarcoma (RAF) signaling pathway (Wilhelm et al., 2004) and the platelet-derived growth factor receptor-β (PDGFR-β) (Mody and Abou-Alfa, 2019) as well as the beneficial effects, particularly in cancer and its complications, which is likely due to the inhibition of nuclear factor kappa B (NF-κB) and production of the proresolution mediators at the molecular level (Freitas and Campos, 2019).
In addition to the known anti-angiogenic and anti-proliferative effects of sorafenib it has also been described to have effects on the soluble epoxide hydrolase (sEH) which showed similar antiinflammatory effects as conventional sEH inhibitors in lipopolysaccharide-induced inflammation models in mice (Liu et al., 2009).Sorafenib is a potent inhibitor of sEH compared with conventional urea-based sEH inhibitors (Hwang et al., 2013).The sEH is expressed in numerous human tissues with the main distinction between microsomal epoxide hydrolase (mEH) and sEH (Morisseau and Hammock, 2013).The sEH metabolization is the dominant pathway in humans, so that it can be assumed that sEH inhibition has a stabilizing effect on endogenous epoxy metabolites in tissues (Spector and Kim, 2015).
The role of epoxidized LC-PUFAs is well established in several biological processes, such as angiogenesis, inflammation, and tumor growth: In different animal models, it has been shown that sEHinhibition has a positive influence on cardiovascular and liver abnormalities (Iyer et al., 2012), liver fibrosis and portal hypertension (Zhang et al., 2018), fatty liver (Yao et al., 2019) and non-alcoholic steatohepatitis (Wang et al., 2019).The EETs are known to affect blood pressure, inflammation, pain sensation, and regeneration (Arnold et al., 2010;Morisseau et al., 2014).However, a proangiogenic effect of 11,12-EET and 14,15-EET, the main EET regioisomers in mammals (Spector and Kim, 2015), has been described via the epidermal growth factor (EGF) and VEGF pathways (Zhang et al., 2014), which may explain the finding that EETs promote tumor growth.The n-6 AA-derived regioisomers 5,6-EET and 8,9-EET were found to increase cell proliferation and de novo vascularization (Yan et al., 2008), whereas 11,12-EET and 14,15-EET promote tumor angiogenesis through endothelial cell proliferation (Panigrahy et al., 2012;Zhang et al., 2013).An increase in 14,15-EET through sEH-inhibition led to increased tumor growth and metastasis through cell invasion in experimental studies (Panigrahy et al., 2012).In summary, through mechanisms of cell proliferation, de novo vascularization and endothelial proliferation several EET-regioisomers promote tumor growth and metastasis (Yan et al., 2008;Panigrahy et al., 2012;Zhang et al., 2013;Zhang et al., 2014).The sEH-inhibitory effect of sorafenib might thus carry clinically relevant consequences by increasing these pro-tumorigenic EET mediators.
In contrast, n-3 PUFA-derived regioisomers show antiangiogenic effects both via the VEGF and FGF-2 pathway (Zhang et al., 2014).In a tumor mouse model, it was shown, that low dose sEH-inhibition led to an increase in n-3 DHAderived 19,20-EDP and thereby reduced tumor angiogenesis and cell invasion and thus inhibition of tumor growth (Zhang et al., 2013).Furthermore, a protective effect in obesity and obesityrelated comorbidities, such as fatty liver disease, of n-3 epoxy PUFA was found in animal models (López-Vicario et al., 2015).The beneficial effects of n-3 PUFA regarding to cancer and its complications are probably due to their anti-inflammatory and pro-resolution mediators (Freitas and Campos, 2019).In the context of abundant DHA, the sEH-inhibitory effect of sorafenib might thus lead to higher levels of 19,20-EDP, mediating antitumor effects.
Currently, n-6 PUFA are found in a ratio of approximately 20 times more than n-3 PUFA in the human diet (Harris, 2006); therefore humans have low n-3 PUFA tissue levels, and a shift of the competitive n-6 and n-3 PUFA metabolism towards n-6 PUFA derived lipid metabolites (Spector and Kim, 2015).
We therefore aimed to investigate levels of n-6 and n-3 PUFA in HCC patients, as well as n-6 PUFA-and n-3 PUFA-derived epoxide and corresponding dihydroxy compounds in HCC patients without and during sorafenib therapy.Based on the data of our pilot study (Leineweber et al., 2020), we hypothesized that sorafenib treatment, due to sEH-inhibitory and possibly CYP-modulating effects, might increase the presence of potentially tumor growth-suppressing DHA-derived EDPs, as well as of potentially tumor growthpromoting EETs.

Patients and blood sampling
The study population evaluated in this sub-analysis comprised patients within the palliative treatment arm of the randomized, controlled, multicenter phase II SORAMIC study, which evaluated sorafenib alone compared to selective internal radiation therapy (SIRT) combined with sorafenib on overall and progression-free survival in patients with advanced HCC (Ricke et al., 2019).
Patients were included in this analysis if they received study treatment in the palliative arm of SORAMIC and signed an informed consent form, so blood samples were collected and analyzed at baseline (BL) and at the first follow-up visit (FU) after approximately 7-9 weeks, and samples were stored for subsequent analyses at −80 °C.Of the 424 randomized patients assigned to the palliative arm, we performed lipidomic analysis of 43 patients from the intention to treat (ITT) population, all of whom received sorafenib, characterized in terms of gender distribution, age, body mass index (BMI), presence of cirrhosis, liver function, biomarker, or tumor stage according to Barcelona Clinic Liver Cancer (BCLC) stage, except in the expression of the Child-Pugh points between 5 and 6/7 (p < 0.0453) as shown in Table 1.

Sample preparation and GC
Plasma samples were analyzed for determination of fatty acids using the gas chromatography (GC) technology as described previously (Wang et al., 2022).75 μL of EDTA plasma per sample was used for the GC preparation.Methylation and extraction of FAs were carried out on the basis of an established protocol (Kang and Wang, 2005).Briefly, frozen samples were thawed at room temperature.All samples were then mixed with 50 μL pentadecanoic acid (PDA, 1 mg/mL in ethanol, Merck Schuchardt OHG, Hohenbrunn, Germany) as internal standard, 500 μL borontrifluoride (BF 3 , Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) in 14% methanol (Merck KGaA, Darmstadt, Germany), and 500 μL n-hexane (Merck KGaA, Darmstadt, Germany) in glass vials and were tightly closed.After vortexing, samples were incubated for 60 min in a preheated block at 100 °C.After cooling down to room temperature, the mixture was added to 750 μL water, vortexed, and extracted for 4 min.Then all samples were centrifuged for 5 min (RT, 3,500 rpm).From each sample, 100 μL of the upper n-hexane layer was transferred into a micro-insert (placed in a GC glass vial), tightly closed, and analyzed by GC.
The prepared samples were analyzed using an Agilent 1290 HPLC system with a binary pump, an autosampler, and a column thermostat with a Agilent Zorbax Eclipse plus C18 column 150 mm × 2.1 mm, 1.8 µm using a solvent system of aqueous acetic acid (0.05%) and acetonitrile:methanol (50:50).The multiple step elution gradient started at 95% aqueous phase, which was increased within 18 min-98% organic phase and held there for 10 min.The flow rate was set at 0.3 mL/min, and injection volume was 20 μL.The HPLC was coupled with an Agilent 6495 Triple Quad mass spectrometer with an electrospray ionization source.Analysis of lipid mediators was performed with the Multiple Reaction Monitoring in the negative mode, limit of quantitation (LOQ) was 0.01 ng/mL.

Statistical analysis
Statistical analysis was performed using GraphPad Prism 9 (GraphPad Software, La Jolla, CA, United States).The comparison was made using the Wilcoxon matched-pairs signedrank test.The correlation was made using linear regression.All values are presented as the mean ± standard error of the mean.Statistical significance was assumed when p < 0.05.(*0.01 ≤ p < 0.05; **0.001 ≤ p < 0.01; ***p < 0.001).

Results
Blood samples from a total of n = 43 HCC patients were analyzed in a paired fashion with blood taken without and undergoing 7-9 weeks sorafenib treatment.This is a sub-analysis of the well-characterized study population of the randomized controlled, multicenter phase II SORAMIC trial (Ricke et al., 2019).Patients from this study population, from which suitable amounts of blood samples were available for the analysis of fatty acids and their lipid metabolites before and during sorafenib treatment, were chosen for the analysis performed and presented here.The patients received sorafenib 200 mg twice a day for 1 week before increasing the dose to 400 mg twice a day.Based on disease progression and clinical condition, the sorafenib dose was escalated to 600-800 mg or reduced to 0-200 mg.The patients' general characteristics are shown in Table 1.

N-6 and n-3 epoxides and dihydroxy metabolites are higher in patients undergoing sorafenib treatment
To investigate the effect of sorafenib treatment on the n-6 and n-3 PUFA epoxide formation, we measured the concentrations of the epoxymetabolites and corresponding dihydroxy metabolites by quantitative LC-ESI-MS/MS analysis in plasma samples of HCC patients without and during sorafenib therapy.As a result of sorafenib treatment, the levels of n-6 AA-derived epoxymetabolites 5,6-EET and 8,9-EET increased significantly.Levels of epoxymetabolites derived from the n-3 PUFAs tended to increase as well, but for DHA-and EPA-derived epoxides failed to reach significance (Figures 2A-C; Supplementary Figures S1A-C).The concentrations of EETs were higher compared to the n-3 PUFA-derived EEQs.EDP metabolite concentrations were approximately half of those observed for the EETs, while the concentrations of the EEQs were the lowest in this patient cohort.
The dihydroxy-PUFA products of the epoxy-PUFA formed via the sEH, respectively dihydroxyeicosatrienoic acids (DHETs) from EETs, dihydroxydocosapentaenoic acids (DiHDPAs) from EDPs and dihydroxyeicosatetraenoic acids (DiHETEs) from EEQs increased whilst on sorafenib treatment as well (Supplementary Table S1).When comparing absolute amounts of AA-, DHA-and EPA-derived epoxy-plus dihydroxy-PUFA significantly higher levels of metabolites derived from all three PUFAs were found (Figures 2D, E).

N-6 and n-3 fatty acid levels decrease during sorafenib treatment
To explore the presence of LC-PUFA, the fatty acid composition in plasma from patients with HCC was analyzed without and during sorafenib treatment by gas chromatography.Interestingly, we found a decrease in the relative content of n-6 (AA) and n-3 (DHA) PUFAs, with an increasing n-6/n-3 ratio (Figure 3A; Supplementary Table S2).

N-6 and n-3 cytochrome P450 epoxy and dihydroxy product ratios do not support the hypothesis of increased sEH inhibition during sorafenib treatment
To determine whether the sEH inhibitory effect of sorafenib is detectable from the lipid metabolites assessed here, we analyzed whether sorafenib treatment increases plasma content of epoxymetabolites as compared to their dihydroxy products.As a marker for the enzyme activity in the CYP epoxygenase/sEH axis the plasma ratio of EET to DHET as characterization of the sEH inhibition has been used (Liu et al., 2009).We adapted this approach using the following equation for the AA-as well as the DHA-and EPA derived epoxy-and dihydroxy compounds: However, in contrast to lower ratios that would indicate lower sEH activity, we found higher dihydroxy/epoxy product ratios in the sorafenib treated patients (Figure 4).This does not support lower sEH activity in the sorafenib-treated patients.Interestingly, these higher ratios were only significant for the n-3 PUFA DHA-and EPA-derived metabolites.
3.4 Metabolization of AA and EPA to their derived cytochrome P450 epoxy and dihydroxy products is similar, while metabolization of DHA to 19,20-EDP and 19,20-DiHDPA is markedly higher In order to assess total CYP-epoxide and corresponding dihydroxy formation as a function of their respective substrate fatty acids we analyzed the epoxy and corresponding dihydroxy concentrations as a ratio with their respective substrate PUFA:

CYP − products epoxy + dihydroxy PUFAs PUFAs
We found higher CYP product/PUFA ratios due to sorafenib treatment, providing evidence of increased presence of bioactive epoxy-PUFA from AA, EPA, and DHA in patients undergoing sorafenib treatment (Figures 5A-C).Furthermore, when analyzed as a ratio of DHA-derived CYP-products versus DHA as substrate, the 19,20-metabolites were found to be the predominant metabolites formed (Figure 5B).We found significant differences in plasma fatty acid composition in patients with HCC without sorafenib compared to during sorafenib treatment.Relative levels of AA and DHA were significantly lower during sorafenib treatment.Furthermore, we were able to demonstrate significantly higher EET levels and a trend towards increased n-3 CYP metabolites especially 19,20-EDP in this study population with HCC receiving sorafenib treatment.
When taking into account the different levels of the precursor n-3 PUFAs EPA and DHA we were able to establish that EPA is metabolized by CYP enzymes to a similar extent as AA, while DHA utilization was higher, leading to significantly increased levels of the 19,20-metabolites derived from DHA during sorafenib treatment (Figures 2E, 5B).Given that previous data from mouse models show inhibition of tumor angiogenesis and reduced cell invasion by increasing 19,20-EDP (Zhang et al., 2013) and to dampen and alleviate inflammation in the liver (López-Vicario et al., 2015), this could be a beneficial effect of sorafenib that could be harnessed in HCC therapy by supplementing DHA.
Generally, a beneficial role of n-3 PUFAs to dampen development of HCC is described both in animal models (Lim et al., 2009;Weylandt et al., 2011;Inoue-Yamauchi et al., 2017) and human observation studies (Sawada et al., 2012;Gao et al., 2015;Koh et al., 2016).This and the disbalance of the n-6/n-3 ratio in the Western Diet (Harris, 2006) suggests that supplementation of n-3 PUFAs could balance the n-6/n-3 ratio and may reduce tumor progression in HCC patients, regardless of sorafenib treatment.With all the described effects of n-3 PUFAs as receptor agonists, modulators of molecular signalling pathways and inflammatory responses, and data indicating that n-3 PUFA increase the efficacy of chemotherapies and consequently the overall survival of cancer patients, n-3 PUFAs can thus be considered as pharmaceutical nutrients (Bougnoux et al., 2009;Chagas et al., 2017;Paixão et al., 2017).However, in the population studied here the n-3 PUFA baseline was not associated with the overall survival (Supplementary Figure S2).
Prior studies showed that dietary increase of baseline n-3 PUFA concentrations can enhance formation of n-3 PUFAderived CYP epoxy-PUFA (Fischer et al., 2014;Sarparast et al., 2020;Weylandt et al., 2022).Higher levels of n-3 PUFA may thus potentially increase anti-tumor n-3 PUFA-derived epoxymetabolites as well as decrease pro-tumor n-6 PUFAderived metabolites (Zhang et al., 2014).Interestingly, in this study we found lower levels of DHA in patients treated with sorafenib (Figure 3), further supporting the concept to increase DHA in the daily diet in order to increase also levels of 19,20-EDP in HCC patients treated with sorafenib.
Many classes of currently used drugs can block or modify pathways of lipid mediator formation.Particularly wellestablished are non-steroidal anti-inflammatory drugs inhibiting the cyclooxygenase (COX) enzymes as well as numerous clinically well-established substances that modify (induce, inhibit) CYP enzymes and thereby modify lipid mediator formation.In general, by using quantitative LC-MS/MS oxylipin analysis in the context of established pharmacotherapy (pharmacolipidomics) as shown in this paper we hope to identify oxylipins that might be used to stratify and possibly also modify and improve response to treatments: Experimental data indicate strong biological effects of specific lipid mediators, particularly with regard to inflammationdampening oxylipins from n-3 PUFA (n-3 IDOs) (Weylandt et al., 2022) in contrast to often inflammation-promoting oxylipins from n-6 PUFA (n-6 IPOs).As concentrations of these can be modified by changes in fatty acids substrates, as well as established drugs such as sorafenib, there could be a rationale for targeted modifications of the n-6/n-3 PUFA ratio in the diet in the context of established pharmacotherapy to harness these effects.
In systemic HCC therapy, the combination of immune checkpoint inhibitors, and VEGF pathway inhibitors such as sorafenib could promote an immune-permissive environment, thereby enhancing the response to immune therapeutic approaches.Immunotherapy is effective in only approximately 1/ 3 of cancer patients, and targeting the TME to decrease tumor cell evasion is regarded as an opportunity to improve response to immunotherapy, as conversion of "cold" tumors to "hot" tumors with T cell infiltration is associated with a better response rate to cancer treatment (Bonaventura et al., 2019).TME-modifying properties of lipid mediator levels may therefore enhance antitumor effects by transforming the immune landscape.Aspirin dampens tissue inflammation via the metabolism of arachidonic acid in the COX enzymatic pathways and can reduce cellular growth in hepatocellular carcinoma (Tao et al., 2018;Refolo et al., 2020).Whether a CYP/sEH-dependent effect on lipid mediators-which could be modulated/enhanced by sorafenib as described here-could also play a role in the TME, remains a topic for future studies that are directly analyzing effects of n-3 IDO and n-6 IPO levels and formation, as well as immune cells in liver and liver tumor tissue.
In our data presented here, we were not able to discern an sEH inhibitory effect of sorafenib (Figure 4).Non-etheless we established significantly higher levels of CYP-derived epoxy and dihydroxy metabolites in patients undergoing sorafenib treatment.While previous results from one of us show that storage at −80 °C should be sufficient to yield stable CYP derived oxylipin readings comparable to other oxylipins (Gladine et al., 2019;Koch et al., 2020), there might have been variations in the process of blood sampling and storage leading to changes in epoxy-PUFA levels.Another explanation would be that sorafenib may have more complex effects on PUFA-derived metabolites in humans, with increased formation of epoxy-PUFAs.We did not analyse expression of sEH directly, therefore there is a possibility of increased sEH expression, as described in an animal models of high fat diet induced liver disease (López-Vicario et al., 2015) which might compensate for an inhibitory effect of sorafenib.Interestingly, in our pilot study we found an increase of 8,9-EET, 11,12-EET and 14,15-EET levels in HCC patients treated with sorafenib (Leineweber et al., 2020) while we found a significant increase only of 5,6-EET and 8,9-EET levels here.We believe this might be due to the analytical limitations due to differences in sample taking and storage, and not a mechanistic difference in the effect observed.Indeed, we suggest to use the combined analysis of epoxides and dihydroxy compounds (as in Figures 2D-F) to assess epoxy metabolite formation.However future studies could better address this question by analyzing the samples under defined conditions and at different time intervals after blood drawing to measure the effects of these variations on epoxy-and dihydroxy-PUFA levels.

Conclusion
In this study, we investigated the effect of sorafenib treatment on PUFA formation and epoxy lipid mediator concentrations in peripheral blood plasma in a group of 43 HCC patients as a sub-analysis of the randomized, controlled, multicenter phase II SORAMIC study.We were able to demonstrate markedly increased epoxy plus dihydroxy PUFA concentrations in the peripheral blood of HCC patients undergoing sorafenib therapy.These results support previous findings that sorafenib treatment induces a change in epoxy-/dihydroxy-PUFA concentrations.
Given the anti-tumor effects described in experimental models for the n-3 PUFA-derived 19,20-EDP, these data further support the hypothesis that dietary n-3 PUFA supplementation in addition to sorafenib treatment could contribute anti-tumor effects due to n-3 epoxy-PUFA.Frontiers in Pharmacology frontiersin.org11 Leineweber et al. 10.3389/fphar.2023.1124214

FIGURE 1
FIGURE 1 CYP-dependent lipid metabolite formation from AA/DHA/EPA, and the potential effects of sorafenib.

FIGURE 4
FIGURE 4Ratio of n-6 and n-3 PUFA-derived dihydroxy to epoxy-PUFA as a marker for sEH activity in n = 43 patients with HCC without and undergoing sorafenib treatment (***p < 0.001).
FIGURE 5 N-3 and n-6 PUFA-derived epoxides plus dihydroxy compounds as a marker for the presence of CYP metabolites in plasma from n = 43 patients with HCC without and undergoing sorafenib treatment.(A) Ratio of AA-derived products divided by AA plasma content, (B) ratio of DHA-derived products divided by DHA plasma content, (C) ratio of EPA-derived products divided by EPA plasma content (*p < 0.05, **p < 0.01; ***p < 0.001; ****p < 0.0001).
DiHETE dihydroxyeicosatetraenoic acid EDP epoxydocosapentaenoic acid EEQ epoxyeicosatetraenoic acid EET epoxyeicosatrienoic acid EGF epidermal growth factor EPA eicosapentaenoic acid FGF fibroblast growth factor FID flame ionization detector GC gas chromatography HCC hepatocellular carcinoma HPLC/ESI high performance liquid chromatography electrospray ionization LC-PUFA long-chain polyunsaturated fatty acid mEH microsomal epoxide hydrolase MRI magnetic resonance imaging MS mass spectrometry NAFLD fatty acid non-alcoholic fatty liver disease NASH non-alcoholic steatohepatitis NF-κB nuclear factor kappa B n-3 PUFA omega-3 polyunsaturated fatty acids n-6 PUFA omega-6 polyunsaturated fatty acids OS overall survival PDA pentadecanoic acid PDGFR-β platelet-derived growth factor receptor β PFS progression-free survival RAF rapidly accelerated fibrosarcoma sEH soluble epoxide hydrolase SIRT selective internal radiation treatment SPE solid phase extraction VEGFR vascular endothelial growth factor receptor

TABLE 1
Patient characteristics of the n = 43 HCC patients receiving sorafenib treatment.Data are presented as mean ± standard error of the mean.