|Home | About | Journals | Submit | Contact Us | Français|
The advent of multi-kinase inhibitors targeting the VEGF-receptor has revolutionized the treatment of highly angiogenic malignances such as renal cell carcinoma. Interestingly, several such inhibitors are commercially available, and they each possess diverse specific beneficial and adverse effect profiles. In examining the structure of sorafenib, it was hypothesized that this compound would possess inhibitory effects on the soluble epoxide hydrolase (sEH), an enzyme with pleiotropic effects on inflammation and vascular disease. We now show that sorafenib, but not another VEGF-receptor targeted inhibitor sunitinib, is a potent inhibitor of the human sEH in vitro (KI = 17 ± 4 nM). Furthermore, sorafenib causes the expected in vivo shift in oxylipid profile resulting from sEH inhibition, evidence of a reduction in the acute inflammatory response. Lipopolysaccharide (LPS)-induced hypotension was reversed with sorafenib, but not sunitinib, treatment, suggesting that sEH inhibition accounts for at least part of the anti-inflammatory effect of sorafenib. The pharmacokinetic studies presented here in light of the known potency of sorafenib as a sEH inhibitor indicate that the sEH will be largely inhibited at therapeutic doses of sorafenib. Thus it is likely that sEH inhibition contributes to the beneficial effects from the inhibition of the VEGF-receptor and other kinases during treatment with sorafenib.
With the advent of more complete knowledge of the molecular biology of cancer, new therapies have recently been designed which target mechanisms by which the disease escapes standard therapy. For example, the multi-kinase and VEGF-receptor inhibitors, such as sorafenib and sunitinib (1, 2), interrupt the pathway by which angiogenesis becomes established and promulgated, resulting in inadequate nourishment of metastatic disease thereby leading to a higher degree of treatment success. In certain malignancies, such as kidney cancer whose mechanism of oncogenesis generally involves disrupted hypoxia pathways and thus is highly angiogenic, these agents have had the effect of revolutionizing treatment.
Upon viewing the recently described X ray-crystal structure of B-Raf complexed with sorafenib (3), we noted a structural similarity between this drug (Fig. 1A) and the class of urea-based compounds (Fig. 1B) that inhibit the soluble epoxide hydrolase (sEH) (4). The sEH converts epoxyeicosatrienoic acids (EETs) to the less active dihydroxyeicosatrienoic acids (DHETs) (5). The EETs have been demonstrated to be vasodilators in various animal models (6-10) and play an important role in regulation of blood pressure as well as control and prevention of heart disease (6, 11-14). In addition, EETs are potently anti-inflammatory through mediating the nuclear factor kappa B (NF-κB) and IκB kinase system (15-17). The sEH inhibitors have been shown to stabilize the EET levels and thus have beneficial effects on hypertension (18), nociception (19), atherosclerosis (20), and inflammation (21) through increasing endogenous levels of EETs and other lipid epoxides. Thus, we asked whether sorafenib possesses sEH inhibitory activity and, consequently, whether this effect influences the known beneficial properties or adverse effects of this drug.
Although there are several VEGF-R inhibitors commercially available (2), there exist differences in effect profiles for the various clinically available agents. While there are minimal published studies comparing sunitinib to sorafenib in renal cell carcinoma (RCC) (22), with the one available study showing a greater frequency of gastrointestinal symptoms with sunitinib (23), there is an ongoing NCI clinical trial to attempt to address this issue (ClinicalTrials.gov Identifier: NCT00326898). In an attempt to explain differences in VEGF-R inhibitor treatment outcomes as well as adverse events, we have examined the sEH inhibitory properties of these RCC therapeutics.
We now show that sorafenib, but not another clinically available VEGF-R inhibitor sunitinib, possesses powerful sEH inhibitory activity. While the sEH inhibition demonstrated by sorafenib shows the expected alteration in oxylipin profile in mice, such activity does not contribute to kinase inhibition or sorafenib-induced cytotoxicity shown by the VEGF inhibitors. Thus, sorafenib, in addition to its advantageous effects on tyrosine kinases, may also result in salutary effects on hypertension, inflammation, or nociception based on its property of sEH inhibition.
The ERK, p-ERK, and β-actin antibodies were from Sigma (St. Louis, MO), and the PARP, VEGF-R, and p-VEGF-R were from Cell Signaling (Danvers, MA). Goat anti-mouse and goat anti-rabbit horseradish peroxidase-conjugated IgG were obtained from Bio-Rad (Richmond, CA), and ECL Western Blotting Detection Reagents were obtained from Amersham Biosciences (Buckinghamshire, United Kingdom). Recombinant human sEH enzymes were produced in a baculovirus expression system and purified by affinity chromatography (24, 25).
ACHN and A498 human RCC cells (from ATCC) were maintained in MEM 1X media supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 1 mM sodium pyruvate, 0.1 mM non-essential amino acid, and 0.75% sodium bicarbonate at 37 °C in a humidified incubator containing 5% CO2 in air.
Protein concentration was quantified using the Pierce BCA assay with bovine serum albumin (BSA) as calibrating standard. The IC50s were determined by a fluorescent assay using cyano(6-methoxy-naphthalen-2-yl)methyl trans-[(3-phenyloxyran-2-yl)methyl] carbonate (CMNPC) as a fluorescent substrate (26). Human sEH was used at a concentration of 1 nM that gave linear generation of product with both time and protein concentration. Human sEH was incubated with inhibitors for 5 min in pH 7.0 Bis-Tris/HCl buffer (25 mM) containing 0.1 mg/mL of BSA at 30°C prior to substrate introduction ([S] = 5 μM). Activity was measured by determining the appearance of the 6-methoxy-2-naphthaldehyde with an excitation wavelength of 330 nm and an emission wavelength of 465 nm for 10 minutes. Assays were performed in triplicate. IC50 is the concentration of inhibitor that reduces enzyme activity by 50%. IC50 was determined by regression of at least five datum points with a minimum of two points in the linear region of the curve on either side of the IC50. Results are presented as the average ± standard deviation of three separate measurements.
Dissociation constants were determined following the method described by Dixon (27) for competitive tight binding inhibitors, using [3H]-1,3-diphenyl-trans-propene oxide (t-DPPO) as substrate (28). Inhibitor at concentrations between 0 and 50 nM for sorafenib [0-25 nM for 12- (3-adamantan-1-yl-ureido)-dodecanoic acid (AUDA); 0-5 nM for trans-4-[4-(3-adamantan-1-yl- ureido)-cyclohexyloxy]-benzoic acid (t-AUCB); 0-200 nM for 1-trifluoromethoxyphenyl-3-(1- acetylpiperidin-4-yl) urea (TPAU); 0-50 nM for 1- trifluoromethoxyphenyl-3 (1- methylsulfonyl)piperidin-4-yl) urea (TUPS)] was incubated in triplicate for 5 min in pH 7.4 sodium phosphate buffer at 30°C with 100 μL of the enzyme (1 nM of human sEH). Substrate (3.6 ≤ [S]final ≤ 30 μM) was then added. Velocity was measured as described (28). For each substrate concentration, the plots of the velocity as a function of the inhibitor concentration allow the determination of an apparent inhibition constant (KIapp) (27). The plot of these KIapp as a function of the substrate concentration allows the determination of KI when [S] = 0. Results are presented as average ± standard error of KI calculation.
Molecular modeling was performed using “BioMedCAChe 5.0” software (Fujitsu Computer Systems Corporation). The atomic coordinates of the crystal structure of human sEH complexed with the sEH inhibitor were retrieved from Protein Data Bank (PDB) (entry 1ZD3) (29). The sEH inhibitor was removed from the active site, and sorafenib was manually docked into the ligand-binding pocket by superposition with the parent molecule sEH inhibitor. The ligand and the amino acid residues within 8.0 Å from the ligand were minimized on MM geometry (MM3).
A 200 μL aliquot of cells (1 × 103 cells in quiescent media) was added to a 96 well plate and incubated for 18 hr at 37 °C in a humidified incubator containing 5% CO2 in air. After incubation, of the test compounds at the indicated concentration were added into each well for 48 hr. Control cultures were treated with DMSO. After incubation, a 20 μL MTT solution (5 mg/mL in phosphate buffer) was added to each well and the incubation continued for 4 hr, after which time the solution in each well was carefully removed. The blue crystalline precipitate in each well was dissolved in DMSO (200 μL). The visible absorbance at 560 nm of each well was quantified using a microplate reader.
The CaspACE assay kit (Promega, Madison, WI) was used to measure the protease activity of caspase-1 (ICE) and caspase-3 (CPP32), per the manufacturer's instructions. Briefly, 2 × 106 cells per 10 cm dish were treated as described in the text. Positive control cells were treated with 4 μM Camptothecin. The cells were harvested, washed, and equal protein quantities were incubated with the DEVD-pNA caspace-3 substrate in Caspase Assay Buffer. The plates were covered with parafilm and incubated at 37°C overnight. Color development was measured at 405 nm.
Male Swiss-Webster mice (8 week) were purchased from Charles River Laboratories and experiments were performed according to protocols approved by the Animal Use and Care Committee of University of California-Davis. Sorafenib tosylate (1 mg) was suspended in 1 mL trioleine and then strongly mixed on a mini vortexer for 5 minutes. Blood (10 μL) was collected from the tail vein using a pipette tip rinsed with 7.5% EDTA(K3) at 0, 0.5, 1, 1.5, 2, 4, 6, 8, 24 hours after oral dosing with the inhibitor. The blood samples were prepared and analyzed according to the methods detailed in our previous study (30). Sorafenib was detected by negative model electrospray ionization tandem quadrupole mass spectrometry in multiple reaction-monitoring mode (MRM) with the precursor and dominant daughter ions of 463.0 and 193.8, respectively.
Sorafenib and the sEH inhibitort-AUCB were dissolved in trioleine containing 10% ethanol and 10% polyethylene glycol (average molecular weight: 400, PEG 400) to give a clear solution while sunitinib gave a suspension in the same formulation solution. Male, Swiss Webster mice (8 week) were randomly assigned to be injected i.p. with LPS (10 mg/kg body weight) and immediately orally administered with sorafenib tosylate (20 mg/kg body weight, equal to 30 μmol/kg), sunitinib malate (17 mg/kg body weight, equal to 30 μmol/kg) or t-AUCB (1 mg/kg body weight, equal to 2.5 μmol/kg), respectively. Animals receiving oral gavage of trioleine containing 10% ethanol and 10% PEG 400 immediately after i.p. injection of LPS or saline served as positive or negative controls, respectively. Systolic blood pressure was determined before and 24 h after treatment with non-invasive tail cuff methods using a CODA6 system (Kent Scientific Co., CT). For each measurement set, 3 acclimation cycles and 25 data cycles were used. The reported systolic blood is presented as a mean of at least 5 cycles of the measurement set. If the systolic blood pressure was under the detection limit, we recorded it as the detection limit (60 mmHg) for further analysis.
Another set of animals was assigned in random to be repeatedly treated as above described. Animals were sacrificed 24 hours after drug treatment. Blood was collected to separate plasma for oxylipin analysis and cytokine assay as previous reported methods (30). Plasma samples were extracted for oxylipin analysis using the methods described in (30). Plasma oxylipin were measured using an Agilent 1200 Series HPLC (Agilent Technologies, Inc. Santa Clara, CA) coupled with an Applied Biosystems 4000 QTRAP hybrid, triple-quadrupole MS instrument (Applied Biosystems, Foster City, CA). The instrument was equipped with a linear ion trap and a Turbo V ion source and was operated in negative MRM mode. Plasma oxylipins were separated using a 2.1 mm × 150 mm Pursuit XRs-C18 5 μm column (Varian Inc, Palo Alto, CA) held at 40 °C. The gradient conditions are given in Table S1. The injection volume was 10 μL and the samples were kept at 10 °C in the auto sampler. The mass spectrometers were set with a negative electrospray mode with following parameters. CUR= 20 psi, GS1=50 psi, GS2=30 psi, IS=-4500 V, CAD= HIGH, TEM=400°C, ihe=ON, DP= - 60 V. The collision energies used for CAD (-18 to -38 eV) varied according to molecular species and were individually maximized for each compound. The parent and daughter ions for monitoring the target oxylipin and the corresponding parameters are presented in Table S2. The plasma ratio of EETs to DHETs was calculated for each individual animal.
Based on the structural similarity between sorafenib (3) and sEH inhibitors (sEHIs) (Fig. 1) (29), we suspected that sorafenib might be a sEH inhibitor. To test this hypothesis, we first measured the inhibition potency (IC50) of sorafenib with the recombinant affinity purified human and murine sEHs using a fluorescent assay. As shown in Table 1, we found an IC50 for sorafenib of 12 nM for the human and 30 for the murine sEH enzymes, which is very similar to that of previously reported potent sEH inhibitors, such as AUDA, t-AUCB, TPAU, or TUPS (31-33). Interestingly, sunitinib and dasatinib, another other VEGF-R and multikinase inhibitor respectively (2), are not potent inhibitors of the human sEH. To define the potency of sorafenib as a sEH inhibitor, we determined its dissociation constant (KI; see Fig. 2) using a radioactivity-based assay. Sorafenib has a KI of 17 nM, which is approximately 10-fold higher than that of the very potent t-AUCB, but similar to the KI of other potent sEH inhibitors (Table 1). Interestingly, the sEH inhibitors TPAU and TUPS appear relatively less potent when determining KI with a radioactive assay than determining IC50 with a fluorescent assay. Such discrepancy has been observed before (26).
Because sorafenib showed good inhibitory activity against sEH, we next sought to understand how it binds to this enzyme. Sorafenib was manually docked at the active site of the human sEH using an X-ray structure previously published (Fig. 3) (Gomez et al., 2006). Sorafenib was noted to be bound through H-bonding interactions with the residues Tyr381, Tyr465, and Asp333 as observed for other urea-based inhibitors (34, 35). However, unlike t-AUCB, sorafenib did not establish a hydrogen bond with Met418 (33). The lack of this H-bonding could explain the slightly decreased inhibitory activity of sorafenib compared to t-AUCB. Conversely, t-AUCB and other sEH inhibitors were manually docked at the active site of the B-Raf X-ray structure (3). Interestingly, the latter compounds were able to bind by making favorable H-bonds with the urea groups in t-AUCB as does sorafenib, without any unfavorable interactions with other residues at the active site (data not shown), suggesting that sEH inhibitors could inhibit B-Raf and other kinases. However, the sEH inhibitors studied here only show the urea interaction observed between sorafenib and B-Raf (3) and not the additional favorable interactions between sorafenib and B-Raf used to explain sorafenib's tight binding with the target kinase. Thus there appears to be different structure-activity relationships between the sEH inhibitors and kinase inhibitors with sorafenib being somewhat unique in being a potent inhibitor of both enzymes. The structures of many sEH inhibitors are close enough to sorafenib that a cautious approach with inhibitor design suggests screening possible sEH inhibitors against a library of kinases.
In addition to the sEH inhibitory activity described above, sorafenib has well-established inhibitory activity against several kinases including Raf, MAPK/ERK, and VEGF-R, which by conventional wisdom accounts for its potent anti-proliferative and anti-angiogenic effects in vivo (36-38). Thus, we sought to determine whether sEH inhibitors acted on several oncogenically-relevent kinases. While both sorafenib and the MEK inhibitor PD98059 (as control) attenuate ERK phosphorylation as expected in two RCC cell lines, five sEHIs selected because of their varied IC50's (Table 1) do not decrease ERK phosphorylation at a similar concentration (Fig. 4A, 4B). In addition, while sorafenib attenuates phospho-VEGF and causes apoptosis as is evidenced by PARP cleavage, there was no effect by three sEHIs with widely variable structures, KI's and IC50's on these properties (Fig. 4C).
Sorafenib is known to decrease cell growth and tumor vascularization and induce apoptosis; all of these are presumed mechanisms of sorafenib's therapeutic effect in kidney cancer (39). We next asked whether the sEH inhibitory activity of sorafenib accounts for its apoptosis or growth inhibitory effects in RCC cells. We utilized the MTT assay to assess cell growth and an assay of caspase-3 activite to measure apoptosis. Both RCC cell lines were incubated with sorafenib or five sEH inhibitors for 48 h. While sorafenib markedly decreased cell growth (by 65-70% as compared to serum-stimulated cells), the effect of the sEH inhibitors on cell growth was quite variable and considerably less pronounced (Fig. 5a). Furthermore, cell growth was reduced more with the weaker sEHIs suggesting that the sEH inhibitory activity does not correspond to RCC cell viability (r2 < 0.10 between cell viability and inhibition potency). Sorafenib incubation also resulted in apoptosis as evidenced by activation of caspase-1 and caspase-3 activity, as expected, while there was no consistent such effect with the sEH inhibitors (Fig. 5b).
If the intrinsic sEH inhibitory activity of sorafenib contributes to its pro-apoptotic activity, it would be expected that sEH inhibitors would synergize with sorafenib with respect to this property. To examine this possibility, both RCC cell lines were incubated with sorafenib alone, sunitinib alone (which has anti-VEGF and anti-raf activity but not sEH inhibitory activity; see Table 1), or with one of two sEHI's having disparate IC50's; these cells were subjected to an MTT assay (Fig. 6A). The experiment was repeated with the addition of exogenous EETs, the substrate for sEH (Fig. 6B). Consistent with the data described above, both sEHIs and both raf/VEGF inhibitors had apoptosis-inducing activity (the latter of a higher magnitude than the former), but there was no synergistic effect when the raf/VEGF inhibitors and sEH inhibitors were added together to the cells both in the presence and absence of exogenous EETs. Thus, it is unlikely that the sEHI activity inherent in sorafenib is contributing substantially to sorafenib-induced apoptosis.
Published data from our group have shown that sEH inhibitors possess marked anti-inflammatory activity in an LPS-challenged mouse model of acute inflammation (21). To determine whether sorafenib possesses similar anti-inflammatory activity which may be due to its sEHI property, we first examined the pharmacokinetics of sorafenib tosylate in a murine model by oral administration at 5 mg/kg. Blood levels of sorafenib reached the maximum concentration (Cmax) of ~120 nM 2 h after administration, and then slowly decrease such that no sorafenib could be detected after 24 h (Fig. 7). In this system, 1-compartmental analysis was demonstrated to be the best fit model. We calculated an AUC 0-24 of 630 nM h, and an elimination half life of 2.6 hours for sorafenib. The sorafenib pharmacokinetic profile is compared with that of t-AUCB used at a roughly 5x lower concentrations due to its high potency (Fig. 7). The sEHI t-AUCB afforded an larger AUC0-24 of 2580 nM h and a longer elimination half life of more than 24 hours at an oral dose of 1 mg/kg (30). The compounds showed AUC0-24/IC50 (an estimate of exposure and potency) of 21 and 645 for sorafenib and t-AUCB, respectively. Sorafenib and t-AUCB had blood levels above their murine IC50s for 8 and 24 hours respectively. Compared with the results from murine model, sorafenib in man provides a higher Cmax (more than 1500 nM), later Tmax (more than 2 h), longer elimination half life (more than 20 hours), larger AUC0-12 (more than 13000 nM h) and larger AUC0-12/IC50 (more than 1080) in its phase I study even at the similar doses (40, 41).
To confirm that sorafenib inhibits sEH in vivo, the EETs and DHETs were measured by LC-MS/MS in plasma samples taken 24 h after treatment with LPS and then sorafenib. The termination time (24 h) was established by a time-dependent study of sorafenib reversing the LPS-challenged hypotension. The plasma ratio of EETs to DHETs characterizes the sEH inhibition. Upon LPS-challenge, the production of EETs dramatically decreased as compared to non-treated animals, resulting in a significant decrease in the plasma ratio of EETs to DHETs (Fig. 8A-B). Sorafenib inhibited the production of DHETs, reversing the ratio of EETs to DHETs to normal level. Administration of sorafenib alone inhibited the production of DHETs, resulting in a dramatic increase in the plasma ratio of EETs to DHETs. As a control, administration of the sEH inhibitor t-AUCB increased the production of EETs resulting in an increase in the ratio of EETs to DHETs as expected. Sunitinib, which does not possess sEH inhibitory activity (see Table 1), inhibits the production of both EETs and DHETs resulting in a decrease in the ratio of EETs to DHETs. Thus, administration of sorafenib, but not sunitinib, to mice has the oxylipid signature of sEH inhibition, consistent with the in vitro data described earlier.
As discussed above, sEH inhibition has been shown to be anti-inflammatory in an LPS-challenged murine hypotension model (21), yet it would not be expected that kinase inhibition (as from sorafenib) would have a similar global effect (42). To determine whether sorafenib has a salutary effect on LPS-induced hypotension, we treated LPS-challenged mice with sorafenib. While the sEH inhibitor t-AUCB at 2.5 μmol/kg attenuates the hypotensive response of LPS to some degree, LPS-induced hypotension was completely reversed with sorafenib treatment (Fig. 8C), suggesting that sEH inhibition accounts for at least part of the anti-inflammatory effect of sorafenib. Sunitinib attenuated the LPS-induced hypotensive effect to a smaller degree. However, the reduction in hypotension and the alteration in oxylipin profiles by sunitinib are much greater than that can be accounted for by its negligible sEH inhibition, suggesting more complex mechanisms of action.
The treatment of kidney cancer has been revolutionized by the appearance and surprising efficacy of the multi-kinase inhibitors sorafenib and sunitinib in heretofore chemotherapy-resistant malignancies (1). While these drugs target many tyrosine kinases, given the high degree of angiogenesis accompanying kidney cancer, the VEGF-receptor tyrosine kinase is likely the most important target of these agents in this disease. This physiology is dictated by the constitutive activation of the hypoxia-inducible factor (HIF) pathway seen frequently in kidney cancer (43). However, while the VEGF-R inhibitors possess important salutary effects on the malignancies targeted, there exist serious adverse effects associated with these treatments, most likely due to the fact that VEGF is necessary for vascular homeostasis (44). Thus, it is important to investigate other properties of the VEGF-R inhibitors and thereby tease out mechanistic differences which can be exploited in future refinement of this therapeutic paradigm. In this study, we have used modeling of the crystal structures to identify an unexpected property of sorafenib which begs further investigation.
Docking studies with both murine and human sEH indicated that sorafenib but not several other kinase inhibitors should be a potent inhibitor of the enzyme. Docking of several sEH inhibitors in the B-Raf kinase catalytic site failed to show adverse interactions, cautioning that sEH inhibitors should be screened on kinase libraries. However, the sEH inhibitors docked failed to show the favorable interactions thought to result in sorafenib's potent inhibition of some kinases (3). As predicted from docking studies, we demonstrated that sorafenib, but not sunitinib, causes inhibition of the sEH with the potency similar to the best available sEH inhibitors, while these sEH inhibitors failed to inhibit the kinases studied. This sEH inhibition, which is likely unique among the VEGF-R and raf inhibitors, increases the serum and/or tissue levels of EET and affects other lipid mediators which are known to have a variety of clinical effects, including the reduction of inflammation and hypertension. While all of the ramifications of this finding have not yet been worked out, it is clear from previous work in our and other laboratories that sEH inhibition leads to varied consequences. In the case of treatment of cancers, many of which are now considered inflammatory diseases (45, 46), sEH inhibition may be of benefit due to its anti-inflammatory properties (21).
There also is a caution that further attenuation of inflammation by sEH inhibition, in the presence of standard immunosuppressive compounds sometimes used in chemotherapy, may make a patient more prone to overwhelming infection. However, even complete sEH inhibition only increases EET levels several fold. Given the available data on sEH inhibition and blood pressure (18, 47, 48), it is possible that the sEH inhibitory property of sorafenib may ameliorate hypertension sometimes observed with kinase inhibitor therapy. In addition, the beneficial effects of sEH inhibition in atherosclerosis (20) may lead to beneficial end-organ protection during sorafenib therapy, and the antinociceptive effects of sEH inhibition (19) could add to the favorable effect profile of sorafenib in cancer patients.
Our finding that sorafenib fits into the active site of sEH (Fig. 1), based on the X-ray crystal structures previously published (3, 29) was serendipitous. Despite targeting inhibition of VEGF-R, the other kinase inhibitors dasitinib and sunitinib do not display sEH inhibitory activity either in vitro or in vivo, as is evidenced by their decreased levels of blood pressure and cytokine modulation as compared to that of sorafenib. Along these lines, it is indeed provocative to consider that it is the sEH inhibitory property of sorafenib that accounts for its differences from the other VEGF-R inhibitors with respect to inflammation (42) and hypertension (49). The sEH inhibitors are known to reduce gastrointestinal inflammation from our unpublished studies, thus the reduction in gastrointestinal and possibly cardiac side effects observed with sorafenib over sunitinib may result in part from sEH inhibition (50). The sEH inhibitors have been shown to dramatically reduce renal inflammation and failure for example in mice treated with the cancer drug cisplatin (51) and rats with high angiotensin levels (47). The sEH inhibitors also are known to synergize and be synergized by low dose of indomethacin and COXIBs through the cyclooxygenase pathways of the arachidonic acid cascade (19). Thus, one may be able to reduce side effects of sorafenib further with such drug combinations. However, a caution is that the stabilization of EETs and other endogenous chemical mediators by sEH inhibitors could also lead to unwanted side effects from sorafenib and other kinase inhibitors. For example the EETs stabilized by sEH inhibitors are reported to be mildly angiogenic (52, 53), and inflammation is known to be beneficial in the homeostatic response to sepsis.
While our work shows reasons why sorafenib has a different effect profile than the other VEGF-R inhibitors commercially available, we could find few studies directly comparing these compounds in cancer (reviewed in (22)), although a clinical trial to assess this is in progress. In the one available study, it was shown that both kinase inhibitors have similar side effects, although sunitinib has a higher frequency of gastrointestinal symptoms (23). However, consistent with our inflammation data, it has been shown that sorafenib, but not sunitinib, inhibits the secretion of cytokines and expression of CD1a from dendritic cells (42). Thus, it is possible that, in addition to are findings concerning TNFα, there may exist more strongly differential heretofore unidentified cytokine effects with these two drugs in RCC which result from sEH inhibition.
Following oral administration to mice, both the sEH inhibitor t-AUCB and sorafenib resulted in blood levels of active compound that are consistent with near total inhibition of the sEH. Because the Ki of sorafenib is 10 times higher than that of t-AUCB, it was administered at a higher dose. As expected from their inhibitory potency both sorafenib and t-AUCB increased the ratio of EETs to DHETs to above those seen in normal control animals even when the mice were treated with LPS. The data suggest that the very high therapeutic doses of sorafenib used for renal cancer are likely to be far excess of what is needed for inhibitor of the sEH. A novel sEH inhibitor is now in phase II clinical trials. The high potency of sorafenib as a sEH inhibitor indicates that it possibly could be used for this therapeutic target at doses lower than those needed to control renal cancer.
In summary, we report a novel property of the VEGF-R and Raf inhibitor sorafenib which likely accounts for at least a substantial portion of its benefit and adverse effect profile. Further investigation, including in vivo studies, may lead to novel indications for this drug as explaining apparent benefits over other pharmaceuticals targeted kinases.
This work was supported in part by NIEHS Grant R37 ES02710, NIEHS SBRP Grant P42 ES04699 and NIH HL85727 (J. Y. L., C.M., B.D.H.), and Grant 5UO1CA86402 (Early Detection Research Network, National Cancer Institute, U.S. National Institutes of Health) and grant D06CA-065 from the Morris Animal Foundation (S.-H. P., R.H.W.).