Discovery of ML358, a Selective Small Molecule ... - ACS Publications

May 6, 2015 - (9-16) Xenobiotic detoxification is a common mechanism of drug ... inhibitors of SKN-1 dependent detoxification genes in nematodes, we ...
0 downloads 0 Views 2MB Size
Articles pubs.acs.org/acschemicalbiology

Discovery of ML358, a Selective Small Molecule Inhibitor of the SKN‑1 Pathway Involved in Drug Detoxification and Resistance in Nematodes Satyamaheshwar Peddibhotla,† Pauline Fontaine,⊥ Chi K. Leung,⊥ Patrick Maloney,† Paul M. Hershberger,† Ying Wang,⊥ Michelle S. Bousquet,§,∥ Hendrik Luesch,§,∥ Arianna Mangravita-Novo,† Anthony B. Pinkerton,‡ Layton H. Smith,† Siobhan Malany,*,† and Keith Choe*,⊥ †

Conrad Prebys Center for Chemical Genomics at Sanford-Burnham Medical Research Institute, Orlando, Florida 32827, United States Conrad Prebys Center for Chemical Genomics at Sanford-Burnham Medical Research Institute, La Jolla, California 92037, United States § Department of Medicinal Chemistry, University of Florida, Gainesville, Florida 32610, United States ∥ Center for Natural Products, Drug Discovery and Development, University of Florida, Gainesville, Florida 32610, United States ⊥ Department of Biology and Genetics Institute, University of Florida, Gainesville, Florida 32611, United States

Downloaded by UNIV OF SHEFFIELD on September 14, 2015 | http://pubs.acs.org Publication Date (Web): May 22, 2015 | doi: 10.1021/acschembio.5b00304



S Supporting Information *

ABSTRACT: Nematodes parasitize ∼1/3 of humans worldwide, and effective treatment via administration of anthelmintics is threatened by growing resistance to current therapies. The nematode transcription factor SKN-1 is essential for development of embryos and upregulates the expression of genes that result in modification, conjugation, and export of xenobiotics, which can promote resistance. Distinct differences in regulation and DNA binding relative to mammalian Nrf2 make SKN-1 a promising and selective target for the development of anthelmintics with a novel mode of action that targets stress resistance and drug detoxification. We report 17 (ML358), a first in class small molecule inhibitor of the SKN-1 pathway. Compound 17 resulted from a vanillamine-derived hit identified by high throughput screening that was advanced through analog synthesis and structure−activity studies. Compound 17 is a potent (IC50 = 0.24 μM, Emax = 100%) and selective inhibitor of the SKN-1 pathway and sensitizes the model nematode C. elegans to oxidants and anthelmintics. Compound 17 is inactive against Nrf2, the homologous mammalian detoxification pathway, and is not toxic to C. elegans (LC50 > 64 μM) and Fa2N-4 immortalized human hepatocytes (LC50 > 5.0 μM). In addition, 17 exhibits good solubility, permeability, and chemical and metabolic stability in human and mouse liver microsomes. Therefore, 17 is a valuable probe to study regulation and function of SKN-1 in vivo. By selective targeting of the SKN-1 pathway, 17 could potentially lead to drug candidates that may be used as adjuvants to increase the efficacy and useful life of current anthelmintics.

N

nematodes would provide much needed tools for studying multidrug resistance and may greatly increase the utility of current and future anthelmintics with minimal side effects in humans and other mammals. The transcription factor SKN-1 regulates the expression of ∼65 genes predicted to promote drug modification, conjugation, or export in Caenorhabditis elegans including some associated with anthelmintic resistance and confers resistance to an array of reactive small molecules including reactive oxygen species used in the mammalian immune response.10,23−26 SKN-1 is also essential for development of C. elegans embryos.27 Orthologs of SKN-1 are present in most nematode clades with

ematode parasites cause a substantial amount of human mortality and morbidity.1,2 They also cause an estimated $80 billion loss of plant-based food production annually and are a major burden to animal husbandry.3 Helminth targeting drugs, or anthelmintics, have been used widely for decades to control parasitic nematodes,4,5 and many species are evolving multidrug resistance, a widespread and growing problem6−9 further complicated by incomplete understanding at the molecular and genetic level.9−16 Xenobiotic detoxification is a common mechanism of drug resistance,17−22 and generic inhibitors of glutathione synthesis and drug transport such as buthionine, sulfoxamine, and verapamil are available but have two important limitations: they only affect a single enzyme or class of enzymes, and they also inhibit homologous proteins in mammalian hosts.10 Pharmacological compounds that target multiple detoxification and transport mechanisms unique to © 2015 American Chemical Society

Received: October 22, 2014 Accepted: May 6, 2015 Published: May 6, 2015 1871

DOI: 10.1021/acschembio.5b00304 ACS Chem. Biol. 2015, 10, 1871−1879

Articles

ACS Chemical Biology highest homology within clade V, which includes C. elegans and many important intestinal parasites such as hookworm and the common ruminant parasite Hemonchus contortus.6,10,28 We have identified and characterized a principal SKN-1 regulatory protein named WDR-23 that is distinct from KEAP1, the principle regulatory protein for the homologous mammalian transcription factor Nrf2.29 SKN-1 also binds to target DNA by a unique monomeric mechanism relative to all other known basic leucine zipper factors.10,30 Therefore, SKN-1 is a promising target for the development of drugs that selectively decrease stress resistance and inhibit xenobiotic detoxification in nematodes without affecting homologous pathways in humans and agricultural animals. However, no small molecule inhibitors of the SKN-1 pathway have been reported to date.

Figure 1. Compound 17. Compound 17 is biologically active and noncytotoxic in C. elegans. Concentration-dependent percent inhibition of 17 on Pgst-4::GFP/Pdop-3::RFP ratios after acrylamide induction (black circles, n = 6), Phsp-16.2::GFP/Pdop-3::RFP ratios after heat shock (open circles, n = 3), and worm health (red circles, n = 6).

Downloaded by UNIV OF SHEFFIELD on September 14, 2015 | http://pubs.acs.org Publication Date (Web): May 22, 2015 | doi: 10.1021/acschembio.5b00304



RESULTS AND DISCUSSION HTS, Hit-to-Lead and Medicinal Chemistry Efforts Identified Compound 17. In order to identify small molecule inhibitors of SKN-1 dependent detoxification genes in nematodes, we recently carried out a whole-organism based HTS of a 363 827 compound collection from the NIH Molecular Libraries Small Molecule Repository (MLSMR) in a 1536-well format.31 Toward this aim, we used a C. elegans strain expressing a green fluorescent protein (GFP) reporter driven by the promoter of gst-4 (Pgst-4::GFP), which encodes a Glutathione S-transferase strongly activated by SKN-1 under a variety of conditions.10,25,26,29,31−33 As an internal control, a constitutive red fluorescent protein reporter (Pdop-3::RFP) was coexpressed and used to normalize for variation in number of worms per well.32 The HTS screen yielded a hit set comprised of 13 distinct chemical scaffolds which were prioritized based on the following criteria: (1) inhibition of the SKN-1 target pathway measured by following SKN-1 driven gene expression (Pgst-4::GFP/Pdop-3::RFP) induced by acrylamide, (2) selectivity for inhibition of SKN-1-dependent genes assessed in a heat-shock protein assay (Phsp-16.2::GFP) in which induction by heat is independent of SKN-1,34 and (3) overall adult worm health was assessed by imaging analysis, and compounds were manually scored for toxicity based on three phenotypes (lethality, paralysis, and sterility) as described in the Methods. Out of 23 confirmed hits, compound 3, a vanillamine derivative was prioritized as the best hit (IC50 of 4.5 μM; Emax of 100%) in the primary Pgst-4::GFP assay; it did not inhibit the heat-shock pathway (IC50 > 64 μM) and was nontoxic to live adult worms (LC50 > 64 μM). Subsequent medicinal chemistry optimization and structure−activity relationship (SAR) studies resulted in the discovery of 17. Compound 17 is the first small molecule inhibitor (IC50 = 0.24 μM; Emax of 100%) of SKN-1-dependent detoxification genes in C. elegans. The compound was inactive in the heat shock assay and had no observable effects on adult worm health at concentrations effective against Pgst-4::GFP (Figure 1). Here, we report the design and synthesis of SAR analogs that led to the discovery of 17, as well as the compound’s in vitro and in vivo pharmacological profiling. Analogs were synthesized from substituted benzaldehyde derivatives in two steps (Scheme 1). Alkylation of aldehydes 1a−1c (see Supporting Information for details) under basic conditions with 4-fluorobenzyl bromide or methyl iodide afforded intermediates 2a and 2c−2e, respectively, which on reductive amination with cyclopentylamine afforded compounds 3 and 5−7. Amide analog 4 was obtained by alkaline KMNO4 oxidation of 2a to the acid 2b followed by HATU mediated coupling with cyclopentylamine (Scheme 1A).

Reductive amination of intermediate 2a with diverse amines resulted in analogs 8−13. Benzylation of 1a with 2-, 3-, and 4-substituted benzyl halides followed by reductive amination with either cyclopentyl or cyclopropylamine yielded compounds 15−27 (Scheme 1B). Similarly, analogs 28−32 were generated from commercially available aldehydes via alkylation with 4-chlorobenzyl bromide followed by reductive amination of intermediates 2o−2s with cyclopentylamine (Scheme 1C). Aldehyde intermediate 2h was reacted with various primary amines to afford analogs 33−37 and with secondary cyclic amines to yield analogs 40−43 (Scheme 1D). Reductive amination of cyclopentanone with 33 and N-acetylation of 17 afforded compounds 38 and 39, respectively (Scheme 1D). Initial hit 3 was resynthesized and confirmed with improved activity (IC50 = 1.0 μM, Table 1). Our SAR strategy focused on three distinct regions of 3, the benzyl ether, the central aromatic ring, and the cyclopentyl amine to expand the series. The corresponding amide analog was inactive, highlighting the importance of a basic amine toward inhibition of the SKN-1 pathway (Table 1, entry 4, X = −CO). Replacement of the benzyl ether with a methyl ether resulted in an inactive analog (Table 1, entry 5) showing the importance of an aryl group in this portion of the molecule. Lack of activity of the methyl ether suggests that the activity of this scaffold may not be not derived from any potential formation of orthoquinone or quinonemethide metabolites. Replacement of the OEt (R3) group in the central aryl ring with OMe did not affect activity (Table 1, entries 6). In contrast, removal of the Cl (R2) substituent from the initial hit led to a 10-fold loss in activity (Table 1, entry 7). Next, we explored varying the size and nature of the amine substituent (cycloalkyl vs aryl). The larger cyclohexyl group led to a ∼10-fold decrease in activity compared to the cyclopentyl group in the initial hit (Table 1, entries 3, 8). Among the smaller ring replacements, the cyclopropyl group resulted in an analog with closely comparable activity to the initial hit (Table 1, entries 9, 10). Aryl or branched alkyl replacements were not favored and resulted in analogs that were >15-fold less active or completely inactive (Table 1, entries 11−13). To study the effect of aromatic substituents on the benzyl ether portion of the scaffold, a host of cyclopentyl and cyclopropylamine derived analogs were generated (Table 2). Electron withdrawing substituents at the 4-postion of the aryl ring led to the most active compounds in both series. Removal of the 4-F substituent led to a 10-fold loss in activity (Table 2, entries 3, 14). Replacement with electron rich 4-Me or 4-OMe groups led to a >25-fold less active and an inactive analog, 1872

DOI: 10.1021/acschembio.5b00304 ACS Chem. Biol. 2015, 10, 1871−1879

Articles

ACS Chemical Biology

Downloaded by UNIV OF SHEFFIELD on September 14, 2015 | http://pubs.acs.org Publication Date (Web): May 22, 2015 | doi: 10.1021/acschembio.5b00304

Scheme 1. Synthesis of SKN1 Pathway Inhibitor 17 and Analogsa

Conditions: (i) K2CO3, CH3CN, 80 °C, 15 h, 10−88% (ii) a. Sodium triacetoxyborohydride, THF-acetic acid (4:1), 23 °C, 24 h. b. 4 M HCl (dioxane), dioxane, 23 °C, 10−88%. (iii) a. KMnO4, 2N NaOH, 100 °C, 2 h, 38%. b. HATU, diisopropylethylamine, dichloromethane, 23 °C, 15 h, 50%. (iv) Cyclopentanone, compound 33 (free base precursor), Na2SO4, 1,4-dioxane, 23 °C, 1.3 h, sodium triacetoxyborohydride, 17 h, 84%. (v) Acetyl chloride, triethylamine, dichloromethane, 23 °C, 5 h, 80%. See Supporting Information for other details.

a

Table 1. SAR of Vanillamine derivatives as SKN1 pathway inhibitors, compounds 3−13a

compound

R1

R2

R3

X

R4

IC50 (μM)

Emax (%)

3 4 5 6 7 8 9 10 11 12 13

4-fluorophenyl 4-fluorobenzyl Me 4-fluorophenyl 4-fluorobenzyl 4-fluorophenyl 4-fluorophenyl 4-fluorophenyl 4-fluorophenyl 4-fluorophenyl 4-fluorophenyl

Cl Cl Cl Cl H Cl Cl Cl Cl Cl Cl

OEt OEt OEt OMe OEt OEt OEt OEt OEt OEt OEt

CH2 CO CH2 CH2 CH2 CH2 CH2 CH2 CH2 CH2 CH2

cyclopentyl cyclopentyl cyclopentyl cyclopentyl cyclopentyl cyclohexyl cyclobutyl cyclopropyl isopropyl phenyl benzyl

1.0 ± 0.3 >64 >64 0.83 ± 0.27 10.3 ± 3.4 9.7 ± 3.2 8.7 ± 2.9 2.0 ± 0.7 16 ± 5 >64 17 ± 6

100 NA NA 100 100 100 100 80 100 NA 100

a

Target activity (IC50 and Emax) for inhibition of the SKN-1 pathway was measured by following SKN-1 driven gene expression (Pgst-4::GFP/ Pdop-3::RFP) induced by acrylamide. All compounds exhibited an IC50 > 64 μM against heat shock activated Phsp-16.2::GFP and no toxicity (IC50 > 64 μM) in C. elegans.

central aromatic ring of the molecule (Table 3). Having OMe or OEt at the R2 position resulted in equipotent molecules (Table 3, entries 17, 28). However, replacement of OEt with a Me or electron withdrawing Cl group led to a ∼4−5 fold loss in activity against the SKN-1 pathway (Table 3, entries 29, 30). Replacing the Cl at the R1 position with either a smaller F or a larger Br led to a ∼3−5 fold loss in activity (Table 3, entries 31, 32). Further exploration of the amine substituents (R3 and R4; Table 3) showed that secondary aliphatic amines derived from methyl, propyl, allyl, or propargyl amines were inactive or weakly active compared to the alicyclic cyclopentyl and cyclopropyl analogs (Table 3, entries 17, 24, and 33−36). Replacement of the cyclopentyl group with cyclopropyl was tolerated

respectively (Table 2, entries 15, 16). Among the electron withdrawing substituents, 4-Cl was the most active analog identified (Table 2 entry 17) with an IC50 = 0.24 μM, a 4-fold improvement in potency over the initial hit (Table 2, entries 3, 14−23). For the cyclopropylamine derivatives, the 4-Br substituent yielded the highest activity (IC50 = 0.24 μM, Table 2, entries 10, 24−27). Interestingly, at the 2-position of the aryl ring, the smaller F atom was favored over Cl (Table 2, entries 14, 20, and 21). The analogs with both 2-F and 4-Cl substituents did not offer any further improvement in potency or efficacy (Table 2, entries 17, 23, and 27). With the more active compound 17 in hand, we next examined the effect of replacing the OEt and Cl groups in the 1873

DOI: 10.1021/acschembio.5b00304 ACS Chem. Biol. 2015, 10, 1871−1879

Articles

ACS Chemical Biology

derivative was inactive, confirming that a basic amine is critical for activity (Table 3, entries 38, 39). Finally, a small set of cyclic tertiary amines were generated (Table 3, entries 40−43) among which the piperidine derivative was the most active (IC50 = 0.5 μM, Table 3, entry 41). The result with tertiary amines (Table 3, entries 38 and 41) is significant, because it allows an opportunity to avoid the potential of biological reactivity (nucleophilicity) of 17 while retaining the basic nature of the amine, which is critical for activity and will be investigated further in the future. Activity does not correlate directly with clogP values (entries 17, 18, and 19); furthermore, lack of activity also does not track with clogP values (entries 4, 5, 16, and 39). A complete list of clogP values is provided in the Supporting Information. Based on the results from the above SAR studies, 17 (ML358) was the best analog in terms of potency and efficacy in the series and met the probe criteria established in the MLPCN program (Molecular Libraries Probe Production Centers Network).35 In Vitro Profiling of Compound 17 Confirms SKN-1 Pathway Activity and Inactivity against the Human SKN-1 Homologue. To characterize its activity and assess its potential for broader utility, 17 was profiled further in a set of assays in nematodes and human cell lines. Shown in Figure 2A on the left axis, concentration−response curves against the same target strain (Pgst-4::GFP) but induced by 25 μM juglone instead of acrylamide confirmed inhibition of SKN-1 with IC50 = 0.65 μM. Compound 17 also inhibited a different SKN-1-dependent promoter (Pgst-30::GFP) with acrylamide with IC50 = 0.13 μM. In previous studies, we found that some paralytic compounds can inhibit activation of Pgst-4::GFP by acrylamide33 presumably indirectly by decreasing consumption and hence exposure to the inducers such as acrylamide and juglone. However, 17 had only a 6.6% effect (10 μM, P = 0.029, n = 34) on the rate of pumping by the pharynx (feeding organ) in C. elegans, indicating that 17

Downloaded by UNIV OF SHEFFIELD on September 14, 2015 | http://pubs.acs.org Publication Date (Web): May 22, 2015 | doi: 10.1021/acschembio.5b00304

Table 2. SAR of Vanillamine Derivatives as SKN1 Pathway Inhibitors, Compounds 14−27a

compound

R1

R2

IC50 (μM)

Emax (%)

3 14 15 16 17 18 19 20 21 22 23 10 24 25 26 27

4-F H 4-Me 4-OMe 4-Cl 4-Br 4-CN 2-Cl 2-F 3-F 2-F; 4-Cl 4-F 4-Cl 4-Br 4-CN 2-F; 4-Cl

cyclopentyl cyclopentyl cyclopentyl cyclopentyl cyclopentyl cyclopentyl cyclopentyl cyclopentyl cyclopentyl cyclopentyl cyclopentyl cyclopropyl cyclopropyl cyclopropyl cyclopropyl cyclopropyl

1.0 ± 0.3 9.9 ± 3.3 26 ± 9 >64 0.24 ± 0.08 0.4 ± 0.1 0.4 ± 0.1 12 ± 4 0.8 ± 0.3 7.2 ± 2.4 0.9 ± 0.3 2.0 ± 0.1 0.51 ± 0.03 0.23 ± 0.05 1.6 ± 0.2 0.52 ± 0.06

100 114 100 NA 100 88 80 100 100 100 90 100 92 93 88 94

a

Target activity (IC50 and Emax) for inhibition of the SKN-1 pathway was measured by following SKN-1 driven gene expression (Pgst-4::GFP/ Pdop-3::RFP) induced by acrylamide. All compounds exhibited an IC50 > 64 μM against heat shock activated Phsp-16.2::GFP and no toxicity (IC50 > 64 μM) in C. elegans.

well; however, a more angular cyclopropylmethyl group led to a ∼27 fold loss in activity (Table 3, entries 24, 37). N-methylation was tolerated, but the corresponding tertiary amine was ∼2 fold less active than 17, whereas the N-acetyl

Table 3. SAR of Vanillamine Derivatives as SKN1 Pathway Inhibitors, Compounds 28−43a

compound

R1

R2

17 (ML358) 28 29 30 31 32 33 34 35 36 24 37 38 39 40 41 42 43

Cl Cl Cl Me F Br Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl

OEt OMe Cl Me OMe OMe OEt OEt OEt OEt OEt OEt OEt OEt OEt OEt OEt OEt

R3

R4

H cyclopentyl H cyclopentyl H cyclopentyl H cyclopentyl H cyclopentyl H cyclopentyl H methyl H propyl H allyl H propargyl H cyclopropyl H cyclopropylmethyl Me cyclopentyl −COCH3 cyclopentyl −CH2CH2CH2CH2− −CH2CH2CH2CH2CH2− −CH2CH2OCH2CH2− −CH2CH2N(Me)CH2CH2−

IC50 (μM)

Emax (%)

0.24 ± 0.08 0.3 ± 0.1 0.9 ± 0.3 1.1 ± 0.4 1.2 ± 0.4 0.7 ± 0.2 >64 8.7 ± 0.8 12.3 ± 1.1 8.1 ± 0.4 0.51 ± 0.03 6.7 ± 0.8 0.42 ± 0.05 >64 4.0 ± 0.5 0.50 ± 0.04 1.7 ± 0.1 3.4 ± 0.3

100 88 90 81 84 83 NA 100 100 72 93 108 86 NA 70 70 76 70

a

Target activity (IC50 and Emax) for inhibition of the SKN-1 pathway was measured by following SKN-1 driven gene expression (Pgst-4::GFP/Pdop-3::RFP) induced by acrylamide. All compounds exhibited an IC50 > 64 μM against heat shock activated Phsp-16.2::GFP and no toxicity (IC50 > 64 μM) in C. elegans. 1874

DOI: 10.1021/acschembio.5b00304 ACS Chem. Biol. 2015, 10, 1871−1879

Articles

Downloaded by UNIV OF SHEFFIELD on September 14, 2015 | http://pubs.acs.org Publication Date (Web): May 22, 2015 | doi: 10.1021/acschembio.5b00304

ACS Chemical Biology

For therapeutic use, SKN-1 inhibitors would ideally be selective for SKN-1 over Nrf2, the mammalian homologue of SKN-1, to prevent unwanted activity in animal and human hosts. We used a Nrf2-dependent luciferase reporter in the human IMR-32 neuroblastoma cell line to measure Nrf2 activity.37,38 This reporter system was induced 44.2 ± 3.85- and 16.5 ± 2.80-fold by sulforaphane or tert-butylhydroquinone (TBHQ), respectively, in the absence of 17 (Figure 2B). In the presence of 17, no significant concentration-dependent inhibition of luciferase reporter activity was observed with activation by either inducer. Compound 17 had significant activity with sulforaphane (1 μM, P < 0.05), but no clear dose− response was observed (R2 < 0.140). In addition, we assessed cytotoxicity of 17 in IMR-32 and LNCaP (human prostate adenocarcinoma) cells and observed >30-fold index in MTT cytotoxicity (CellTiter96, Promega) versus potency against Pgst-4::GFP inhibition in nematodes (Table 4). Gene Expression Profiling Confirms Inhibition of Core SKN-1 Target Genes. We next used quantitative PCR assays for six skn-1-dependent detoxification genes to verify inhibition of endogenous mRNA expression levels.10 Four of these were induced over 2-fold by 7 mM acrylamide (Figure 3A), and compound 17 inhibited all of them including gst-4 (Figure 3A, P < 0.01). These results confirm our GFP reporter data above and demonstrate that inhibition is broader than just gst-4. To investigate how broadly the compound affects skn-1dependent gene expression, we performed mRNA-seq on L4 larval stage skn-1(k1023) worms with and without a 6 h exposure to 10 μM of 17. Using a 2-fold cutoff and false-discovery rate of 5%, we found that skn-1(k1023) activated 1532 genes without 17 and 1201 genes with it, a difference of 331. Conversely, 17 had essentially no effect on the number of genes down-regulated by skn-1(k1023) (496 with and 498 without 17). These data are consistent with inhibition of a broad subset of genes activated by SKN-1 with little, or no, effect on gene inhibition by SKN-1. Compound 17 significantly inhibited a total of 399 genes in the skn-1(k1023) mutant at least 2-fold; 120 of these were also upregulated at least 2-fold by skn1(k1023) (Figure 3B), which is a highly significant 5.3-fold enrichment over random chance (P = 9.73 × 10−54). DAVID (Database for Annotation, Visualization and Integrated Discovery) analysis of these 120 genes finds high enrichment of the functional categories Glutathione S-transferase, lipid glycosylation, and oxidation/reduction (Figure 3C, all genes listed in Table S1). Consistent with the Phsp-16.2::GFP data

Figure 2. In vitro profiling of 17 in C. elegans and human cells. (A) Concentration-dependent percent inhibition of 17 on Pgst-4::GFP/ Pdop-3::RFP ratios after juglone induction (black circles, n = 3) and Pgst-30::GFP/Pdop-3::RFP ratios after acrylamide induction (open circles, n = 3) and percent response of 17 on Pgst-4::GFP/Pdop-3::RFP ratios in skn-1(k1023) gain-of-function worms (red circles, n = 3). (B) Compound 17 does not exhibit significant concentration-dependent inhibition in human IMR-32 neuroblastoma cells expressing Nrf2dependent antioxidant response element (ARE) luciferase activity induced by tBHQ (black circles) or sulforaphane (open circles). Compound 17 had significant activity with sulforaphane at 1 μM (P < 0.05), but no clear dose−response was observed (R2 < 0.140).

does not act via decreasing consumption of the inducers (data not shown). We recently reported identification of a skn-1 gain-offunction allele (skn-1(k1023)) that strongly and constitutively activates many detoxification genes including gst-4.36 Importantly, 17 decreased Pgst-4::GFP activity with IC50 = 0.53 μM when applied to skn-1(k1023), suggesting that it can act on the SKN-1 pathway to regulate Pgst-4::GFP expression independent of stress (Figure 2A, right axis).

Table 4. Summary of in Vitro ADME Properties of SKN-1 Inhibitor 17a aqueous solubility (μM) in pION’s buffer pH 5.0/6.2/7.4 aqueous solubility (μM) in 1× PBS, pH 7.4 a chemical stability (% remaining after 48 h) PBS pH 7.4 only a chemical stability (% remaining after 48 h) PBS pH 7.4/ACN (1:1) PAMPA permeability, Pe (×10−6 cm/s) donor pH: 5.0/6.2/7.4 acceptor pH: 7.4 plasma protein binding (% bound) human 1 μM/10 μM plasma protein binding (% bound) mouse 1 μM/10 μM plasma stability (% remaining at 3 h) human/mouse hepatic microsome stability (% remaining at 1 h) human/mouse cytotoxicity C. elegans LC50 (μM) cytotoxicity Fa2N-4 immortalized human hepatocytes LC50 (μM) cytotoxicity human prostate adenocarcinoma cells (LNCaP) (μM) cytotoxicity human neuroblastoma cells (IMR-32) (μM)

>494 />494/103 >54 24.96 100 561/1250/870 99.89/99.85 99.86/99.86 72.88/72.18 38.06/42.56 >64 >50 8.8 10

a

Detailed experimental procedure described in the Supporting Information. The apparent lack of stability in PBS is associated with precipitation over 48 h which is mitigated by the addition of ACN. 1875

DOI: 10.1021/acschembio.5b00304 ACS Chem. Biol. 2015, 10, 1871−1879

Articles

Downloaded by UNIV OF SHEFFIELD on September 14, 2015 | http://pubs.acs.org Publication Date (Web): May 22, 2015 | doi: 10.1021/acschembio.5b00304

ACS Chemical Biology

like ivermectin reduce or eliminate nematode parasite infections by targeting ion channels that promote feeding and attachment behaviors;42 worms treated chronically with ivermectin grow poorly and display arrested development. Levamisole is a nicotinic receptor agonist that causes nematode paralysis and development arrest.43 As shown in Figure 4C,D, compound 17 increased the incidence of larval development arrest when combined with ivermectin or levamisole in C. elegans. Conversely, 17 had no effect on sensitivity to albendazole, a representative of a class of anthelmintic that inhibits β-tubulin assembly and egg production44 (data not shown). In vitro pharmacology profiling reveals that compound 17 exhibits good solubility and permeability properties and moderate stability in liver microsomes. In vitro pharmacology screening (Table 4) showed that 17 achieved concentrations >400 × IC50 in aqueous buffer between a pH range of 5.0−7.4 with increasing solubility as pH is lowered as expected for a basic amine (−NH) containing compound. In PBS buffer, 17 attained a solubility >54 μM (>225 × IC50). Consistent with its solubility data, 17 exhibited very good permeability with increasing pH of the donor compartment. Compound 17 was highly plasma protein bound and showed good stability in both human plasma and mouse plasma. Additionally, 17 showed moderate stability and was ∼60% metabolized in human and mouse liver microsomes within 1 h and showed no toxicity (>50 μM) toward immortalized Fa2-N4 human hepatocytes.



CONCLUSIONS A high throughput screen of the MLSMR collection in C. elegans based on a SKN-1 driven GFP reporter assay identified the vanillamine derived primary hit 3, which was optimized via medicinal chemistry to 17. Compound 17 is a potent and selective inhibitor of the nematode SKN-1 detoxification pathway and is inactive against Nrf2, the human homologue of SKN-1. Compound 17 reduced transcription of xenobiotic detoxification and conjugating enzymes in a SKN-1 dependent manner without inhibiting any heat-shock proteins demonstrating pathway specificity. Compound 17 sensitized worms to oxidants that are released by the mammalian immune response, and therefore it has the potential to affect nematode parasite infections on its own. Compound 17 also sensitized worms to ivermectin and levamisole, two broad-spectrum anthelmitics that have become highly susceptible to resistance.45 Drug efflux inhibitors and glutathione synthesis inhibitors have been shown to enhance ivermectin sensitivity in C. elegans,13 but our results are the first to demonstrate increased sensitivity with a compound developed against a nematode-specific target that functions upstream of detoxification genes. The compound is nontoxic to C. elegans and thus may have low susceptibility to resistance. Compound 17 is nontoxic in Fa2N-4 immortalized human hepatocytes and has a >30-fold difference in cytotoxicity in LNCaP and IMR-32 cell lines compared to its potency in C. elegans. In addition, 17 has good solubility, permeability, and chemical and metabolic stability in human and mouse liver microsomes. Compound 17 is the first small molecule inhibitor developed against the SKN-1 pathway and may lead to better understanding of regulation and function of this core xenobiotic detoxification and longevity pathway in C. elegans and parasitic nematodes. By selective targeting of the SKN-1 pathway present in diverse nematode clades, 17 may lead to drug candidates that could be used as adjuvants to increase the efficacy of current anthelmintics. Efforts to optimize 17 further and improve its in vivo profile are underway.

Figure 3. Compound 17 inhibits skn-1-dependent genes. (A) Relative C. elegans mRNA levels of skn-1-dependent detoxification genes induced at least 2-fold by acrylamide (**P < 0.01, ***P < 0.001, n = 2−4 populations of worms). (B) Venn diagram of genes up-regulated in skn-1(k1023) and down-regulated by 17 at least 2-fold at a false discovery rate of 5% (genes listed in Table S1). (C) Functional categories significantly enriched (DAVID) in the 120 genes that are up-regulated in skn-1(k1023) and down-regulated by 17. (D,E) Relative C. elegans mRNA levels of skn-1-dependent detoxification genes in wildtype N2 (D) or skn-1(k2013) (E) worms (*P < 0.05, **P < 0.01, ***P < 0.001, and, n = 4 populations of worms).

(Figure 1), 17 did not inhibit expression of any heat shock proteins confirming specificity to its inhibitory activity. Similar to other bioactive small molecules,39,40 17 also induced many predicted cytochrome P450, lipid glycosylation, and glutathione s-transferase genes (Table S2), suggesting that it initiates a general drug response in addition to inhibiting a subset of skn-1 dependent genes and could have broad effects on drug metabolism, which will be characterized in future studies. We also used quantitative PCR assays to help confirm these RNA-seq data. As shown in Figure 3D, compound 17 at 10 μM reduced mRNA levels for three out of four gst genes known to be regulated by SKN-1 in wildtype N2 worms under basal conditions. All four gst mRNA levels were induced in skn1(k1023) worms and compound 17 at 10 μM reduced levels significantly (Figure 3E), consistent with the RNA-seq data (Table S1). Compound 17 Sensitizes C. elegans to Oxidants and Anthelmintics. Given that SKN-1 promotes oxidative stress resistance, we wanted to test if compound 17 could sensitize C. elegans to oxidants. Hydrogen peroxide is an endogenous oxidant that is toxic to C. elegans, and paraquat is a pesticide that generates superoxide within cells.41 As shown in Figure 4A,B, compound 17 reduced survival in the presence of either 5 mM hydrogen peroxide or 13 mM paraquat. Macrocyclic lactones 1876

DOI: 10.1021/acschembio.5b00304 ACS Chem. Biol. 2015, 10, 1871−1879

Articles

Downloaded by UNIV OF SHEFFIELD on September 14, 2015 | http://pubs.acs.org Publication Date (Web): May 22, 2015 | doi: 10.1021/acschembio.5b00304

ACS Chemical Biology

Figure 4. Compound 17 sensitized C. elegans to oxidants and anthelmintics. (A) Survival of wildtype N2 worms exposed to 5 mM hydrogen peroxide and compound 17 (n = 268−348 worms from four plates in two trials). (B) Lethality of N2 worms exposed to 13 mM paraquat and compound 17 (n = 4 wells of worms from two trials). (C) Percent larval arrest of N2 worms exposed to 3 ng/mL ivermectin and compound 17. (D) Percent larval arrest of N2 worms exposed to 500 μM levamisole and compound 17. (B−D) *P < 0.05, **P < 0.01, and ***P < 0.001 relative to treatments without oxidant or anthelmintic; †P < 0.05, ††P < 0.01, and †††P < 0.001 relative to treatments with only oxidant or anthelmintic. n = 4 plates of worms from two trials.



Burnham Biomedical Research Institute. The pipeline for processing was as follows: mapped with TopHat splice-aware aligner,47 transcript abundance estimated with the expectation-maximization approach,48 normalization as reads per kb per million (RPKM),49,50 and differential testing with a generalized linear model likelihood ratio test implemented in edgeR.51 Enrichment of functional gene categories was identified with the web-based DAVID tool (Database for Annotation, Visualization and Integrated Discovery v6.7).52 Nrf2 Transcriptional Activity Assays. Cell culture assays for Nrf2 transcriptional activity were completed as described previously.37,38 The luciferase reporter assay relies on activation of the Nrf2-responsive antioxidant response element (ARE) found upstream of many phase II detoxification enzymes. An ARE-luciferase reporter construct was transfected into LNCaP and IMR32 cells. Twenty-four hours after transfection, cells were treated with compound for 1.5 h before addition of 10 μM ARE-activators, sulforaphane (SF) or tert-butylhydroquinone (tBHQ). After an additional 24 h of treatment, the ability of compounds to inhibit SF/tBHQ-driven activation of the ARE-luc reporter was measured based on luminescence output. Cell Viability with ATP-lite. Hepatic toxicity of compounds was determined with Fa2N-4 immortalized human hepatocytes using the ATP-lite 1-step assay (PerkinElmer) assay according to the manufacturer’s instructions. Fa2N-4 cells (XenoTech, Kansas City, KS) were seeded at 50 000 cells/well and incubated with a range of concentrations of the test compound (0.01 μM-50 μM) in MFE support media for 24 h at 37 °C, 5% CO2. At the end of the experiment, cell viability was determined by cellular ATP levels using the ATP-lite kit according to the manufacturer’s instructions. Luminescence was measured on the Infinite M200 plate reader (Tecan US). The concentration of each compound that killed 50% of the cells (LC50) was calculated by nonlinear regression analysis using a log (inhibitor) versus response equation with a variable slope, using the statistic software package Prism6 (GraphPad, San Diego, CA). Cell Viability with MTT. The MTT (3-(4,5-dimethylthiazol-2-yl)2,5-diphenyltetrazolium bromide) assay is a cytotoxicity assay that relies on the cellular conversion of a tetrazolium salt to a formazan compound. The formation of the formazan product can be measured using a 96-well plate reader at an absorbance of 570 nm. Cell viability was measured according to the manufacturer’s instructions (Celltiter 96, Promega).

METHODS

C. elegans Strains. The following strains were used: wild-type N2 Bristol, VP596 dvIs19[pAF15(Pgst-4::GFP::NLS)];vsIs33[Pdop-3::RFP], QV65 gpIs1[Phsp-16.2::GFP];vsIs33[Pdop-3::RFP], QV87 kIs41[GST30::GFP]X;vsls33[Pdop-3::RFP], QV129 skn-1(k1023);dvIs19;vsIs33, QV122 skn-1(k1023);zjEx53[Pgst-4::tdTom], and QV212 skn-1(k1023). Unless noted otherwise, worms were cultured at 20 °C using standard methods.46 Gene Reporter Assays and Screening. The primary Pgst-4::GFP and secondary Pgst-30::GFP and Phsp-16.2::GFP assays were conducted as described previously.31,32 To assess overall worm health, three phenotypes (four categories) were manually scored with a dissecting microscope in each well: lethality, paralysis (penetrance and expressivity), and sterility. Worms were considered alive if they displayed any movement in response to tapping of plates or repeated prodding with an eyebrow hair and scores were recorded as estimates of percentage of worms dead. Penetrance of paralysis was scored as estimates of the percentage of the worm population that displayed slow or uncoordinated movement in response to tapping of plates. Expressivity was scored as the relative degree of paralysis in individual worms (0% = full movement similar to no compound controls and 100% = only slight movements when the plate is tapped). Sterility was estimated based on the number of offspring (eggs and L1 larvae) in each well compared to the control wells without inhibitors (0% = similar numbers of offspring as no compound controls and 100% = no offspring). All assays were run in triplicate by the same researcher. Scores of the four categories for each treatment were summed to achieve a composite score and compared to an average score for untreated wells. Gene Expression Analysis. Quantitative real-time RT-PCR was used to measure mRNA levels as described previously.29 For RNA-seq, total RNA was isolated from synchronized L4 stage C. elegans with an Ambion RNAqueous micro total kit (Life Technologies, Grand Island, NY) and submitted to the Yale Center for Genomic Analysis for library generation and sequencing. mRNA was then isolated and cDNA synthesized and end-labeled. Labeled cDNA libraries were sequenced (75 bp paired ends) with an Illumina HiSeq 2000 according to manufacturer’s recommendations (San Diego, CA). Libraries were sequenced to a depth of 13.5−26.0 million reads each. Raw sequence data were then sent to the Applied Bioinfomatics Core at the Sanford 1877

DOI: 10.1021/acschembio.5b00304 ACS Chem. Biol. 2015, 10, 1871−1879

Articles

Downloaded by UNIV OF SHEFFIELD on September 14, 2015 | http://pubs.acs.org Publication Date (Web): May 22, 2015 | doi: 10.1021/acschembio.5b00304

ACS Chemical Biology C. elegans Sensitivity Assays. For hydrogen peroxide, synchronized N2 L1s were grown on OP50 agar plates until they were at the L3 larval stage and then pretreated overnight (12 h) with probe 17 (10 or 30 μM). The next day, worms were washed in NGM buffer containing 1% LB-broth and loaded into a 96-well plate containing 5 mM hydrogen peroxide and probe 17 (10 or 30 μM). For paraquat, synchronized N2 L4 larvae were transferred to agar plates containing paraquat (13 mM) and probe 17 (3, 10, or 30 μM). Survival was assessed by movement and response to tapping of the plate. For anthelmintic sensitivity assays, synchronized N2 L1 larvae were transferred to agar media containing OP50 bacteria and ivermectin (3 ng/mL) or levamisole (500 μM) with probe 17 (0.5, 1, or 2 μM). Larval arrest was scored 3 (ivermectin) or 4 days later (levamisole). Dimethyl sulfoxide (DMSO) was used as the vehicle control in all bioassays containing reagents diluted in this solvent. Statistical Analyses. Statistical significance was determined with a one-way ANOVA followed by a Dunnett or Bonferroni Multiple Comparison post hoc test when three or more different treatments were compared or a two-way ANOVA with a Bonferroni Multiple Comparison posthoc test when treatments were compared across two factors. Chi-square tests were used to evaluate categorical data, and log rank tests were used to compare survival curves. P values of