Novel 1,3,4-Selenadiazole-Containing Kidney-Type Glutaminase

Dec 13, 2018 - To improve the in vivo activity, we explored a bioisostere replacement of the sulfur atom in bis-2-(5-phenylacetamido-1,2,4-thiadiazol)...
0 downloads 0 Views 1MB Size
Subscriber access provided by AUSTRALIAN NATIONAL UNIV

Article

Novel 1,3,4-Selenadiazole Containing Kidney-Type Glutaminase Inhibitors Showed Improved Cellular Uptake and Antitumor Activity Zhao Chen, li di, Ning Xu, Jinzhang Fang, Yan Yu, Wei Hou, Haoqiang Ruan, Panpan Zhu, Renchao Ma, Shiying Lu, Danhui Cao, Rui Wu, Mowei Ni, Wei Zhang, Weike Su, and Benfang Helen Ruan J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b01198 • Publication Date (Web): 13 Dec 2018 Downloaded from http://pubs.acs.org on December 13, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Novel 1,3,4-Selenadiazole Containing Kidney-Type Glutaminase Inhibitors Showed Improved Cellular Uptake and Antitumor Activity

Zhao Chen1a, Di Li 1a, Ning Xu1a, Jinzhang Fang1a, Yan Yu1, Wei Hou1, Haoqiang Ruan1, Panpan Zhu1, Renchao Ma1, Shiying Lu1, Danhui Cao1, Rui Wu1, Mowei Ni2, Wei Zhang3, Weike Su1, Benfang Helen Ruan1* aEqual

first author

1College

of Pharmaceutical Science, Collaborative Innovation` Center of Yangtza River Delta

Region Green Pharmaceuticals, IDD & CB, Zhejiang University of Technology, Hangzhou, China 2 Center

for Cancer Research, Zhejiang Cancer Hospital

3 Department

of Urology, Tongde Hospital of Zhejiang Province

*Corresponding author Benfang Helen Ruan (Ph. D) Professor of Pharmaceutical Science College of Pharmaceutical Science, Collaborative Innovation Center of Yangtza River Delta Region, IDD & CB, Green Pharmaceuticals, Zhejiang University of Technology E-mail: [email protected]; [email protected] Tel: 86-18357023608 Fax: (0086) 571-88871098 ¤Current address: 18 Chaowang Road, Xiachengqu, Hangzhou, Zhejiang, China, 310014 Running title: 1,3,4-Selenadiazole containing KGA allosteric inhibitors Key word: 1,3,4-selenadiazole, kidney type glutaminase inhibitors, KGA, gls1, EZMTT non-toxic cell viability assay, potency, efficacy, ROS, biomolecular interaction assay, xenograft model

1

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ABSTRACT: Kidney-type glutaminase (KGA/isoenzyme glutaminase C [GAC]) is becoming an important tumor metabolism target in cancer chemotherapy. Its allosteric inhibitor, CB839, showed early promise in cancer therapeutics but limited efficacy in in vivo cancer models. To improve the in vivo activity, we explored a bioisostere replacement of the sulfur atom in bis-2-(5-phenylacetamido-1,2,4-thiadiazol ethyl sulfide (BPTES) and CB-839 analogs with selenium using a novel synthesis of the selenadiazole moiety from carboxylic acids or nitriles. The resulting selenadiazole compounds showed enhanced KGA inhibition, more potent induction of reactive oxygen species (ROS), improved inhibition of cancer cells, and higher cellular and tumor accumulation than the corresponding sulfur-containing molecules. However, both CB839 and its selenium analogs show incomplete inhibition of the tested cancer cells, and a partial reduction in tumor size was observed in both the glutamine-dependent HCT116 and aggressive H22 liver cancer xenograft models. Despite this, tumor tissue damage and prolonged survival were observed in animals treated with the selenium analog of CB839. INTRODUCTION Targeting KGA (or GAC)1-3 is expected to selectively inhibit the growth of tumor cells growth, but not normal cells. This is because, in tumor cells, the Warburg effect increases glycolysis 200-fold compared to normal cells, but the majority of glucose carbon is converted to lactate for excretion and cannot be used in the mitochondrial tricarboxylic acid (TCA) cycle. To compensate for this, cancer cells develop glutamine dependence, and KGA is highly elevated4-5 to hydrolyze glutamine to glutamate and then to α-ketoglutarate for energy production throughout the TCA cycle. Glutaminase has two isoforms with very different structural features: liver-type glutaminase (LGA) and kidney-type glutaminase (KGA). The glutaminase activity of LGA is constant, whereas tetrameric KGA (or its isoenzyme GAC) is activated by phosphate6. In addition, their tissue distributions are different; LGA exists predominantly in the liver, while

2

ACS Paragon Plus Environment

Page 2 of 40

Page 3 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

KGA is mainly found in the kidney and brain and is highly elevated in tumors4-5. Interestingly, in cases of liver cancer, down-regulation of LGA (GLS2 gene) and up-regulation of KGA (GLS1 gene) were reportedly associated with poor outcome5-6. Therefore, KGA is an important isoform to target in cancer therapeutics. So far, various KGA inhibitors have been developed for cancer, several of which are shown in Figure 1. The active site inhibitors (Acivicin and DON)7 showed good in vivo anticancer activity but high toxicity, perhaps due to the lack of selectivity against LGA, which is important for liver function. Specific targeting of KGA is important to lower its toxic side effects. BPTES8-9 is known to be a KGA allosteric inhibitor, and its x-ray co-crystal structure shows that it binds at the KGA tetramer interface. Since BPTES showed poor solubility (0.01 µM), several series of BPTES derivatives were prepared to improve solubility, including Compound 69, UPGL0000410, thiazolidine-2,4-dione derivatives11, and CB-83912, as shown in Figure 1. Thiazolidine-2,4-dione derivatives contain a Michael acceptor and react to KGA protein residues. They demonstrated good in vivo anti-tumor effects, whereas the other compounds demonstrated only limited in vivo efficacy. CB-839 was the most potent KGA allosteric inhibitor; it inhibited a triple-negative breast cancer cell line with an IC50 value of 20–30 nM, but at an oral dose of 200 mg/kg, it achieved only 50% tumor growth suppression in an in vivo xenograft model12. To develop more potent KGA inhibitors, we investigated the possibility of improving KGA inhibition and anticancer activity with the 1,3,4-selenadiazole group. In the co-crystal structure of KGA and BPTES/CB8398,13,14, the thio-diazo is near Y394 at the allosteric site. Y394 could provide an aromatic ring for π stacking or –OH for hydrogen bond formation. Since hydrogen bonds only form among N, O, and F atoms, Y394 residue is more likely to interact with the sulfur of the thiodiazole ring through polar or non-polar interactions. In addition, due to the increase in electro-donation from O and N to S, the electro-density increased from furan and pyrrole to thiophene15. Therefore, replacing sulfur with selenium

3

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

could increase the electro-density of the selenophene ring and might affect the interaction between Y394 and a 1,3,4-selenadiazole ring. The thiadiazol group of CB839 is also involved in a water-mediated interaction with ASP327, and replacing sulfur with bulky selenium might increase this interaction. In addition, the allosteric inhibitor BPTES was reported to bind to KGA through an induced fit mechanism8, and it is difficult to predict changes in inhibitory activity based on structure docking. As a bioisostere for sulfur and oxygen atoms, selenium can produce unique biological activity. For example, selenocysteine is an essential amino acid that is required for the biological activity of many critical human enzymes, such as glutathione peroxidase (GSH-Px), thioredoxin reductase (TrxR), and deiodinase16. Various selenium-containing molecules showed good anticancer activity, including isoselenocyanates17, benzoselenadiazole18, methylseleninic acid19, di-selenium ether20, selenoureas21, ethylselen22, and hexylselen23. Therefore, it seems likely that replacing the thio-diazo group with a selenio-diazo group in this series of compounds would generate improved anti-cancer or other interesting biological activity. However, the mechanism still remains to be investigated. Chemical synthesis of 1,3,4-selenadiazole was reported previously in harsh chemical synthetic conditions24–30 that are not suitable for chemical synthesis of compounds containing sensitive functional groups, such as amine, hydroxyl, ester, or amide. Therefore, we explored chemical synthesis using amino selenourea and nitrile under mild conditions, successfully incorporating the 1,3,4-selenadiazole moiety into BPTES and CB839 with high yield. As shown in Figure 2, various 1,3,4-selenadiazole derivatives of BPTES and CB839 were synthesized for structure activity relationship (SAR) analysis. The effects of selenium substitution and different acyl groups were evaluated using a reliable KGA enzyme activity assay, Biolayer interference (BLI)-based biophysical assay, and precise cell-based assay. Our results demonstrated that some 1,3,4-selenadiazole derivatives feature enhanced inhibition of KGA, inhibition of cancer cell growth, and tumor reduction in xenograft models.

4

ACS Paragon Plus Environment

Page 4 of 40

Page 5 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Figure 1. Examples of KGA inhibitors reported in the literature.

RESULTS AND DISCUSSION Synthesis of 1,3,4-selenadiazole derivatives of BPTES and CB839 Prior studies synthesized 1,3,4,-selenadiazoles under very harsh conditions, such as {PhP(Se)(l-Se)}24, Woollins’ reagent (WR)]29, or refluxing hydrazinecarboselenoamide with carboxyl compounds in phosphorous oxychloride28. Since thiadiazole can be readily synthesized from hydrazinecarbothioamide with a carboxylic acid or nitrile using P2O5 or trifluoroacetic acid (TFA) as a catalyst8, we attempted to react hydrazinecarboselenoamide derivatives31-32 with nitriles in the presence of P2O5 or TFA to synthesize various 1,3,4selenadizole derivatives, as shown in Figure 2. Hydrazinecarboselenoamide (A1) was synthesized according to the method described in the literature33. Phenyl- or alkyl-substituted hydrazinecarboselenoamides were synthesized from an isoselencynate reaction with suitable amines31. Phenylamines containing a strong electron-withdrawing group, such as NO2 or CF3, or a bulky group, like tri-phenyl methylamine, yielded no reaction product, as shown in Figure 2 (compound A series).

5

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 2. Synthesis of 1,3,4-selenadiazole analogs: (a) selenium urea derivatives; (b) bis-cyanide derivatives; (c) bis-1,3,4-selenadiazole; (d) symmetric and asymmetric BPTES derivatives; (e) mono-cyanide derivatives; and (f) CB839 derivatives.

The resulting amino seleniumurea compound (A1) reacted well with the cyano compounds (B or E). In general, for bisnitriles, both cyano groups were derivatized to 1,3,4-selenadiazoles; when a large excess of bisnitrile was present, monosubstituted 1,3,4-selenazole could be obtained, providing more flexibility in chemical synthesis. In addition, the reaction conditions are mild enough to yield selenadiazoles in the presence of various functional groups, such as amine, amide, ether, and thioether. For example, the 1,3,4-selenadiazole analogs of BPTES and CB839 were synthesized with a good yield. The BPTES analogs were synthesized using 3,3'-thiodipropanenitrile or adiponitrile as a starting material, and then amide formation was performed using various acid derivatives9. CB839 analogs were synthesized from N-(6-(4-cyanobutyl)pyridazin-3-yl)-2-(3-(trifluoromethoxy)phenyl)acetamide according to the method described in the literature34 for SAR analysis. The resulting 1,3,4-selenadiazole-containing BPTES and CB839 derivatives were 6

ACS Paragon Plus Environment

Page 6 of 40

Page 7 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

subject to comparison of in vitro and in vivo activity with the corresponding 1,3,4-thiadiazole compounds. SAR comparisons of selenadiazole analogs and their thiadiazole counterparts Many previously reported KGA allosteric inhibitors were tested for activity in multiple enzyme-coupled biochemical assays to assess SAR. These coupled assays include GluO-horseradish peroxidase measuring resorufin formation at ex530/em590 nm35-36, ammonia formation measured with Nessler’s reagent by absorbance at 450 nm36, glutamate dehydrogenase (GDH)-diaphorase measuring resorufin formation at Ex544/Em590 nm36, and [3H]-glutamate detection assay10. The complicated nature of these assays has resulted in great variations in IC50 determination and, consequently, misinterpretation of SAR analysis. For example, the IC50 values reported for BPTES in the literature range from 0.1–3.3 µM in GluO-horseradish peroxidase-coupled assays. In addition, the solubility of BPTES was determined to be 10 nM10,36; ebselen was mis-identified as a potent 9 nM KGA inhibitor23,36; and there were discrepancies in the optimal bridge length (4- or 6-carbon) in BPTES analogs.10,23 Since GluO-horseradish peroxidase-coupled KGA assay36 is a three enzyme coupled assay detecting H2O2 formation which is problematic in excluding false positives, we performed our KGA assay using a stable 2-(3-(2-methoxy-4-nitrophenyl)-2-(4-nitrophenyl)-2H-tetrazol-3-ium-5-yl)benzenesulfonate sodium salt (EZMTT) dye for NAD(P)H detection. The coupled enzyme GDH was always tested at the same time to serve as a blank control for identification of false positives. Because KGA is activated by phosphate, the KGA enzyme was preincubated with the test compound for 30 min at room temperature in a phosphate-free buffer, and then substrates and 200 mM phosphate were added to initiate hydrolysis. After 4 h of incubation at room temperature, the resulting glutamate was quantified by GDH-EZMTT detection reagents. As shown in Table 1, the 1,3,4-selenadiazole compounds did not show inhibition of GDH. In addition, the C series (CPD1-3) did not show KGA inhibition, perhaps due to the

7

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 40

lack of an essential amide linkage, which is in agreement with the co-crystal structure of KGA-BPTES, where the amide group forms via hydrogen bonding with multiple residues at the allosteric site9.

Table 1. Structure activity relationships of 1,3,4-selenadiazole and 1,3,4-thiadiazole derivatives of BPTES and CB839 R3

No.

R2

X

X N N

N N

R3

R4

R4

H N

CPD1

R2

X

GDH

KGA

A549

IC50

IC50

IC50

Max-INH

μM

μM

μM

%

R3=R4

CH2CH2SCH2CH2

Se

>13

>13

>13

0±2%

R3=R4

CH2CH2SCH2CH2

Se

>13

>13

>13

0±3%

R3=R4

CH2CH2SCH2CH2

Se

>13

>13

>13

0±2%

-NH2

CH2CH2CH2CH2

S

>13

>13a

>13

0±3%

R3=R4

CH2CH2CH2CH2

S

>13

>13a

>13

0±3%

-NH2

CH2CH2SCH2CH2

>13

1.1±0.03

>13

-NH2

CH2CH2SCH2CH2

>13

1.4±0.05

>13

R3=R4

CH2CH2SCH2CH2

>13

0.027±0.003

0.11±0.02

81±4%

R3=R4

CH2CH2SCH2CH2

>13

0.28±0.05

2.9±0.3

80±3%

-NH2

CH2CH2SCH2CH2

>13

2.1±0.3

>13

H3CO

N H

CPD2

H N

CPD3

H N

CPD4

O H N

CPD5

O O N H

CPD6

Se

N

O N H

CPD7

S

N

O N H

CPD8

Se

N

O N H

CPD9

S

N

O

CPD10

HN

Se

N

8

ACS Paragon Plus Environment

Page 9 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

O

CPD11

HN

-NH2

CH2CH2SCH2CH2

R3=R4

CH2CH2SCH2CH2

R3=R4

CH2CH2SCH2CH2

-NH2

CH2CH2SCH2CH2

-NH2

CH2CH2SCH2CH2

R3=R4

S

>13

5.1±0.4

>13

>13

0.06

1±0.3

82±4%

>13

0.13±0.02

2.2±0.4

80±3%

>13

2.8±0.2

>13

S

>13

5±0.5

>13

CH2CH2SCH2CH2

Se

>13

0.06±0.01

0.23±0.06

82±3%

R3=R4

CH2CH2SCH2CH2

S

>13

0.13±0.03

0.42±0.05

72±2%

-NH2

CH2CH2SCH2CH2

Se

>13

8±0.4

>13

0±2%

R3=R4

CH2CH2SCH2CH2

Se

>13

1.7±0.2

>13

0±2%

-NH2

CH2CH2SCH2CH2

Se

>13

0.3±0.05

5±0.3

60±3%

>13

0.001

0.017

82±2%

>13

0.002±0.0003

0.04±0.005

82±2%

>13

0.001±0.0003

0.017±0.004

80±3%

N

O

CPD12

HN

Se

N

O

CPD13

HN

S

N

H N

CPD14

O

H N

CPD15

O H N

CPD16

O

H N

BPTES O

H N

O

CPD17

Se

H

H H

H N

O

CPD18

H

H H

O N H

CPD19 H

H H

CPD20

N

O F3CO

N

NH

Se

N N O F3CO

F3CO

N

N H

CB-839

CPD21

O

N N

N

N

S

O

N

NH

N H

O N H

N N N

O

N N Se

N H

9

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

F

CPD22

F3CO

O

N N

N H

CPD23

a:

F3CO

O N H

O

N N

O NH

N

N H

Se

Page 10 of 40

>13

0.02±0.002

0.46

91±2%

>13

0.01±0.002

0.01±0.002

90±3%

O

N N N

O

N N Se

N H

OCH3

The shape of the binding curve is abnormal, as shown in Figure 3.

Most D-series compounds (CPD4-19) showed KGA inhibition activity, except for CPD4 and CPD5. CPD5 was a BPTES derivative with a 4-carbon bridge between the two thioazoles (R2)10 that was previously reported to be a better KGA inhibitor (IC50 = 1.6 µM) than BPTES (IC50 = 3.3 µM; tested at the same time) in the GluO-horseradish peroxidase-coupled assay. In our assay, CPD5 did not inhibit GDH and showed essentially no sigmoid dose response curve with KGA (Table 1; Figure 3). However, we repeated the experiments several times with different people. The KGA inhibition curves showed great variation in shape, and all lacked a dose response. In a BLI biomolecular interaction assay (Figure 3), BPTES, CB839, CPD8, and CPD20 prevented KGA from binding to BPTES immobilized on the chip, but CPD5 did not. In addition, the solubility of CPD5 was very poor, possibly leading to non-specific interaction with KGA and causing a noisy signal. Taken together, the optimal bridge for bis-thiadiazole- or bis-selenadiazole-type KGA allosteric inhibitors is the 5-atom thioethyl ether or the 6-carbon straight chain.

10

ACS Paragon Plus Environment

Page 11 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Figure 3. Biochemical and biophysical analyses demonstrated that CPD8 and CPD20 have better KGA inhibitory activity than the corresponding BPTES and CB839. (a) GDH inhibition assay; (b) KGA/GDH-coupled assay; (c) BLI subtraction assay to show compound binding to KGA. Sensorgrams were plotted by subtracting two sensorgrams of KGA binding to the immobilized BPTES in the presence or absence of a compound. One sample contained only 100 nM KGA, and the other contained 100 nM KGA and a compound (BPTES, CB-839, CPD8, CPD20, or CPD5). The subtracted sensorgram represents the amount of KGA bound to the compound (BPTES, CB-839, CPD8, CPD20, or CPD5). The sensorgram showed no binding of the compound (CPD5) to the KGA enzyme, because the enzyme could bind to immobilized BPTES equally well in the presence or absence of the compound. When a compound (10 µM BPTES, CB839, CPD8, or CPD20) can bind to the KGA enzyme, the subtracted sensorgram will show the difference, shown as signals in pink or blue. Even though researchers suggested that asymmetric BPTES analogs will have better solubility and in vivo activity10, SAR analysis of CPD6-CPD17 demonstrated that, by removing one of the corresponding phenylacetyl moiety (such as CPD6, CPD7, CPD10, CPD11, CPD14, CPD15, or CPD17), the asymmetric analogs (CPD7 & CPD9; CPD8 & CPD10) lost more than 10-fold activity in both the KGA assay and glutamine-dependent A549 cancer cell proliferation assay. Also, the methylene next to the amide group is important to activity; CPD17 and CPD19 showed a 10-fold difference in IC50 values. This is supported by the co-crystal structure of BPTES and KGA8; since the KGA allosteric site

11

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

exists at the interface of the tetramer, both thiodiazole and the amide bond extension of BPTES showed important hydrogen interactions with tetrameric KGA, which may have led the symmetric BPTES analogs to interact more than the asymmetric ones. Interestingly, the BPTES selenadiazole analogs demonstrated more than two times better potency than the corresponding 1,3,4-thiadiazole analogs in the KGA enzyme assay. SAR analysis suggests that 1,3,4-selenadiazole fits better in the KGA allosteric site. In addition, replacing the amide group with an amine lowered its KGA inhibitory activity, which is in agreement with the co-crystal structure of KGA-BPTES, where the amide group forms via hydrogen bonding with multiple residues at the allosteric site. To increase solubility, heteroaromatic groups were used to replace the phenyl group. SAR analysis of seven pairs of 1,3,4-selenadiazoles and their 1,3,4-thiadiazole analogs demonstrated that the 1,3,4-selenadiazoles compound showed better activity than the corresponding 1,3,4-thiadiazoles compound in a KGA biochemical assay and cancer cell growth inhibition assay. In particular, the indole-substituted 1,3,4-selenadiazole compound showed five-fold improvement in potency and complete inhibition of KGA activity. Additionally, the heteroaromatic group achieved complete inhibition of KGA enzyme activity and increased compound solubility. Selenadiazole analogs of CB839 were also synthesized (CPD20, CPD21, CPD22, CPD23). CPD20 is selenium-substituted CB839, and it showed two times better potency than CB839 in the KGA inhibition assay. Comparison of the growth inhibitory activity of BPTES, CB839, and their 1,3,4-selenadiazole analogs in glutamine-dependent cell lines Although BPTES and CB839 are potent KGA allosteric inhibitors, they achieved a maximum of only approximately 50% reduction in tumor size in vivo. The poor solubility of BPTES was blamed for its poor in vivo outcome. To select a suitable cell line for the assay, we measured the glutamine dependence of several cell lines and their inhibition by KGA inhibitors. Using non-toxic EZMTT reagents, we discovered that all the tested cell lines

12

ACS Paragon Plus Environment

Page 12 of 40

Page 13 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

showed severe growth reduction in the absence of glutamine for five days, but they remained alive. Interestingly, even though complete inhibition of KGA was achieved by CPD8, CB839, and the selenium analogs in a biochemical assay, these compounds did not achieve complete inhibition of glutamine-dependent cell lines such as A549 and Caki-1 (maximal inhibition: 70–90%) (Figure 4; Table 2). It is not likely that this is due to the solubility issue because the water solubility was measured at 0.1 µg/ml for BPTES and 3 µg/ml for CB839. Thus, a question remains: Why do potent and apparently efficacious KGA inhibitors (CB839 and CPD20 compounds) show limited efficacy in a cell-based assay; is this a KGA-target-related issue or a cellular penetration problem?

Figure 4. KGA allosteric inhibitors (BPTES or CB839 type) only partially inhibit cancer cell growth, whereas a KGA/GDH dual allosteric inhibitor (hexylselen) showed complete inhibition.

Table 2. Inhibition of cancer cell growth by 1,3,4-selenadiazole and 1,3,4-thiadiazole derivatives of BPTES and CB839 H22

Caki-1

HCT116

IC50

Maximal

IC50

Maximal

IC50

Maximal

(μM)

Inhibition %

(μM)

Inhibition %

(μM)

Inhibition %

BPTES

>10

29±3

1.16±0.12

73±3

0.54±0.08

69±4

CB839

>10

44±3

0.04±0.009

88±2

0.028±0.005

90±3

CPD9

>10

42±2

3.92±0.35

85±3

2.2±0.13

93±2

CPD8

1.38±0.12

71±3

0.4±0.05

87±4

0.32±0.09

94±2

13

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 40

CPD20

6.78±0.21

75±4

0.019±0.004

89±2

0.009±0.002

90±4

CPD21

>10

10±2

0.014±0.003

90±2

0.009±0.003

92±3

Hexylselen

0.84±0.09

100±2

1.29±0.16

100±2

2.17±0.23

100±2

CPD23

4.05±0.25

70±4

0.02±0.003

90±3

0.02±0.003

92±2

Comparison of cellular concentrations of 1,3,4-thiadiazole compounds and their corresponding 1,3,4-selenadiazole analogs by Raman analysis37 As shown in Figures 3 and 4, KGA inhibitors showed complete inhibition of enzyme activity, but only partial inhibition of cancer cell lines. However, the bis-selenadiazole containing compounds showed improved efficacy in comparison with the corresponding bis-thiadiazole compounds. We were interested in investigating whether the cellular uptake or metabolism of the compounds played a role in cellular efficacy. CPD6, CPD8, and CPD9 were subjected to a substrate uptake assay using Raman analysis (Figure 5a), which was possible because their 1:1 mixtures with nanogold particles showed strong SERS spectra. As shown in Figure 5, within wavelengths of 1100–1800 cm-1, the 1:1 mixtures showed a dose-dependent increase in absorbance up to a compound concentration of 2 µM, whereas in the absence of nanogold particles, no absorbance was observed at these wavelengths. The intensity at 1100–1800 cm-1 wavelengths started to drop completely at compound concentrations above 2 µM. Therefore, for our compound uptake assay, we used premade 1:1 compound/nanogold mixtures up to a concentration of 2 µM. In addition, because the peak at 1352 cm-1 wavelength does not exist in the control cells in the absence of the compounds, we selected the 1352 cm-1 wavelength as the compound marker for the cellular compound uptake assay. After treating the glutamine-dependent Caki-1 cells with 1:1 compound/nanogold mixture particles for 3 h, 9 h, 16 h, and 24 h, SERS spectra were collected to compare the cellular concentration of the compounds. Interestingly, as shown in Figure 5c, the most potent selenadiazole compound (CPD8) showed the highest cellular concentration; the intensity at 1100–1800 cm-1 was three to four times greater than its corresponding thiadiazole analog (CPD9) or asymmetrical analog (CPD7). Overall, the

14

ACS Paragon Plus Environment

Page 15 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

bis-1,3,4-selenadiazole compounds accumulated to a much greater extent than the corresponding bis-thiadiazole and mono-1,3,4-selenadiazole compounds.

Figure 5. Raman analysis of the compounds (CPD6, CPD9, and CPD8) before and after their cellular uptake. (a) SERS spectra show the dose response of the 1:1 mixtures of compound with nanogold particles. (b) The dose response of the 1:1 mixtures at an absorbance of 1352 cm-1. (c) Cellular uptake and distribution of the 1:1 mixtures of 2 µM compound with nanogold particles. (d) Light microscope pictures of the corresponding cell images in C.

1,3,4-selenadiazole analogs showed significantly enhanced ROS in A549 cancer cells ROS is a key regulator of cancer cell growth. Many selenium-containing compounds showed activation of ROS, although most reported selenium compounds are reactive16-23. Even though the mechanism of action remains to be investigated, as shown in Figure 6, CPD20 and CPD22 generated more ROS than CB839 at a concentration of 10 µM.

15

ACS Paragon Plus Environment

Journal of Medicinal Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 6. ROS activity. 10000/well A549 cells treated or not treated (Ctrl) with 100 µg/ml ROSUP (positive control) or different concentration of CB839 or CPD20 or CPD23 (0, 3, or 10 µM) for 12 h. At a compound concentration of 10 µM, the selenium-containing analogs showed improved ROS activity in comparison with CB839.

1,3,4-selenadiazole analog CPD20 showed improved efficacy in the glutamine-dependent HCT116 tumor model Because the growth of HCT116 cancer cells is inhibited by 1,3,4-selenadiazoles (CPD20) more than the corresponding CB839 compound (Table 2 and Figure 4), we used the HCT116 cancer cell xenograft model to compare their in vivo efficacy. We subcutaneously injected 10 mg/kg of the compounds (CB839 and CPD20) into young nude mice (n=6). As shown in Figure 7, CB839 and CPD20 reduced the size and weight of the HCT116 tumor, and statistical analysis showed that the 40% reduction in tumor weight by CPD20 is statistically significant. In addition, H&E staining of the end-point tumor tissues was performed to assess the tumor tissue damage. In the untreated control group (Figure 7e), the tumor cells were arranged in sheets with large cell size, large nuclei, and a high nucleus/cytoplasm ratio, and they were infiltrated with scattered lymphocytes. In the CPD20- or CB839-compound-treated groups (Figure 7d), the tumor cells could not be clearly seen under a microscope, and the CPD20-compound-treated tumors showed more necrosis and deterioration than the 16

ACS Paragon Plus Environment

Page 16 of 40

Page 17 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

CB839-compound-treated ones. Overall, the in vivo efficacy is in good agreement with the cell growth inhibition results reported in Figures 4 and 6).

Figure 7. Comparison of the in vivo efficacy of three different groups of HCT116 xenograft models (n=6 mice per group). The selenium analog (CPD20) showed better in vivo activity than CB839 after 15 injections of N.S. control, 10 mg/kg CB839, or 10 mg/kg CPD20 compound. (a) Comparison of the tumor weights of untreated and treated groups was performed using an unpaired t test (*: P