Chemical Space and Biological Target Network of Anti-inflammatory

Subscriber access provided by TULANE UNIVERSITY

Chemical Information

Chemical Space and Biological Target Network of Anti-inflammatory Natural Products Ruihan Zhang, Jing Lin, Yan Zou, Xing-Jie Zhang, and Wei-Lie Xiao J. Chem. Inf. Model., Just Accepted Manuscript • DOI: 10.1021/acs.jcim.8b00560 • Publication Date (Web): 12 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 31 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 Chemical Information and Modeling

1

Chemical Space and Biological Target Network

2

of Anti-inflammatory Natural Products

3

AUTHOR NAMES

4

Ruihan Zhang, Jing Lin, Yan Zou, Xing-Jie Zhang, Wei-Lie Xiao*

5

AUTHOR ADDRESS

6

2 Rd. Cuihubei, 650091 Kunming, China, Key Laboratory of Medicinal Chemistry for

7

Natural Resources, Ministry of Education and Yunnan Province, School of Chemical

8

Science and Technology, Yunnan University

1

ACS Paragon Plus Environment

Journal of Chemical Information and Modeling 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

1

Abstract

2

Natural products (NPs) are a promising source of anti-inflammatory molecules for

3

the development of drugs. Despite there being an abundance of reports of large numbers

4

of NPs having bioactivity in preliminary cell-based assays of anti-inflammatory potential,

5

their further optimization and exploration is limited by lack of a comprehensive

6

understanding of their effective scaffold structure or biological targets. To facilitate

7

target-based studies of anti-inflammatory NPs, the details of 665 NPs reported to have

8

anti-inflammatory activity were extracted from the literature and compiled into a dataset

9

we termed InflamNat. The physicochemical properties of the NPs were analyzed and the

10

distribution of their structures and scaffolds presented. A compound-target network was

11

constructed from data in the PubChem Bioassay database. The results demonstrated that,

12

compared to natural anti-cancer compounds in the NPACT database, compounds from

13

the InflamNat dataset contained a comparable distribution of compound type, but with a

14

higher proportion satisfying Lipinski’s rule. The all-atom structures and scaffold of the

15

compounds were diverse and barely convergent, with flavonoids and triterpenoids as the

16

groups in highest abundance. The biological targets of the InflamNat compounds were

17

identified as belonging to a variety of protein families that had varied function. Seventy

18

two percent of InflamNat compounds involved in the network were identified as having

19

more than one biological target, highlighting the potential for multi-target

20

anti-inflammatory drug development. In conclusion, anti-inflammatory NPs provide a

21

good library for the screening of target-based leads or fragment-based drug design. Thus

22

elucidation of their biological targets is fundamental for either a specific single- or

23

multi-target drug development strategy. Meanwhile, a large proportion of the chemical 2


Page 2 of 31

Page 3 of 31 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


1

space of anti-inflammatory NPs is still unexplored, with novel active scaffolds remaining

2

to be discovered.

3



1

Page 4 of 31

Introduction

2

Inflammation can be defined as the “complex reaction of vascularized tissue to

3

infection, toxin exposure, or cell injury that involves extravascular accumulation of

4

plasma proteins and leukocytes.”1, and is a critical system for protection from external or

5

internal injury. However, an abnormal, or over-activated inflammatory reaction might be

6

harmful for cells, tissues or organs. In addition to acute inflammation or autoimmune

7

disease, inflammation as a pathological condition is involved in cancer2, diabetes3,

8

Alzheimer’s

9

anti-inflammation becomes an increasingly important topic in drug development.

disease4,

depression5

and

cardiovascular

diseases6.

Therefore,

10

Natural products (NPs) hold many advantages as a source of compounds for

11

anti-inflammatory drug development. Firstly, a long history of application in folk

12

medicine provides clinical indications for the discovery and development of

13

anti-inflammatory compounds from natural resources. Secondly, NPs inspire a large

14

number of novel scaffolds, with complex structure in terms of ring-systems and

15

stereochemistry7. Thirdly, NPs have potential for the development of multi-target

16

anti-inflammatory agents, which would be beneficial for the treatment of complex and

17

systemic diseases such as inflammation8.

18

A number of databases that are focused on anti-cancer NPs have been published,

19

including NPACT9, AfroCancer10 and NPCARE11, whereas no database focusing on

20

anti-inflammatory NPs has yet been compiled. In the Dictionary of Natural Products12,

21

886 entries were annotated with “antiinflammation” (accessed in October 2018),

22

however, the definition of anti-inflammatory activity might be ambiguous in literature, 4


Page 5 of 31 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


1

and a number of the compounds have also been reported to be cytotoxic, suggesting that

2

suppression of pro-inflammatory factors might be as a result of the death of cells rather

3

than inhibition of inflammatory pathways. Due to the lack of a dataset, the landscape of

4

structures and biological targets of anti-inflammatory NPs is still unexplored. During a

5

review of the literature we found that the majority of anti-inflammatory NPs were

6

reported with only preliminary bioassay results such as inhibition of nitric oxide (NO) or

7

interleukins (ILs), whereas the identity of biological molecular targets responsible for

8

their anti-inflammatory activity remains unclear, limiting further exploitation and

9

optimization of these naturally-derived scaffolds.

10

To facilitate target-based anti-inflammatory NP research, we present in this study a

11

dataset of anti-inflammatory NPs (InflamNat) collected from the literature, with a

12

comprehensive analysis of their physicochemical properties, chemical space and

13

biological

14

anti-inflammatory and anti-cancer compounds within the InflamNat dataset, and to

15

evaluate their structural diversity, comparisons were made with compounds in the

16

Naturally occuring Plant-based Anti-cancerous Compound-Activity-Target (NPACT)

17

database9. NPACT contains comprehensive information on 1574 plant-derived

18

compounds tested for anti-cancer activity both in vitro and in vivo.

targets.

To

ascertain

the

commonalities

and

differences

19

20

Results and Discussion

21

Overview of the anti-inflammatory natural products dataset (InflamNat) 5


between


1

Following data collection as described in the methods Section, details of 665

2

anti-inflammatory NPs were extracted from 362 research articles and collated as the

3

InflamNat dataset. The data from compounds within InflamNat were compared with the

4

1574 molecules from the NPACT database in this study.

5

Among the anti-inflammatory compounds collected, only 106 molecules (15.9%)

6

were duplicated in the NPACT dataset, indicating that the InflamNat and NPACT

7

datasets were diverse. The InflamNat dataset comprised almost all types of NP

8

compound, with terpenoids as the largest group (especially triterpenoids), followed by

9

flavonoids (Figure 1A). The distribution of compound type in the datasets was found to

10

be similar, with ratios in InflamNat to NPACT of terpenoids, flavonoids and alkaloids

11

being 39%/33%, 19%/21% and 8%/7%, respectively. We have summarized the

12

originating species for each compound reported in the literature (details in supporting

13

information), with 86% of the collected compounds derived from terrestrial plants (see

14

supporting information for details), 9% from marine life, 5% from terrestrial fungi and

15

bacteria and two compounds from terrestrial animal metabolites (Figure 1B).

16 6


Page 6 of 31

Page 7 of 31 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


1 2

Figure 1. Distribution of compound types (A) and origins (B) in the InflamNat dataset

Physicochemical properties of the anti-inflammatory natural products

3

Lipinski’s rule defines physicochemical conditions for orally-active drugs and is the

4

best-known criteria of drug-likeness evaluation: molecular weight (MW) > 500 Da, the

5

oil-water partition coefficient log P > 5, hydrogen bond donors (HBD) > 5 and hydrogen

6

bond acceptors (HBA) > 10. Orally-active drugs should not violate any more than one of

7

these four conditions. The physicochemical parameters for each InflamNat compound

8

with reference to Lipinski’s rule were calculated and compared with those in the NPACT

9

dataset. Four hundred and forty two compounds in the InflamNat dataset (66.4%) and

10

831 compounds in the NPACT (52.7%) dataset were in compliance. As displayed in

11

Figure 2A, 79.1% of the InflamNat compounds were within the MW range 200-500,

12

higher than that of the NPACT compounds, the proportion of which with a MW > 500

13

was significantly higher. The distribution of ALogP (Figure 2B) suggested that there

14

were fewer hydrophobic molecules with an ALogP > 6 in the InflamNat dataset than in

15

NPACT. Figures 2C and D demonstrate that the majority of the compounds in both

16

datasets did not have values for hydrogen bond donors or acceptors that were

17

unfavorable. Overall, a higher proportion of InflamNat compounds fitted Lipinski’s rule

18

than did those in the NPACT dataset.

7



1

2

Figure 2. Distribution of physicochemical parameters for InflamNat and NPACT compounds.

3

(A) Molecular weight; (B) ALogP; (C) Number of hydrogen bond donors; (D) Number of

4 5

hydrogen bond acceptors.

6

Chemical Space of the anti-inflammatory natural products

7

To investigate the chemical space of the InflamNat compounds, the all-atom

8

structures were first clustered in a hierarchical manner. The similarity scores (Tanimoto

9

coefficients) were calculated pairwise for the InflamNat compounds. The two large

10

regions in the heatmap presented in Figure 3 represent phenolic scaffolds (flavonoids,

11

benzopyranoids) and terpenoid scaffolds. For a similarity cutoff of 0.6 (the value used in 8


Page 8 of 31

Page 9 of 31 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


1

the following discussion), a total of 368 clusters were generated. Representative

2

structures of the 10 largest clusters are displayed in Figure 3. The largest cluster

3

comprised flavonoids, which share similar structural features (phenol and ketone groups)

4

with nearby benzopyranoids. These fragments increased the interaction and reactivity

5

between the compound and biological macromolecules, e.g. π-π stacking (aromatic rings)

6

and hydrogen bonding (phenolic hydroxyl and ketone groups), contributing to the vast

7

diversity of biological activity exhibited in flavonoids and benzopyranoids. Another

8

significant cluster is the triterpenoids, represented by the pentacyclic triterpenoids. The

9

anti-inflammatory activity of triterpenoids might be due to their structural similarity with

10

steroid hormones, which control the immune system and mediate inflammatory reactions.

11

Glycosides of the flavonoids and triterpenoids are the major clusters comprising

12

anti-inflammatory NPs. Nevertheless, sugar moieties are usually considered essential for

13

NPs to exhibit beneficial pharmacokinetic properties, but are irrelevant to their

14

pharmacophore13. Another two major structural clusters in the InflamNat dataset have the

15

framework of sesquiterpenoids.

9



1

2

Figure 3. Hierarchical structural clustering of the InflamNat compounds. Representative

3

compounds of the major clusters are shown.

4 5

Compared with the NPACT dataset, the all-atom structures of the InflamNat

6

compounds were far from convergent (Figure 4). The 10 largest clusters of structures

7

occupied only 23.1% of the total number of compounds, and 74.9% of the clusters had

8

only one member (i.e. 25.1% of the clusters had more than one member, as demonstrated

9

by the coordinates shown in Figure 4A), indicating that the compounds outside the major

10

clusters are quite diverse. To further investigate the chemical space of the InflamNat

11

compounds, Bemis-Murcko (BM) scaffolds of the compounds were generated, excluding

12

the purely linear molecules (without rings). As a result, 654 and 1445 scaffolds were

13

generated and clustered into 188 and 190 groups, for the InflamNat and NPACT 10


Page 10 of 31

Page 11 of 31 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


1

compounds, respectively, using a similarity cutoff of 0.6. The higher diversity of the

2

chemical space of the BM scaffolds of InflamNat compounds than those from the

3

NPACT dataset are reflected in the cumulative curves (Figure 4B). Apart from the BM

4

scaffolds presented in the major all-atom structure clusters, the majority of the top-ranked

5

scaffold clusters were quite simple (Figure 4C), being molecules with only one or two

6

rings but a variety of side-chains. In other words, after simplifying the all-atom structures

7

to scaffolds, it was principally the single and double ring-systems that contributed to the

8

increased clustering convergence, whereas there was still diversity among the complex

9

frameworks.

10 11



1

Figure 4. Scaffold Analysis. Clustering results (similarity cutoff = 0.6), comparing InflamNat and

2

NPACT datasets and shown as cumulative frequency curves of (A) all-atom structures and (B)

3

Bemis-Murcko scaffolds. Coordinates are shown for the turning points from which the clusters

4

have only one member. (C) Representatives of the scaffold clusters with five or more members

5

and reported in at least two different publications are presented for the InflamNat dataset, ranked

6 7

from left to right, then top to bottom.

8

Network of InflamNat Compounds and their identified biological targets

9

A majority of the research articles collected in this study evaluated the efficacy of

10

anti-inflammatory NPs using cell or animal inflammation models, but without identified

11

molecular targets. We therefore generated a compound-target network based on the data

12

from PubChem14. Our interest was about the activity of compounds to directly modulate

13

human targets, either inhibitory or stimulatory, thus records detailing gene expression or

14

those relating to other organisms were disregarded. In total 1293 entries were obtained,

15

including those from 176 InflamNat compounds and 258 human targets (including the

16

same target expressed in murine, pig, bovine, etc.). The network was streamlined by

17

classifying the targets into eight groups, with a number of protein subtypes merged into

18

one node (Figure 5). The complete data table is attached as SI which can be imported into

19

Cytoscape to generate the entire network. The network presented in Figure 5 contains 179

20

compound nodes and 259 target nodes. The compound number and a list of targets with

21

IDs of the cited PubChem Bioassay (AID) are presented in the SI. The distribution of the

22

degree of nodes (number of directly connected neighbors) is summarized in Figure 6A,

23

demonstrating that 72% of the 176 InflamNat compounds have been reported with more

24

than one molecular target and suggesting that NPs could be promising agents for 12


Page 12 of 31

Page 13 of 31 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


1

multi-targets in anti-inflammation therapy8. The compounds with the most targets are

2

represented by the flavonoids, as displayed in Figure 6B. Although the InflamNat

3

compounds were collected based on their anti-inflammation activity, their biological

4

targets, however, are related to various disease states, e.g. immune system disease,

5

cancer, metabolic disease or neurological disorders.

6

The targets with most active compounds from InflamNat are Hydroxysteroid

7

dehydrogenase (HSD), Tyrosyl-DNA phosphodiesterase 1 (TDP1), DNA polymerase,

8

Tau

9

domain-containing proteins (JMJD), Vitamin D receptor (VDR), Cyclooxygenase (COX)

10

and Retinoid-related orphan receptor γ (ROR γ ). Notably, not all these high-ranking

11

targets are directly related to anti-inflammatory activity. For instance, TDPs and DNA

12

polymerase are important for DNA repair and replication, respectively, and both are

13

known as anti-tumor drug targets. CYP1/2/3 are major enzymes participating in drug

14

metabolism in humans, but are also widely represented in plants, fungi, animals and other

15

organisms. CYP19 (aromatase), is a key enzyme for biosynthesis of estrogens and

16

considered a therapeutic target for breast cancer. Other than that, HSD, Tao protein,

17

JMJD, COX and RORγ are closely related to inflammation. HSDs catalyze the

18

conversion of 17-hydroxysteroids to 17-ketosteroids, thus performing an essential role in

19

the production and function of steroid hormones involved in inflammatory reactions15.

20

Tau protein is crucially involved in the pathology of Alzheimer′s disease, with abnormal

21

hyperphosphorylation of Tau protein closely related to neuroinflammation16. JMJDs are

22

histone demethylases that regulate gene expression through post-translational

protein,

Cytochrome

P450

family

member

13


(CYP)

1/2/3/19,

Jumonji


Page 14 of 31

1

modification, with JMJD3 particularly involved in inflammation in macrophages,

2

mononuclear cells and bone marrow cells17. COX is required for the biosynthesis of

3

prostanoids, such as Prostaglandin E2 (PGE2), an important regulator of inflammation

4

response. Inhibition of COX reduces inflammatory reactions, and COX2, in particular, is

5

the principal target of nonsteroidal anti-inflammatory drugs (NSAIDS)18. Vitamin D

6

regulates gene transcription related to calcium metabolism and the immune response via

7

binding to VDR, making VDR a potential drug target for autoimmune diseases19. RORγ

8

is a key regulator of the production of pro-inflammatory cytokine IL-17 in lymphocytes.

9

Selective inhibition of RORγ has been validated as a therapy for inflammatory

10

disorders20. Overall, for more than three quarters of the InflamNat compounds,

11

anti-inflammatory targets with direct interaction remain to be determined. In this study,

12

the

13

anti-inflammatory

14

pharmacophores, revealing the mechanism of action and predicting off-target effects. On

15

the other hand, more data from target-based assays is urgently required to improve the

16

quality of this network.

compound-target NP

network

provided

development,

important

regarding

clues

inspiring

14


for

active

target-based scaffolds

and

Page 15 of 31 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


Figure 5. Compound-target network 15



Figure 6. (A) Degree distribution of nodes, defined as the number of directly connected neighbors. Gold: compound nodes; Turquoise: target nodes. (B) The 10 InflamNat compounds with the greatest number of reported targets. (C) Targets with the most reported active modulators in the InflamNat dataset.

Conclusions This work aimed to facilitate the development of anti-inflammatory NPs by presenting a landscape of their chemical space and biological targets. A series of cheminformatic analyses was performed on the InflamNat dataset with data extracted from the literature. Comparisons with the NPACT dataset showed that the InflamNat 16


Page 16 of 31

Page 17 of 31 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


dataset was more drug-like, with fewer large molecules with a MW > 500 and highly hydrophobic molecules with ALogP > 6. Flavonoids and triterpenoids were the principal structural types in InflamNat. The chemical space of the InflamNat compounds was far from convergent, suggesting that there was potential to discover novel anti-inflammatory scaffolds from NPs. The known biological targets of the InflamNat compounds are not limited to those in the anti-inflammation field but encompassing a variety of target types that are related to different physiological functions. Of the InflamNat compounds with target-based activity data, 72% have multiple targets, as represented by flavonoids. The development of anti-inflammatory NPs as drugs face challenges in addition to opportunity. Firstly, it is difficult to identify a key therapeutic target for systemic disorders such as inflammation. In this case, with a much larger chemical space than approved drugs21, NPs inspire novel scaffolds and pharmacophores in the design of multi-target agents. As shown in our study, a number of known anti-inflammatory NPs target more than one biological macromolecule. Secondly, further exploration of the pharmacological effect of NPs, including target identification, is limited to the quantity of compounds that can be obtained. Therefore, in silico approaches could play an important role in improving the efficiency of the research, such as modeling the structure-activity relationship to guide structure optimization, simplifying active scaffolds to increase the availability of synthesis, predicting biological targets so as to provide guidance for the laboratory, etc7. From a practical point of view in the study of anti-inflammatory NPs, this work provides a statistical assessment of the structures and targets of current anti-inflammatory NPs, and a fundamental dataset for further investigation.

17



Methods Data Collection and general analysis Six hundred and sixty five anti-inflammatory natural compounds (InflamNat) were extracted from the literature up to October 2018, by searching the Web of Science with the search term natural product plus the following keywords: arthritis, autoimmune*, immunomodulat*, inflammat* and NF-kappaB. In addition, the search excluded reviews, non-English articles, articles without ‘*inflam*’ or ‘immu’ in the abstract, and articles focused on raw extracts or a mixture of NPs. By reading through the full text of each article, anti-inflammatory NPs satisfying following criteria were extracted and collected into the InflamNat dataset: (1) Bioassays performed in inflammation cell models (LPS-stimulated macrophages, microglial cells, mast cell etc.) in addition to cells isolated from inflammatory animal models (ear edema, paw edema, pulmonary inflammation etc.); (2) Compounds able to inhibit the production of nitrite oxide (NO), prostaglandin E2 (PGE2) or pro-inflammatory cytokines (e.g. interleukins-1, 6 or 8), with an EC50/IC50 < 50 μM, or an inhibition rate > 50% at a tested concentration less than 50 μM or 50 μg/ml; (3) cell viability > 90% was required near the EC50/IC50, to exclude cytotoxicity-related reduction in NO, PGE2 or pro-inflammatory cytokine generation; (4) a threshold of p-value < 0.05 was set for the statistical significance of the experimental data. The 2D structure of 665 anti-inflammatory natural compounds were obtained from PubChem14 (505 compounds) or were manually generated. The SMILES description of molecular structure was collected, as shown in the supporting information. Stereochemical information reported in the referenced article or stored as isomeric 18


Page 18 of 31

Page 19 of 31 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


SMILES in the PubChem database (if available) was preserved. The type of compound was as described in the introductory document of the Dictionary of Natural Products. The Ligand Preparation module in Discovery Studio (DS) was employed to generate tautomers at a pH 6.5-8.5 and isomers for undefined stereocenters in order to compare InflamNat and NPACT compounds. The physiochemical properties related to drug-likeness were analyzed with the QSAR module in DS, including MW, ALogP (log P predicted using the method published by Viswanadhan et. al. in 1989)

22,

HBA and

HBD. Structure clustering and scaffold generation Structure clustering and Bemis-Murcko (BM) scaffold generation were performed using the ChemmineR23 and rcdk24 packages for R25, respectively. Binning and hierarchical clustering were achieved using a single linkage method and atom pair descriptor. The similarity of molecules was described by the Tanimoto coefficient. A single BM scaffold was generated for each compound, with the minimum fragment size set to 3. Compound-target network generation The bioactivity targets of the selected compounds were obtained from PubChem Bioassay14,

disregarding

anti-bacterial,

anti-fungal,

anti-viral,

insecticidal

and

anti-feedant targets, and in addition to assays of gene expression of the targets. Nomenclature and classification were checked against the neXtProt database (https://www.nextprot.org/)26. The compound-target network was generated using 19



Cytoscape27. The edge of the network simply refers to the compound being active against modulation of the target, either inhibitory or stimulatory, with “active” as defined by PubChem Bioassay, i.e. with an IC50/EC50/Ki/Kd under 50 μM. To display the network with clarity, targets were divided into seven categories. If Bioassay did not specify the target subtype, protein subtypes were merged as one node.

ASSOCIATED CONTENT

Supporting Information. SI _InflamNat_Compounds (XLSX): name, type, Pubchem CID, SMILES and reference of the InflamNat compounds. SI_Network (XLSX): Compound-Target network table of InflamNat. Import the table into Cytoscape to view the network. SI_Target_AID (PDF): Targets and the PubChem Assays (AID) referenced by the Compound-target network

AUTHOR INFORMATION

Corresponding Author *E-mail: [email protected]

20


Page 20 of 31

Page 21 of 31 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


Funding Sources

This project was supported financially by grants from the Yunnan Applicative and Basic Research Program (2018FY001 and 2018FA048), Natural Science Foundation of China (81422046 and 21762048).

Notes The authors declare no competing financial interest.

ACKNOWLEDGMENT

We thank Dr. Zhongjie Liang and Dr. Junyan Lu for valuable discussion and suggestions. We are grateful to the high-performance computing center of Yunnan University for providing the calculation resources and software.

REFERENCES 1.

Abbas, A. K.; Lichtman, A. H.; Pillai, S., Basic Immunology: Functions and Disorders of the

Immune System, the Fourth edition. Peking University Medical Press and Elsvier(Singapore) Pte Ltd: 2014. 2.

Crusz, S. M.; Balkwill, F. R., Inflammation and Cancer: Advances and New Agents. Nat.

Rev. Clin. Oncol. 2015, 12 (10), 584-596. 21



3.

Donath, M. Y., Targeting Inflammation in the Treatment of Type 2 Diabetes: Time to Start.

Nat. Rev. Drug Discov. 2014, 13 (6), 465-476. 4.

Heppner, F. L.; Ransohoff, R. M.; Becher, B., Immune Attack: the Role of Inflammation in

Alzheimer Disease. Nat. Rev. Neurosci. 2015, 16 (6), 358-372. 5.

Miller, A. H.; Raison, C. L., The Role of Inflammation in Depression: from Evolutionary

Imperative to Modern Treatment Target. Nat. Rev. Immunol. 2016, 16 (1), 22-34. 6.

Siti, H. N.; Kamisah, Y.; Kamsiah, J., The Role of Oxidative Stress, Antioxidants and

Vascular Inflammation in Cardiovascular Disease (a Review). Vascul. Pharmacol. 2015, 71, 40-56. 7.

Rodrigues, T.; Reker, D.; Schneider, P.; Schneider, G., Counting on Natural Products for

Drug Design. Nat. Chem. 2016, 8 (6), 531-541. 8.

Koeberle, A.; Werz, O., Multi-Target Approach for Natural Products in Inflammation. Drug

Discov. Today 2014, 19 (12), 1871-1882. 9.

Mangal, M.; Sagar, P.; Singh, H.; Raghava, G. P. S.; Agarwal, S. M., NPACT: Naturally

Occurring Plant-based Anti-cancer Compound-Activity-Target Database. Nucleic Acids Res. 2013, 41 (D1), D1124-D1129. 10. Ntie-Kang, F.; Nwodo, J. N.; Ibezim, A.; Simoben, C. V.; Karaman, B.; Ngwa, V. F.; Sippl, W.; Adikwu, M. U.; Mbaze, L. M., Molecular Modeling of Potential Anticancer Agents from African Medicinal Plants. J. Chem. Inf. Model. 2014, 54, 2433-2450. 11. Choi, H.; Cho, S. Y.; Pak, H. J.; Kim, Y.; Choi, J. Y.; Lee, Y. J.; Gong, B. H.; Kang, Y. S.; Han, T.; Choi, G.; Cho, Y.; Lee, S.; Ryoo, D.; Park, H., NPCARE: Database of Natural Products and Fractional Extracts for Cancer Regulation. J. Cheminformatics 2017, 9. 12. Dictionary of Natural Products 27.1. http://dnp.chemnetbase.com/(accessed October 1, 2018). 13. Ertl, P.; Schuffenhauer, A., Cheminformatics Analysis of Natural Products: Lessons from Nature Inspiring the Design of New Drugs. Prog. Drug. Res. 2008, 66, 217, 219-35. 14. Kim, S.; Thiessen, P. A.; Bolton, E. E.; Chen, J.; Fu, G.; Gindulyte, A.; Han, L. Y.; He, J. E.; He, S. Q.; Shoemaker, B. A.; Wang, J. Y.; Yu, B.; Zhang, J.; Bryant, S. H., Pubchem Substance and Compound Databases. Nucleic Acids Res. 2016, 44 (D1), D1202-D1213. 15. Chapman, K. E.; Coutinho, A. E.; Gray, M.; Gilmour, J. S.; Savill, J. S.; Seckl, J. R., The Role and Regulation of 11 Beta-Hydroxysteroid Dehydrogenase Type 1 in The Inflammatory Response. Mol. Cell. Endocrinol. 2009, 301, 123-131. 16. Zilka, N.; Kazmerova, Z.; Jadhav, S.; Neradil, P.; Madari, A.; Obetkova, D.; Bugos, O.; Novak, M., Who Fans the Flames of Alzheimer's Disease Brains? Misfolded Tau on the 22


Page 22 of 31

Page 23 of 31 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


Crossroad of Neurodegenerative and Inflammatory Pathways. J. Neuroinflamm. 2012, 9. 17. Salminen, A.; Kaarniranta, K.; Hiltunen, M.; Kauppinen, A., Histone Demethylase Jumonji D3 (JMJD3/KDM6B) at the Nexus Of Epigenetic Regulation of Inflammation and the Aging Process. J. Mol. Med.-JMM 2014, 92, 1035-1043. 18. Mitchell, J. A.; Larkin, S.; Williams, T. J., Cyclooxygenase-2: Regulation and Relevance in Inflammation. Biochem. Pharmacol. 1995, 50, 1535-42. 19. Ying, K. Y.; Chintalacharuvu, S. R.; Lu, H. F.; Nagpal, S., Vitamin D Receptor Modulators for Inflammation and Cancer. Mini-Rev. Med. Chem. 2005, 5, 761-778. 20. Dhar, T. G. M.; Zhao, Q.; Markby, D. W. Targeting the Nuclear Hormone Receptor ROR Gamma T for the Treatment of Autoimmune and Inflammatory Disorders. In Annual Reports in Medicinal Chemistry, Vol 48, Desai, M. C., Ed.; 2013; Vol. 48, pp 169-182. 21. Chen, Y.; de Lomana, M. G.; Friedrich, N. O.; Kirchmair, J., Characterization of the Chemical Space of Known and Readily Obtainable Natural Products. J. Chem. Inf. Model. 2018, 58, 1518-1532. 22. Viswanadhan, V. N.; Ghose, A. K.; Revankar, G. R.; Robins, R. K., Atomic Physicochemical

Parameters

for

Three

Dimensional

Structure

Directed

Quantitative

Structure-Activity Relationships. 4. Additional Parameters for Hydrophobic and Dispersive Interactions and Their Application for an Automated Superposition of Certain Naturally Occurring Nucleoside Antibiotics. J. Chem. Inf. Comp. Sci. 1989, 29, 163-172. 23. Cao, Y.; Charisi, A.; Cheng, L. C.; Jiang, T.; Girke, T., Chemminer: a Compound Mining Framework for R. Bioinformatics 2008, 24 (15), 1733-1734. 24. Guha, R., Chemical Informatics functionality in R. J. Stat. Softw. 2007, 18 (5). 25. Team, R. c. R: A Language and Environment for Statisctical Computing, R Foundation for Statistical Computing: Vienna, Austria, 2016. 26. Gaudet, P.; Michel, P. A.; Zahn-Zabal, M.; Britan, A.; Cusin, I.; Domagalski, M.; Duek, P. D.; Gateau, A.; Gleizes, A.; Hinard, V.; Rech de Laval, V.; Lin, J.; Nikitin, F.; Schaeffer, M.; Teixeira, D.; Lane, L.; Bairoch, A., The Nextprot Knowledgebase on Human Proteins: 2017 Update. Nucleic Acids Res. 2017, 45, D177-D182. 27. Shannon, P.; Markiel, A.; Ozier, O.; Baliga, N. S.; Wang, J. T.; Ramage, D.; Amin, N.; Schwikowski, B.; Ideker, T., Cytoscape: a Software Environment for Integrated Models of Biomolecular Interaction Networks. Genome Res. 2003, 13 (11), 2498-2504.

23



For Table of Contents use only

24


Page 24 of 31

Page 25 of 31 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


Table of Content 254x127mm (150 x 150 DPI)



Figure 1. Distribution of compound types (A) and origins (B) in the InflamNat dataset 332x141mm (150 x 150 DPI)


Page 26 of 31

Page 27 of 31 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. Distribution of physicochemical parameters for InflamNat and NPACT compounds. (A) Molecular weight; (B) ALogP; (C) Number of hydrogen bond donors; (D) Number of hydrogen bond acceptors. 175x134mm (300 x 300 DPI)



Figure 3. Hierarchical structural clustering of the InflamNat compounds. Representative compounds of the major clusters are shown. 261x181mm (150 x 150 DPI)


Page 28 of 31

Page 29 of 31 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 4. Scaffold Analysis. Clustering results (similarity cutoff = 0.6), comparing InflamNat and NPACT datasets and shown as cumulative frequency curves of (A) all-atom structures and (B) Bemis-Murcko scaffolds. Coordinates are shown for the turning points from which the clusters have only one member. (C) Representatives of the scaffold clusters with five or more members and reported in at least two different publications are presented for the InflamNat dataset, ranked from left to right, then top to bottom. 176x154mm (300 x 300 DPI)



Figure 5. Compound-target network 215x160mm (300 x 300 DPI)


Page 30 of 31

Page 31 of 31 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. (A) Degree distribution of nodes, defined as the number of directly connected neighbors. Gold: compound nodes; Turquoise: target nodes. (B) The 10 InflamNat compounds with the greatest number of reported targets. (C) Targets with the most reported active modulators in the InflamNat dataset. 173x149mm (300 x 300 DPI)


Chemical Space and Biological Target Network of Anti-inflammatory

Recommend Documents