Host dihydrofolate reductase (DHFR)-directed cycloguanil analogues endowed with activity against influenza virus and respiratory syncytial virus

Accepted Manuscript

Host dihydrofolate reductase (DHFR)-directed cycloguanil analogues endowed with

activity against influenza virus and respiratory syncytial virus

Michele Tonelli, Lieve Naesens, Sabrina Gazzarrini, Matteo Santucci, Elena Cichero,

Bruno Tasso, Anna Moroni, Maria Paola Costi, Roberta Loddo

PII:

S0223-5234(17)30344-6

10.1016/j.ejmech.2017.04.070

EJMECH 9421

DOI:

Reference:

To appear in: European Journal of Medicinal Chemistry

Received Date: 24 March 2017

Revised Date: 12 April 2017

Accepted Date: 25 April 2017

Please cite this article as: M. Tonelli, L. Naesens, S. Gazzarrini, M. Santucci, E. Cichero, B. Tasso, A.

Moroni, M.P. Costi, R. Loddo, Host dihydrofolate reductase (DHFR)-directed cycloguanil analogues

endowed with activity against influenza virus and respiratory syncytial virus, European Journal of

Medicinal Chemistry (2017), doi: 10.1016/j.ejmech.2017.04.070.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to

our customers we are providing this early version of the manuscript. The manuscript will undergo

copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please

note that during the production process errors may be discovered which could affect the content, and all

legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Graphical abstract

Host dihydrofolate reductase (DHFR)-directed cycloguanil analogues endowed with activity

against influenza virus and respiratory syncytial virus

Michele Tonelli,^a,*Lieve Naesens,^bSabrina Gazzarrini,^cMatteo Santucci,^dElena Cichero,^aBruno

Tasso,^aAnna Moroni,^cMaria Paola Costi,^dRoberta Loddo^e

ACCEPTED MANUSCRIPT

Host dihydrofolate reductase (DHFR)-directed cycloguanil analogues endowed with

activity against influenza virus and respiratory syncytial virus

Michele Tonelli,^a,*Lieve Naesens,^bSabrina Gazzarrini,^cMatteo Santucci,^dElena Cichero,^a

Bruno Tasso,^aAnna Moroni,^cMaria Paola Costi,^dRoberta Loddo^e

^aDipartimento di Farmacia, Università di Genova, Viale Benedetto XV 3, 16132 Genova,

b

Italy; Rega Institute for Medical Research, KU Leuven, Minderbroedersstraat 10, B-3000

Leuven – Belgium;

^cDepartment of Biosciences and National Research Council (CNR) Biophysics Institute

(IBF), University of Milan, Via Celoria 26, 20133 Milan, Italy;

^dDepartment of Life Sciences, University of Modena and Reggio Emilia, Via Campi 103,

41100 Modena, Italy

^eDipartimento di Scienze Biomediche, Sezione di Microbiologia e Virologia, Università di

Cagliari, Cittadella Universitaria, 09042 Monserrato (CA), Italy

In memory of Federica Novelli.

*To whom correspondence should be addressed:

Dr. Michele Tonelli:

tonelli@difar.unige.it.

tel:

+390103538867;

fax:

+390103538399. E-mail:

ACCEPTED MANUSCRIPT

Abstract

We have identified a series of 1-aryl-4,6-diamino-1,2-dihydrotriazines, structurally related to

the antimalarial drug cycloguanil, as new inhibitors of influenza A and B virus and

respiratory syncytial virus (RSV) via targeting of the host dihydrofolate reductase (DHFR)

enzyme. Most analogues proved active against influenza B virus in the low micromolar

range, and the best compounds (11, 13, 14 and 16) even reached the sub-micromolar potency

of zanamivir (EC₅₀= 0.060 µM), and markedly exceeded (up to 327 times) the antiviral

efficacy of ribavirin. Activity was also observed for two influenza A strains, including a virus

with the S31N mutant form of M2 proton channel, which is the most prevalent resistance

mutation for amantadine. Importantly, the compounds displayed nanomolar activity against

RSV and a superior selectivity index, since the ratio of cytotoxic to antiviral concentration

was >10,000 for the three most active compounds 11, 14 and 16 (EC₅₀~0.008 µM), far

surpassing the potency and safety profile of the licensed drug ribavirin (EC₅₀= 5.8 µM,

SI>43).

Keywords: 1-aryl-4,6-diamino-1,2-dihydrotriazine derivatives; anti-influenza A and B

viruses activity; anti-RSV activity; host (human) DHFR inhibition.

ACCEPTED MANUSCRIPT

1. Introduction

The Ortho- and Paramyxoviridae families comprise important respiratory pathogens, i.e.

influenza A and B viruses and respiratory syncytial virus (RSV), respectively. The acute

respiratory illnesses caused by these viruses represent major medical problems, given their

significant morbidity and potential mortality, particularly in vulnerable populations such as

small infants, elderly people or patients with underlying medical conditions [1]. Besides, the

threat for new influenza A virus pandemics (such as that of 2009 [2]) is a reason for global

and constant concern. Since the current arsenal of antiviral drugs to treat or prevent influenza

or RSV infections is quite limited [1,3], new therapeutics are highly needed. According to a

recommendation by the World Health Organization [4], attention should be given to

innovative agents with broad activity against diverse respiratory viruses.

Viruses, as obligate intracellular parasites, encode multiple virus-specific proteins

essential for replication, which also depends on critical interactions with host cell proteins.

Most approved antiviral drugs target unique proteins encoded by one virus or a range of

closely related viruses. This strategy is prone to selecting drug-resistance, particularly for

viruses, which possess high mutability (such as influenza virus) or require long-term therapy.

An alternative and relevant approach is to address host factors involved in the viral life cycle.

This type of inhibitors is anticipated to possess a markedly higher barrier for selecting drug-

resistant viruses and, furthermore, may display broad-spectrum antiviral activity when

dealing with a cellular target that is recruited by different viruses. Two host-directed antiviral

drugs are maraviroc, a CCR5 receptor antagonist approved for HIV therapy, and alisporivir, a

cyclophylin inhibitor that is undergoing Phase III tests for hepatitis C treatment [5]. Specific

host proteins were proven to be critical for the replication of diverse unrelated viruses [6], yet

ACCEPTED MANUSCRIPT

the array of possible cellular targets (the ‘virus-host interactome’) is continuously growing, as

recently reviewed for influenza [7] and RSV [8].

The first example of a broadly-acting antiviral drug is ribavirin, a nucleoside analogue

that was proposed to act directly at the level of the viral polymerase, although an indirect

effect via inhibition of the host-cell IMP dehydrogenase and depletion of the GTP pool seems

more plausible [9]. Another enzyme of the purine and pyrimidine pathways is dihydrofolate

reductase (DHFR) which catalyzes the reduction of dihydrofolate (DHF) to tetrahydrofolate

(THF), a crucial cofactor for the biosynthesis of IMP and thymidylate. Folate antagonists

interfering with DHFR can be applied in diverse pharmacological (i.e. antimalarial,

antibacterial and antineoplastic) settings [10-13]. The licensed antifolates trimethoprim [14],

pyrimethamine [15] and cycloguanil are potent inhibitors of bacterial and protozoal DHFR,

respectively, but only weak inhibitors of mammalian DHFR enzymes. On the other hand, the

drug methotrexate (MTX) is a potent unselective DHFRs inhibitor (Ki= 0.01 to 0.2 nM) [16],

because of its close structural similarity with dihydrofolic acid, the natural substrate of the

enzyme [17]. MTX shows a binding affinity to human DHFR (hDHFR) 1000-fold higher

than that of folic acid [16], explaining its clinical application as anticancer, anti-inflammatory

and immunosuppressive agent. Indeed, the MTX capability of affecting different intracellular

pathways has been very recently described, highlighting a rather complex mechanism of

action besides the most important therapeutic activity related to hDHFR inhibition [18].

Ongoing research efforts to develop novel antifolates for cancer chemotherapy and microbial

infections continue to be extensively reviewed [19]. Cycloguanil is the active metabolite of

the antimalarial drug proguanil (Paludrine® or Malarone®), that is approved for prophylaxis

and treatment of infections by Plasmodium vivax or falciparum. The species-selective activity

of cycloguanil (and pyrimethamine) has traditionally been attributed to higher affinity of the

drug for Plasmodium bifunctional dihydrofolate reductase-thymidylate synthetase (DHFR-

ACCEPTED MANUSCRIPT

TS) than for hDHFR [20]. Since 1991, cycloguanil and related 1-aryl-4,6-diamino-1,2-

dihydrotriazines were studied with the aim at treating Pneumocystis Carinii pneumonia [21],

searching for more selective inhibitors for P. Carinii DHFR over host DHFR (especially

human enzyme). Indeed, trimethoprim, the antifolate most widely used for that kind of

infection, was a poor inhibitor of P. Carinii DHFR (Ki= 280 µM) and showed about 6-fold

greater selectivity for hDHFR (Ki= 48 µM). Some 1-aryl-4,6-diamino-1,2-dihydrotriazines

exhibited a selective P. Carinii DHFR inhibition, while cycloguanil and some related

analogues (two of them corresponding to our compounds 11 and 14) were disclosed to bind

slightly stronger to hDHFR (cycloguanil, Ki= 43.0 µM) than to P. Carinii enzyme

(cycloguanil, Ki= 109.0 µM). Finally, the Author suggested that not only the expected

selective fungal enzyme inhibitors, but even compounds with higher species-selectivity

profile for hDHFR showed improvement over agents currently used to treat P. Carinii

infections.

In our previous studies, we focused on the design of antiviral agents by exploring

diverse and original chemotypes [22-23]. In the search of novel promising derivatives, in this

manuscript we reported for the first time the intriguing antiviral profile of cycloguanil (1).

Based on this information, we deemed interesting to proceed our work directed to the design,

synthesis and evaluation of the antiviral activity and cytotoxicity played by further

compounds, against a wide range of RNA and DNA viruses. In particular, we explored the

most relevant structure-activity relationship (SAR) exhibited within a series of structural

analogues of cycloguanil (1), including a number of 1-aryl-4,6-diamino-1,2-dihydrotriazines.

These compounds proved to inhibit virus reproduction targeting the host cell DHFR. In

particular, this new class of host-directed antiviral agents displayed promising dual activity

against influenza and respiratory syncytial virus (RSV).

ACCEPTED MANUSCRIPT

Selected compounds were also tested against hDHFR, in order to gain knowledge on their

efficacy versus the recombinant protein. Finally, docking studies were performed, in order to

support and disclose the most probable binding mode for these derivatives as hDHFR ligands

and foster lead optimization process.

2. Results and discussion

2.1. Design of host-directed antiviral agents

The compounds, object of the present study, are characterized by the 1-aryl-4,6-

diamino-1,2-dihydrotriazine scaffold, such as the cited cycloguanil, which was identified by

us as prototype (1) of a new class of antiviral agents exploiting a host DHFR inhibition

mechanism. In order to better clarify the intriguing effectiveness of tuning an adequate host

DHFR inhibition for the development of antiviral compounds endowed with optimized

antiviral activity and safety profiles, we proceeded with the design, synthesis and evaluation

of a series of cycloguanil (1) analogues. Furthermore, several studies about compound (I)

[27], containing a pyridopyrimidine system bioisostere of the (1) triazine core, as hDHFR-

targeting derivative, prompted us to deepen some SAR requirements within the series of

congeners (Fig. 1).

ACCEPTED MANUSCRIPT

Fig. 1. Chemical structure of the reference hDHFR inhibitor I and of the investigated series

of 1-aryl-4,6-diamino-1,2-dihydrotriazines

.

Thus, different functionalized 1-aryl-4,6-diamino-1,2-dihydrotriazines (2-28) were

designed (Fig. 2), by exploring the effect on biological activity as a result of the chemical

variation of the para-Cl substituent (R₃) on the phenyl ring and/or of the two methyl groups

(R₁, R₂) at C(2) of cycloguanil (1) with smaller/bulkier alkyl groups. Five compounds are

newly synthesized, while the others have been re-synthesized, to be tested as new antiviral

agents capable of inhibiting the host cell DHFR.

We also included in our investigation the cycloguanil isomeric anilinodihydrotriazine

29; proguanil (30), the metabolic precursor of cycloguanil; 1-(4-chlorophenyl)biguanide (31),

the second major metabolite of proguanil; and two 2,4-diaminopyrimidine planar analogues,

pyrimethamine (32) and trimethoprim (33) which are potent inhibitors of protozoal and

bacterial DHFR, respectively.

ACCEPTED MANUSCRIPT

R₁

R₂

H₂N

N

3

2

4

N^{5 1}N

R₁

R₂

R₃

6

R₃

NH₂

1 (Cycloguanil)

CH₃

C₂H₅

iPr

4-Cl

H

4-CH₃

4-OCH₃

4-F

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

4-Br

4-NO₂

4-COOH

4-COOEt

2-Cl

3-Cl

3-F

3-Br

3-CF₃

2,4-Cl₂

3,4-Cl₂

3,5-diCF₃

4-Cl

3-Cl

3-CF₃

4-Cl

-(CH₂)₄-

-(CH₂)₅-

H

3-Cl

3-CF₃

4-Cl

CH₃

nPr

C₆H₅

H

Fig. 2. Structures of the investigated compounds 1-33.

All the compounds underwent cell culture evaluation for cytotoxicity and antiviral

activity against a wide range of RNA and DNA viruses. With the aim of determining if the

host DHFR was targeted by our compounds series along the virus replication pathway, the

most promising compounds were tested against the recombinant protein of the hDHFR

enzyme. Successive docking studies allowed to find a molecular rationale for the mechanism

by which our compounds could inhibit the hDHFR.

ACCEPTED MANUSCRIPT

2.2. Chemistry

We deemed interesting to synthesize and evaluate the antiviral activity of a series of

4,6-diamino-1,2-dihydrotriazines, bearing in position 1 an aromatic ring, variously

substituted, and in position 2 one or two alkyl or aromatic moieties (Fig. 2).

Most of the tested dihydrotriazines were already described and re-prepared according to

cited references: syntheses of compounds 1-3, 15, 16, 18, 20, 23 and 28 [28], 4 [29], 5 [30],

14 [31], 19 [32] were achieved by a one-step acid-catalyzed cyclocondensation among an

aromatic amine, dicyandiamide and a carbonyl compound (3-component syntheses); for 26

and 27 was used the two-component syntheses which progresses in two steps by

condensation of a preformed arylbiguanide and a carbonyl derivative [33]. The progress of

the reaction was monitored by means of the negative result of colored copper complex

formed characteristically by biguanide with a freshly cuprammonium sulphate solution.

By the treatment of cycloguanil hydrochloride with an excess of alkali at reflux, it

undergoes an intramolecular rearrangement to the isomeric anilinodihydrotriazine compound

(29) [28].

Compounds 12 and 17 were known as free base, while in our synthetic route they

crystallize directly from the reaction mixture as pure hydrochlorides through the

aforementioned three component synthesis, thus they were experimentally described together

with the novel compounds 21, 22 and 25 (Scheme 1).

O

a

C

NH

C

12, 17

NH₂

R

O

+

NC N

NH₂

H

a

( X )

21, 22, 25

R= 3-Cl; 3-F; 3-CF₃;

3,5-diCF₃

X= CH₂, CH₂-CH₂

Scheme 1. ^aReagents and conditions: a) 1 equiv. conc. HCl, r.t., 24 h, 34-68%.

ACCEPTED MANUSCRIPT

All new compounds showed excellent analytical and spectroscopic data, in good

1

agreement with their structures (see Experimental Part). In the H NMR spectra the

protonated nitrogen of 4,6-diamino-1,2-dihydrotriazine nucleus gives rise to a net singlet near

δ 9 to 10, while the other amino groups yield signals in the range δ 7 to 8.

3. Biological evaluation

3.1. Inhibition of influenza virus. The antiviral activities of compounds 1-33 against influenza

A (H1N1 subtype) and B viruses are presented in Table 1. Our procedure [34] uses

exponentially growing MDCK cells in which the virus-induced cytopathic effect (CPE) is

monitored by either microscopy or cell viability testing. As shown in Table 1, there was

overall correlation between the antiviral EC₅₀values obtained by either method, indicating

the reliability of the observed antiviral effect. On the other hand, microscopic evaluation

revealed that several compounds produced cytostatic effect, yielding a minimum cytotoxic

concentration (MCC) of, for instance, 4 µM for cycloguanil (1). By comparison, its CC₅₀

value by cell viability assay was 52 µM, indicating that a true cytotoxic effect with manifest

cell killing was only seen at higher concentrations of 1. It is worth noting that cycloguanil is

the active metabolite of the prodrug proguanil, an antimalarial drug that is considered as safe

even when used during pregnancy [35]. Furthermore, most of the compounds were cytotoxic

only in MDCK cells while leaving unaffected (MCC > 100 µM) the human HeLa (Table 2)

and primate Vero cell lines (Table 3) used to grow the other viruses under investigation. In

addition, some selected compounds (1, 11, 13, 14, 16, 25 and 32) were assayed for

cytotoxicity against human airway epithelial Calu-3 cells confirming MCC values higher than

100 µM (data not shown).

ACCEPTED MANUSCRIPT

Table 1. Antiviral activity of compounds 1-33 in influenza virus-infected MDCK^acell cultures.

Anti-influenza activity: EC₅₀^b(µM)

A/Virginia/ATCC3/2009 (H₁N₁) A/Ned/378/05 (H₁N₁)

Cytotoxicity (µM)

Compound

B/Ned/537/05

d

CPE

MTS

CPE

MTS

CPE

MTS

MCC^c

CC₅₀

28

40

17

28

33

43

29

14

2.2

4.0

1.1

1.2

≥4

52

1

>100

53

2

53

34

48

34

11

1.2

≥4

3

59

49

63

37

8.2

2.0

14

4

59

36

41

24

8.4

1.5

≥4

5

39

19

21

7.3

2.9

1.0

≥4

6

>83

34

>83

0.67

21

>100

1.1

>100

0.85

nt

35

7.8

≥100

100

>100

≥100

≥0.2

≥0.8

≥0.03

≥0.2

>100

≥0.2

≥0.8

≥20

≥4

>100

1.4

7

>100

56

>100

19

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

1.0

0.10

1.7

0.040

0.2

>100

1.0

Nt

>100

29

>100

1.0

Nt

nt

0.016

0.015

>100

0.080

1.8

0.015

0.007

>100

0.030

0.8

1.1

0.60

>100

0.35

nt

6.1

≥83

1.4

>83

0.72

36

>100

0.55

Nt

>100

6.7

>100

>83

33

>100

>83

24

>100

45

>100

68

8.3

2.2

2.8

1.0

>83

>100

10

>83

49

>100

Nt

>100

nt

18

6.1

≥20

≥0.8

>100

≥20

>100

4.5

0.4

>100

49

Nt

nt

1.6

0.33

36

>100

Nt

>100

nt

>83

10

>100

>83

>100

>83

3.0

Nt

nt

19

4.4

>100

17

4.9

29

>100

ACCEPTED MANUSCRIPT

>83

124

>400

8.4

>83

211

>400

19

>100

15

>100

35

>100

2.0

>100

1.1

62

>100

41

28

>100

100

29

30

>100

≥4

31

32

>100

2.3

>100

2.2

>100

>400

4.9

>100

>400

3.9

>100

>400

≥100

>100

>400

>100

33

Amantadine

Rimantadine

Ribavirin

Zanamivir

0.32

10

0.050

7.6

2.3

6.8

0.80

0.20

0.060

0.055

b

^aMDCK: Madin-Darby canine kidney cells. EC₅₀: 50% effective concentration producing 50% inhibition of virus-induced cytopathic

c

effect (CPE), as determined by microscopy or by measuring the cell viability with the colorimetric formazan-based MTS assay. MCC:

d

minimum compound concentration producing a microscopically detectable alteration in normal cell morphology. CC₅₀: 50% cytotoxic

concentration, as determined by measuring the cell viability with the colorimetric formazan-based MTS assay. Values shown are the

mean of 3-6 determinations. nt = not tested.

ACCEPTED MANUSCRIPT

Table 2. Antiviral evaluation of compounds 1-33 in HeLa cell cultures.

Antiviral activity: EC₅₀^a(µM)

Cytotoxicity:

Compound

MCC^b(µM)

Respiratory

Vesicular

Coxsackie B4

syncytial virus

0.55

1.4

stomatitis virus

virus

>100

20

>100

1.8

>100

≥100

1

2

1.2

3

1.5

4

1.2

5

0.16

27

6

7

>100

48

8

9

10

0.0080

0.01

0.08

0.010

>100

0.0075

0.27

5.4

11

12

13

≥100

>100

≥100

>100

≥100

>100

≥100

14

15

16

17

18

0.80

8.0

19

20

0.22

0.20

>100

10.2

8.0

21

22

23

24

25

6.8

>100

20

26

>100

0.75

>100

2.0

27

28

>100

≥100

>100

>250

29

30

31

32

33

DS-10,000^c

5.8

50

85

Ribavirin

^aEC₅₀: 50% effective concentration producing 50% inhibition of virus-induced cytopathic effect (CPE), as determined

b

by microscopy. MCC: minimum compound concentration producing a microscopically detectable alteration in normal

cell morphology. ^cFor DS-10,000 (dextran sulphate of MW 10,000) concentrations are in µg/mL. Values shown are the

mean of two determinations.

ACCEPTED MANUSCRIPT

Table 3. Antiviral evaluation of compounds 1-33 in Vero cell cultures.

Antiviral activity: EC₅₀^a(µM)

Cytotoxicity:

Compound

MCC^b(µM)

Para-

influenza-3

virus

Reovirus-1 Sindbis

virus

Coxsackie

B4 virus

Punta

Yellow

Toro virus fever virus

>100

1

2

3

4

5

6

7

8

9

>100

6.8

9.5

>100

≥100

>100

≥100

10

11

12

13

14

2.1

>100

6.8

>100

15

>100

≥100

>100

16

17

18

>100

3.8

>100

19

20

21

22

>100

23

24

25

>100

20

>100

112

>100

5.4

>100

>250

>100

8.9

>100

46

>100

45

>100

0.40

>100

20

26

27

28

29

30

>100

≥20

>100

≥100

>250

31

32

33

>100

2.3

>100

2.3

DS-10,000^c

>250

29

>250

Ribavirin

Mycophenolic

acid

0.40

1.8

1.4

>100

20

2.3

>100

^aEC₅₀: 50% effective concentration producing 50% inhibition of virus-induced cytopathic effect (CPE), as determined

b

by microscopy. MCC: minimum compound concentration producing a microscopically detectable alteration in normal

c

cell morphology. For DS-10,000 (dextran sulphate of MW 10,000) concentrations are in µg/ml. Values shown are the

mean of two determinations.

ACCEPTED MANUSCRIPT

Analysis of the data in Table 1 revealed the following trends. Twenty-four out of thirty-three

compounds (73%) proved active against at least one influenza virus strain, and 30% displayed

activity against all three influenza virus strains.

Cycloguanil (1), the prototype of this class of 4,6-diamino-1,2-dihydrotriazine derivatives,

exhibited modest activity against influenza A/H1N1 viruses, while being 15-fold more potent

against influenza B virus (EC₅₀= 2.2 µM). Based on this interesting observation, the study was

oriented towards design and synthesis of more effective agents. Firstly, replacement of the chlorine

atom in position 4 of the aromatic ring was explored: the introduction of either electron-

withdrawing group isosteres of chlorine (F, 5; Br, 6) or electron-donor groups (CH₃, 3; OCH₃, 4)

was permitted without changing the activity, whereas polar electron-withdrawing groups (NO₂,

COOH and COOEt: 7-9) were detrimental. Having assumed that chlorine is a well-suited

substituent, we investigated its influence on the mono-substituted cycloguanil-related 2-Cl (10) and

3-Cl (11) isomers, and on the di-substituted 2,4-dichloro (15) and 3,4-dichloro derivatives (16).

Compounds 11 and 16, bearing a lipophilic electron-withdrawing group in position 3 of the

aromatic ring, proved to be strong inhibitors with sub-micromolar potency against influenza B virus

(EC₅₀= 0.10 and 0.080 µM, respectively) and only 2-fold lower activity against influenza A viruses.

In contrast, compound 10 was only poorly active against influenza B (EC₅₀= 56 µM), while the

2,4-dichloro substitution abolished the activity (15). Since these data clearly indicated that a

lipophilic electron-withdrawing atom in position 3 properly modulates the antiviral activity, we next

synthesized the 3-F (12), 3-Br (13) and 3-CF₃(14) analogues. Compound 14 showed an activity

profile comparable to that of 11, whereas bromine derivative 13 was found to be the best influenza

B inhibitor among the entire series, having an EC₅₀of 0.016 µM. In the next step, we varied the two

methyl groups on C(2) of the 4,6-diamino-1,2-dihydrotriazine scaffold, by introducing bulkier alkyl

groups than the two ones derived from acetone. This was achieved by performing condensation

with methyl ethyl ketone (for 18), methyl isopropyl ketone (for 19), cyclopentanone (for 20-22), or

ACCEPTED MANUSCRIPT

cyclohexanone (for 23-25). The progressive increase in the substituent’s size led to a proportional

decrease in the activity against influenza A virus, while having much less impact for influenza B,

even when combined with 3-Cl or 3-CF₃substitutions.

Among the three compounds with mono-substitution at position C(2), the CH₃group (26,

EC₅₀= 17 µM) was tolerated to keep influenza B inhibition, whilst the n-propyl chain (27) and

phenyl ring (28) were detrimental.

Since in vivo application of cycloguanil is through its prodrug proguanil (30), we checked

whether also the prodrug and 1-(4-chlorophenyl)biguanide (31) possess anti-influenza activity. The

two main metabolites of the antimalarial drug proguanil are the active form cycloguanil and the

inactive 1-(4-chlorophenyl)biguanide, which constitute about 30% and 23%, respectively, of the

total drug concentration in plasma.³²Both compounds 30 and 31 were shown inactive, alike the

anilinedihydrotriazine isomer 29 of cycloguanil. The 2,4-diaminopyrimidines trimethoprim and

pyrimethamine were chosen as fully unsaturated cycloguanil analogues: only pyrimethamine (32)

was able to inhibit one of the two influenza A viruses plus influenza B virus (EC₅₀= 15 and 2.0 µM,

respectively), thus matching cycloguanil in terms of potency.

Of the two influenza A virus strains tested, the A/Ned/378/05 strain carries a wild type M2

protein. Against this strain, the three most active cycloguanil analogues (11, 14 and 16) displayed

comparable potency as zanamivir and rimantadine, while surpassing ribavirin and amantadine. The

A/Virginia/ATCC3/2009 strain, which has an S31N mutant M2 protein alike most currently

circulating influenza A strains, had marginal to no sensitivity to the adamantane-based M2-

blockers. This strain was strongly susceptible to compounds 11, 14 and 16 with EC₅₀values of

about 1 µM, which is 2-fold and 8-fold better than the values of zanamivir and ribavirin,

respectively.

Whereas the antiviral EC₅₀values were nicely comparable for the two influenza A strains

tested, all active analogues displayed higher activity (lower EC₅₀value) against influenza B virus. In

ACCEPTED MANUSCRIPT

fact, 73% of all compounds inhibited influenza B virus with EC₅₀values in the range of 0.016 to 29

µM. Hence, compared to influenza A, influenza B virus appears more sensitive to a reduction in the

THF pool. This could point to a higher sensitivity of influenza B virus to nucleotide imbalances

[perhaps related to the slower growth kinetics of influenza B compared to influenza A virus [36],

although this seems contradicted by the finding that the GTP-depleting agent ribavirin displayed

similar EC₅₀values for influenza A and B viruses (Table 1).

Recently, Balgi et al. [37] used a high-throughput yeast growth restoration assay to screen

250,000 compounds. They identified three tetrahydrotriazines with antiviral activity in a plaque

reduction assay and, with only one exception, efficacy (when assessed at 100 µM) against the wild

type A/M2 proton channel in a two-electrode voltage clamp (TEVC) assay. Based on the chemical

resemblance, we deemed interesting to assess if our 4,6-diamino-1,2-dihydrotriazines (1, 14 and 16)

were also able to inhibit the A/M2 proton channel expressed in Xenopus oocytes using the TEVC

assay. At 100 µM, neither of the three compounds significantly inhibited the wild type or S31N

mutant M2 channel [38], thus excluding M2 inhibition as the antiviral mechanism of action in

virus-infected MDCK cells (data not shown).

3.2. Inhibition of RSV. Several compounds produced promising antiviral effect in a similar CPE

reduction assay for RSV in HeLa cells (Table 2). Since most compounds were not cytotoxic at 100

µM (the highest concentration tested), the selectivity index, i.e. ratio of MCC to EC₅₀, was

calculated to be at least 10,000 for the most active analogues 11, 14 and 16. These compounds

inhibited RSV replication at nanomolar concentrations, far surpassing the antiviral activity of

ribavirin (EC₅₀= 5.8 µM, SI> 43). The latter is the only small molecule drug to treat RSV

infections, but, as a consequence of its limited efficacy, the need of prolonged aerosol

administration and the risk of toxicity, its use is limited to children at high risk [39]. Therefore,

effective and safe drugs are strongly needed to treat severe RSV-linked respiratory pathologies

ACCEPTED MANUSCRIPT

which can also affect adults and, particularly, elderly. Thus, the highly potent compounds 11, 14

and 16 could be interesting lead compounds for further chemical derivatization and development of

improved RSV inhibitors. Similar to what was observed for influenza virus, the drugs cycloguanil

(1) and pyrimethamine (32) were found to be equipotent RSV-inhibitors with EC₅₀values of 0.55

and 0.75 µM, respectively, which is nearly one order of magnitude superior to the value of

ribavirin.

The compounds with no activity against RSV (i.e. 8, 9, 15, 27-31, 33) were also inactive

against influenza. The three most active RSV inhibitors, i.e. 11, 14 and 16 (EC₅₀~0.008 µM; Table

2) also had the best EC₅₀values (Table 1) for influenza. Hence, it is clear that the same biochemical

mechanism explains the inhibitory effect of this class of compounds towards RSV and influenza

virus.

As for the other RNA viruses tested in either HeLa (Table 2) or Vero cells (Table 3), 11-14

and 16 also inhibited Reovirus-1, a member of the Rotaviridae. They were found active in the low

micromolar range without visible cytotoxicity (MCC > 100 µM). Compound 32 exhibited broader

antiviral activity but only at concentrations quite close to those producing cytotoxicity (MCC ≥20

µM) (Tables 2 and 3). Neither of the compounds inhibited the replication of yellow fever virus

(Table 3), a member of the Flaviviridae which was described in another report as sensitive to

methotrexate, the prototype inhibitor of mammalian DHFR enzymes [40]. Finally, we evaluated in

human embryonic lung (HEL) fibroblast cells, several DNA viruses of the herpes-, adeno- or

poxvirus families, plus coronavirus 229E (an RNA virus), but neither of the compounds 1-33

proved active against any of these viruses (data not shown).

3.3. Dihydrofolic acid reverses the antiviral effect. Since host cell DHFR inhibition seemed the

plausible explanation for the observed antiviral effect, we performed a combination experiment in

which RSV-infected HeLa cells were exposed to compound 14 in combination with different

ACCEPTED MANUSCRIPT

concentrations of the natural DHFR substrate dihydrofolic acid. The antiviral EC₅₀value of 14

gradually increased in function of the concentration of added dihydrofolic acid, this increase being

as high as 267-fold when 14 was combined with 100 µM dihydrofolic acid. Intriguingly, the

cytotoxicity of 14 was not affected, since the MCC value was 20 µM independent of whether

dihydrofolic acid was present or not. This suggests that disruption of the host DHFR reaction has a

much higher impact on virus replication than on cell growth, which concurs with the promising

antiviral selectivity of our host-directed DHFR inhibitors.

Table 4. Drop in anti-RSV activity of compound 14 when combined with dihydrofolic acid

Concentrationof

dihydrofolic acid

Antiviral activity of compound 14 Cytotoxicity of compound 14

MCC^c(µM)

EC₅₀^a(µM)

Fold-increase^b

(µM)

400

40

0.80

0.40

267

133

20

16

6.4

2.6

1.0

0.4

0

0.090

0.030

0.006

0.004

0.003

30

10

2

1

20

1

^aEC₅₀: 50% effective concentration producing 50% inhibition of virus-induced cytopathic

b

c

effect (CPE), as determined by microscopy. Ratio of EC₅₀(i) to EC₅₀( ); MCC: minimum

0

compound concentration producing a microscopically detectable alteration in normal cell

morphology. Values shown are the mean of three determinations.

3.4. Inhibition of hDHFR enzyme. Finally we assayed some selected compounds (1, 11, 13, 14, 16,

25 and 32) against hDHFR in order to confirm that the observed antiviral activity against RSV was

clearly influenced by this enzyme inhibition. Concerning the inhibition experimental data of the test

compounds on hDHFR (Table 5), SARs moved with the same trend compared to antiviral activity

in cell-based assays: the Ki values ranged between 0.07-0.13 µM for the best antiviral compounds

(13, 11, 16, 14 in decreasing order) while the bulkier spiro-compound 25 was less effective with Ki

ACCEPTED MANUSCRIPT

value equal to 13.17 µM. Cycloguanil (1) and pyrimethamine (32) displayed the same degree of

potency, about 5-fold lower than that of the most potent compounds (11, 13, 14 and 16).

Interestingly, for almost all the compounds tested (1, 11, 13, 14, 16, 25 and 32) a similar trend can

be observed between the Ki against hDHFR and the antiviral activity against Influenza B virus and

Respiratory Syncytial Virus, thus supporting the concept that the observed antiviral effect is

sustained also by the hDHFR inhibition (Fig. 3).

Moreover, kinetic inhibition studies of cycloguanil (1) were performed and a competitive

inhibition pattern was observed towards the substrate for binding to the human enzyme (Fig. 1S).

Thus, these 1-aryl-4,6-diamino-1,2-dihydrotriazine derivatives proved to efficiently work

weakening the virus replication machinery by a direct inhibitory effect versus the host (human)

DHFR.

Table 5. Inhibition constant (K_i) of compounds 1, 11, 13, 14, 16, 25 and 32 on hDHFR enzyme.

Compound

1

Ki (µM)

0.41

0.10

11

0.07

13

0.13

14

0.11

16

13.17

0.47

0.011

25

32

Compound I^a

^aPyrido[2,3-d]pyrimidine antifolate (I) (Figure 1) co-crystallized with hDHFR [27] (pdb code = 4QHV; see molecular

modelling studies).

ACCEPTED MANUSCRIPT

Fig. 3. Scatter Plot of Ki values (hDHFR) versus EC₅₀towards Influenza B virus and Respiratory

Syncytial Virus. Data plotted are reported in Table 1 and Table 2. All the values are expressed in

logarithm scale.

4. Molecular modelling studies

Molecular modelling studies were performed on the hDHFR inhibitors identified (1, 11, 13,

14, 16, 25 and 32) to explore the structural basis of the interaction between the mentioned

compounds and the human enzyme. The docking studies were performed using the X-ray

crystallographic structure of the hDHFR, in complex with a pyridopyrimidine-based inhibitor (I)

(pdb code = 4QHV; resolution = 1.61 Å) [24]. The human DHFR inhibitor, compound I, was

considered as positive control (Fig. 1).

The main issues to be addressed were to clarify, through docking studies of 1, 11, 13, 14, 16

and 25, the role played by the 4,6-diamino-1,2-dihydrotriazine nucleus and of the hydrophobic

framework, including the phenyl ring and the R1-R3 substituents, with respect to the 2,4-

diaminopyrimidine core and to the phenyl-substituted pyridine-3-amine moiety of I. In addition,

pyrimethamine (32) was also submitted to docking calculations, having shown a similar antiviral

behaviour to cycloguanil.

ACCEPTED MANUSCRIPT

As shown in Figure 4A, I was engaged in H-bonds between the two NH₂substituents placed

onto the pyrimidine core and the I7 and E30 carbonyl group and side-chain, respectively. The

whole, planar bicyclic ring was involved in π−π stacking with Y33 and F34, while the isopropyl

phenyl ring was projected towards F31, I60, P61, displaying Van der Waals contacts.

As consequence, these kinds of contacts efficiently stabilized I within the hDHFR binding

site, leading the complex to a favourable value of the estimated binding affinity (hDHFR-I ∆G= -

28.0 kJ/mol), in harmony with the high compound potency profile (Ki= 11 nM, Table 5).

Fig. 4. A) details of the X-ray crystallographic complex hDHFR - inhibitor I (C atom; yellow); B)

Docking pose of cycloguanil (1) (C atom; white) within the X-ray crystallographic structure of the

human DHFR in complex with the inhibitor I (C atom; yellow). The most important residues are

labelled and coloured by atom type.

Based on our docking calculations (Table 6), when small bulky alkyl groups are chosen for

R1 and R2, the presence of a meta substituent placed in R3, rather than at the ortho and para

positions of the phenyl ring, proved to be preferred. Accordingly, cycloguanil (1) was characterized

by weak H-bonds with I7 and E30, while the 4-chlorophenyl ring moved towards T56 and Y121, on

the opposite side of the cavity occupied by the isopropyl phenyl ring of the reference inhibitor (Fig.

ACCEPTED MANUSCRIPT

4B). A comparable docking mode was observed for pyrimethamine (32). As shown in Table 6, the

related complexes displayed quite adequate and comparable values of predicted binding affinity

energies (hDHFR-1 ∆G= -7 kJ/mol; hDHFR-32 ∆G= -8 kJ/mol), turning in moderate affinity

toward the human enzyme (1 Ki = 0.41 µM; 32 Ki = 0.47 µM).

Interestingly, the most promising derivatives 11, 13, 14, 16 shared a common docking mode,

exhibiting the required H-bonds with I7 and E30, by means of the two NH₂groups of 1,2-

dihydrotriazine scaffold (as shown for compound 13 in Fig. 5A). In addition, the dimethyl

substitution in R1 and R2 of 13, and also the presence of a 3-Br-phenyl ring, proved to be

particularly effective to properly mimic the same positioning displayed by I within the X-ray

crystallographic structure of hDHFR.

Fig. 5. A) Docking pose of compound 13 (C atom; pink) within the X-ray crystallographic structure

of the human DHFR in complex with the inhibitor I (C atom; yellow); B) Docking pose of

compound 25 (C atom; tan) within the X-ray crystallographic structure of the human DHFR in

complex with the inhibitor I (C atom; yellow). The most important residues are labelled and

coloured by atom type.

ACCEPTED MANUSCRIPT

As consequence, all of them were efficiently characterized by the most important and crucial

bonds with the biological target, as previously discussed for compound I, and therefore experienced

the most favourable predicted binding affinity profiles, being in agreement with the experimental

data. In particular, compound 13 (hDHFR-13 ∆G= -19 kJ/mol) proved to be the most promising (Ki

= 0.07 µM).

Conversely, the presence of bulky group or the introduction of a cycloalkyl ring in R1 and

R2, moved the derivative toward a quite turned docking mode (if compared with I), as shown for

compound 25 (Fig. 5B). Indeed, only one of the two NH₂groups onto the dihydrotriazine ring was

able to mimic the role played by those of I, detecting only one H-bond with E30. In addition, the

bioisosteric replacement of the pyridopyrimidine scaffold with the smaller dihydrotriazine one,

inevitably impaired the ability of the compound to display the same pattern of hydrophobic

contacts, previously mentioned for I. This compound was the only one of the series here proposed

to exhibit this kind of docking mode, losing key-contacts such as one H-bond with I7 and several

Van der Waals and π-π stacking. Conceivably, this positioning compromises the ability of

compound 25 to gain effective and crucial bonds with the human biological target, turning in lower

affinity values (25 Ki = 13.17 µM) if compared with the other congeners (1, 11, 13, 14, 16: Ki =

0.07-0.41 µM). A perspective of the predicted binding affinity values for the enzyme in complex

with any selected docking pose also supports the experimental data since the hDHFR-25 complex

exhibited the worse predicted value (hDHFR-25 ∆G= -1 kJ/mol) with respect to those including 1,

11, 13, 14, 16 (∆G from -20.0 to -7.0 J/mol).

On all these basis, it was expected that the 4,6-diamino-1,2-dihydrotriazine nucleus,

properly decorated with small, rigid and planar groups, could represent an interesting scaffold able

to fulfil the minimal pharmacophore requirements to exert hDHFR inhibition.

ACCEPTED MANUSCRIPT

Table 6. Binding affinity values obtained by molecular docking studies of compounds 1, 11, 13, 14,

16, 25, 32.

Receptor-Ligand

Complex (LeadIT)

Binding Affinity Energy

Receptor-Ligand

Complex (LeadIT)

Binding Affinity Energy

∆G (kJ/mol)

hDHFR-1

hDHFR-11

hDHFR-13

hDHFR-14

-7.0

hDHFR-16

hDHFR-25

hDHFR-32

hDHFR-I

-20.0

-1.0

-15.0

-19.0

-17.0

-8.0

-28.0

5. Conclusions

This report describes the discovery of a new class of host-directed antiviral agents characterized by

a 1-aryl-4,6-diamino-1,2-dihydrotriazine scaffold, responsible for a host (human) DHFR inhibition

mechanism. Host-targeting antivirals represent an alternative and emerging strategy to address host

factors involved in virus life cycle. This type of inhibitors could show a markedly higher barrier for

selecting drug-resistant viruses and, furthermore, display broad-spectrum antiviral activity when

interacting with a common cellular target that is recruited by different viruses. The interesting dual

activity of the 1-aryl-4,6-diamino-1,2-dihydrotriazines against influenza and respiratory syncytial

viruses, via inhibition of the cellular (human) DHFR enzyme, points to this host factor as a new

therapeutic target for these two respiratory viruses. In fact, reversal effect on antiviral activity has

been demonstrated in RSV-infected HeLa cells exposed to compound 14 in combination with

different concentrations of dihydrofolic acid, such as natural DHFR substrate. The most promising

compounds, tested against the recombinant protein of the hDHFR, also confirmed to bind this

enzyme in the sub-micromolar range. Kinetic inhibition studies of cycloguanil showed a

competitive inhibition behavior, and docking studies disclosed the most probable binding mode for

this class of compounds as hDHFR ligands. Notably, the antiviral activity against RSV and

influenza B viruses in cell-based assays, the Ki values on the recombinant hDHFR enzyme and,

also the estimated binding affinity (∆G) shared an interesting comparable SAR trend.

ACCEPTED MANUSCRIPT

The possibility to suppress influenza virus by interfering with the purine or pyrimidine

pathway was proposed for a few other enzymes [9,41] but our study is the first to identify the

relevance of host (human) DHFR in antiviral therapy.

Most of the compounds proved effective inhibitors of influenza B virus, giving sub-

micromolar activity in case of the most potent analogues 11, 13, 14 and 16, which compare

favorably with the licensed antiviral drugs zanamivir, even exceeding the antiviral efficacy of

ribavirin. Compounds 11, 14 and 16 also possessed low micromolar activity against influenza A

virus and, even more importantly, nanomolar activity against RSV with a selectivity index of at

least 10,000, far surpassing the antiviral activity of the reference drug ribavirin.

These novel host-directed 1-aryl-4,6-diamino-1,2-dihydrotriazine derivatives combine

promising antiviral activity with low cost and easily accessible chemical synthesis. Therefore, our

results provide the foundation to further explore the structure-activity relationships of this class of

compounds versus DHFR (especially human enzyme), and the apparently important role of this host

enzyme in influenza and RSV virus replication.

ACCEPTED MANUSCRIPT

6. Experimental section

6.1. Chemistry.

6.1.1. General methods. Chemicals, solvents and commercially available compounds [Proguanil

(30), 4-chlorophenylbiguanide (31), pyrimetamine (32) and trimethoprim (33)] were purchased

1

from Sigma-Aldrich (Milan, Italy). Mps: Büchi apparatus, uncorrected. H NMR and ¹³C NMR

spectra were recorded on a Varian Gemini-200 instrument at 200 and 50 MHz, respectively;

DMSO-d6; δ in ppm rel. to Me₄Si as internal standard. J in Hz. Elemental analyses were performed

on a Carlo Erba EA-1110 CHNS instrument in the Microanalysis Laboratory of the Department of

Pharmacy of Genoa University.

6.1.2. General method for the synthesis of 4,6-diamino-1,2-dihydrotriazine derivatives.

A solution of the substituted aniline hydrochloride (4.36 mmol) in 40 mL of acetone and 5 mL of

MeOH was reacted at r.t. with dicyandiamide (4.58 mmol, 1.05 equiv.) with stirring for 24 h.

Compounds separated directly from reaction mixture, thus they were collected by filtration, washed

with acetone and recrystallized from the same solvent. Only in the case of compound 17, the

solution was evaporated to dryness under vacuum and the residue was recrystallized from MeOH.

6.1.3. 4,6-Diamino-2,2-dimethyl-1-(3-fluorophenyl)-1,2-dihydrotriazine hydrochloride (12)

Yield 68%. Mp 209-210°C (acetone). ¹H NMR (200 MHz, DMSO-d₆): 9.45 (s, 1H, ⁺NH, exchanges

with D₂O), 7.92-7.17 (m, 4 arom H and 7.74 br s, 3H, amino groups, exchange with D₂O,

superimposed signals), 6.46 (br s, 1 H, amino group, exchanges with D₂O), 1.34 (s, 6 H, 2Me). ¹³C

NMR (50 MHz, DMSO-d₆): 164.6, 159.7, 157.4, 156.6, 136.0, 135.8, 131.3, 131.1, 126.0, 117.5,

117.0, 116.7, 116.3, 69.3, 26.8. Anal. Calcd for C₁₁H₁₄FN₅• HCl: C, 48.62; H, 5.56; N, 25.77.

Found: C, 48.70; H, 5.70; N, 25.96.

ACCEPTED MANUSCRIPT

6.1.4. 4,6-Diamino-2,2-dimethyl-1-(3,5-ditrifluoromethylphenyl)-1,2-dihydrotriazine hydrochloride

(17)

Yield 51%. Mp 195-198°C (MeOH). ¹H NMR (200 MHz, DMSO-d₆): 9.50 (s, 1H, ⁺NH, exchanges

with D₂O), 8.26 (s, 1 arom H), 8.20 (s, 2 arom H), 7.96-7.20 (br s, 3H, amino groups, exchange

with D₂O), 6.76 (br s, 1H, NH₂, exchanges with D₂O), 1.38 (s, 6 H, 2Me). ¹³C NMR (50 MHz,

DMSO-d₆): 157.3, 156.6, 136.6, 132.0, 131.6, 131.3, 125.2, 123.4, 119.7, 69.6, 27.0. Anal. Calcd

for C₁₃H₁₃F₆N₅• HCl: C, 40.06; H, 3.62; N, 17.97. Found: C, 39.77; H, 3.91; N, 18.30.

6.1.5. 1-(3-Chlorophenyl)-2,2-cyclotetramethylene-4,6-diamino-1,2-dihydrotriazine hydrochloride

(21)

1

+

Yield 38%. Mp 241°C (acetone). H NMR (200 MHz, DMSO-d₆): 9.34 (s, 1H, NH, exchanges

with D₂O), 7.90-7.05 (m, 4 arom H and 3H, amino groups, exchange with D₂O), 6.56 (br s, 1H,

13

amino group, exchanges with D₂O), 1.94-1.20 (m, 8 H, (CH₂)₄). C NMR (50 MHz, DMSO-d₆)

δ

157.7, 157.4, 136.0, 133.8, 131.4, 129.6, 128.5, 78.6, 35.8, 20.1. Anal. Calcd for C₁₃H₁₆ClN₅• HCl:

C, 49.69; H, 5.45; N, 22.29. Found: C, 49.42; H, 5.55; N, 22.09.

6.1.6.

2,2-Cyclotetramethylene-4,6-diamino-1-(3-trifluoromethylphenyl)-1,2-dihydrotriazine

hydrochloride (22)

Yield 37%. Mp 225-226°C (acetone). ¹H NMR (200 MHz, DMSO-d₆): 9.44 (s, 1H, ⁺NH, exchanges

with D₂O), 7.98-7.23 (m, 4 arom H and 3H, amino groups, exchange with D₂O), 6.61 (br s, 1H,

13

amino group, exchanges with D₂O), 1.98-1.25 (m, 8 H, (CH₂)₄). C NMR (50 MHz, DMSO-d₆):

157.8, 157.5, 135.5, 134.0, 131.1, 126.7, 126.3, 78.6, 35.9, 20.1. Anal. Calcd for C₁₄H₁₆F₃N₅• HCl:

C, 48.35; H, 4.93; N, 20.14. Found: C, 48.11; H, 5.15; N, 20.33.

Article Doi

Host dihydrofolate reductase (DHFR)-directed cycloguanil analogues endowed with activity against influenza virus and respiratory syncytial virus

DOI: 10.1016/j.ejmech.2017.04.070

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.ejmech.2017.04.070

Full text of DOI:10.1016/j.ejmech.2017.04.070

Products guided by the article

R&D Labs maybe for 5405-65-2

Relevant to this article

Hot Product