Design and evaluation of dihydroisoquinolines as potent and orally bioavailable human cytomegalovirus inhibitors

Article abstract of DOI:10.1016/S0960-894X(00)00265-1

Following the identification of first pass metabolism issues with our recently described anti-HCMV compounds, the naphthyridines and isoquinolines, we have designed a class of novel metabolically stable and orally bioavailable anti-HCMV agents, the dihydroisoquinolines. (C) 2000 Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00265-1

Bioorganic & Medicinal Chemistry Letters 10 (2000) 1477±1480
Design and Evaluation of Dihydroisoquinolines as Potent and
Orally Bioavailable Human Cytomegalovirus Inhibitors
Laval Chan,* Tomislav Stefanac, Nathalie Turcotte, Zhuhan Hu, Yanmin Chen,
Jean Bedard, Suzanne May and Haolun Jin*
Â
BioChem Pharma Inc., 275 Armand-Frappier Blvd., Laval, Quebec, Canada H7V 4A7
Received 23 February 2000; accepted 4 May 2000
AbstractÐFollowing the identi®cation of ®rst pass metabolism issues with our recently described anti-HCMV compounds, the
naphthyridines and isoquinolines, we have designed a class of novel metabolically stable and orally bioavailable anti-HCMV
agents, the dihydroisoquinolines. # 2000 Elsevier Science Ltd. All rights reserved.
Infections by HCMV can have debilitating e€ects par-
ticularly in individuals whose immune system has been
weakened either by disease such as AIDS or by immu-
nosuppressive therapy following organ transplant. These
individuals are very often a‚icted with a variety of pro-
blems such as encephalopathy and retinitis. Although
current therapies can be e€ective, they are all associated
with various toxicities such as nephrotoxicity (foscarnet
and cidofovir, CDV) and myelotoxicity (ganciclovir,
GCV). In addition, administration of these drugs require
long intravenous iv infusion times; the negative impact
of these therapies on quality of life of HCMV patients
has been recognized as a problem.1 There has also been
growing evidence, albeit circumstantial, that HCMV
plays an important role in the process of atherogenesis.2
There is therefore a clear medical need for HCMV drugs
with a more desirable pro®le.
to ®rst pass metabolism.4 In vivo studies in mice con-
®rmed our hypothesis; no parent compound was detected
20 min after 1 was given orally. The same primary
metabolite was detected in vivo and in the S9 incubation
experiment. Following a preparative scale incubation
with mouse S9 homogenate, this metabolite was isolated,
characterized (NMR and MS) and was assigned the 1,6-
naphthyrid-5-one 3 structure. Unfortunately, 3 was devoid
of any anti-HCMV activity thus indicating that neither 1,
or 2, or their metabolites can serve as orally bioavailable
anti-HCMV agents.
Our recent e€ort in this area led to the identi®cation of
two novel classes of potent and selective anti-HCMV
agents,3 the 1,6-naphthyridines 1 and the 6-isoquinolines
2. We were also encouraged by the fact that both 1 and
2 lacked features that are associated with poor oral
bioavailability such as a guanine (GCV) or a phosphonic
acid (CDV and foscarnet) moiety; thus both 1 and 2 were
likely candidates in the identi®cation of oral HCMV
drugs. However, during preliminary pharmacological
evaluation of 1 and 2, it was found that these compounds
were not stable in mouse or monkey S9 liver preparations
indicating that these compounds are likely to be subject
In this communication, we will describe our e€orts to
circumvent ®rst pass metabolism issues while attempting
to retain antiviral activity. Our initial approach was to
block metabolism by installing a substituent at the site of
oxidation; the synthesis of these compounds is depicted
in Scheme 1. MCPBA oxidation of 1 gave N-oxide 4 in
good yields and reaction with phosphorus oxychloride5 or
with triphenylphosphonium bromide gave the chloride 5,
or bromide 7, respectively. Reaction of chloro derivative
*Corresponding author. Tel.: +1-450-978-7802; fax: +1-450-978-
7777; e-mail: chanl@biochempharma.com
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00265-1

1478
L. Chan et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1477±1480
fully and partially saturated systems were therefore
considered; in the latter case, at least some degree of pla-
narity can be maintained. Such strategy precludes the 1,6-
naphthyridine class of compounds, but the isoquinoline
analogues are amenable to this approach as the com-
pounds resulting from this modi®cation are the tetrahydro
or dihydroisoquinolines.
The synthesis of the desired compounds is depicted in
Scheme 2. The enolate of the known isoquinolinone 810
was prepared from reaction with LiHMDS and quenched
with Mander's reagent11 giving the b-keto ester as enol 9
in 52% yield. Reduction of enol 9 to b-hydroxy ester 10
was readily achieved by palladium catalyzed hydrogena-
tion, 10 was obtained in quantitative yield as a 4:1 mixture
of isomers (determined by 1H NMR). Conversion of
alcohol 10 to mesylate 11 followed by DBU catalyzed b-
elimination gave the a,b-unsaturated ester 12 in 88% yield.
Finally, the methyl ester was hydrolyzed with lithium
hydroxide and following acidi®cation, the desired acid 13
was obtained in 70% yield. The requisite amides were
then prepared either by the mixed anhydride method (iso-
propyl chloroformate) or by the use of EDCI and HOBt as
coupling agents. The reduction of the a,b-unsaturated
amide to the corresponding 5,6,7,8-tetrahydroisoquino-
line was readily achieved by palladium catalyzed hydro-
genation.
Scheme 1. Synthesis of 5-substituted naphthyridines. (a) 5 (R=Cl)
POCl3, rt, 66%; 7 (R=Br) PPh3, Br2, NEt3, CH2Cl2, 31%; (b) 6A
(R=F) KF, sulfolane, 230 ꢀC, 43%; 6B (R=NH2); (i) NaN3, DMF,
80 ꢀC, 100%; (ii) PPh3, xylene, 150 ꢀC then HCl, MeOH, 83%; 6C
(R=Me) Me4Sn, Pd(PPh3)4, LiCl, DMF, 20%; (c) 6D (R=OMe)
NaOMe, THF, 75%.
5 with potassium ¯uoride6 at 230 ꢀC yielded ¯uoro com-
pound 6A in 43% yield. Similarly, displacement of chlo-
ride with azide anion followed by the Staudinger7 reaction
gave amine 6B in good yields. The 5-methyl analogue 6C
was obtained in low yields under Stille8 conditions.
Finally, the methoxy 6D was obtained from bromide 7 in
75% yield. The monkey S9 stability of the methoxy 6D
and amino 6B analogues were determined and they were
found to be indeed resistant to oxidation. However, all
the 5 substituted 1,6-naphthyridines 5±7 as well as N-oxide
4 were devoid of HCMV activity (Table 1) indicating that
substitution cannot be tolerated at this position and this
approach was therefore abandoned. Interestingly, N-
oxide 4 was also found to be stable to monkey S9 thereby
indicating 4 is not an intermediate during the process of
oxidation.
The anti-HCMV activity and cytotoxicity of the com-
pounds were determined by plaque reduction assay and
inhibition of cell proliferation,12 respectively, and the
results are summarized in Table 2. The tetrahydroiso-
quinoline analogue 14 is devoid of anti-HCMV activity
suggesting that planarity of the left-hand part is indeed
a signi®cant factor. The corresponding 7,8-dihydroiso-
quinoline derivative 15 shows activity comparable to
parent compound. However, as with the isoquinoline
class, potency cannot be enhanced by the introduction of
bulkier alkoxy groups on the 20 position of the phenyl
ring. The absence of a hydrogen bond acceptor at position
The second strategy adopted was more successful; we
attributed the facile oxidation of the 5-position to the
presence of the bicyclic aromatic system and the nitrogen
atom at position 6. The reactivity of 1,6-naphthyridines
and similar systems at this position is well documented.9
We reasoned that if the right hand ring is no longer
aromatic, the reactivity of the 5-position would be sub-
stantially decreased and oxidation might be prevented.
We were also aware that the bicyclic portion of the lead
compounds 1 and 2 is planar and disruption of this
planarity may be detrimental to antiviral activity. Both
Table 1. Antiviral and cytotoxicity of 5-substituted naphthyridine
a
Compound
IC50 (mg/mL)
CC50b (mg/mL)
4
5
>25
>1 0
ꢁ10
ꢁ10
>1 0
>2 5
>2 5
100
ꢁ3.2
ꢁ 10
ꢁ12.5
>2 5
>5 0
ꢁ 25
6A
6B
6C
6D
7
Scheme 2. Synthesis of isoquinoline analogues. (a) LiHMDS, THF,
78 ꢀC, methyl cyanoformate, 52%; (b) Pd/C, MeOH, H2, 100%; (c)
MsCl, NEt3, CH2Cl2, 0 ꢀC, 100%; (d) DBU, CH2Cl2, 88%; (e) LiOH,
THF/water then aq HCl, 70%; (f) isopropyl chloroformate, NEt3,
RNH2, THF, 0 ꢀC or RNH2, EDCI, HOBt, DMF, 70±80%.
aMean of duplicate values (SD<15%), all experiments were per-
formed at least twice.
bMean of triplicate values (SD<15%).

L. Chan et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1477±1480
Table 2. Antiviral activity and cytotoxicity of in Hs68 cell line
Table 3. Pharmacokinetic studies in mice14,a
1479
a
Compound
R
IC50 (mg/mL) CC50b (mg/mL)
Dose (mg/Kg)
Route
Cmax
(mg/mL)
AUC(0-1)
(mg/mL min)
t1=2
(min)
BA
(%)
2
Ð
0.9
43
Compd 15
50
5
po
iv
12.47
6.94
734
57
42.3
14.2
14
15
16
17
18
25
100
Compd 20
5
50
5
po
po
iv
0.60
15.1
5.56
57.8
21
1762
94
15.2
72.5
14.2
43.9
22.5
60.4
0.64
0.68
0.71
>50
50
iv
2914
aMean of triplicate data (SD<8%).
12.5
In conclusion, it was found that the 7,8-dihydroiso-
quinoline class of compounds o€er a metabolically stable
and orally bioavailable alternative to the 1,6-napthyr-
idines and isoquinolines. The favorable pharmacoki-
netic pro®le of this class of compounds will allow us to
evaluate these compounds as potential oral anti-HCMV
agents. Further work to determine the mode of action
of this novel class of HCMV agents and to improve
potency is ongoing.
25
10
>25
19
20
0.23
0.1
>50
Acknowledgements
We wish to thank Annie St-Pierre for performing the
anti-HCMV plaque reduction and cytotoxicity assays
and Lyne Marcil for her assistance in the preparation of
this manuscript.
5
21
0.17
0.3
15
References and Notes
GCV
12.5
aMean of duplicate values (SD<15%), all experiments were per-
formed at least twice.
1. For recent reviews, see: (a) de Jong, M. D.; Galasso, G. J.;
Gazzard, B.; Griths, P. D.; Jabs, D. A.; Kern, A. R.;
Spector, S. A. Antiviral Res. 1998, 39, 141. (b) de Jong, M. D.;
Boucher, C. A. B.; Danner, S. A.; Gazzard, B.; Griths, P.
D.; Katlama, C.; Lange, J. M. A.; Richman, D. D.; Vella, S.
Antiviral Res. 1998, 37, 1. (c) Bernstein, K. BioCentury 1996, 4,
83. (d) Perry, C. M.; Davis, R. Pharmacoeconomics 1997, 12,
209.
bMean of triplicate values (SD<15%).
5 is believed to account for this lack of enhancement of
potency.13 We therefore turned our attention to non-
benzylic amides in an e€ort to obtain the right combination
for higher potency. Phenethylamines and other systems
with a 2 carbon linker between the amide moiety and an
aryl ring were found to provide good potency and selec-
tivity. For example, the phenethylamine 17 and the 2,5-
dimethoxy phenethylamine 19 showed activity similar to
that of 15 and a selectivity index of greater than 200 but
surprisingly the 2-methoxy derivative 18 was not active.
Other aryl systems such the tryptamine 20 and the
thiophene 21 derivatives were also found to be active.
2. Bruggeman, C. A.; Marjorie, H. J.; Nelissen-Vrancken, G.
Antiviral Res. 1999, 43, 135.
3. (a) Chan, L.; Jin, H.; Stefanac, T.; Lavallee, J.-F.;
Falardeau, G.; Wang, W.; Bedard, J.; May, S.; Yuen, L. J.
Med. Chem. 1999, 42, 3023. (b) Chan, L.; Jin, H.; Stefanac, T.;
Wang, W.; Lavallee, J.-F.; Bedard, J.; May, S. Bioorg. Med.
Chem. Lett. 1999, 9, 2583.
4. Parkinson, A. In Casarett and Doull's Toxicology: The
Basic Science of Poisons; 5th ed.; Klaassen, C. D., Ed.;
McGraw-Hill: New York, 1995, Chapter 2.
The stability of 15 and 20 towards monkey and mouse S9
homogenate was then determined. Both compounds were
found to be stable to these liver preparations, almost no
metabolites were detected under these conditions. Pre-
liminary in vivo experiments also indicated that these
compounds were stable and well absorbed when adminis-
tered po in mice (Table 3). In contrast, under similar con-
ditions, 1,6-naphthyridine 1 was virtually undetectable
after oral dosing. The oral bioavailability in mice (at a
dose of 50 mg/kg) of compound 20 was determined to
be 60.4%.14
5. Kobayashi, Y.; Kumada, I.; Sato, H. Chem. Pharm. Bull.
1969, 17, 1045.
6. Dmowski, W.; Wielgat, J. J. Fluorine Chem. 1988, 41, 241.
7. Kappe, Th.; Pfa€enschlager, A.; Stadlbauer, W. Synthesis
1989, 666.
8. Echavarren, A. M.; Stille, J. K. J. Am. Chem. Soc. 1987,
109, 5478.
9. Lowe, P. A. In Comprehensive Heterocyclic Chemistry;
Katritzcky A. R.; Rees, C. W., Ed.; Pergamon: Oxford; Vol. 2.
pp 581±627.
10. Lardenois, P.; Frost, J.; Dargazanali, G.; George, P.
Synth. Commun. 1996, 26, 2305.

1480
L. Chan et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1477±1480
11. (a) Mander, L. N.; Sethi, P. Tetrahedron Lett. 1983, 24,
5425. (b) Yanmato, M.; Yoshida, H.; Ikezawa, K.; Kohashi,
Y. Chem. Pharm. Bull. 1986, 34, 71.
12. Bedard, J.; May, S.; L'Heureux, L.; Stamminger, T.;
Copsey, A.; Drach, J.; Hu€man, J.; Chan, L.; Jin, H.; Rando,
R. F. Antimicrob. Agents Chemother. 2000, 44, 929.
13. Falardeau, G.; Chan, L.; Stefanac, T.; May, S.; Jin, H.;
Lavallee, J.-F. Manuscript in preparation.
concentrations of triplicate data at each time point. The area
under the concentration-time curve (AUC) was estimated by the
linear trapezoidal method from time zero to the last measured
sample time with extrapolation to in®nity by using the term-
inal slope (lz) generated by WinNonlin non-compartmental
model. The oral bioavailability (BA%) was calculated from
AUCpo;0-1=AUCiv;0-1at the corresponding dose. Half-life
(t1=2) was calculated from 0.693/lz. Chen, Y.; Mongrain, I.;
Richard, A.; Chan, L.; Winocour, P.; Hu, Z. Antiviral Chem.
Chemother. submitted.
14. Pharmacokinetic parameters were generated by WinNon-
Lin (Scienti®c Consulting Inc. Apex, N.C.) using mean plasma