Pyridoxine hydroxamic acids as novel HIV-integrase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2016.01.028

A series of pyridoxine hydroxamic acid analog bearing a 5-aryl-spacers were synthesized. Evaluation of these novel HIV integrase complex inhibitors revealed compounds with high potency against wild-type HIV virus.

Full text of DOI:10.1016/j.bmcl.2016.01.028

Bioorganic & Medicinal Chemistry Letters xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Pyridoxine hydroxamic acids as novel HIV-integrase inhibitors
Brent R. Stranix a, , Jinzi J. Wu a,y, Guy Milot a, Françis Beaulieu a,§, Jean-Emanuel Bouchard a,
⇑
Kristine Gouveia a, André Forte a,à, Seema Garde a, Zhigang Wang a, Jean-François Mouscadet b,–
,
Olivier Delelis b, Yong Xiao a
a Ambrilia Biopharma Inc., 1000 Ch. du Golf, Verdun, Montreal, QC H4P 2R2, Canada
b BPA, ENS de Cachan, CNRS, Cachan, France
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of pyridoxine hydroxamic acid analog bearing a 5-aryl-spacers were synthesized. Evaluation of
Received 5 November 2015
Revised 7 January 2016
Accepted 11 January 2016
Available online xxxx
these novel HIV integrase complex inhibitors revealed compounds with high potency against wild-type
HIV virus.
Ó 2016 Published by Elsevier Ltd.
Keywords:
HIV integrase inhibitor
Pyridoxine
Hydroxamic acid
Active site binding
Human immunodeficiency virus (HIV) integrase inhibitors are a
new component to the anti-HIV chemotherapy pharmacy, the cur-
rent treatment for acquired immunodeficiency syndrome (AIDS).1
This class of compounds inhibits the integration of the viral gen-
ome into the host’s and thus the infectious process. Although inte-
grase inhibitors have been well received, new problems have
recently been identified.2,3 The rapid emergence of several viral
strains resistant to one or more of the drugs currently available
or in trials for the treatment of AIDS has now become the most
important issue in the treatment of HIV infection.4 We recently
discovered HIV integrase inhibiting compounds of general struc-
ture as shown in Figure 1. We had previously found that pyridox-
ine or pyridoxal phosphates could serve as an efficient backbone
scaffold for the synthesis of such inhibitors albeit by an unknown
mechanism of action.5
account for possible mutational changes in the active site, two dif-
ferent binding modes for the molecules were devised; in the first
mode, the metal chelating ligand would involve the C-3 phenolic
ligand while in the second mode would engage the pyridyl moiety
directly.
One of the concerns about using a heterocyclic hydroxamic acid
was the association of this functional group with toxicity. Studies
in the ability of the hydroxamic acid ability to induce host DNA
mutations have show this to proceed through formation of iso-
cyanates resulting from Lossen rearrangement reaction.6 This iso-
cyanate is presumed to acylate genetic material causing the
mutations. We surmised that using a ortho hydroxyl group would
scavenge this putative isocyanate to form an azabenzoxazolone.
Our first approach to explore new designs was to build a struc-
ture activity relation on spacers and aryl groups at the 5 position of
the pyridoxin binding ligand.7,8 Scheme 1 shows the different syn-
thetic routes taken to create the diversity of this class of molecules
from a pivotal core intermediate. The core intermediate I was syn-
thesized according to literature precedent9 in good yields, giving us
a benzylic alcohol as a point of entry. The coupling to phenols using
Mitsunobu conditions to yield a variety of aryl ethers II. Oxidation
of the benzylic alcohol to the corresponding aldehyde III was per-
formed using MnO2 in CHCl3 with high yield of crystalline com-
pound. This aldehyde could then be further coupled to activated
carbanions to yield alkenes IV and alkanes V, or amines to give ani-
lides and alkylamines VI. Mild oxidation of the aldehyde to yield
the carboxylic acid VII was effected using phosphate buffered
The inhibitors discussed in the present letter were designed to
bind the active site of the HIV integrase complex, known to bear
a two-divalent metal ion catalytic motif. Furthermore, in order to
⇑
Corresponding author at present address: Champlain Exploration Inc., 1100-
2500 Rene-Levesque W., Montreal, QC H3B C6L, Canada.
y
Present address: Ascletis, 1785 Jianghai Road, Building No. 1, Suite 804, Binjiang
District, Hangzhou, China.
à
Present address: Cirion Biopharma Inc., 3150 Delaunay, Laval, QC H7L 5E1,
Canada.
§
Present address: Omegachem, 480, rue Perreault Lévis, QC G6W 7V6, Canada.
Present address: Bio-Rad Laboratories, Marnes-la-Coquette, France.
–
http://dx.doi.org/10.1016/j.bmcl.2016.01.028
0960-894X/Ó 2016 Published by Elsevier Ltd.
Please cite this article in press as: Stranix, B. R.; et al. Bioorg. Med. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.bmcl.2016.01.028

2
B. R. Stranix et al. / Bioorg. Med. Chem. Lett. xxx (2016) xxx–xxx
or by chromatography. All compounds were tested in the IN strand
transfer assay as well as a cell-based antiviral assay using the MT-4
cell line. The data are summarized in Table 1. Our SAR study clearly
shows that the spacer geometry between the pyridoxine hydroxy-
mate binding moiety and a fluorobenzene are critical determinants
for the activity of these inhibitors. Table 1 compares HIV integrase
inhibition results found through varying the spacer geometry
between the pyridoxin hydroxamate binding moiety and a fluo-
robenzene. The table lists results on the purified enzyme, and cell
based assay. Enzyme inhibition strand-transfer assays (IC50) were
performed using a well known methodology using 30 processed
oligonucleotides.12,13 Anti-viral activity (EC50) was measured ini-
tially by a single cycle infection assay in MT-4 cells followed by a
7 day multi-cycle assay in parallel with the Cytotoxicity assay
(CC50).14
A good correlation is found between these for each compound.
The comparison of the saturated dimethylene spacer and the
styryl analog shows that a rigid planar spacer does not place
the compound in a favorable position. Extension of the fluo-
rophenyl by one or more atoms also seems to be unfavorable
for binding. Two or three atom spacers containing sulfur gave
Figure 1. General structure of inhibitors.
(pH 9) NaOCl in phase-transfer conditions. This could then be
coupled to a variety of amines to yield amides VIII. The benzylic
alcohol was converted to a leaving group using methanesulfonyl
chloride IX and reacted with alkanoates, sulfides, sulfinic acid salts
to give compounds X, XI and XII. The acid VII was reacted in a Cur-
tius reaction to give the 5-amino compound XIII which provided a
point of entry for reversed amides XIV. Scheme 2 shows the
method used to convert the intermediates discussed above to the
phenolic hydroxamates. First deprotection was effected using
70% formic acid neat, followed by reaction of aqueous hydroxy-
lamine in pyridine. The resulting compounds were isolated in puri-
fied form by either recrystallizing from acetonitrile water mixtures
Scheme 1. Reagents and conditions: (a) ArOH, Ph3P, DEAD; (b) MnO2 CHCl3 99%; (c) NaOClaq tetrabutylammonium iodide, DCM; (d) MS-Cl, DCM TEA rt 95%; (e) ArCH2OH,
NaH 85–95%; (f) ArSH, base 75%; (g) ArSO2–Na, DMF; (h) Ar–CH2–P(Ph)3, THF reflux; (i) Ar–NH2, NaCNBH4, MeOH 60–80%; (j) H2/Pdquant.; (k) CDI, R–NH2; (l) (PhO)3PN3,
toluene reflux; (m) ArCO–X, base.
Scheme 2. Reagents and conditions: (a) HCOOH; (b) NH2OH, Py.
Please cite this article in press as: Stranix, B. R.; et al. Bioorg. Med. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.bmcl.2016.01.028

B. R. Stranix et al. / Bioorg. Med. Chem. Lett. xxx (2016) xxx–xxx
3
Table 1
Table 2
HIV integrase inhibition constants for compounds 1–14
Effect of Aryl substituents
a,b
Compd
Ar
EC50 (nM)
CC50
54
(lM)
a
b
Compd
IC50 (nM)
EC50 (nM)
CC50 (lM)
14
100
40
1
2
327
76
21
50
>100,000
>10,000
7
50
51
3
4
5
25
20
50
27
50
15
100
41
7
10,000
9300
16
17
25
25
50
6
7
2300
328
2000
40
23
51
1860
18
19
1500
4500
25
32
8
9
250
550
500
695
15
2
20
11
16
20
21
40
50
50
10
11
12
550
80
11,000
2300
4350
22
400
31
13
300
6300
50
Mean of at least three experiments.
a
Luciferase-bearing envelope defective (envꢀ) NL-4.3 virus pseudotyped with
a
HIV-1 env (HXBc2).
Mean of at least three experiments.
Anti-viral activity using a single-cycle infection where the cells were infected
b
b
Anti-viral activity using a single-cycle infection where the cells were infected
with HIV-1 env (HXBc2).
with a luciferase-bearing envelope defective (envꢀ) NL-4.3 virus pseudotyped with
HIV-1 env (HXBc2).
weak binding inhibitors. Indeed the best spacers appear to be
the aminomethylene, amides, oxymethylene and the benzylic
ethers.
Table 2 compares the effect of varying the aryl group on a series
of 5-amides. The effect of electron withdrawing groups on the phe-
nyl ring is notable. Methoxy and methylene dioxy compounds
showed a major loss of activity in comparison to the phenyl ring,
whereas the halogenated rings displayed the most activity. A sig-
nificant loss of activity was also found when replacing the phenyl
by a cyclohexane ring.
All the compounds were tested in several assays however we
chose to show the results on the most promising candidate,
compound 16. Table 3 shows the favorable pK profile and selec-
tivity of this compound in animal models and in vitro tests.
Compound 16 was selected due to its excellent therapeutic index
(CC50/EC50) of >1000. The HIV integrase strand-transfer inhibition
of this compound was measured bypassing 30 processing activity
as 30 processing inhibitors are known to remain tightly com-
plexed and do not favor a high catalytic turnover making inhibi-
tors of this mechanism less desirable as therapeutics. This
compound showed specificity when compared to assays involv-
ing HCV polymerase and HIV Reverse Transcriptase. Two distinct
assays were used to determine the mechanism of action of these
compounds. The first was accumulation of 2-LTR circles observed
when inhibitors were added during infection and is considered
as a hallmark of these inhibitors. Non-integrated genomes form
episomic 2-LTR circles.10,11 The second method used active site
mutation N155H which displayed a 10-fold drop in inhibition.
We also tested these compounds in protein binding assays and
found a relatively low binding in general. For example we found
a protein binding value of 89% for compound 16. Animal tests
showed this family of compounds to have favorable pharmacoki-
netics in rats. The bioavailability for the example compound 16
was 25% over 24 h with a t1/2 of 0.73 h. The compound plasma
concentration after a 50 mg/kg dosage remained above the EC50
for 18 h.
In conclusion, we report the design, synthesis and SAR of a new
series of pyridoxine hydroxamic acids as potent anti-HIV agents.
The SAR illustrated that anti-viral potency of these IN inhibitors
depended on the aryl substitution and aryl-spacer at the 5-position
of core pyridoxine heterocycle. Amongst them, compound 18
showed favorable pharmacological data to warrant further devel-
opment. Further studies are ongoing to assess the inhibitory activ-
ity against resistant HIV strains.
Please cite this article in press as: Stranix, B. R.; et al. Bioorg. Med. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.bmcl.2016.01.028

4
B. R. Stranix et al. / Bioorg. Med. Chem. Lett. xxx (2016) xxx–xxx
Table 3
Potency and pharmacokinetics of 16
IC50
29 nM
25 nM
128 nM
Reverse transcriptase
LogP
Protein binding
F% (24 h)
>50
0.6
89%
25%
lM
Mutagenicity
Negative Ames test
7 fold
10 fold
EC50 (single cycle assay
EC50 (multicycle assay)
CC50
Accumulation of 2-LTR at EC50
Accumulation of 2-LTR at EC90
N155H mutant EC50
22
lM
>300 nM
HCV EC50
>50
l
M
T1/2
0.73 h
a Mean of at least three experiments.
b Anti-viral activity using the molecular clone HIV-1 NL4-3 in MT-4 cells.
Anti-viral activity using a single-cycle infection where the cells were infected with a luciferase-bearing envelope defective (envꢀ) NL-4.3 virus pseudotyped with HIV-1 env
(HXBc2).
Protein binding was tested by Rapid Equilibrium Dialysis (RED) method (The RED (Rapid equilibrium Dialysis) (Pierce, Rockford Ill.)).
Female Sprague-Dawley rats were randomly selected and assigned to two groups. A group of 13 rats were administered 5 mg/kg of compound 2 intravenously. A second
group of 8 rats were administered 50 mg/kg of compound 2 orally. Following the dosing, blood was collected at 7 different time points. Plasma samples obtained from the
blood samples were analyzed by LC/MS/MS and the bioavailability of compound 16 was determined.
7. Plewe, M. B.; Butler, S. L.; Dress, K. R., et al. J. Med. Chem. 2009, 52, 7211. http://
Acknowledgements
dx.doi.org/10.1021/jm900862n.1.
8. Stranix, B.; Bouchard, J. E.; Milot, G. Patent WO 2,011,088,549 A1, 2011.
9. Stranix, B.; Beaulieu, F.; Bouchard, J. E.; Milot, G.; Zhigang, W.; Ruel, R. WO
2,009,146,555 A1, 2009; US Patent 8,742,123, 2014.
10. Korytnyk, W.; Angelino, N. J. Med. Chem. 1977, 20, 745. http://dx.doi.org/
10.1021/jm00216a002.
We gratefully acknowledge M. Ge, T. Oudanonh for their contri-
bution in bioanalysis. Dr. Rejean Ruel and Mr. Serge Dandache for
their suggestions and input.
11. Hazuda, D. J.; Felock, P.; Witmer, M., et al. Science 2000, 287, 646. http://dx.doi.
org/10.1126/science.287.5453.646.
12. Brussel, A.; Sonigo, P. J. Virol. 2004, 78, 11263.
References and notes
13. Hazuda, D. J.; Felock, P.; Hastings, J. C.; Pramanik, B.; Wolfe, A. J. Virol. 1997, 71,
7005. The results of the integrase strand transfer assay are reported as IC50
1. Lennox, J. L. Curr. Opin. HIV AIDS 2012, 7, 409. http://dx.doi.org/10.1097/
COH.0b013e3283562a27.
2. Eron, J. J., Jr. Clin. Infect. Dis. 2000, 30, S160.
values. IC50 values were determined using
a sigmoidal dose–response
equation. The formula used for calculating % inhibition was: % Inhibition =
[1 ꢀ (sample counts/average of positive control)] * 100. The percent inhibition
of HIV-1 integrase activity was graphed against the log of the compound
concentration (M). Using GraphPad Prism or Activity Base (IDBS) software IC50
was determined using following sigmoidal dose–response equation: Y = (A
+ ((B ꢀ A)/(1 + ((C/X) ^ D)))), where A is the lower plateau (~0%), B is the higher
plateau (~100%), C is the IC50, D is the slope, X is the compound concentration
(M), and Y is the % inhibition.
3. Quashie, P. K.; Sloan, R. D.; Wainberg, M. A. BMC Med. 2012, 10, 34. http://dx.
doi.org/10.1186/1741-7015-10-34.2; Vella, S. Handb. Exp. Pharmacol. 2000, 140,
23.
4. Mouscadet, J.-F.; Delelis, O.; Marcelin, A.-G.; Tchertanov, L. Drug Resist. Update
2010, 13, 139. http://dx.doi.org/10.1016/j.drup.2010.05.001.1.
5. Quashie, P. K.; Mesplède, T.; Wainberg, M. A. Curr. Infect. Dis. Rep. 2013, 15, 85.
http://dx.doi.org/10.1007/s11908-012-0305-1.
6. Wu, J. J.; Milot, G.; Beaulieu, F.; Bouchard, J. -E.; Xiao, Y.; Dandache, S.; Gouveia,
K.; Forte, A.; Garde, S.; Wang, Z.; Ge M.; Ounadah, T.; Stranix, B. Small-Molecule
Inhibitors of HIV-1 Integrase. In XVII International Resistance Workshop; Vol
Sitges, 2008; p A18.1.
14. Japour, A. J. et al Antimicrob. Agents Chemother. 1993, 37, 1095; Pauwels, R. et al
J. Virol. Methods 1988, 20, 309.
Please cite this article in press as: Stranix, B. R.; et al. Bioorg. Med. Chem. Lett. (2016), http://dx.doi.org/10.1016/j.bmcl.2016.01.028