Synthesis and evaluation of new dinitrobenzamide mustards in human prostate cancer

Article abstract of DOI:10.1016/j.bmcl.2020.127697

Tumor hypoxia has been widely explored over the years as a diagnostic and therapeutic marker. Herein, we have reported the design and synthesis of a series of dinitrobenzamide mustards (DNBM) based on the PR-104A hypoxia-selective prodrug. Specifically, we explored the impact of various leaving groups and the introduction of a carboxylic acid group on the biological performance of the DNBM constructs. Once in hand, the Log D values, cytotoxicity in PC-3 and DU-145 human prostate cancer cells lines and the hypoxia selectivities of the DNBM analogs were examined. Overall, the DNBM constructs were found to be tolerant to modifications with none of the explored modifications substantially degrading the cytotoxic potential of the constructs.

Full text of DOI:10.1016/j.bmcl.2020.127697

Journal Pre-proofs
Synthesis and Evaluation of New Dinitrobenzamide Mustards in Human Pros‐
tate Cancer
Alireza Basiri, Wenting Zhang, Jered Garrison
PII:
S0960-894X(20)30808-8
DOI:
Reference:
https://doi.org/10.1016/j.bmcl.2020.127697
BMCL 127697
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date:
Revised Date:
Accepted Date:
2 September 2020
11 November 2020
13 November 2020
Please cite this article as: Basiri, A., Zhang, W., Garrison, J., Synthesis and Evaluation of New Dinitrobenzamide
Mustards in Human Prostate Cancer, Bioorganic & Medicinal Chemistry Letters (2020), doi: https://doi.org/
10.1016/j.bmcl.2020.127697
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. 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.
© 2020 Published by Elsevier Ltd.

Synthesis and Evaluation of New Dinitrobenzamide
Mustards in Human Prostate Cancer
Alireza Basiri†,ǁ, Wenting Zhang†,ǁ, Jered Garrison†,§, ǁ,¦, *
†. Department of Pharmaceutical Sciences, College of Pharmacy, University of Nebraska Medical
Center, 985830 Nebraska Medical Center, Omaha, NE 68198, United States
§. Department of Biochemistry and Molecular Biology, College of Medicine, University of Nebraska
Medical Center, 985870 Nebraska Medical Center, Omaha, NE 68198, United States
ǁ. Center for Drug Delivery and Nanomedicine, University of Nebraska Medical Center, 985830
Nebraska Medical Center, Omaha, NE 68198, United States
¦. Eppley Cancer Center, University of Nebraska Medical Center, 985950 Nebraska Medical Center,
Omaha, NE 68198, United States
ABSTRACT: Tumor hypoxia has been widely explored over the years as a diagnostic and therapeutic
marker. Herein, we have reported the design and synthesis of a series of dinitrobenzamide mustards
(DNBM) based on the PR-104A hypoxia-selective prodrug. Specifically, we explore the impact of various
leaving groups and the introduction of a carboxylic acid group on the biological performance of the
DNBM constructs. Once in hand, the Log D values, cytotoxicity in PC-3 and DU-145 human prostate
cancer cells lines and the hypoxia selectivities of the DNBM analogs were examined. Overall, the DNBM
constructs were found to be tolerant to modification with none of the explored modifications substantially
degrading the cytotoxic potential of the constructs.
Keywords: DNBM, prodrug, prostate cancer, hypoxia
Corresponding Author
* Phone: +1-402-559-3453. Fax: 001-402-559-9365. E-mail: jcgarrison@unmc.edu.

The prevalence of hypoxic tissue, areas with oxygen partial pressure (pO2) values significantly lower
than normal, is a characteristic of many solid tumors. It has been shown that up to 50-60% of tumors in
cancers of the breast, head and neck, prostate, pancreas and soft tissue sarcomas exhibit significant
amounts of hypoxia and/or anoxia.1-3 Also, hypoxia has been linked as a major driving force behind the
aggressiveness and treatment resistance of many human cancers. This has provided the impetus to the
development of therapeutic and diagnostic agents that target tumor hypoxia.4 Several strategies for
exploiting tumor hypoxia are now in preclinical or clinical development, with the main focus on prodrugs
that are activated by metabolic reduction under hypoxic conditions to form cytotoxins.5
Dinitrobenzamide mustards (DNBMs) represent a class of hypoxia-activated prodrugs.6 This class of
mustards contains a latent nitrogen mustard moiety, which can be activated upon reduction of the
molecule under hypoxic conditions. PR-104, a DNBM, is a phosphate ester “pre-prodrug” currently in
clinical trials. 7, 8 When administered, this prodrug is dephosphorylated to its alcohol analog PR-104A, see
Fig. 1. The nitro groups are subsequently reduced under hypoxic conditions via PR-104N to the
corresponding hydroxylamine or amine (i.e., PR-104H and PR-104M).9 This reduction-activated
‘‘electronic switch’’ triggers a substantial increase in the alkylating power of the nitrogen mustard.10
These activated mustards, then induce the formation of DNA interstrand cross-links, which is generally
considered the therapeutic action of this class of chemotherapeutics.11, 12
Fig. 1: Mechanism of metabolic activation of PR-104A, catalyzed by one-electron reduction under hypoxia,
modified from 13

In the following, we describe the synthesis and in vitro investigation of seven DNBM prodrugs
resembling the PR-104A paradigm. The design of these mustard prodrugs, depicted in Scheme 1,
followed two major aims. Firstly, we hypothesized that by increasing the leaving group potential (i.e., Br
< mesylate ~ tosylate << triflate) of the mustard moiety, an increase in cytotoxicity would be observed. To
that end, we synthesized compounds 11, 16 (PR-104A), 17 and 18. Previously, we have utilized hypoxia-
selective prodrugs and other adduct-forming constructs as trapping agents to increase the long term
residualization of receptor-targeted constructs in tumors.14, 15 In the same context, our secondary goal was
to include a chemical handle on the prodrugs, to allow future conjugations to fluorophores, radioisotopes
and/or receptor-targeted agents. With this purpose in mind, the alcohol functional group of the above-
mentioned analogs was replaced with a carboxylic acid, yielding compounds 12, 19, 20 and 21. With
these compounds in hand, we examined what impact these structural modifications had on the
physiochemical properties, hypoxia selectivity and biological activities in PC-3 and DU-145 human
prostate cancer cell lines.
We initiated the synthesis by the acid-catalyzed esterification of 2-chloro-3,5-dinitrobenzoic acid, 1,
in methanol. The protected dinitrobenzoate, 2, was subsequently reacted with bis(chloroethyl)amine, 3, to
furnish compound 4, which was isolated through flash column chromatography in 86% yield. The
structure of 4, depicted in Scheme 1, was confirmed by single-crystal x-ray diffraction analysis. In the
next step, compound 4 was reacted with an aqueous solution of KOH (3N) in dioxane to remove the
protecting methyl group furnishing the free benzoic acid product, 5. Compound 5 was subsequently
converted to the acid chloride, 6, using oxalyl chloride under mild conditions. In situ, the acid chloride, 6,
was reacted with ethanolamine (7), or β-Alanine t-butyl ester (8), to yield the DNBMs 9 and 10,
respectively, in reasonable yields after flash column chromatography.
Halide exchange was carried out on compounds 9 and 10 using LiBr in 3-methyl-2-butanone to
provide the corresponding crude dibromo analogs 11 and 12. The crude material was further purified
using silica gel chromatography to yield pure forms of 11 and 12 in 72 and 82% yields, respectively.
Specifically, for compound 12, the bromide exchange reaction led to the removal of the t-Bu protecting
group from the side chain. Compound 16 (PR-104A) was prepared by refluxing 11 and silver mesylate
(13) in acetonitrile. Unreported sulfonate esters 17-21, were prepared by refluxing 11 or 12 with the
corresponding sulfonate silver salts (i.e., silver mesylate, 13; silver tosylate, 14; and silver triflate, 15) in
acetonitrile and isolated through flash column chromatography. For compounds 19-21, due to the
presence of the acid moiety on the side chains, 0.1% formic acid was used as a modifier for better
separation during column chromatography. The structures of PR-104A, as well as the newly synthesized
1
prodrug analogs were confirmed by mass spectrometry, H and 13C NMR spectral analyses (supporting
information).

Partition or distribution coefficients (Log P or Log D) are known to impact the absorption,
distribution, metabolism, excretion and toxicity (ADMET) properties of drugs. These values are widely
used in drug discovery for quantitative structure-activity relationships and quantitative structure-property
relationships.16 Generally, compounds with moderate lipophilicity (log D = 0-3) are thought to provide a
good balance between solubility and cellular permeability, which is needed to access intracellular
targets.17 To examine the impact of the structural modifications on the distribution properties, log D7.4
values of the synthesized prodrugs were measured using the shake-flask method (Table 1).18
Owing to the presence of the ionizable carboxylic acids present in 12, 19, 20 and 21, these
compounds were expected to have lower distribution coefficient values at physiological pH, compared to
their alcohol side chain counterparts 11, 16 (PR-104A), 17 and 18. The distribution coefficients indicated
a more dichotomic result. For compounds 11 and 16, replacement of the alcohol with a carboxylic acid
resulted in an order of magnitude increase in hydrophilicity for the resulting 12 and 19. However, this
same alteration for 17 and 18 had a negligible impact on hydrophilicity for the resulting 20 and 21.
Replacement of the bromide in 11 and 12 with a methyl sulfonate furnished analogs 16 and 19, which
displayed lower distribution coefficients in aqueous media. For instance, log D7.4 values of 11 shifted
from 1.7 ± 0.2 to 0.80 ± 0.12 for analog 16. Interestingly, when the side chain is an alcohol, the tosylate
(17) and triflate (18) leaving groups have substantially higher hydrophilicities than the bromide (11).
Conversely, when the side chain is a carboxylic acid, the bromide (12) yields higher hydrophilicity
compared to the tosylate (20) and nearly analogous hydrophilicity relative to the triflate (21) derivative.

Scheme 1. Synthesis of mustard DNBM prodrugs 11-21
The activation and therapeutic mechanism of hypoxia-activated DNBM prodrugs has been well-
established.19, 20 Under hypoxic conditions and in the presence of nitroreductase enzymes, the electron-
withdrawing 4-nitro group transforms into an electron-donating 4-hydroxylamine or amine moiety (see
Fig. 1). This “electronic switch” triggers the formation of the highly reactive aziridinium species via an
intramolecular SN2 reaction.21 These aziridinium groups can subsequently alkylate DNA to yield their
cytotoxic/ therapeutic effect.22
To examine the cytotoxicities and hypoxia-selectivities of the prodrugs 11, 12, and 16-21, these
compounds were evaluated in PC-3 and DU-145 human prostate cancer cell lines, under hypoxic and
normoxic conditions using a cell viability assay. In brief, the cells were exposed to the mustard agents, at
a range of concentrations for two hours under hypoxic or normoxic conditions, thereafter, washed with
fresh media and incubated under normoxic conditions for three days. The determined cytotoxicity
potential of these agents was based on cell viability assays using almarBlue® and reported as half-

maximal inhibitory concentration (IC50) values. The results of these studies are summarized in Table 1.
The established prodrug PR-104A (16) was used as a positive control for comparison against the newly
synthesized analogs.
Table 1. Measured log D7.4, IC50 and hypoxia selectivity ratio values of the synthesized prodrugs 11,12,16-21
IC50
Hypoxia
selectivity
PC-3 cells
DU-145 cells
†
Compd.
Log D7.4
PC-3
DU-
145
N2
Air
N2
Air
7.4
6.5
2.2
1.7± 0.2
26±2
75±4
193±13
490±30
80±4
47±6
177±6
413±6
11
12
0.42 ±
0.09
8.8
0.80 ±
0.12
16
48±4
37±3
42±4
41±3
42±3
90.±3
320±30
146±11
245±15
417±18
460±30
350±30
78±5
28±2
51±4
56±3
54±4
68±2
210±20
108±5
6.7
3.9
2.7
3.9
(PR104A)
0.95±
0.10
17
18
19
20
21
0.33 ±
0.04
5.8
10.
11
9.2
5.8
6.7
7.0
470±30
326±14
363±19
480±30
-0.71 ±
0.12
1.03 ±
0.07
0.15 ±
0.14
3.9
† Log D7.4 (mean ± SEM) are from two or more independent experiments carried out in triplicate; ‡IC50 (mean ±
SEM); hypoxia selectivity = IC50 (air)/ IC50 (N2)
In PC-3 cells, the positive control PR-104A (16), as expected, demonstrated higher cytotoxicity under
hypoxic relative to normoxic conditions, with the IC50 values of 48 ± 4 and 320 ± 30 µM for hypoxia and
normoxia, respectively, and a hypoxia selectivity ratio of 6.7. Patterson et. al. has reported the IC50 value
for PR-104A in PC-3 cells was 7.3 µM under anoxia (< 10 ppm oxygen in gas phase) and a hypoxia-
selectivity ratio of approximately 15. The discrepancy between these values almost certainly stems from
the residual oxygen concentrations used for the hypoxic/anoxic conditions. Tissue hypoxia is generally
defined as a pO2 pressure less than 15 mmHg (19,800 ppm), although this cut-off value is somewhat
subjective when considering the range of physiologically relevant oxygen levels in normal tissues.23, 24 We
opted to examine our compounds under a 0.1% oxygen environment (1,000 ppm) due to it being more in-
line with possible clinical hypoxia conditions. The higher residual oxygen concentrations for our

experiments likely resulted in less efficient hypoxia-activation of the prodrugs leading to higher IC50
values and a lower hypoxia selectivity ratio, compared to that reported under anoxic conditions.
As expected, all the compounds demonstrated greater cytotoxicities under hypoxic relative to
normoxic conditions. In PC-3 cells under our hypoxic conditions, the compounds demonstrated
cytotoxicities that ranged from 26 ± 2 to 90 ± 3 µM. All of the analogs displayed similar or higher
cytotoxicities compared to PR-104A (16) under hypoxic conditions, except for prodrugs 12 and 21.
Compound 11 demonstrated the highest cytotoxicity with an IC50 value of 26 ± 2 µM, followed by
compound 17 with an IC50 value of 37 ± 3 µM (Fig. 2). Hypoxia selectivity ratios for the constructs
ranged from 3.9 to 11 in PC-3 cells. Only 11, 19 and 20 exhibited better hypoxia selectivity ratios than
PR-104A (16), see Fig. 4.
Conversely, in the DU-145 cell line under identical conditions, the cytotoxicities under hypoxia
ranged slightly narrower from 28 ± 2 to 80 ± 4 µM. In this case, PR-104A (78 ± 5 µM) exhibited
cytotoxicity that was statistically analogous to 11 but was significantly lower than all other constructs
investigated. Construct 17 (28 ± 2 µM) gave the lowest IC50 value followed by 12 with 47 ± 6 µM (Fig.
3). In DU-145 cells the hypoxia selectivity ratios ranged from 2.2 to 9.2. With exception to compound 11,
all investigated constructs had better hypoxia selectivity ratios than PR-104A (16), see Fig. 4.
700
600
500
400
300
200
100
PR-104A Min.
0
11
12
17
18
19
20
21 PR-104A
Prodrugs
hypoxia
normoxia
condition
Fig. 2: cytotoxicity activities of prodrugs in PC-3 cells, under hypoxic and normoxic conditions

700
600
500
400
300
200
100
0
PR-104A Min.
11
12
17
18
19
20
21 PR-104A
Prodrugs
hypoxia
normoxia
condition
Fig. 3: cytotoxicity activities of prodrugs in DU-145 cells, under hypoxic and normoxic conditions
Analysis of the data did yield some structure-activity trends concerning the explored modifications
(i.e., carboxylic acid/alcohol and mustard leaving group). Compounds with carboxylic acid moieties
tended to have lower cytotoxicity values under normoxia. At first glance, this may be inferred to a lower
cellular uptake profile due to the negatively charged carboxylate under physiological conditions.
However, the cytotoxicities of the DNBMs containing carboxylic acids were mostly consistent with the
cytotoxicities of the alcohol bearing analogs under hypoxic conditions. If these charged compounds have
lower cellular uptake values, this suggests that the cytotoxicity of the carboxylic acid analogs was
inherently higher than the alcohol derivatives. At any rate, the decrease in cytotoxicity under normoxic
conditions and the maintenance of cytotoxicity under hypoxic conditions led to generally larger hypoxia
selectivity ratios for the carboxylic acid analogs relative to the alcohol derivatives. For instance, in the
PC-3 cell line, prodrug 20 showed hypoxia selectivity of 11, while its alcoholic analog, 17, showed a
hypoxia selectivity of 3.9. Under our conditions, increasing the leaving group potential did not lead to
significant gains in cytotoxicities. However, for DNBM analogs with the triflate leaving group, 18 and 21,
significant declines in cytotoxicities were observed under both normoxic and hypoxic conditions relative
to other analogs. This may well be attributable to the greater instability of the triflate constructs in
solution superseding any potential cytotoxicity gains due to the greater leaving group potential of this
group.

While the cytotoxicities of the constructs in both cell lines were in a relatively narrow range (26 ± 2 to
90 ± 3 µM), there was a significant variance between cell lines for some compounds. For example,
compound 11 had the highest cytotoxicity in PC-3 cells but was the worst-performing construct in the
DU-145 cell line. Relatively large cell-line dependent variation for a DNBM is not uncommon and has
been observed by others.20 Additionally, a linear regression analysis was performed to examine if a
correlation existed between the log D7.4 and IC50 values, but no statistically significant correlation was
found.
Our interest in the DNBM class of hypoxia-activated prodrugs lies in their properties beyond
chemotherapeutics, specifically for the potential development of diagnostic agents and other therapeutic
approaches. Numerous investigators have explored other bio-reductive prodrugs for this purpose.25-28 It
was with this idea in mind that we explored the impact of modifying the alcohol side chain of PR-104A
with a carboxylic acid, to aid in eventual conjugation to other functionalities (i.e., fluorophores,
radiolabels and other constructs). None of the modifications explored above substantially influenced the
ability of the constructs to activate in hypoxic environments, form intracellular adducts and yield their
cytotoxic effects. Due to this tolerance to modification, we intend to further explore the capability of these
DNBMs in the development of diagnostic and radiotherapeutic agents.
16
14
12
10
8
16
14
12
10
8
6
6
4
4
2
2
0
0
1
1
1
1
1
2
2
P
1
1
1
1
1
2
2
P
1
2
7
8
9
0
1
R
1
2
7
8
9
0
1
R
1
-
0
1
4
0
A
4
A
Prodrugs
Prodrugs
Fig. 4: hypoxia selectivity of prodrugs 11-21 in PC-3 (left) and DU-145 cells (right)
To summarize, we have designed and successfully synthesized a series of DNBMs resembling the
hypoxia-selective prodrug PR-104A. With these analogs in hand, we examined to what extent
modification of the mustard leaving-group potential and incorporation of a carboxylic acid moiety affects
their biological performance. All of the synthesized analogs demonstrated significantly higher
cytotoxicities under hypoxic relative to normoxic conditions. This implies that all of the synthesized
DNBMs can undergo reduction under hypoxic conditions and convert to the active cytotoxic form of the
drug. Increasing the leaving group potential of the DNBM did not lead to notable gains in cytotoxicity.

Incorporation of a carboxylic acid group into the DNBM construct did not negatively impact cytotoxicity
and tended to yield better hypoxia-selectivity ratios than analogous alcohol derivatives. In the future, we
intend to exploit these DNBMs for their adduct formation capabilities and to develop diagnostic and
radiotherapeutic agents.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that
could have appeared to influence the work reported in this paper.
Refrences:
1.
2.
Hockel M, Vaupel P. J Natl Cancer Inst. 2001;93(4): 266-276.
Vaupel P, Höckel M. Recombinant Human Erythropoietin (rhEPO) in clinical oncology.
Springer; 2002:127-146.
3.
4.
5.
6.
Vaupel P, Mayer A. Cancer Metastasis Rev. 2007;26(2): 225-239.
Subarsky P, Hill RP. Clin Exp Metastas. 2003;20(3): 237-250.
Brown JM, Wilson WR. Nat Rev Cancer. 2004;4(6): 437.
Wilson WR, Hicks KO, Pullen SM, Ferry DM, Helsby NA, Patterson AV. Radiat Res.
2007;167(6): 625-636.
7.
8.
9.
10.
McKeage MJ, Gu Y, Wilson WR, et al. BMC cancer. 2011;11(1): 432.
McKeage MJ, Jameson MB, Ramanathan RK, et al. BMC cancer. 2012;12(1): 496.
Helsby NA, Wheeler SJ, Pruijn FB, et al. Chem Res Toxicol. 2003;16(4): 469-478.
Palmer BD, Wilson WR, Anderson RF, Boyd M, Denny WA. J Med Chem. 1996;39(13): 2518-
2528.
11.
12.
13.
14.
15.
Hunter FW, Wouters BG, Wilson WR. Br J Cancer. 2016;114(10): 1071.
Stornetta A, Deng K-CK, Danielli S, et al. Biochem Pharmacol. 2018;154: 64-74.
Phillips RM. Cancer Chemoth Pharm. 2016;77(3): 441-457.
Zhou Z, Wagh NK, Ogbomo SM, et al. J Nucl Med. 2013;54(9): 1605-1612.
Wagh NK, Zhou Z, Ogbomo SM, Shi W, Brusnahan SK, Garrison JC. Bioconjugate Chem.
2012;23(3): 527-537.
16.
Comer JE. Drug bioavailability: Estimation of solubility, permeability, absorption and
bioavailability. 2003:21-45.
17.
18.
19.
20.
21.
22.
23.
24.
Di L, Kerns EH. Curr Opin Chem Biol. 2003;7(3): 402-408.
Läppchen T, Rossin R, van Mourik TR, et al. Nucl Med Biol. 2017;55: 19-26.
Helsby NA, Wheeler SJ, Pruijn FB, et al. Chem Res Toxicol. 2003;16(4): 469-478.
Patterson AV, Ferry DM, Edmunds SJ, et al. Clin Cancer Res. 2007;13(13): 3922-3932.
Williams EM, Little RF, Mowday AM, et al. Biochem J. 2015;471(2): 131-153.
Brookes P, Lawley P. Biochem J. 1961;80(3): 496.
Ortiz-Prado E, Dunn JF, Vasconez J, Castillo D, Viscor G. Am J Blood Res. 2019;9(1): 1-14.
McKeown SR. Br J Radiol. 2014;87(1035): 20130676.

25.
Duszynski CC, Avati V, Lapointe AP, Scholkmann F, Dunn JF, Swain MG. Hepatology.
2020;71(4): 1408-1420.
26.
27.
Hoigebazar L, Jeong JM, Lee J-Y, et al. J Med Chem. 2012;55: 3155-3162.
Bejot R, Kersemans V, Kelly C, Carroll L, King RC, Gouverneur V. Nucl Med Biol. 2010;37:
565-575.
28.
Das T, Chakraborty S, Banerjee S, Sarma HD, Samuel G, Venkatesh M. Radiochima Acta.
2006;94: 375-380.