Systematic Variation of Pyrrolobenzodiazepine (PBD)-Dimer Payload Physicochemical Properties Impacts Efficacy and Tolerability of the Corresponding Antibody-Drug Conjugates

Article abstract of DOI:10.1021/acs.jmedchem.0c00691

Cytotoxic pyrrolobenzodiazepine (PBD)-dimer molecules are frequently utilized as payloads for antibody-drug conjugates (ADCs), and many examples are currently in clinical development. In order to further explore this ADC payload class, the physicochemical properties of various PBD-dimer molecules were modified by the systematic introduction of acidic and basic moieties into their chemical structures. The impact of these changes on DNA binding, cell membrane permeability, and in vitro antiproliferation potency was, respectively, determined using a DNA alkylation assay, PAMPA assessments, and cell-based cytotoxicity measurements conducted with a variety of cancer lines. The modified PBD-dimer compounds were subsequently incorporated into CD22-targeting ADCs, and these entities were profiled in a variety of in vitro and in vivo experiments. The introduction of a strongly basic moiety into the PBD-dimer scaffold afforded a conjugate with dramatically worsened mouse tolerability properties relative to ADCs derived from related payloads, which lacked the basic group.

Full text of DOI:10.1021/acs.jmedchem.0c00691

Subscriber access provided by UNIV OF WESTERN ONTARIO
Article
Systematic Variation of Pyrrolobenzodiazepine (PBD)-dimer
Payload Physicochemical Properties Impacts Efficacy and
Tolerability of the Corresponding Antibody-Drug Conjugates
Leanna R. Staben, Jinhua Chen, Josefa dela Cruz-Chuh, Geoffrey Del Rosario, Mary
Ann Go, Jun Guo, Siamak Cyrus Khojasteh, Katherine R. Kozak, Guangmin Li, Carl
Ng, Gail Lewis Phillips, Thomas H. Pillow, Rebecca K. Rowntree, John Wai, Binqing
Wei, Keyang Xu, Zijin Xu, Shang-Fan Yu, Donglu Zhang, and Peter S. Dragovich
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00691 • Publication Date (Web): 31 Jul 2020
Downloaded from pubs.acs.org on August 1, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Systematic Variation of Pyrrolobenzodiazepine (PBD)-dimer
Payload Physicochemical Properties Impacts Efficacy and
Tolerability of the Corresponding Antibody-Drug Conjugates
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†
‡
†
†
†
Leanna R. Staben, Jinhua Chen, Josefa dela Cruz-Chuh, Geoff del Rosario, Mary Ann Go,
†
†
†
†
†
Jun Guo, S. Cyrus Khojasteh, Katherine R. Kozak, Guangmin Li, Carl Ng, Gail D. Lewis
†
†
†
‡
†
†
Phillips, Thomas H. Pillow, Rebecca K. Rowntree, John Wai, BinQing Wei, Keyang Xu,
‡
†
†
,†
Zijin Xu, Shang-Fan Yu, Donglu Zhang, Peter S. Dragovich*
†Genentech, Inc., 1 DNA Way, South San Francisco, California 94080, USA
‡WuXi AppTec Co., Ltd, 288 Fute Zhong Road, Waigaoqiao Free Trade Zone, Shanghai
00131, China.
2
RECEIVED DATE (to be automatically inserted after manuscript is accepted)
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT: Cytotoxic pyrrolobenzodiazepine (PBD)-dimer molecules are frequently utilized
as payloads for antibody-drug conjugates (ADCs), and many examples are currently in clinical
development. In order to further explore this ADC payload class, the physicochemical properties
of various PBD-dimer molecules were modified by the systematic introduction of acidic and
basic moieties into their chemical structures. The impact of these changes on DNA binding, cell
membrane permeability, and in vitro antiproliferation potency was respectively determined using
a DNA alkylation assay, PAMPA assessments, and cell-based cytotoxicity measurements
conducted with a variety of cancer lines. The modified PBD-dimer compounds were
subsequently incorporated into CD22-targeting ADCs, and these entities were profiled in a
variety of in vitro and in vivo experiments. Introduction of a strongly basic moiety into the PBD-
dimer scaffold afforded a conjugate with dramatically worsened mouse tolerability properties
relative to ADCs derived from related payloads which lacked the basic group.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
ACS Paragon Plus Environment

Page 3 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
Pyrrolobenzodiazepine (PBD)-dimers are compounds which form covalent DNA
interstrand cross-links in a sequence-dependent manner and which exhibit broad-spectrum sub-
nanomolar antiproliferative activities against a variety of cancer cell lines.1 Because of their
exceptional cell potencies, PBD-dimers have been extensively employed as antibody-drug
conjugate (ADC) payloads, and ADCs bearing such entities have been studied in advanced
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
preclinical experiments and/or human clinical trials. For example, SG2057 {(11aS,11a'S)-8,8'-
(pentane-1,5-diylbis(oxy))bis(7-methoxy-2-methylene-1,2,3,11a-tetrahydro-5H-
benzo[e]pyrrolo[1,2-a][1,4]diazepin-5-one), compound 1} was incorporated into ADCs designed
2
3
to treat acute myeloid leukemia (AML) or HER2-expressing cancers while SG3199
{
(11aS,11a'S)-8,8'-(pentane-1,5-diylbis(oxy))bis(7-methoxy-2-methyl-1,11a-dihydro-5H-
benzo[e]pyrrolo[1,2-a][1,4]diazepin-5-one), compound 2} was employed as a component of
rovalpituzumab tesirine (“Rova-T”) which progressed to phase 3 clinical trials for the treatment
4
of small-cell lung cancer. In addition, SG1882 {(S)-2-(4-aminophenyl)-7-methoxy-8-(3-(((S)-7-
methoxy-2-(4-methoxyphenyl)-5-oxo-5,11a-dihydro-1H-benzo[e]pyrrolo[1,2-a][1,4]diazepin-8-
yl)oxy)-propoxy)-1,11a-dihydro-5H-benzo[e]pyrrolo[1,2-a][1,4]diazepin-5-one, compound 3}
was the cytotoxic payload associated with vadastuximab talirine which was assessed in phase 3
5
trials to treat acute myeloid leukemia (AML). Several additional PBD-dimer-containing ADCs
are currently undergoing clinical evaluation against a variety of cancers including loncastuximab
6
7
8
9
tesirine (ADCT-402), ADCT-601, MEDI-2228, and camidanlumab tesirine (ADCT-301). In
1
0
spite of these advancements, the late-stage clinical failures of both Rova-T and vadastuximab
11
talirine suggest that additional optimization of the PBD-dimer entities may be required in order
to fully realize the therapeutic potential of this ADC payload class.
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 1. Representative cytotoxic PBD-dimers employed as ADC payloads.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
As part of our own explorations of PBD-dimer-containing ADCs which also employed
linkers that afforded lysosomal cleavage,2,3 we wished to better understand how modifications to
the physicochemical properties of the cytotoxic payloads impacted the in vitro and in vivo
performance of the associated conjugates. In particular, we were curious to learn how the
incorporation of acidic or basic moieties into the chemical structures of typical PBD-dimers
might alter efficacy and/or tolerability outcomes associated with ADCs constructed from the
4
ACS Paragon Plus Environment

Page 5 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
modified payloads. We also desired to understand how the results of these alterations compared
to those associated with other well-known ADC payload classes. Prior studies of related ADC
payload modifications include comparison of cell-permeable and cell-impermeable auristatin
1
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
derivatives (MMAE and MMAF, respectively), modification of maytansinoid-derived payload
1
3
physicochemical properties, and examination of topoisomerase inhibitor payloads with varying
cell-permeability characteristics.14 However, there are currently few published examples that
study the impact of physicochemical property modification of the PBD-dimer ADC payload
class.15,16 In this report, we describe our efforts to systematically modify PBD-dimer
physicochemical properties via the introduction of various ionizable functional groups as part of
a focused effort to better understand such impacts. We also detail in vitro and in vivo testing
results of antibody-drug conjugates derived from the new PBD-dimer entities which explore how
these payload modifications alter ADC bioactivity in both in vitro and in vivo settings. In
addition, we compare the physicochemical property effects observed for the modified PBD-
dimers and associated conjugates with those previously reported for other ADC payload
classes.12-15
RESULTS and DISCUSSION
We began our explorations by identifying a location within the PBD-dimer chemical
structure that could be employed for the introduction of the desired diverse ionizable
functionalities without significantly impacting the ability of the molecules to bind to the DNA
minor groove and subsequently cross-link the nucleic acid.17 We were inspired by our previous
studies which examined ADCs constructed from non-cleavable linker-drugs bearing a centrally-
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
located phenyl moiety in the PBD-dimer structure (e.g., compound 4, Figure 2).18 Such ADCs
were highly active in both in vitro and in vivo experiments, and these results indicated that the
alkyne/PEG fragment, which would remain attached to the PBD-dimer core structure following
lysosomal catabolism of the associated antibody, did not significantly impede DNA binding
and/or alkylation by the cytotoxic compound. We therefore sought to utilize a related alkyne-
based attachment strategy to derivatize the PBD-dimer core structure with the desired ionizable
functional groups. Accordingly, we designed compounds 5-9 (Table 1) in which the various
structural changes were incorporated at the terminus of the phenyl-alkyne moiety present in the
center of the PBD-dimer molecules. As shown in Table 1, the functionalities employed in 5-9
were purposely selected to span a range of acidic and basic moieties with the methyl group
present in 7 serving as a non-ionizable (neutral) comparator. In addition, the degree to which
these acidic and basic groups ionized was expected to vary as the compounds transitioned
between various biological environments as a result of their antibody-mediated intracellular
delivery (e.g., compound 6 was anticipated to be predominantly ionized in the neutral cytosol but
significantly less so in the weakly acidic lysosome). As part of these designs, we also envisioned
subsequently connecting the new PBD-dimer entities to antibodies using well-precedented
dipeptide-containing linkers that could be attached to the N-10 positions of the PBD structures
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(Table 1, Figure 2).
6
ACS Paragon Plus Environment

Page 7 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 2. Alkyne-containing PBD-dimer linker-drug employed in our previous studies (ref. 18).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
.
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 1. Structures and physicochemical and biological properties of compounds 5-9.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
DNA Alkylation
% remain, 1 h)a
PAMPA
Pe (x10 cm/s)
Physico-Chem
Cell Antiproliferation IC50 (pM)c
-6
b
(
Calcd
Class
Other Lines
Cmpd.
R
pKad
(Average)e
Strand 1 Strand 2 pH = 4.0 pH = 7.4
BJAB
KPL4
Strong
Acid
>20,000
(n = 6)
1
5
.5
.5
62
0
58
0
0.68
10
0.60
>20,000
>20,000
5
Weak
Acid
8,500
0.70f
4,300
3.3
15,000
23
(n = 5)g
6
7
8
9
24
(n = 6)
Neutral NA
Weak
30
30
0
27
30
0
37
48
36
108
(n = 7)
6
.0h
2.4
0.34
8.6
65
Base
Strong
Base
264
(n = 6)
8
.2h
2.0
29
86
1
00
100
0
68
15
61
11
5,000
14
72,000
53
ND
1
0
NA
NA
NA
NA
NA
NA
6
7
4
1
(
n = 7)
8
ACS Paragon Plus Environment

Page 9 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
aPercent of unalkylated single-strand DNA remaining following independent exposure of double-
strand DNA to 100 µM of each compound for 1 hour. 0 = strong alkylation observed, 100 = no
b
(weak) alkylation observed. PAMPA effective permeability determined for each compound at
c
indicated pH; >20 = high, 10-20 = medium, <10 = low. Antiproliferation activity determined in
cell culture experiments using indicated cell line. Data are presented as geometric means (n = 4
d
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
per group); 95% confidence intervals are listed in Table S1. pKa value of the R-group appended
e
to each compound calculated using the MoKa software package (version 2.6.6). Average of
antiproliferation values determined for each compound using a panel of 5-7 cell lines. See Table
f
g
S2 for more details. pH = 6.2. Value does not include two tested cell lines with IC50 outcomes
h
>
100,000 pM. Conjugate acid. NA = not applicable. ND = not determined.
1
8
As was observed in our prior work, modelling studies performed with 5-9 suggested
that the alkyne fragments would protrude away from the DNA minor groove into solvent and
thereby minimize the impact of the various terminal functional groups on the ability of the new
molecules to bind to DNA (Figure 3). To confirm that this was indeed the case, we assessed the
ability of each compound to alkylate a fragment of double stranded DNA that contained a
sequence known to be efficiently recognized by this class of cytotoxic agents.19 In this semi-
quantitative assessment, the methyl-containing compound 7 exhibited DNA alkylation activity
that was somewhat weaker than that exhibited by the reference compound 1 which lacked the
centrally located phenyl-alkyne moiety (Table 1). This outcome likely reflects altered
interactions of the two PBD-dimers with the double-stranded oligonucleotide that presumably
result from the different linkers used to connect the PBD-monomer units.
We next compared the DNA alkylation properties of compounds 5, 6, 8, and 9 to those of
7
to determine the impact of the various ionizable groups that resided in the former molecules.
As shown in Table 1, the weakly-basic pyridine-containing compound 8 was equipotent with 7,
while the strongly basic 9 was more active (similar to reference compound 1). In contrast, the
strongly acidic molecule 5 was somewhat less active than 7. These outcomes suggested that the
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
positioning of the strongly ionizable moieties present in 5 and 9 by the designed alkyne fragment
did not completely eliminate electrostatic interactions between those entities and the negatively
charged phosphate backbone of the test oligonucleotide. The weaker alkylation activity
exhibited by compound 8 relative to 9 likely results from only minimal ionization of the pyridine
moiety at the neutral pH of the assay as compared to the considerable protonation expected for
the dimethylamino group. Somewhat surprisingly, the tetrazole-containing compound 6
displayed strong DNA alkylation properties that were analogous to those exhibited by reference
compound 1. The origin of this activity is not known at this time, but may involve interaction of
the (presumably ionized) tetrazole moiety with the test oligonucleotide and/or its associated
water network in a manner that somehow compensates for the anticipated unfavorable
electrostatic interactions with the DNA phosphates.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
An important caveat associated with the above assessments is that a single fragment of
test DNA was used to compare the molecules, and this fragment may not perfectly represent all
nucleic acid sequences that are expected to be targeted by the molecules in cells. The semi-
quantitative nature of the alkylation assay employed also adds some uncertainty to the described
comparisons. In spite of these limitations, the DNA-binding assessments indicated that
compounds 5-9 all retained measurable DNA alkylation properties that were considerably
differentiated from a minor groove binding reference compound that was incapable of covalently
modifying the nucleic acid (compound 10, Figure 4, Table 1). For this reason, we chose to
continue with the assessment of 5-9 as ADC payloads and the associated biological data will be
reported later in this work. However, although we attempted to normalize DNA affinity among
the new PBD-dimer molecules via the above assessments, this variable may still partially
10
ACS Paragon Plus Environment

Page 11 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
contribute to the in vitro and in vivo outcomes described for the associated compounds and
conjugates (see below for specific instances where this situation may apply).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Composite model of compounds 5-9 binding to the DNA minor groove. Pink atom
indicates location of ionizable groups attached to terminus of central phenyl-alkyne moiety.
.
Figure 4. Reference compound 10 employed in DNA alkylation assessments as inactive control.
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
We expected the ionizable groups present in compounds 5, 6, 8, and 9 to impact their cell
membrane permeability relative to the non-ionizable methyl fragment incorporated into
compound 7. Given the nature and diversity of the ionizable moieties, we also expected the pH
2
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
of the assay media to influence the outcomes. Accordingly, we employed PAMPA testing as a
surrogate for cell membrane permeability and assessed the new molecules at two different pH
values. The first value (pH 4.0) was selected to mimic the lysosomal environment that the
molecules were expected to encounter shortly after their release from a catabolized antibody-
drug conjugate. The second value (pH 7.4) was intended to duplicate the cellular cytosol as well
as the extracellular environment. As shown in Table 1, the measured PAMPA Pe parameters
correlated well with the calculated pKa data for the various compounds at both pH values tested.
In particular, the strongly acidic 5 exhibited low PAMPA Pe outcomes relative to neutral
compound 7 at both pH 4.0 and 7.4. In contrast, the weakly acidic 6 displayed a much larger Pe
at the former pH compared to the latter value that was consistent with its reduced ionization
under the more acidic conditions. PAMPA results for the two basic molecules 8 and 9 were also
predictably influenced by the pH employed in the assessments with both compounds exhibiting
smaller Pe values at the more acidic pH relative to neutral conditions. Collectively, the PAMPA
data indicated that PBD-dimers 5-9 might exhibit differing permeabilities toward biological
membranes and that such differences would likely be dependent on the pH of the associated
biological environment.
We next assessed the ability of compounds 5-9 to exhibit antiproliferation effects in in
vitro cell-based experiments. This testing was conducted at near-neutral pH (7.4), and two
diverse cell lines were employed: the B-cell-derived BJAB line and the KPL4 breast cancer line.
As shown in Table 1, compound 5, which contained the most acidic functional group and which
12
ACS Paragon Plus Environment

Page 13 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
displayed the lowest PAMPA Pe values, was not active in the cell-based assessments when
tested to relatively high concentrations (20 nM). In stark contrast, the neutral molecule 7
exhibited picomolar antiproliferation effects in both cell lines that were similar to those
displayed by the well-characterized control compound 1 (Table 1). These data suggested that, as
a result of sulfonic acid ionization at physiological pH, compound 5 was not able to effectively
transition through the outer membranes of either employed cell line (differences in DNA
binding/alkylation between 5 and 7 may also have contributed somewhat to the observed
outcomes). The weakly acidic 6 also exhibited significantly impaired activity in these cell assays
relative to compound 7 but was surprisingly more active than compound 5 in spite of a similarly
low pH 7.4 PAMPA Pe value. The improved DNA alkylation properties of 6 relative to 5 may
have contributed to the latter outcome. In contrast, the molecules bearing basic functional
groups (8 and 9) displayed only slightly attenuated antiproliferation potencies relative to 7 (Table
1). The latter results were consistent with both 8 and 9 exhibiting PAMPA values at pH 7.4 that
were predictive of good or moderate membrane permeability characteristics. In addition, the
stronger DNA alkylation properties of 9 as compared to 7 may have helped compensate for the
lower pH 7.4 PAMPA Pe value (and presumed poorer cell membrane permeability) displayed by
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
the former molecule.
Reassuringly, the BJAB and KPL4 outcomes also paralleled
antiproliferation results obtained by testing compounds 5-9 against panels of five, six, or seven
additional diverse cancer cell lines (Table 1 and Supporting Information).
Having profiled the new PBD-dimer compounds in the various in vitro assessments
described above in unconjugated form, we sought to attach them to antibodies to further
characterize their biological properties. Accordingly, we prepared CD22-targeting antibody
conjugates 11a-15a and the corresponding HER2 ADCs 11b-15b (Table 2). The well-known
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Valine-Citrulline (Val-Cit) protease-cleavable dipeptide trigger was incorporated into the design
of these conjugates, and the PBD-dimer payloads were connected to this moiety via an N10
carbamate functional group and a para-amino-benzyloxy (PAB) self-immolative spacer.21,22
ADCs bearing related PBD-dimer payloads that were linked to the antibodies via similar N10
carbamates displayed strong antigen-dependent anticancer activity in both in vitro and in vivo
assessments.4,22 These prior outcomes suggested that efficient delivery of the associated PBD-
dimer payloads occurred following treatment of cells and tumors with the conjugates, and we
were therefore confident that equally favorable delivery outcomes would occur with the new
molecules under study. In addition, the new conjugates 11a-15a and 11b-15b were constructed
via maleimide derivatization of cysteine residues that were engineered into the LC-K149 location
of the antibody structures [drug-antibody ratios (DARs) of ~2.0].23 This location and connection
approach were known from our prior studies to afford conjugates that exhibited highly favorable
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
4
in vitro and in vivo stability properties. Significant retro-Michael-related deconjugation was
therefore not expected to be observed with the new ADCs. In addition, the CD22 antigen was
known to be highly expressed on BJAB cells,25 and conjugates which target this entity had
previously shown strong in vitro and in vivo biological activities.26 Similarly, HER2 was well-
precedented as an ADC target, and the antigen was highly expressed on KPL4 breast cancer cells
(
IHC 3+).27 Conversely, the CD22 antigen was not highly expressed on KPL4 cells and the
HER2 receptor was not significantly present on the surface of B-cell lines (e.g., BJAB). These
differential expression patterns enabled the CD22 and HER2 conjugates to serve as respective
negative (i.e., non-targeted) controls for KPL4 and B-cell ADC experiments.
14
ACS Paragon Plus Environment

Page 15 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 2. Structures and biological properties of conjugates 11a-15a and 11b-15b.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Cell Antiproliferation IC50 (ng/mL)c
Calcd
pKaa
BJAB
(CD22)
WSU-DLCL2
KPL4
(HER2)
Jurkat
(neg. control)
DARb
R
Class
ADC Antigen
(CD22)
11a
CD22
HER2
CD22
HER2
CD22
HER2
CD22
HER2
CD22
HER2
CD22g
2.0
2.0
1.9
1.8
2.0
2.0
2.0
2.0
1.9
2.0
1.7
1.6
12
>20,000
67d
5.4
>20,000
14d
>10,000
6.2
>20,000
>20,000
15,000
>20,000
450
Strong
Acid
1.5
1
1b
2a
1
>1,000
0.54
Weak
Acid
5
.5
12b
3a
13b
>20,000
6.4
>20,000
1.6
1
>1,000
0.68
Neutral NA
Weak
2,200
5.4
560
820
14a
14b
15a
15b
2.2
>10,000
0.51
1,900
1,900
1,600
810
6
.0e
Base
4,200
4.2
2,500
1.9
>700
0.27
Strong
Base
.2e
8
1,300
14
870
Control 1f
Control 2f
NA
NA
NA
NA
7.1
ND
290
16a
16b HER2g
ND
ND
4.5
ND
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
pKa value of the R-group appended to each PBD-dimer payload calculated using the MoKa
b
c
software package (version 2.6.6). Drug-antibody ratio. Antiproliferation activity determined in
cell culture experiments using indicated cell line. As these assessments were not all
simultaneously performed, an appropriate control conjugate (16a or 16b) was included in every
determination to serve as an activity benchmark. Data are presented as geometric means (n = 4
per group); 95% confidence intervals are listed in Table S3. The targeted antigen expressed by
each cell line (CD22 or HER2) is listed at the top of each column. Control ADC (16a) IC50
was weaker than typically observed (BJAB = 74 ng/mL; WSU-DLCL2 = 20 ng/mL). Conjugate
acid. Control ADC which releases compound 1 PBD-dimer payload (see Figure 6). LC-V205
antibody attachment site; the alternate attachment site is not anticipated to impact in vitro
potency relative to the LC-K149 site. NA = not applicable. ND = not determined. Star symbol
= second copy of attached linker drug.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
d
e
f
g
As shown in Table 2 and Figure 5, all of the CD22-targeting conjugates 11a-15a
exhibited strong antiproliferation activities when tested against the BJAB cell line that were
similar to the potency displayed by a control conjugate which incorporated the well-known PBD-
dimer 1 (16a, Figure 6). In contrast, the corresponding HER2 conjugates 11b-15b afforded
drastically weaker outcomes in the BJAB assessments (Table 2, Figure 5). These results were
consistent with antigen-selective delivery of the new PBD-dimer payloads to the BJAB cells via
the CD22-targeting ADCs. Related antiproliferation outcomes were also observed when the
conjugates were tested against a second B-cell-derived and CD22-expressing cell line (WSU-
DLCL2, Table 2, Figure 5; see Figure S1 for CD22 expression in WSU-DLCL2 cells). The
near-equal activity of ADCs 11a-15a against the B-cell lines was notable given the disparate
antiproliferation potencies of the unconjugated PBD-dimer payloads depicted in Table 1 above
(see Figure 7 as well). The strong activities of conjugates 11a and 12a against the BJAB line
were particularly striking in light of the dramatically attenuated potencies observed for the
corresponding unconjugated payloads against the same cell line (Tables 1 and 2, Figure 7). The
described results paralleled those reported earlier for the auristatin-derived, acid-containing ADC
16
ACS Paragon Plus Environment

Page 17 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
payload MMAF which is a relatively weak in vitro antiproliferative agent in unconjugated form
but which exhibits significantly more potent cell-based activity when appropriately conjugated to
an antibody.12a Note that 12a exhibited anti-BJAB activity that was similar to control conjugate
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
6a when both conjugates were tested side-by-side in the same antiproliferation assessment
(Table 1). Thus, when normalized based on the activity of control compound 1, 12a is believed
to be approximately equipotent with 11a and 13a-15a against the B-cell lines.
Figure 5. In vitro antiproliferation activities of PBD-dimer conjugates as a function of cell line.a
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
aAntiproliferation data are taken from Table 2. The targeted antigen expressed by each cell line
(
CD22 or HER2) is listed below each name. pKa values refer to the R-group appended to each
PBD-dimer payload and were calculated using the MoKa software package (version 2.6.6); red
acidic, blue = basic (conjugate acid). *Control ADC (16a) IC50 was weaker than typically
=
observed; see Table 2 for additional details.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. Structure of control conjugates 16a (CD22) and 16b (HER2). Star symbol = second
copy of attached linker drug.
These results suggested that the PBD-dimer payloads containing acidic functional groups
(including highly ionizable moieties such as the sulfonic acid present in 5) were able to
efficiently transition from the cellular lysosome to the cytosol and/or nucleus following
antibody-mediated delivery. In the case of 5, this transition occurred in spite of the payload
exhibiting a low PAMPA Pe value at lysosomal pH that was suggestive of poor membrane
permeability (Table 1). The origin of the potent in vitro antiproliferation activity of conjugates
1
1a and 12a is not known with certainty at this time. Some possibilities include (1) active
transport of the released payloads across the lysosomal membrane,28 (2) the lysosomal
membrane being a more porous barrier with respect to small-molecule passive permeability
18
ACS Paragon Plus Environment

Page 19 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
relative to the outer cell membrane, and (3) underestimation of payload membrane permeability
properties via the PAMPA assessment. Additional experiments are thus required to further
define the precise mechanism(s) by which these modified PBD-dimer ADC payloads exert their
in vitro biological activities.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. BJAB or KPL4 in vitro antiproliferation activities of PBD-dimer conjugates as a
a
function of free payload in vitro antiproliferation potency.
aAntiproliferation data are taken from Tables 1 and 2. The targeted antigen expressed by each
cell line (CD22 or HER2) is listed below each name. pKa values refer to the R-group appended
to each PBD-dimer payload and were calculated using the MoKa software package (version
2
.6.6); red = acidic, blue = basic (conjugate acid). *Control ADC (16a) IC50 was weaker than
typically observed; see Table 2 for additional details.
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Strong in vitro antiproliferation potencies were also noted for the CD22-targeting
conjugates whose PBD-dimer payloads contained added basic moieties. Both ADCs 14a and
15a were equipotent with conjugate 13a for which the payload was the control molecule bearing
the neutral methyl R-group (Table 2, Figures 5 and 7). As with conjugate 11a above, the in vitro
activity of 15a was unexpected since the corresponding unconjugated payload associated with
the latter ADC exhibited a low PAMPA Pe value at lysosomal pH (Table 1). Like conjugate
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1a, the potent activity of 15a suggested efficient transition of the released payload from the
lysosome to the cellular cytosol and/or nucleus following antibody-mediated delivery by
mechanisms that are not fully understood at this time (see above for some possibilities).
Importantly, potent antigen-dependent antiproliferation activities were also observed
when the HER2-targeting conjugates 11b-15b were tested against the KPL4 cell line (Table 2,
Figure 5). The majority of the ADCs whose payloads contained ionizable groups were
equipotent in the KPL4 assessments to the conjugate for which the payload lacked any added
ionizable functionality (compare 12b, 14b, and 15b, with 13b; Table 2, Figure 5). The one
possible exception to this trend was conjugate 11b which exhibited an approximately 10-fold
weaker KPL4 IC50 value relative to 13b (Table 2, Figure 5). This result may reflect reduced
lysosome-cytosol transition efficiency in the KPL4 cells for the sulfonic acid-containing PBD-
dimer payload (5) corresponding to 11b relative to what was observed in the BJAB line.
However, the outcome may also be related to the reduced DNA alkylation activity displayed by
payload 5 compared to the PBD-dimer associated with conjugate 13b (compound 7, Table 1).
The described data mirrored results observed previously with maytansine-containing
trastuzumab-MCC-DM1 (T-DM1) which generates a cell-impermeable catabolite but is
nevertheless highly potent against the KPL4 cell line.27c,29 Importantly, the KPL4 outcomes also
20
ACS Paragon Plus Environment

Page 21 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
confirmed that the potent in vitro antiproliferation activities displayed by conjugates bearing the
new PBD-dimer payloads were associated with cancers in addition to those of B-cell origin.
Reassuringly, none of the tested conjugates were particularly active when tested against the
Jurkat cell line which did not express either the CD22 or HER2 antigen to a significant degree
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
Table 2, Figure 5).25,30
We were curious to understand how the in vitro biological activities observed for the
ADCs described above would translate into in vivo anti-tumor efficacy studies. As shown in
Figure 8A, single intravenous administration of CD22-targeting conjugate 13a (0.3 mg/kg) to
mice bearing WSU-DLCL2 xenografts afforded strong and durable anti-tumor efficacy over an
extended observation time period. Alternatively, administration of a matched dose of ADC 12a
resulted in considerably weaker in vivo activity against the same WSU-DLCL2 model. These
outcomes contrasted with the in vitro studies described in Table 2 above in which the two
conjugates exhibited highly similar WSU-DLCL2 antiproliferation potencies (when normalized
relative to control conjugate activity). One possibility for the differing in vitro and in vivo anti-
cancer results is a greater dependency on bystander effects for efficacy in the xenograft studies
relative to the cell-culture environment. Based on the PAMPA (pH 7.4) and antiproliferation
characterizations depicted in Table 1, the PBD-dimer released from 12a (compound 6) was not
expected to permeate from cell to cell as effectively as the corresponding payload delivered by
1
3a (compound 7). Such intercellular transition differences may be magnified in the
environment of the WSU-DLCL2 tumors relative to cell-culture assessments since a greater
proportion of the xenografted tumor cells are likely not exposed to direct contact by the
administered ADCs. Encouragingly, the in vivo WSU-DLCL2 efficacy of conjugate 12a could
be increased by administering a higher dose to the test animals (compare 1.0 mg/kg efficacy
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
outcomes with the 0.3 mg/kg results, Figure 8A). Consistent with the involvement of bystander
effects in WSU-DLCL2 efficacy outcomes, the activity of conjugate 11a (bearing the least cell-
permeable PBD-dimer payload 5; c.f., Table 1) was attenuated relative to that observed for ADC
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2a when 1.0 mg/kg doses of the two conjugates were compared (Figure 8A). Importantly, all
the conjugates tested in these initial in vivo assessments exhibited minimal effects on mouse
body weights during the course of the studies (Figure 8B).
Figure 8. In vivo activity of conjugates 11a, 12a, and 13a in the WSU-DLCL2 xenograft model.
A. Efficacy (tumor growth inhibition). B. Impact on mouse body weights. Blue traces: 11a,
red traces: 12a, green traces: 13a, black traces: vehicle control. Numbers next to traces
indicate doses of each conjugate (in mg/kg) that were administered once IV at the day 0 time
point. Group mean tumor volumes or body weight changes (+/- SEM) are plotted over the
duration of study (n=8/group). See Experimental section for additional details.
22
ACS Paragon Plus Environment

Page 23 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 9. In vivo activity of conjugates 11a, 11b, 13a, and 13b in the WSU-DLCL2 xenograft
model.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A. Efficacy (tumor growth inhibition). B. Impact on mouse body weights. Blue traces (solid
line): 11a, Blue traces (dotted line): 11b, green traces (solid line): 13a, green traces (dotted
line): 13b; black traces: vehicle control. Numbers next to traces indicate doses of each
conjugate (in mg/kg) that were administered once IV at the day 0 time point. Group mean tumor
volumes (+/- SEM) are plotted over the duration of study (n=5/group). See Experimental section
for additional details.
Intrigued by the above observations, we further probed the impact of the payload
modifications on in vivo activity by conducting a second xenograft study using the WSU-DLCL2
model. Figure 9A depicts the efficacy outcomes that were observed when conjugates 11a and
1
3a were tested at the 1.0 mg/kg dose level in this new experiment. As was observed in our
initial in vivo assessment of the two conjugates, the anti-tumor efficacy displayed by 11a
(bearing the less cell-permeable payload) was attenuated relative to the activity exhibited by 13a.
We also compared the performance of HER2-targeting control conjugates 11b and 13b in this
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
experiment (both administered at 1.0 mg/kg) and were surprised to observe the highly disparate
outcomes depicted in Figure 9A for the two ADCs. Specifically, strong in vivo efficacy was
noted for 13b that was similar to that observed for ADC 13a at the later time points of the study.
In contrast, the HER2-targeting 11b displayed no in vivo efficacy, and the poor performance of
this conjugate was easily distinguished from that of the related and much-more-active CD22-
targeting ADC (11a). Stability assessments performed with conjugates 11a and 13a confirmed
minimal deconjugation of the attached linker-drugs seven days after their administration to mice
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(Figure S2). In addition, no obvious biotransformations of the PBD-dimer payloads were
observed in these studies. Similar excellent in vivo stability properties are expected to be
associated with all conjugates described in this work. Collectively, these data suggested that the
permeability differences associated with the PBD-dimer payloads delivered by conjugates 11a
and 13a (compounds 5 and 7, respectively; Table 1) may have influenced the antigen-
independent activities of the corresponding HER2 conjugates (11b and 13b) in this WSU-
DLCL2 in vivo experiment. However, we did not compare the in vivo activities of 11b and 13b
at lower doses (a 0.3 mg/kg dose of the related CD22 conjugate 13a is sufficient to regress the
WSU-DLCL2 tumor model; Figure 8A), and we cannot rule out that the outcomes observed with
1
3b are simply a result of over-dosing the conjugate. As was the case with our initial
assessments in this tumor model with conjugates 11a, 12a, and 13a, it is possible that bystander
effects contributed to the observed non-specific outcomes shown in Figure 9 and such
phenomena were minimized by the poor cell permeability associated with the PBD-dimer
payload delivered by conjugates 11a and 11b. Interestingly, the activity displayed by conjugates
1
1b and 13b in the in vivo experiment paralleled the in vitro antiproliferation outcomes observed
for the ADCs (and the corresponding CD22-targeting conjugates) in non-specific activity
24
ACS Paragon Plus Environment

Page 25 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
assessments employing Jurkat cells which did not express either the HER2 or CD22 antigens
(i.e., 13a/13b were more potent than 11a/11b, Table 2, Figure 5). As was also observed in our
initial in vivo assessments, all the conjugates tested in these subsequent WSU-DLCL2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
experiments did not significantly impact mouse body weights over the duration of the studies
(although minor body weight loss was noted for conjugate 13a when administered at the 1.0
mg/kg dose level, Figure 9B).
Figure 10. In vivo activity of conjugates 13a, 14a, and 15a in the WSU-DLCL2 xenograft
model.
A. Efficacy (tumor growth inhibition). B. Impact on mouse body weights. Green traces: 13a,
orange traces: 14a, purple traces: 15a, black traces: vehicle control. Numbers next to traces
indicate doses of each conjugate (in mg/kg) that were administered once IV at the day 0 time
point. Group mean tumor volumes (+/- SEM) are plotted over the duration of study (n=5/group).
See Experimental section for additional details.
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
To complete our in vivo assessments of the conjugates bearing the new PBD-dimer
payloads, we profiled ADCs 13a, 14a, and 15a in the WSU-DLCL2 xenograft model. All of the
tested conjugates exhibited strong efficacy in these assessments following single IV
administrations at relatively low doses (Figure 10A). Consistent with the efficacy outcomes
associated with ADCs whose PBD-dimer payloads contained acidic functional groups, conjugate
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
5a (bearing the least-permeable and strongly basic PBD-dimer) displayed somewhat attenuated
in vivo antitumor activity relative to 13a and 14a at the 0.3 mg/kg dose (Figure 10A; c.f., Table
). More strikingly, conjugate 15a also had a much more deleterious impact on mouse body
1
weights as compared to 13a and 14a when the ADCs were administered at the 1.0 mg/kg dose
level (Figures 9B and 10B). Collectively, these data indicated that introduction of a strongly
basic moiety into the PBD-dimer payload structure simultaneously worsened the in vivo efficacy
and mouse tolerability outcomes of the corresponding ADCs as compared to conjugates bearing
unmodified (neutral) PBD-dimer molecules. These outcomes may be related to the propensity of
strongly basic small-molecules to indiscriminately bind to acidic tissue constituents (e.g.,
membrane phospholipids) and/or accumulate in acidic cellular compartments such as the
lysosome.31 Similarly, basic radioactive labels were observed to be widely distributed to the
tissues of xenografted mice following administration of the corresponding radiolabeled
antibodies.32 Additional studies involving other PBD-dimer scaffolds, alternate in vivo efficacy
models, and more-complete toxicology assessments are required to fully assess the generality of
these observations. However, the current data suggest that improvements in the preclinical
performance of PBD-dimer-containing ADCs can be realized by not including strongly basic
moieties in the associated payloads.
26
ACS Paragon Plus Environment

Page 27 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 1. Synthesis of unconjugated PBD-dimers 5-9.a
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
aReagents and conditions: (a) alkyne, Et
b) TFA, 0-15 °C, 1 h, 15-80%; (c) Et N, MsCl, DCM, 15 °C, 0.5 h; (d) LiBr, THF, 10 °C, 16 h,
94%; (e) Na SO , 3:1 MeOH:H O, 60 °C, 6 h, 31%. THP = 2-tetrahydropyran.
NH, Pd(Ph
P) , CuI, DMF, 100 °C, 10-20 min, 40-67%;
2
3
4
(
3
2
3
2
The unconjugated PBD-dimers described in this work were prepared by the methods
depicted in Scheme 1. For the synthesis of compounds 6, 8, and 9, iodo-containing intermediate
1
7 (prepared as depicted in Scheme 4) was coupled to various alkynes (19-21) using well-
precedented protocols.33 Subsequent removal of the Boc protecting groups present in the
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 69
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
coupling products (22-24) afforded the desired PBD-dimers. The 1-dimethylamino-2-propyne
(19) required for the synthesis of 22 was obtained from commercial sources while the
preparations of alkynes 20 [5-(but-3-yn-1-yl)-1H-tetrazole] and 21 [4-(but-3-yn-1-yl)pyridine]
are described in Scheme 5 below. Alternatively, the known THP-containing intermediate 1818
was similarly coupled to commercially available propargyl alcohol (25) or 1-butyne (26) to
respectively give products 27 and 28. The latter entity was deprotected under acidic conditions
to afford PBD-dimer 7. The former compound was transformed into sulfonic acid 29 (via
propargyl bromide 28), and the THP moieties present in 29 were subsequently removed to give
PBD-dimer 5.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 2. Preparation of antibody-drug conjugates 11a-15a and 11b-15b.a
aReagents and conditions: (a) DMF, Tris pH 6.5 to 7.5, 25 °C, 3-24 h.
28
ACS Paragon Plus Environment

Page 29 of 69
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The ADCs described in this work were prepared by conjugating anti-CD22 10F4v3 LC-
K149C or anti-HER2 7C2 LC-K149C mAbs to linker-drugs 31-35 via the engineered LC-K149C
cysteine residues according to previously published protocols (Scheme 2).34 ADCs produced
using the conjugation reaction conditions and purification schemes typically afforded protein
yields of 25 to 95%. All of the conjugates were characterized in regards to aggregation (SEC
HPLC), drug to antibody ratio (LCMS), and amount of unconjugated (free) drug present in the
final conjugates (LCMS). Detailed characterization data for each conjugate are provided in
Table S4 (Supporting Information).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The linker-drugs required for the preparation of the ADCs described in this work were
synthesized by the methods depicted in Scheme 3. The nitro groups present in intermediate 36
(prepared as shown in Scheme 4) were reduced to afford the corresponding bis-aniline
compound (37), and this entity was mono-protected in moderate yield by exposure to one
equivalent of di-tert-butyl dicarbonate to give 38. Transformation of the unprotected aniline
contained in 38 to the corresponding isocyanate (or possibly the carbamic chloride) followed by
condensation with commercially available Boc-Val-Cit-PAB-OH afforded carbamate 39 in
reasonable yield. The silyl protecting groups present in 39 were removed under acidic
3
5
conditions to generate diol 40, and this intermediate was oxidized with IBX to give compound
4
1. As detailed in Scheme 1 for the unconjugated PBD-dimers, coupling of 41 with alkynes 19,
3
3
2
0, 21, and 25 (Scheme 1) was accomplished using Sonogoshira protocols to afford compounds
43b-43e in good-to-moderate yields. The sulfonic acid-containing linker-drug 43a was prepared
by directly coupling prop-2-yne-1-sulfonic acid 42 (prepared as described in Scheme 5 below) to
4
1 in contrast to the related multi-step synthesis employed in Scheme 1 to generate the
corresponding unconjugated PBD-dimer 5. Removal of the Boc protecting groups present in
29
ACS Paragon Plus Environment