Isoquinoline-based biaryls as a robust scaffold for microtubule inhibitors

Article abstract of DOI:10.1016/j.ejmech.2019.111865

We here report the discovery of isoquinoline-based biaryls as a new scaffold for colchicine domain tubulin inhibitors. Colchicinoid inhibitors offer highly desirable cytotoxic and vascular disrupting bioactivities, but their further development requires improving in vivo robustness and tolerability: properties that both depend on the scaffold structure employed. We have developed isoquinoline-based biaryls as a novel scaffold for high-potency tubulin inhibitors, with excellent robustness, druglikeness, and facile late-stage structural diversification, accessible through a tolerant synthetic route. We confirmed their bioactivity mechanism in vitro, developed soluble prodrugs, and established safe in vivo dosing in mice. By addressing several problems facing the current families of inhibitors, we expect that this new scaffold will find a range of in vivo applications towards translational use in cancer therapy.

Full text of DOI:10.1016/j.ejmech.2019.111865

European Journal of Medicinal Chemistry xxx (xxxx) xxx
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Isoquinoline-based biaryls as a robust scaffold for microtubule
inhibitors
Yvonne Kraus 1, Carina Glas 1, Benedikt Melzer, Li Gao, Constanze Heise, Monique Preuße,
Julia Ahlfeld, Franz Bracher, Oliver Thorn-Seshold*
Department of Pharmacy e Center for Drug Research, Ludwig-Maximilians University of Munich, Butenandtstrasse 5-13, Munich, 81377, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
We here report the discovery of isoquinoline-based biaryls as a new scaffold for colchicine domain
tubulin inhibitors. Colchicinoid inhibitors offer highly desirable cytotoxic and vascular disrupting bio-
activities, but their further development requires improving in vivo robustness and tolerability: prop-
erties that both depend on the scaffold structure employed. We have developed isoquinoline-based
biaryls as a novel scaffold for high-potency tubulin inhibitors, with excellent robustness, druglikeness,
and facile late-stage structural diversification, accessible through a tolerant synthetic route. We
confirmed their bioactivity mechanism in vitro, developed soluble prodrugs, and established safe in vivo
dosing in mice. By addressing several problems facing the current families of inhibitors, we expect that
this new scaffold will find a range of in vivo applications towards translational use in cancer therapy.
© 2019 Elsevier Masson SAS. All rights reserved.
Received 9 August 2019
Received in revised form
8 October 2019
Accepted 6 November 2019
Available online xxx
Keywords:
Colchicine
Cytoskeleton
Isoquinoline
Microtubule dynamics
Tubulin polymerisation inhibitor
1. Introduction
target the tumour exterior but are poorly effective in the interior
[5].
The protein tubulin is a prime target for development of cancer
therapeutic inhibitors, most famously since the microtubule cyto-
skeleton it forms plays a crucial role in mitosis, which is required
for the progression of all cancer types [1,2]. Tubulin-inhibiting
Antimitotic cytotoxins from the classes of taxanes and of vinca al-
kaloids have become blockbuster drugs, and have been used in the
treatment of millions of patients [3]. A third class of tubulin in-
hibitors are the colchicine domain inhibitors (CDIs). CDIs with low-
or sub-nanomolar antiproliferative/cytotoxic activity in cellulo have
been extensively developed. However, in vivo their main bioactive
mechanism at tolerated doses is not cytotoxicity mediated through
tubulin disruption in target (tumour) cells, but instead vascular
disrupting agent (VDA) activity, mediated by disrupting microtu-
bule integrity in the endothelial cells lining the vascular system. In
the tumour neovascular network this disruption is poorly with-
stood, leading to staunching of tumoural blood flow and necrosis of
the tumour interior [4]. VDA activity is thus highly desirable as a
complementary mechanism to typical cytotoxins that usually
The prototypical requirements for CDI bioactivity have been well
studied [6]. CDIs typically display pharmacophore-relevant groups
on two aryl rings, here denoted as north and south rings, that are
usually twisted with respect to one another [7]. The south ring
typically features a 3,4,5-trimethoxy substitution pattern and the
north ring a monomethoxy substituent with a hydrogen or an
optional small polar group (OH, NH2, F) in the ortho position [8,9].
These substituents ensure potent binding to tubulin when correctly
arranged on the scaffold “exterior”. Combretastatin A-4 (CA4;
Fig. 1a) [10] illustrates these features. It is a Z-stilbene that acts as a
low nanomolar cytotoxin in cellulo, and its E-isomer - whose
methoxy groups do not project in the correct directions - is several
orders of magnitude less bioactive [11].
A crucial result observed during CDI development is that
although in cellulo a broad range of scaffold “interiors” for CDIs can
deliver good cytotoxic potency as long as the “exterior” sub-
stituents are correctly arranged, in vivo the scaffold choice often
determines medicinal chemistry factors critical for therapeutic
applications. For example, CA4 reached Phase III clinical trials as a
VDA in the form of its phosphate prodrug CA4P, while colchicine
(which has a similar disposition of substituents and similar in cel-
lulo potency) is not tolerated at the doses required for VDA activity
(Fig. 1a) [12,13]. Typical scaffold-dependent factors decisive for
* Corresponding author.
E-mail address: oliver.thorn-seshold@cup.lmu.de (O. Thorn-Seshold).
These authors contributed equally to this work.
1
https://doi.org/10.1016/j.ejmech.2019.111865
0223-5234/© 2019 Elsevier Masson SAS. All rights reserved.
Please cite this article as: Y. Kraus et al., Isoquinoline-based biaryls as a robust scaffold for microtubule inhibitors, European Journal of Medicinal
Chemistry, https://doi.org/10.1016/j.ejmech.2019.111865

2
Y. Kraus et al. / European Journal of Medicinal Chemistry xxx (xxxx) xxx
Fig. 1. (a) Archetypal literature-known CDIs. (b) Selected literature-reported biaryl-type CDIs. (c) IQTubs and their prodrugs. (d) Synthetic route to IQTubs, shown exemplarily for
IQTub4 and IQTub4P. Ts: para-toluenesulfonyl, TMP: 2,2,6,6-tetramethylpiperidinyl, Bn: benzyl, TFA: trifluoroacetic acid. (e,f) Control IQTubs, and other control compounds similar
to IQTubs, used in this study (IQTub6[29], IQTub8[30], 8 and 9[31], and 10[32] were previously reported but not biologically evaluated).
in vivo use include robustness to spontaneous deactivation in vivo
(e.g. stilbenes suffer cis to trans isomerisation [14]), as well as
traditional ADME/PK parameters including metabolism (e.g. cyto-
chrome degradation of, or inhibition by, the stilbene double bond
[15,16], or metabolic processing of aryl aryl ketone bridges [17,18])
and overall plasma residence time [14], as well as their ease of
synthetic access.
CA4-inspired CDIs with north and south rings held at a twist are
the class of CDI VDAs which has progressed furthest in clinical
trials, and these have been excellently reviewed [6,7,13]. A variety of
such scaffolds have been tested, aiming to improve upon the in vivo
properties of CA4P. Beyond stilbenes (like CA4P analogues CA1P
[19] and AVE8062[20]), colchicinoids such as ZD6126 [21], aryl aryl
ketones such as BNC105P [22], and a range of other scaffolds have
been used for CDIs, several of which feature in vivo-confirmed VDA
activity (Fig. 1a) [4]. Other classes of CDIs, such as the quasi-planar
indanocine, plinabulin, nocodazole and indanorin [13], are less
developed for in vivo applications and generally have not been
confirmed to display VDA activity. Improved techniques for
screening for VDA activity [23] are currently opening new per-
spectives for VDA development. While no CDIs have yet been
approved as VDAs [24], the value of investigating new CDI scaffolds
with potentially improved VDA activity has been elegantly stated
by Fojo [25]; we here report our progress in this area [26].
2. Results
Design and Synthesis. We primarily desired to develop CA4-
inspired CDIs with VDA potential, featuring a molecular scaffold
that would be highly robust to in vivo inactivation and/or metabolic
degradation mechanisms that affect the major VDA scaffolds in
clinical trials (stilbenes and aryl aryl ketones). By contrast to the
Please cite this article as: Y. Kraus et al., Isoquinoline-based biaryls as a robust scaffold for microtubule inhibitors, European Journal of Medicinal
Chemistry, https://doi.org/10.1016/j.ejmech.2019.111865

Y. Kraus et al. / European Journal of Medicinal Chemistry xxx (xxxx) xxx
3
range of twisted bis-aryl structures with one- or two-atom bridging
groups between the rings (e.g. BNC105P or CA4P, respectively) and
the CDIs with non-aromatic bicyclic rings (e.g. dihydronaph-
thalenes, benzosuberenes, benzodiazepines) [6,27,28], we found
that fully aromatic biaryls are sparsely reported and under-
explored as CDIs. This presented an opportunity, as a biaryl bond
promises to be more metabolically resistant than other bridging
bond types; biaryls can be druglike and do not require stereospe-
cific synthesis; a biaryl is permanently in a ring-twisted state that is
usually crucial for CDI bioactivity [6] (in contrast to stilbenes that
can isomerise to a planar state); and biaryls can be constructed by
easily diversifiable late-stage cross-couplings. Motivated by similar
logic, a very limited number of CDI biphenyls such as MP5-F9 [14]
have previously been developed (Fig. 1a), but their potency was
approximately 500-fold less than of the cognate stilbene-type CDIs
and as far as we are aware they were unfavourable for further
development. Alternatively, Miller et al. [17,18] reported a series of
triaryl CDIs (e.g. 1; Fig. 1b). However, we considered that triaryl
designs introduce undesirable synthetic complexity and molecular
weight, and they do not promise solubility high enough for in vivo
application (typically, > 10 mM injection concentration in aqueous
media for small animal studies). We therefore determined to
explore biaryls that occupy similar space to colchicine itself, i.e. that
feature one bicyclic and one monocyclic aromatic ring, in the hope
that this would allow high potency, good solubilisation, and mini-
mal complexity in our designs.
It was not obvious whether the north or the south ring inside
the CDI pharmacophore would be more suitable for replacement by
such a bicyclic aromatic ring. Cell-active CDIs with a large north
ring are well attested (e.g. 2 and 3; Fig. 1b) but the south ring has
also been reported to tolerate replacement with bicyclics [13],
which our recent experience [33e35] has supported. We therefore
determined to synthesise and evaluate both “south” and “north”
ring replacement types. We focused on isoquinolines as the bicyclic
ring (rather than e.g. naphthalenes) due to isoquinolines’
straightforward and flexible retrosynthesis, as well as the solubility
enhancement expected from including a ring nitrogen. Within the
south and north sets we also wanted to explore both hydrogen and
small polar group substituents at the variable position. This ligand-
based design process yielded south ring replacements IQTub1-2
and north ring replacements IQTub3-5. According to established
SAR [7,13] we desired that IQTub2 and/or IQTub4 would prove the
most potent members of their sets, according to preference of the
isoquinoline for occupying the south or the north position. To
render them potentially in vivo suitable we also designed their
phosphate prodrugs IQTub2P/IQTub4P (Fig. 1c). The phenolic
phosphate prodrug strategy has been particularly successful for CDI
VDAs (CA4P and CA1P [36], ZD6126 [21], BNC105P22) in conferring
suitable water-solubility for in vivo application (allows i.v. injection
at high concentration in aqueous media [22] and reduces plasma
protein binding compared to the otherwise lipophilic phenols [21])
[13]. Although the phosphates are not cell permeable they can be
efficiently and nonspecifically cleaved extracellularly by diverse
phosphatases to liberate the active, membrane-permeable phenols
which diffuse passively across membranes to exert bioactivity by
binding to tubulin at the lipophilic colchicine binding domain [37].
Noting that a CDI’s bioactivity is typically abolished if its variable
position is not occupied by small groups (but instead by OMe, Et,
etc) [13], we designed O-methylated IQTub6 in the hope it would
be inactive (a “designed-inactive” control [33,34]). This probes
whether IQTubs reproduce the general CDI structure-activity
relationship (SAR): if so, it would support the conjecture that
IQTubs are also CDIs. To explore other polarity-altering alkylations,
we also designed N-methylated quaternary IQTub7 and benzo-
dioxolo IQTub8, to extend our SAR in case the north ring
replacements proved bioactive (Fig. 1e). Searching the literature
across a similar scaffold space showed that only IQTub6, first re-
ported in the 1930s [29] as a papaverine analogue, and IQTub8,
reported by Reeve and Eareckson in 1950 [30] during a study to-
wards di- and tetrahydroisoquinoline analogues of podophyllo-
toxin, had previously been reported. However neither of these
compounds had been evaluated for bioactivity nor were they used
in SAR studies, so we determined to test them ourselves (see below
and Supporting Information).
The isoquinoline building blocks were synthesised by a modified
Pomeranz-Fritsch reaction according to Reimann [38]. These were
directly metalated at C-1 using Knochel-Hauser base [39] similarly
to
a published procedure [31], and Suzuki coupled with
appropriately-substituted phenylboronic esters [32], delivering the
targets IQTub1-5 and IQTub8 after optional deprotection. Meth-
ylating IQTub3 and IQTub4 yielded IQTub7 and IQTub6 respec-
tively. This synthetic protocol is tolerant to a range of substituents,
and allows late-stage diversification from a given isoquinoline
scaffold with high overall yields (Fig. 1d and Supporting Informa-
tion). All IQTubs were then taken into cell biological testing for
antiproliferative activity, using the resazurin viability assay in HeLa
cervical cancer cell line [40].
Structure-activity studies of IQTubs. Neither of the south ring
replacements showed strong cytotoxicity (apparent EC50s: IQTub1
~ 15 mM, IQTub2 ~ 50 mM). We controlled in two orthogonal ways
whether re-orientation of the substituents could permit bioactivity
while still retaining the south ring replacement strategy. Firstly,
noting that ketone-bridged compounds such as BNC105P (Fig. 1a)
and phenstatin [13] are potent CDIs, we slightly increased the
flexibility between the aromatic rings by introducing a one-atom
bridge to allow subtle changes in relative ring orientation while
maintaining the positioning of all methoxy groups - an attested
strategy in VDA development [6]. To this end we tested ketone
bridged alkaloid thalimicrinone (8) as well as its sp [3]-configured
carbinol synthetic precursor 9 [31] (Fig. 1e), however, neither was
bioactive in our hands (EC50 > 80 mM). Next, we reoriented the
north ring methoxy group from the meta into the para position, in
case a larger bite angle between the south and north ring methoxy
substituents would be better accommodated. For this we examined
oxoisoaporphine precursor 10 from previous work [32] (Fig.1e), but
this too was inactive in our hands (EC50 > 100 mM). We concluded
that south ring replacement with the isoquinoline was not viable.
Pleasingly though, the designed-active north ring replacements
IQTub3-5 showed strong bioactivity with cellular cytotoxicity EC50
values in the mid-nanomolar range (Fig. 2a). As hoped, the
designed-inactive O-methylated control compound IQTub6 was
indeed inactive (EC50 > 20 mM), suggesting that the north ring
replacement IQTubs obey the SAR expected for colchicine site
binders. We found no literature precedence for expecting either
activity or inactivity of charged compound IQTub7, but in the event
it was inactive (EC50 > 30 mM). Previously reported structure
IQTub8 (see also discussion in Supporting Information) [30]
featuring a benzodioxole north ring e which could be able to bind
similarly to isovanillyl IQTub4 according to literature SAR [13] e
proved inactive in our hands (EC50 > 20 mM; Fig. 2b).
However, as we observed turbidity while handling IQTub8, we
assumed that its apparent cellular bioactivity could have been
limited by poor solubility. We were therefore motivated to cross-
check whether water-soluble prodrugs of both scaffold orientations
would support our choice of the north ring strategy and confirm the
inactivity of the south ring type. We synthesised phenolic phos-
phate prodrug IQTub4P, which was soluble in aqueous buffer to
>10 mM (see Supporting Information). It displayed near-identical
cytotoxicity in cell culture as its free phenol form IQTub4 (Fig. 2)
indicating that O-phosphorylation is reliable as a strategy for
Please cite this article as: Y. Kraus et al., Isoquinoline-based biaryls as a robust scaffold for microtubule inhibitors, European Journal of Medicinal
Chemistry, https://doi.org/10.1016/j.ejmech.2019.111865

4
Y. Kraus et al. / European Journal of Medicinal Chemistry xxx (xxxx) xxx
Fig. 3. Drug property assessments of IQTub4. (a) Metabolic stability against mouse
and human liver microsomes (IQTub4 at 2 M). (b) Inhibition assay against five human
cytochrome P450s (IQTub4 at 10 M). (c) hERG channel inhibition assay.
m
m
Fig. 2. Cell viability assessment of IQTubs. (a) Antiproliferation assays showed strong
dose-dependent cytotoxicity for north ring designs IQTub3, IQTub4, IQTub4P and
IQTub5 (one representative experiment of three independent experiments shown). (b)
EC50 values for the bioactive compounds and permutation controls, as well as refer-
ence compound nocodazole. (HeLa cell line; 48 h incubation; resazurin assays; see also
Supporting Information).
moved to confirm whether the cytotoxic IQTubs’ mechanism of
action was indeed inhibition of tubulin polymerisation. We first
visualised their capacity for microtubule (MT) network perturba-
tion in cell culture. Sustained tubulin polymerisation inhibition
disorganises and ultimately depolymerises cellular MT networks.
We observed dose-dependent MT network depolymerisation with
the active IQTubs (Fig. 4a and b), and z-stack projection images
additionally revealed that cells treated with low doses of active
IQTubs already accumulated into rounded, mitotically arrested
states with defective spindles and chromosome alignments
(Fig. 4c). These are hallmarks of treatment with MT-depolymerising
agents. In contrast, neither designed-inactive SAR control IQTub6
(Fig. 4a) nor non-cytotoxic IQTub7-8 (Fig. S5) gave any MT dis-
organisation in this assay even at the highest tested concentrations,
suggesting that the bioactivity observed for IQTubs is tubulin-
specific.
The cytotoxic effects of CDIs in cellulo typically arise due to
interruption of tubulin reorganisation in the mitotic spindle, which
blocks mitosis, ultimately triggering cell death (as observed under
longer-term treatments in the viability assays). If IQTubs’ tubulin-
disrupting effects were indeed the cause of their observed cyto-
toxicity, treated cells should be measurably arrested in G2/M-
phase. We therefore used flow cytometry to perform cell cycle
analysis of HeLa cells under IQTub4P treatment. Results indeed
showed extensive G2/M-arrest from concentrations as low as
500 nM (Fig. 5a).
IQTubs inhibit tubulin polymerisation dynamics. Aiming to
confirm the molecular mechanism of IQTub bioactivity as one of
direct binding to tubulin (rather than binding to regulatory proteins
or affecting signalling), we then tested for inhibition of tubulin
polymerisation dynamics in cell-free settings using purified
tubulin. We observed closely similar polymerisation inhibiting ac-
tivity for IQTubs as for colchicine itself, which strongly supported
the direct binding interpretation and the mechanistic classification
of the IQTubs as microtubule depolymerisers (Fig. 5b).
Inhibition of tubulin polymerisation dynamics was then directly
assayed in cellulo by live-cell confocal video microscopy in HeLa
cells transfected to express EB3-YFP fusion protein. EB3-YFP is a
fluorescent marker that selectively labels the growing tips of mi-
crotubules and manifests as moving “comets” when MTs are
growing; the comets stop and disappear upon inhibition of tubulin
polymerisation [9]. Within 2 min of applying prodrug IQTub4P,
cellular EB3 comets stopped and vanished, indicating that extra-
cellular dephosphorylation to IQTub4 followed by internalisation
soluble prodrugs (see below); comparison of both compounds’
cytotoxicities in a second cell line (HL-60 human leukaemia line)
confirmed both their mid-nanomolar bioactivity and their close
correspondance of potency (EC50s in HL-60: IQTub4: 340 nM,
IQTub4P: 570 nM). To check whether IQTub1-2’s bioactivity could
have been undercut by poor solubility, we synthesised phosphate
prodrug IQTub2P, but its cytotoxic activity was not measurable
(EC50 [100 mM). We concluded that neither solubility nor sub-
stituent orientation effects had obscured the cytotoxicity of
IQTub1-2, and that the isoquinoline is not well-tolerated as a south
ring replacement. We attributed the weak apparent anti-
proliferative activity of the “inactive” IQTubs (e.g. IQTub1, IQTub6)
as tubulin-independent aggregation-dependent effects, known for
similar motifs [9,33], and confirmed this in mechanistic studies (see
below).
With their bioactivity matching our SAR understanding, we
considered that north ring replacement IQTubs were a viable
approach to a new scaffold for CDIs, and we determined to proceed
with the potent and well-soluble IQTub4P.
Drug-relevant properties of lead IQTub4/IQTub4P. We then
measured selected parameters of IQTub4P that would be relevant
for in vivo applications of the isoquinoline-based biaryl (using
IQTub4 in these assays since it should be the in situ-dephos-
phorylated IQTub4 that penetrates into cells to reach its site of
action [13]). By examining its compound stability in the presence of
mouse and human liver microsomes, we concluded that IQTub4
has good metabolic stability (Fig. 3a). Studies with the five major
human cytochrome P450s showed no significant inhibition of these
CYPs by IQTub4, indicating that it may escape typical drug-drug
interaction pathways (Fig. 3b). A fluorescence polarisation re-
porter assay for hERG channel inhibition showed no significant
effect at the highest tested concentration (Fig. 3c). Lastly, we
experimentally determined the distribution coefficient logD at pH
7.4 to be 2.18, which is within a range (1e3) considered optimum
for general in vivo use [41]. These results suggested that IQTub4
was suited for further pharmacological evaluation (see Supporting
Information for further details and assay benchmarking).
IQTubs inhibit microtubule structure and function. We next
Please cite this article as: Y. Kraus et al., Isoquinoline-based biaryls as a robust scaffold for microtubule inhibitors, European Journal of Medicinal
Chemistry, https://doi.org/10.1016/j.ejmech.2019.111865

Y. Kraus et al. / European Journal of Medicinal Chemistry xxx (xxxx) xxx
5
Fig. 4. Confocal microscopy assessment of cellular microtubule networks after treatment with IQTubs. (a) IQTub3 and IQTub5 induce MT network breakdown (panels on same
scale) while SAR control IQTub6 does not alter MT network integrity (compare to untreated control; panels on same scale). (b) Prodrug IQTub4P is active in cellulo at submicromolar
doses. (c) 120 nM IQTub4P already leads to mitotic arrests and the formation of aberrant multipolar spindles (tubulin panel) with resulting unstructured chromosome alignment
(DAPI panel). Maximum intensity projection along the z-axis of an image stack. (HeLa cells, 24 h treatment, green: a-tubulin stain for microtubule polymer network, blue: DAPI
nuclear counterstain, all scale bars 20
article.)
mm; see also Fig. S5). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this
microtubule depolymerisers with SAR and performance matching
expectations of CDIs. We considered this promising for their po-
tential for in vivo performance, ultimately aiming at development of
IQTubs as a new CDI VDA scaffold. To finish this early compound
development study, we therefore assessed safe dosing parameters
of IQTub4P in vivo in male and female Balb/c mice. Dose-escalation
studies (i.p. and i.v. administration routes) established a single-
administration maximal tolerated dose of 32 mg/kg (i.p.) and
50 mg/kg (i.v.). Repeated dosing (3 administrations, 48 h intervals)
were conducted with 25 mg/kg for both i.v. and i.p. routes and were
tolerated, thus IQTubs may avoid short-term cumulative toxicity,
which is known for the most promising CDI VDAs. We consider
these results favourable for further in vivo progress of this new
scaffold; by comparison the archetypal CDI colchicine has a murine
maximal tolerated dose around only 1 mg/kg [20].
3. Discussion and conclusion
The success of the taxanes and vinca alkaloids in clinical
oncology has stimulated intense interest in novel tubulin in-
hibitors. Colchicine domain inhibitors offer a biological activity
profile which not only includes the antimitotic and pro-apoptotic
effects common to all tubulin inhibitors, but also vascular dis-
rupting effects that are highly desirable for cancer treatment. This
has stimulated much research and development of CDIs and several
have reached late-stage clinical trials for cancer treatment,
although none have yet reached clinical approval.
Fig. 5. (a) Cell cycle analysis of IQTub4P-treated cells shows potent induction of G2/M
arrest. (b) IQTub5 inhibits tubulin polymerisation (cell-free) with similar potency to
the reference tubulin inhibitor colchicine (turbidimetric assay; IQTub5 at 20
colchicine at 16 M). (c) Imaging of MT polymerisation dynamics in live HeLa cells
transfected with EB3-YFP. EB3 comets were tracked, before and then 2 min after
treatment with 10 M IQTub4P (comet tails tinged purple, tips shown black); EB3
mM,
m
m
Here we have performed ligand-based design of isoquinoline-
based biaryls e a simple and robust chemical backbone e as new
scaffolds for potent CDIs. These IQTubs are easily accessible with
straightforward late-stage diversification using short, high-yielding
synthetic routes; and they have shown excellent performance in
early measurements of in vivo-relevant druglike properties. They
display high cellular potency in long-term cellular studies as well as
in short-term live-cell imaging. They deliver robustness to the
comet count (q) statistics were averaged over 7 cells (see Supporting Information for
details).
rapidly gives potent inhibition of cellular tubulin polymerisation
dynamics (Fig. 5c).
IQTubs are well-tolerated in vivo. These assays offered several
independent indications that IQTubs are potent, robust
Please cite this article as: Y. Kraus et al., Isoquinoline-based biaryls as a robust scaffold for microtubule inhibitors, European Journal of Medicinal
Chemistry, https://doi.org/10.1016/j.ejmech.2019.111865

6
Y. Kraus et al. / European Journal of Medicinal Chemistry xxx (xxxx) xxx
scaffold-specific spontaneous deactivation mechanism affecting
stilbene CDIs as well as metabolic liabilities of other major CDI
families, thus offering potentially advantageous ADME/PK
compared to the current CDIs. Through SAR studies as well as
extensive cellular and cell-free assays, we have clarified and sup-
ported their mechanism of action, showing multiple proofs that
their tubulin inhibition matches the bioactivity expected for CDIs.
Lastly, we have determined the in vivo tolerability for the lead
compound IQTub4P to be more than an order of magnitude higher
than that of the archetypal CDI colchicine, which is promising for
future research into the in vivo applications and VDA performance
of this potent, biologically robust IQT scaffold.
HRMS was performed by electron impact (EI) at 70 eV (Thermo
Finnigan MAT 95 or Jeol GCmate II spectrometers) unless stated
otherwise. HPLC-MS: Analytical measurements for determination
of the purities of the final products were performed on an Agilent
1100 SL coupled HPLC-MS system with H2O:MeCN eluent gradients
through a Thermo Scientific Hypersil GOLD™ C18 column (1.9 mm;
3 ꢀ 50 mm) maintained at 25 ꢁC, detected on an Agilent 1100 series
diode array detector and a Bruker Daltonics HCT-Ultra spectrom-
eter (ESI mode, unit m/z). Full experimental details are given in the
Supporting Information.
4.1. Representative synthesis: IQTub4
It should also be noted that VDAs are typically designed for short
plasma residence times (with high dose tolerance), enabling fast
but transient dose delivery to vascular endothelial cells, aiming that
in this way only the more susceptible tumour neovasculature
should lose microtubule-based mechanostasis during the time of
exposure, therefore giving tumour-selective vascular shutdown
[24,42]. In this context, the speed of dephosphorylation-
internalisation-inhibition observed with IQTub4P (Fig. 5c) is
particularly encouraging.
The current IQTubs already feature satisfying cellular potency,
but we fully expect that ongoing SAR studies will deliver analogues
with higher potency and/or significantly tuned biochemical prop-
erties. Still, their therapeutic performance cannot yet be predicted
since cellular potency is known not to be the limiting factor for CDI
VDA activity in vivo. Cardiotoxicity is a major dose- and therapy-
limiting parameter with CDIs, which is known to be scaffold-
dependent in that it is strongly affected by ADME/PK as well as
molecular structural parameters [24]. In this context, enabling
exploration of a new scaffold class with the potential for strongly
differentiated ADME/PK and scaffold-dependent tolerability is a
valuable step for the development of CDI VDAs and investigations
of these parameters are ongoing.
The ease and substituent tolerance of IQTub synthesis is another
important feature of this work, comparing favourably with many
CDI scaffold systems. However, more broadly, we consider that the
design logic of scaffold hopping from stilbene to isoquinoline-based
biaryl scaffolds to improve in vivo-relevant drug properties is an
exciting advance, promising medicinal chemistry applications to
targets well beyond the current CDIs.
In closing, we predict that by unlocking isoquinoline-based
biaryl tubulin inhibitors, this research will open up new possibil-
ities for refinement of tubulin inhibitors, for in vivo VDA develop-
ment, and for broader applications of robust biaryl designs to other
pharmacophores.
6-Benzyloxy-7-methoxyisoquinoline (4) (first reported by a
different procedure [43]). 4-Benzyloxy-3-methoxybenzaldehyde
(4.99 g, 20.6 mmol) was dissolved in toluene (50 mL) and amino-
acetaldehyde dimethyl acetal (2.38 g, 22.7 mmol) was added. Using
a Dean-Stark apparatus the reaction mixture was heated to reflux
for 16 h. After cooling to room temperature the volatiles were
evaporated and the crude redissolved in MeOH (100 mL). The
mixture was cooled to 0 ꢁC and NaBH4 (1.56 g, 41.2 mmol) was
added portionwise over 30 min, then the reaction mixture was
warmed to room temperature and stirred for 4 h. The volatiles were
evaporated and the crude taken up in water (100 mL) then
extracted with DCM (3 ꢀ 100 mL). The combined organic layers
were dried over Na2SO4, filtered and concentrated. The crude
product was redissolved in DCM (100 mL) and NaOH (1.40 g,
35.0 mmol) and tetrabutylammonium hydrogensulfate (0.490 g,
1.44 mmol) were added. After stirring for 10 min at room temper-
ature a solution of p-toluenesulfonyl chloride (4.71 g, 24.7 mmol) in
DCM (60 mL) was added dropwise over 1 h. The reaction mixture
was stirred for an additional hour. Water (100 mL) was added,
phases were separated, the organic phase washed with water
(2 ꢀ 100 mL) and brine (100 mL), dried over Na2SO4 and concen-
trated. The resulting crude product was dissolved in 1,4-dioxane
(150 mL), aq. 6
M HCl (30 mL) was added and the reaction mixture
heated to reflux for 16 h. After cooling to room temperature, the
solution was poured into water (150 mL) and washed with diethyl
ether (2 ꢀ 100 mL). Using a 6 M NaOH solution the aqueous phase
was adjusted to pH > 9 and extracted with DCM (3 ꢀ 150 mL). The
combined organic layers were dried over Na2SO4 and concentrated.
The resulting crude product was purified by flash column chro-
matography
(100%
EtOAc)
to
give
6-benzyloxy-7-
methoxyisoquinoline (4) as a white solid (1.29 g, 4.88 mmol, 24%).
HRMS (EIꢂþ): 265.1103 calculated for C17H15NO2ꢂþ [M]ꢂþ, 265.1110
found. 1H NMR (400 MHz, CDCl3):
d
(ppm) ¼ 9.04 (s, 1H, 1-H), 8.36
(d, J ¼ 5.6 Hz, 1H, 3-H), 7.51e7.47 (m, 2H, 2’-, 60-H), 7.44 (d,
J ¼ 5.6 Hz, 1H, 4-H), 7.43e7.38 (m, 2H, 3’-, 50-H), 7.36e7.31 (m, 1H,
40-H), 7.21 (s, 1H, 8-H), 7.10 (s,1H, 5-H), 5.29 (s, 2H, CH2), 4.03 (s, 3H,
4. Experimental section
Full and detailed chemical and biological protocols can be found
in the Supporting Information.
OCH3). 13C NMR (101 MHz, CDCl3):
d
(ppm) ¼ 152.3 (C-6), 150.8 (C-
7),150.1 (C-1),142.1 (C-3),136.2 (C-10),132.5 (C-4a),128.9 (2C, C-30,-
50), 128.3 (C-40), 127.4 (2C, C-20,-60), 125.0 (C- 8a), 119.4 (C-4), 106.5
(C-5), 105.7 (C-8), 70.9 (CH2), 56.2 (OCH3).
Compound synthesis and characterisation. All reactions were
performed with unpurified, undried, non-degassed solvents and
reagents from commercial suppliers (Sigma-Aldrich, TCI Europe,
Fisher Scientific etc.), used as obtained, under closed air atmo-
sphere without special precautions, unless otherwise described.
Manual flash column chromatography was performed on Merck
6-Benzyloxy-1-iodo-7-methoxyisoquinoline (5). To a solution
of 6-benzyloxy-7-methoxyisoquinoline (4) (531 mg, 2.00 mmol) in
dry THF (8 mL) was slowly added TMPMgCl$LiCl (1.0 M in THF/
toluene; 3.00 mL, 3.00 mmol) dropwise at room temperature. After
4 h the reaction mixture was cooled to 0 ꢁC, a solution of iodine
(761 mg, 3.00 mmol) in dry THF (3 mL) was added dropwise and the
resulting mixture stirred while warming to room temperature over
1 h. Sat. aq. NH4Cl (4 mL) and sat. aq. Na2S2O3 (4 mL) were added
and the organic materials extracted using DCM (3 ꢀ 50 mL). The
combined organic layers were dried over Na2SO4 and concentrated
in vacuo. The resulting crude product was purified by flash column
chromatography (DCM/EtOAc 5:1) to give 6-benzyloxy-1-iodo-7-
silica gel Si-60 (40e63 mm). MPLC flash column chromatography
was performed on a Biotage Isolera Spektra system, using Biotage
prepacked silica cartridges. Thin-layer chromatography (TLC) was
run on 0.25 mm Merck silica gel plates (60, F-254), with visual-
isation under UV light (254 nm and 365 nm). NMR characterisation
was performed by 1H and 13C NMR spectra recorded on an Avance
III HD 400 MHz Bruker BioSpin and Avance III HD 500 MHz Bruker
BioSpin (1H: 400 MHz and 500 MHz, 13C: 101 MHz and 126 MHz).
Please cite this article as: Y. Kraus et al., Isoquinoline-based biaryls as a robust scaffold for microtubule inhibitors, European Journal of Medicinal
Chemistry, https://doi.org/10.1016/j.ejmech.2019.111865

Y. Kraus et al. / European Journal of Medicinal Chemistry xxx (xxxx) xxx
7
methoxyisoquinoline (5) (0.482 g,1.23 mmol, 62%) as a brown solid.
Antiproliferation Assays. Mitochondrial diaphorase activity in
HeLa cell line was quantified by spectrophotometrically measuring
the reduction of resazurin (7-hydroxy-3H-phenoxazin-3-one-10-
oxide) to resorufin. 5,000 cells/well were seeded on 96-well
microtitre plates and treated with IQTubs 24 h later. Following
HRMS (EIꢂþ): 391.0069 calculated for C17H14INO2ꢂþ [M]ꢂþ, 391.0070
found. 1H NMR (400 MHz, CDCl3):
d
(ppm) ¼ 8.09 (d, J ¼ 5.4 Hz, 1H,
3-H), 7.50e7.46 (m, 2H, 2’-,60-H), 7.43e7.34 (m, 5H, 3’-,4’-,5’-, 4-, 8-
H), 7.04 (s, 1H, 5-H), 5.30 (s, 2H, CH2), 4.08 (s, 3H, OCH3). 13C NMR
(101 MHz, CDCl3):
d
(ppm) ¼ 152.7 (C-6), 152.0 (C-7), 142.0 (C-3),
48 h of treatment, cells were incubated with 20 mL of 0.15 mg/mL
135.9 (C-10), 132.5 (C-1), 128.9 (2C, C-30, -50), 128.5 (C-40), 128.3 (C-
8a), 127.5 (2C, C-20, -60), 124.9 (C-4a), 120.3 (C-4), 111.4 (C-8), 106.9
(C-5), 71.1 (CH2), 56.4 (OCH3).
resazurin per well for 3 h at 37 ꢁC. The resorufin fluorescence
(excitation 544 nm, emission 590 nm) was then measured.
HL-60 cells were seeded in 96-well plates at 9,000 cells/well and
incubated for 24 h before treatment with IQTubs. 24 h later, cells
were treated with 0.5 mg/mL (3-(4,5-dimethylthiazol-2-yl)-2,5-
diphenyltetrazolium bromide (MTT) for 2 h; the medium was
aspirated and formazan crystals were re-dissolved in DMSO
6-Benzyloxy-7-methoxy-1-(3,4,5-trimethoxyphenyl)iso-
quinoline (6). To
a
solution of 6-benzyloxy-1-iodo-7-
methoxyisoquinoline (5) (391 mg, 1.00 mmol) in THF (6 mL) was
added 3,4,5-trimethoxyphenylboronic acid (254 mg, 1.20 mmol),
Pd(PPh3)4 (59.0 mg, 0.0500 mmol) and aq. K2CO3 solution (1.0 M;
3.00 mL, 3.00 mmol). The reaction mixture was stirred in a sealed
pressure tube under nitrogen at 90 ꢁC for 16 h. After cooling to
room temperature the mixture was poured into water (50 mL) and
extracted with EtOAc (3 ꢀ 50 mL). The combined organic layers
were dried over Na2SO4 and concentrated in vacuo. The crude
product was purified by flash column chromatography (EtOAc:DCM
2:1) to give 6-benzyloxy-7-methoxy-1-(3,4,5-trimethoxyphenyl)
isoquinoline (6) (395 mg, 0.915 mmol, 92%) as a yellow solid. HRMS
(EIꢂþ): 431.1733 calculated for C26H25NOꢂ5þ [M]ꢂþ, 431.1729 found.
(190 mL) to measure their absorbance at 570 nm.
Fluorescence and absorbance were measured on a FLUOstar
Omega microplate reader (BMG Labtech), averaged over the tech-
nical replicates and normalised as viability by reference to the
cosolvent-only control set as 100%, and to cell-free (resazurin) or
Triton-X®-100-treated (MTT) controls set as 0%. Results are means
of at least three independent experiments.
Cell cycle analysis. IQTubs were added to HeLa cells in 24-well
plates (seeding density: 50,000 cells/well) and incubated for 24 h.
Cells were collected and stained with 2 mg/mL propidium iodide
1H NMR (400 MHz, CD2Cl2):
d
(ppm) ¼ 8.40 (d, J ¼ 5.5 Hz, 1H, 3-H),
(PI) at 4 ꢁC for 30 min. Following PI staining, cells were analysed by
flow cytometry using a BD LSR Fortessa flow cytometer (Becton
Dickinson) run by BD FACSDiva software. 30,000 cells were
measured per condition and the data were transferred to Flowing
software for cell cycle analysis. Cells were sorted into sub-G1, G1, S
and G2/M phase according to DNA content (PI signal).
Immunofluorescence staining. HeLa cells seeded on glass
coverslips in 24-well plates (50,000 cells/well) were left to adhere
for 18 h then treated for 24 h with IQTubs. Cover slips were washed
then fixed with 0.5% glutaraldehyde, quenched with 0.1% NaBH4,
blocked with PBS þ10% FCS, and treated with rabbit alpha-tubulin
primary antibody (Abcam ab18251; 1:400 in PBS þ 10% FCS) for 1 h;
after washing with PBS, cells were incubated with goat-anti-rabbit
Alexa fluor 488 secondary antibody (Abcam, ab150077; 1:400 in
PBS þ 10% FCS) for 1 h. After washing with PBS, coverslips were
mounted onto glass slides using Roti-Mount FluorCare DAPI (Roth)
and imaged with a Zeiss LSM Meta confocal microscope. Images
were processed using Fiji software. For maximum intensity pro-
jections, images were recorded at different focal planes by incre-
7.52e7.48 (m, 3H, 2’’-, 6’’-, 4-H), 7.47 (s, 1H, 8-H), 7.46e7.41 (m, 2H,
3’’-, 500-H), 7.40e7.35 (m, 1H, 400-H), 7.23 (s, 1H, 5-H), 6.93 (s, 2H, 2’-,
60-H), 5.24 (s, 2H, CH2), 3.88 (s, 6H, 3’-, 50-OCH3), 3.87 (s, 3H, 40-
OCH3), 3.86 (s, 3H, 7-OCH3). 13C NMR (101 MHz, CD2Cl2):
d
(ppm) ¼ 158.5 (C-1), 153.8 (2C, C-30, -50), 152.4 (C-6), 151.0 (C-7),
141.7 (C-3), 138.8 (C-40), 136.7 (C-100), 136.2 (C-10), 134.2 (C-4a),
129.2 (2C, C-300, -500), 128.9 (C-400), 128.5 (2C, C-200, -600), 123.0 (C-8a),
119.2 (C-4), 107.5 (2C, C-20, -60), 107.0 (C-5), 106.3 (C-8), 71.3 (CH2),
61.1 (40-OCH3), 56.7 (2C, 3’-, 50-OCH3), 56.4 (7-OCH3).
7-Methoxy-1-(3,4,5-trimethoxyphenyl)isoquinolin-6-ol
(IQTub4). To
a solution of 6-benzyloxy-7-methoxy-1-(3,4,5-
trimethoxyphenyl)isoquinoline (6) (341 mg, 0.790 mmol) in
MeOH (30 mL) was added Pd/C (10%, 100 mg). The mixture was
stirred vigorously under an atmosphere of hydrogen at room
temperature for 24 h, filtered through a pad of celite, and the filtrate
concentrated in vacuo to give IQTub4 (220 mg, 0.645 mmol, 82%) as
a white solid. HRMS (EIꢂþ): 341.1263 calculated for C19H19NOꢂ5þ
[M]ꢂþ
,
341.1256 found. 1H NMR (400 MHz, (CD3)2SO):
d
(ppm) ¼ 10.29 (s, 1H, OH), 8.31 (d, J ¼ 5.6 Hz, 1H, 3-H), 7.53 (d,
mentally stepping through the sample (step size 1e2 mm) and
maximum intensity projections were obtained using Fiji.
J ¼ 5.6 Hz,1H, 4-H), 7.44 (s, 1H, 8-H), 7.22 (s, 1H, 5-H), 6.99 (s, 2H, 2’-
, 60-H), 3.84 (s, 6H, 3’-, 50-OCH3), 3.82 (s, 3H, 7-OCH3), 3.76 (s, 3H, 40-
EB3 comet assay. [45] HeLa cells (12,000 cells/well) were
seeded on 8-well ibiTreat slides (ibidi) 24 h prior to transfection.
OCH3). 13C NMR (101 MHz, (CD3)2SO):
d
(ppm) ¼ 156.9 (C-1), 152.7
m
(2C, C-30, -50), 150.8 (C-6), 149.6 (C-7), 140.4 (C-3), 137.6 (C-40), 135.3
(C-10), 133.6 (C-4a), 121.1 (C-8a), 118.1 (C-4), 108.5 (C-5), 107.0 (2C,
C-20, -60),105.3 (C-8), 60.1 (40-OCH3), 56.0 (2C, 3’-, 50-OCH3), 55.4 (7-
OCH3). HPLC purity: >95%.
Cells were transiently transfected with EB3-YFP plasmid using
jetPRIME reagent (Polyplus) according to the manufacturer’s in-
structions. Cells were imaged 24 h later, under 37 ꢁC and 5% CO2
atmosphere using an UltraVIEW Vox spinning disc confocal mi-
croscope (PerkinElmer) equipped with an EMCCD camera (Hama-
matsu, Japan) and operated with Volocity software. After focussing
on cells on the microscope stage, 5 imaging frames were acquired
to set a reference measure for EB3 comet activity, then IQTub4P
was added cautiously and cells incubated for 2 min before acquiring
another 5 frames to quantify post-treatment EB3 comet activity.
Cells were imaged at 514 nm (20% laser power, 300 ms exposure
time, 45 frames/min). For EB3 comet statistics, 7 cells from three
independent trials were taken. EB3 comets were counted with a Fiji
plugin based on the “Find maxima” function from the NIH.
Tubulin Polymerisation in vitro assay. 99% purity tubulin from
porcine brain was used in polymerisation assays run according to
manufacturer’s instructions (Cytoskeleton Inc., cat. #T240). Tubulin
was pre-incubated for 10 min at 37 ꢁC with either IQTub5 (20
mM)
or colchicine (16
mM) (in buffer (with 3% DMSO, 10% glycerol) as
appropriate; at time zero, GTP (1 mM) was added and the absor-
bance at 340 nm was monitored over time at 37 ꢁC [44].
Cell Culture. HeLa cells were maintained under standard cell
culture conditions in Dulbecco’s modified Eagle’s medium supple-
mented with 10% fetal calf serum (FCS),100 U/mL penicillin and 100
U/mL streptomycin, at 37 ꢁC in a 5% CO2 atmosphere. HL-60 cells
were cultured in RPMI 1640 medium with 10% FCS without anti-
biotics at 37 ꢁC in a 5% CO2 atmosphere. Compounds and cosolvent
(DMSO; 1% final concentration) were added via a D300e digital
dispenser (Tecan).
Author contributions
Y.K. performed immunofluorescence staining, live-cell micro-
scopy and viability studies. C.G. performed synthesis and
Please cite this article as: Y. Kraus et al., Isoquinoline-based biaryls as a robust scaffold for microtubule inhibitors, European Journal of Medicinal
Chemistry, https://doi.org/10.1016/j.ejmech.2019.111865

8
Y. Kraus et al. / European Journal of Medicinal Chemistry xxx (xxxx) xxx
derivatives retain in vitro anti-cancer activity dependent on mitotic arrest,
PLoS One 12 (3) (2017), e0171806, https://doi.org/10.1371/
journal.pone.0171806.
coordinated chemical data assembly. B.M. and L.G. performed
synthesis. C.H. and M.P. performed cell cycle analysis and viability
studies. J.A. supervised biology, performed cell cycle analysis and
viability studies, and coordinated biological data assembly. F.B.
planned and supervised synthesis. O.T.-S. designed the study, su-
pervised synthesis and biology, coordinated data assembly and
wrote the manuscript.
[15] J. McNulty, S. van den Berg, D. Ma, D. Tarade, S. Joshi, J. Church, S. Pandey,
Antimitotic activity of structurally simplified biaryl analogs of the anticancer
agents colchicine and combretastatin A4, Bioorg. Med. Chem. Lett 25 (1)
(2015) 117e121, https://doi.org/10.1016/j.bmcl.2014.10.090.
[16] S. Kim, H. Ko, J.E. Park, S. Jung, S.K. Lee, Y.-J. Chun, Design, synthesis, and
discovery of novel trans-stilbene analogues as potent and selective human
cytochrome P450 1B1 inhibitors, J. Med. Chem. 45 (1) (2002) 160e164,
https://doi.org/10.1021/jm010298j.
Declaration of competing interest
The authors declare no conflict of interest.
Acknowledgments
[17] D.-J. Hwang, J. Wang, W. Li, D.D. Miller, Structural optimization of indole
derivatives acting at colchicine binding site as potential anticancer agents,
ACS Med. Chem. Lett.
acsmedchemlett.5b00208.
6 (9) (2015) 993e997, https://doi.org/10.1021/
[18] Y. Lu, J. Chen, J. Wang, C.-M. Li, S. Ahn, C.M. Barrett, J.T. Dalton, W. Li,
D.D. Miller, Design, synthesis, and biological evaluation of stable colchicine
binding site tubulin inhibitors as potential anticancer agents, J. Med. Chem. 57
(17) (2014) 7355e7366, https://doi.org/10.1021/jm500764v.
This research was supported by funds from the German
Research Foundation (DFG: SFB1032 Nanoagents for Spatiotem-
poral Control project B09 to O.T.-S.; SFB TRR 152 project P24
number 239283807 to O.T.-S; and Emmy Noether grant to O.T.-S.)
and the German Ministry of Education and Research (GO-Bio
grant to O.T.-S.). We thank Rebekkah Bingham (LMU) for per-
forming the tubulin polymerisation assay, Yuliia Holota (Bienta) for
the in vitro ADME and in vivo tolerated dose assessments, Martina
Stadler (LMU) for the HL-60 viability crosscheck, K. T. Wanner
(LMU) for kind support and collegial discussions, and H. Harz and I.
Solvei (LMU microscopy platform CALM) for microscopy support.
[19] S.A. Hill, G.M. Tozer, G.R. Pettit, D.J. Chaplin, Preclinical evaluation of the
antitumour activity of the novel vascular targeting agent oxi 4503, Anticancer
Res. 22 (3) (2002) 1453e1458.
[20] Y. Nihei, M. Suzuki, A. Okano, T. Tsuji, Y. Akiyama, T. Tsuruo, S. Saito, K. Hori,
Y. Sato, Evaluation of antivascular and antimitotic effects of tubulin binding
agents in solid tumor therapy, Jpn. J. Cancer Res. 90 (12) (1999) 1387e1395,
https://doi.org/10.1111/j.1349-7006.1999.tb00724.x.
[21] P.D. Davis, G.J. Dougherty, D.C. Blakey, S.M. Galbraith, G.M. Tozer, A.L. Holder,
M.A. Naylor, J. Nolan, M.R.L. Stratford, D.J. Chaplin, et al., ZD6126: a novel
vascular-targeting agent that causes selective destruction of tumor vascula-
ture, Cancer Res. 62 (24) (2002) 7247e7253.
[22] G. Kremmidiotis, A.F. Leske, T.C. Lavranos, D. Beaumont, J. Gasic, A. Hall,
M. O’Callaghan, C.A. Matthews, B. Flynn, BNC105: a novel tubulin polymeri-
zation inhibitor that selectively disrupts tumor vasculature and displays
single-agent antitumor efficacy, Mol. Cancer Ther. 9 (6) (2010) 1562, https://
doi.org/10.1158/1535-7163.mct-09-0815.
[23] R.P. Mason, D. Zhao, L. Liu, M.L. Trawick, K.G. Pinney, A perspective on vascular
disrupting agents that interact with tubulin: preclinical tumor imaging and
biological assessment, Integr. Biol. (Camb) 3 (4) (2011) 375e387, https://
doi.org/10.1039/c0ib00135j.
[24] J.H. Gill, K.L. Rockley, C. De Santis, A.K. Mohamed, Vascular disrupting agents
in cancer treatment: cardiovascular toxicity and implications for co-
administration with other cancer chemotherapeutics, Pharmacol. Ther. 202
(2019) 18e31, https://doi.org/10.1016/j.pharmthera.2019.06.001.
[25] P. Giannakakou, D. Sackett, T. Fojo, Tubulin/microtubules: still a promising
target for new chemotherapeutic agents, J. Natl. Cancer Inst. 92 (3) (2000)
182e183, https://doi.org/10.1093/jnci/92.3.182.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.ejmech.2019.111865.
References
[1] T.J. Mitchison, The proliferation rate paradox in antimitotic chemotherapy,
Mol. Biol. Cell 23 (1) (2012) 1e6, https://doi.org/10.1091/mbc.e10-04-0335.
[2] J.R. Peterson, T.J. Mitchison, Small molecules, big impact: a history of chemical
inhibitors and the cytoskeleton, Chem. Biol. 9 (12) (2002) 1275e1285, https://
doi.org/10.1016/s1074-5521(02)00284-3.
[26] O. Thorn-Seshold, F. Bracher, B. Melzer, Isoquinoline Biaryl Compounds, 2018.
EP18207030.
[27] M. Sriram, J.J. Hall, N.C. Grohmann, T.E. Strecker, T. Wootton, A. Franken,
M.L. Trawick, K.G. Pinney, Design, synthesis and biological evaluation of
dihydronaphthalene and benzosuberene analogs of the combretastatins as
inhibitors of tubulin polymerization in cancer chemotherapy, Bioorg. Med.
Chem. 16 (17) (2008) 8161e8171, https://doi.org/10.1016/j.bmc.2008.07.050.
[28] Y. Pang, H. Lin, C. Ou, Y. Cao, B. An, J. Yan, X. Li, Design, synthesis, and bio-
logical evaluation of novel benzodiazepine derivatives as anticancer agents
through inhibition of tubulin polymerization in vitro and in vivo, Eur. J. Med.
Chem. 182 (2019) 111670, https://doi.org/10.1016/j.ejmech.2019.111670.
[29] Verfahren Zur Darstellung von 1- (3, 4, 5-Trialkoxyphenyl)-6, 7-
Dialkoxyisochinolinen. DE614703.
[30] W. Reeve, W.M. Eareckson, Synthesis of some isoquinoline derivatives related
to podophyllotoxin, J. Am. Chem. Soc. 72 (11) (1950) 5195e5197, https://
doi.org/10.1021/ja01167a114.
[31] B. Melzer, F. Bracher, A divergent approach to benzylisoquinoline-type and
oxoaporphine alkaloids via regioselective direct ring metalation of alkoxy
isoquinolines, Org. Biomol. Chem. 13 (28) (2015) 7664e7672, https://doi.org/
10.1039/c5ob00926j.
[3] D.G.I. Kingston, Taxol, a molecule for all seasons, Chem. Commun. 10 (2001)
867e880, https://doi.org/10.1039/b100070p.
[4] D.W. Siemann, The unique characteristics of tumor vasculature and preclinical
evidence for its selective disruption by tumor-vascular disrupting agents,
Cancer Treat Rev. 37 (1) (2011) 63e74, https://doi.org/10.1016/
j.ctrv.2010.05.001.
[5] G.M. Tozer, C. Kanthou, B.C. Baguley, Disrupting tumour blood vessels, Nat.
Rev. Cancer 5 (6) (2005) 423e435, https://doi.org/10.1038/nrc1628.
[6] K.G. Pinney, Molecular recognition of the colchicine binding site as a design
paradigm for the discovery and development of vascular disrupting agents, in:
Vascular-Targeted Therapies in Oncology, Wiley, 2006, pp. 95e121.
[7] T.L. Nguyen, C. McGrath, A.R. Hermone, J.C. Burnett, D.W. Zaharevitz, B.W. Day,
P. Wipf, E. Hamel, R. Gussio, A common pharmacophore for a diverse set of
colchicine site inhibitors using a structure-based approach, J. Med. Chem. 48
(19) (2005) 6107e6116, https://doi.org/10.1021/jm050502t.
[8] R. Gaspari, A.E. Prota, K. Bargsten, A. Cavalli, M.O. Steinmetz, Structural basis of
cis- and trans-combretastatin binding to tubulin, Chem 2 (1) (2017) 102e113,
https://doi.org/10.1016/j.chempr.2016.12.005.
[32] B. Melzer, F. Bracher, A novel approach to oxoisoaporphine alkaloids via
regioselective metalation of alkoxy isoquinolines, Beilstein J. Org. Chem. 13
(2017) 1564e1571, https://doi.org/10.3762/bjoc.13.156.
[33] A. Sailer, F. Ermer, Y. Kraus, F. Lutter, C. Donau, M. Bremerich, J. Ahlfeld,
O. Thorn-Seshold, Hemithioindigos as desymmetrised molecular switch
scaffolds: design control over the isomer-dependency of potent photo-
switchable antimitotic bioactivity in cellulo, Chembiochem 20 (2019)
1305e1314, https://doi.org/10.1002/cbic.201800752.
[9] M. Borowiak, W. Nahaboo, M. Reynders, K. Nekolla, P. Jalinot, J. Hasserodt,
M. Rehberg, M. Delattre, S. Zahler, A. Vollmar, et al., Photoswitchable in-
hibitors of microtubule dynamics optically control mitosis and cell death, Cell
162 (2) (2015) 403e411, https://doi.org/10.1016/j.cell.2015.06.049.
[10] G.R. Pettit, S.B. Singh, E. Hamel, C.M. Lin, D.S. Alberts, D. Garcia-Kendal,
Isolation and structure of the strong cell growth and tubulin inhibitor com-
bretastatin A-4, Experientia 45 (2) (1989) 209e211, https://doi.org/10.1007/
bf01954881.
[34] L. Gao, Y. Kraus, M. Wranik, T. Weinert, S.D. Pritzl, J.C.M. Meiring, R. Bingham,
N. Olieric, A. Akhmanova, T. Lohmüller, et al., Photoswitchable microtubule
inhibitors enabling robust, GFP-orthogonal optical control over the tubulin
cytoskeleton, 2019, p. 716233, https://doi.org/10.1101/716233. bioRxiv 2019.
[35] A. Sailer, F. Ermer, Y. Kraus, R. Bingham, F.H. Lutter, J. Ahlfeld, O. Thorn-Ses-
hold, Potent Hemithioindigo-Based Antimitotics Photocontrol the Microtu-
bule Cytoskeleton in Cellulo, ChemRxiv, 2019. https://doi.org/10.26434/
chemrxiv.9176747.v1.
[11] J.A. Woods, J.A. Hadfield, G.R. Pettit, B.W. Fox, A.T. McGown, The interaction
with tubulin of a series of stilbenes based on combretastatin A-4, Br. J. Canc.
71 (4) (1995) 705e711, https://doi.org/10.1038/bjc.1995.138.
[12] G.M. Tozer, C. Kanthou, C.S. Parkins, S.A. Hill, The biology of the com-
bretastatins as tumour vascular targeting agents, Int. J. Exp. Pathol. 83 (1)
(2002) 21e38, https://doi.org/10.1046/j.1365-2613.2002.00211.x.
[13] G.C. Tron, T. Pirali, G. Sorba, F. Pagliai, S. Busacca, A.A. Genazzani, Medicinal
chemistry of combretastatin A4: present and future directions, J. Med. Chem.
49 (11) (2006) 3033e3044, https://doi.org/10.1021/jm0512903.
[36] K.G. Pinney, G.R. Pettit, M.L. Trawick, C. Jelinek, D.J. Chaplin, The discovery and
development of the combretastatins, in: Anticancer Agents from Natural
Products, CRC Press, Boca Raton, 2011, pp. 27e63.
[14] D. Tarade, D. Ma, C. Pignanelli, F. Mansour, D. Simard, S. van den Berg, J. Gauld,
J. McNulty, S. Pandey, Structurally simplified biphenyl combretastatin A4
Please cite this article as: Y. Kraus et al., Isoquinoline-based biaryls as a robust scaffold for microtubule inhibitors, European Journal of Medicinal
Chemistry, https://doi.org/10.1016/j.ejmech.2019.111865

Y. Kraus et al. / European Journal of Medicinal Chemistry xxx (xxxx) xxx
9
[37] National Cancer Institute, NCI Drug Dictionary: fosbretabulin disodium.
https://www.cancer.gov/publications/dictionaries/cancer-drug/def/
fosbretabulin-disodium.
[38] E. Reimann, H. Renz, Protoberberine Aus Reissert-Verbindungen, 2. Mitt.: eine
Neue Synthese von 8-Methyldibenzo[a,g]Chinolizidinen, Arch. Pharm. 326 (5)
(1993) 253e258, https://doi.org/10.1002/ardp.19933260502.
[39] A. Krasovskiy, V. Krasovskaya, P. Knochel, Mixed Mg/Li amides of the type
R2NMgCl,LiCl as highly efficient bases for the regioselective generation of
functionalized aryl and heteroaryl magnesium compounds, Angew. Chem. Int.
Ed. 45 (18) (2006) 2958e2961, https://doi.org/10.1002/anie.200504024.
[40] T.L. Riss, A.L. Niles, R.A. Moravec, Cell Viability Assays, 2013 May 1 [Updated
2016 Jul 1], https://www.ncbi.nlm.nih.gov/books/NBK144065/. (Accessed 8
May 2019).
[42] K.A. Monk, R. Siles, M.B. Hadimani, B.E. Mugabe, J.F. Ackley, S.W. Studerus,
K. Edvardsen, M.L. Trawick, C.M. Garner, M.R. Rhodes, et al., Design, synthesis,
and biological evaluation of combretastatin nitrogen-containing derivatives as
inhibitors of tubulin assembly and vascular disrupting agents, Bioorg. Med.
Chem. 14 (9) (2006) 3231e3244, https://doi.org/10.1016/j.bmc.2005.12.033.
[43] A.R. Battersby, D.J. Le Count, S. Garratt, R.I. Thrift, Synthetic applications of 1,2-
dihydroisoquinolines: synthesis of ( )-Coreximine, Tetrahedron 14 (1) (1961)
46e53, https://doi.org/10.1016/0040-4020(61)80086-0.
[44] C.M. Lin, S.B. Singh, P.S. Chu, R.O. Dempcy, J.M. Schmidt, G.R. Pettit, E. Hamel,
Interactions of tubulin with potent natural and synthetic analogs of the
antimitotic agent combretastatin: a structure-activity study, Mol. Pharmacol.
34 (2) (1988) 200e208.
ꢀ
[45] T. Kleele, P. Marinkovic, P.R. Williams, S. Stern, E.E. Weigand, P. Engerer,
[41] M.J. Waring, Defining optimum lipophilicity and molecular weight ranges for
drug candidatesdmolecular weight dependent lower LogD limits based on
permeability, Bioorg. Med. Chem. Lett 19 (10) (2009) 2844e2851, https://
doi.org/10.1016/j.bmcl.2009.03.109.
R. Naumann, J. Hartmann, R.M. Karl, F. Bradke, et al., An assay to image
neuronal microtubule dynamics in mice, Nat. Commun.
https://doi.org/10.1038/ncomms5827.
5 (2014) 4827,
Please cite this article as: Y. Kraus et al., Isoquinoline-based biaryls as a robust scaffold for microtubule inhibitors, European Journal of Medicinal
Chemistry, https://doi.org/10.1016/j.ejmech.2019.111865