Beyond Pseudo-natural Products: Sequential Ugi/Pictet-Spengler Reactions Leading to Steroidal Pyrazinoisoquinolines That Trigger Caspase-Independent Death in HepG2 Cells

Article abstract of DOI:10.1002/cmdc.202100052

In this work, we describe how stereochemically complex polycyclic compounds can be generated by applying a synthetic sequence comprising an intramolecular Ugi reaction followed by a Pictet-Spengler cyclization on steroid-derived scaffolds. The resulting compounds, which combine a fragment derived from a natural product and a scaffold not found in nature. are both structurally distinct and globally similar to natural products at the same time, and interrogate an alternative region of the chemical space. One of the new compounds showed significant antiproliferative activity on HepG2 cells through a caspase-independent cell-death mechanism, an appealing feature when new antitumor compounds are searched.

Full text of DOI:10.1002/cmdc.202100052

Full Papers
doi.org/10.1002/cmdc.202100052
ChemMedChem
1
2
3
4
5
6
7
8
9
Beyond Pseudo-natural Products: Sequential Ugi/Pictet-
Spengler Reactions Leading to Steroidal
Pyrazinoisoquinolines That Trigger Caspase-Independent
Death in HepG2 Cells
Fernando Alonso,[a, b] Agustín Galilea,[a, b] Pau Arroyo Mañez,[a, c] Sofía L. Acebedo,[a, b]
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Gabriela M. Cabrera,[a, b] Marcelo Otero,[d] Andrea A. Barquero,*[e, f] and Javier A. Ramírez*[a, b]
In this work, we describe how stereochemically complex
polycyclic compounds can be generated by applying
natural products at the same time, and interrogate an
alternative region of the chemical space. One of the new
compounds showed significant antiproliferative activity on
HepG2 cells through a caspase-independent cell-death mecha-
nism, an appealing feature when new antitumor compounds
are searched.
a
synthetic sequence comprising an intramolecular Ugi reaction
followed by a Pictet-Spengler cyclization on steroid-derived
scaffolds. The resulting compounds, which combine a fragment
derived from a natural product and a scaffold not found in
nature. are both structurally distinct and globally similar to
Introduction
which may account for the fact that the extensive use during
the past decades of target-based approaches for finding active
compounds has decreased the pace of discovery of new
drugs.[3–5] Consequently, the phenotypic screening of new
chemotypes on cell-based or model organism-based assays, an
approach reminiscent of the pre-genomics pharmacology, has
reemerged as an alternative platform for drug discovery in the
last years.[6,7]
Nevertheless, this approach is not usually amenable to a
high-throughput screening (HTS) scheme. As a relatively lower
number of candidates can be tested in such assays, the pressure
to develop more focused, biologically relevant libraries has
risen.
In this context, natural products (NPs) have been for
centuries, and continue to be, a remarkable source of bioactive
chemical structures. About half of the small-molecule drugs
approved for clinical use in the last three decades are either
NPs or natural product-derived compounds.[8,9] This fact can be
intuitively understood considering that NPs have evolved by
interacting with multiple proteins: proteins biosynthesize them
within cells and usually target proteins when exerting their
biological functions (e.g., chemical defense and communica-
tion). But concerns about the difficulties associated with the
repeated isolation of known NPs and synthesizing them during
pharmaceutical manufacture are probably the most important
reasons for the industry‘s lack of enthusiasm for using NPs in
drug discovery.[10]
Fortunately, the new chemoinformatic tools developed for
mining the information encompassed in the rapidly growing
chemical databases have shed some light on understanding the
main structural features in NPs that seem to confer them a
higher biological relevance.[8,11–13] In this sense, several ap-
proaches have been proposed for designing new compounds
known as NP mimics, synthetic molecules that, although not
direct analogs of a particular NP, are enriched in the structural
The discovery of lead structures is central to the drug develop-
ment process. Although modern virtual screening techniques
based on known ligands or target biomolecular structures
provide a powerful tool to address this challenging task, there
are still many situations where neither a ligand nor the target
structure is known.[1] Moreover, it is increasingly recognized that
most drugs not only act on a single but on multiple targets,[2]
[a] Dr. F. Alonso, A. Galilea, Dr. P. A. Mañez, Dr. S. L. Acebedo,
Dr. G. M. Cabrera, Dr. J. A. Ramírez
Departamento de Química Orgánica
Universidad de Buenos Aires
Ciudad Universitaria, Buenos Aires, 1428 (Argentina)
E-mail: jramirez@conicet.gov.ar
[b] Dr. F. Alonso, A. Galilea, Dr. S. L. Acebedo, Dr. G. M. Cabrera, Dr. J. A. Ramírez
Unidad de Microanálisis y Métodos Físicos Aplicados a Química Orgánica
(UMYMFOR)
CONICET – Universidad de Buenos Aires
Ciudad Universitaria, Buenos Aires, 1428 (Argentina).
[c] Dr. P. A. Mañez
Departamento de Química Orgánica de la Facultad de Farmacia
Universitat de València
Burjassot, Valencia 46100 (Spain)
[d] Dr. M. Otero
Departamento de Física, Facultad de Ciencias Exactas y Naturales
CONICET – Universidad de Buenos Aires and Instituto de Física de Buenos
Aires (IFIBA)
Ciudad Universitaria, Buenos Aires, 1428 (Argentina)
[e] Dr. A. A. Barquero
Facultad de Ciencias Exactas y Naturales
Universidad de Buenos Aires. Departamento de Química Biológica
Ciudad Universitaria, Buenos Aires, 1428 (Argentina)
E-mail: alecab@qb.fcen.uba.ar
[f] Dr. A. A. Barquero
Instituto de Quimica Biológica de la Facultad de Ciencias Exactas y
Naturales (IQUIBICEN)
CONICET – Universidad de Buenos Aires
Ciudad Universitaria, Buenos Aires, 1428 (Argentina)
Supporting information for this article is available on the WWW under
https://doi.org/10.1002/cmdc.202100052
ChemMedChem 2021, 16, 1–12
1
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cmdc.202100052
ChemMedChem
features of NPs.[14–18] Some of these approaches employ a Results and Discussion
1
2
3
4
5
6
7
8
9
strategy that incorporates scaffolds already present on NPs into
the design process. The widely known biology-oriented syn-
thesis (BIOS) strategy, introduced by Waldmann is one of the
most promising.[19,20]
The BIOS design is based on the structural classification of
NPs into a limited set of scaffolds that can be used for the
design of collections of NPs mimics,[21] more suitable for
phenotypic screening campaigns than the simple synthetic
compounds available from commercial libraries usually used in
target-oriented discovery strategies such as virtual and HT
screenings.
Design and synthesis of the compounds
The steroidal tetracyclic framework is one of the more
prevalent scaffolds found in nature. As their emergence more
than 2.3 billion years ago, evolution selected steroids as a
diverse family of small molecules with many physiological
functions in all Eukaryotes. Moreover, steroids are lipophilic
molecules, a paramount property considering that, as stated
by Lipinski, many types of biological activities require perme-
ability across lipid membrane bilayers,[11] which render steroids
as suitable starting points for designing biologically relevant
molecules.
Particularly, the incorporation of different heterocyclic rings
into the steroid nucleus has been recognized as a useful
strategy to generate new steroidal derivatives for biological
screenings.[26,27] In this sense, we have previously shown that
after the oxidative cleavage of ring A of a 3-oxo-Δ4-steroid, the
resulting secosteroid can participate as a bifunctional building
block in a three-component-4-center Ugi reaction (Ugi-3CR) to
yield an azasteroid in which a nitrogen atom is incorporated
into the steroidal nucleus, with the concomitant formation of a
quaternary stereocenter.[28] Furthermore, the appropriate choice
of the additional components of the Ugi reaction would allow a
further elaboration of the condensation products into more
complex structures.
A thoroughly search literature search revealed that, accord-
ing to the work of Wang et al.,[29] the election of (2,2-dieth-
oxyethyl)amine and suitable electron-rich aromatic isocyanides
as components in the Ugi-3CR would give an intermediate that
could be submitted to a further Pictet-Spengler reaction to
generate two additional fused rings into the final compound.
Moreover, the pyrazinoisoquinoline fragments obtained
through this sequence are found in several synthetic, non-
natural molecules known to possess different bioactivities, as
required by the designing principles advanced in this work
(Scheme 1).
To test our hypothesis, we transformed cholesterol by a
standard procedure[30] into the corresponding oxoacid 1a,
which was reacted with 2,2-diethoxyethylamine and 3-(2-
isocyanoethyl)-1H-indole to yield a single primary product
(Scheme 2), which is an unexpected outcome considering that
according to our previous results two diastereoisomers could
be formed during the intramolecular reaction.[28] The formation
of only one cyclic product contrasted with the generally low
stereoselectivity found for the Ugi reaction.[31] In this case, the
high diastereoselectivity observed is probably due to the
bulkiness of the substituents added to the rigid steroidal
framework, suggesting that steric factors during the intra-
molecular cyclization might control the observed outcome.
After isolation of the main product, which was obtained in a
53% yield, its structure was confirmed to be the desired 4-
azasteroid 2a by NMR. Several nOe’s from NOESY experiments
and a relatively down field shift for H-9 (2.54 ppm) suggested
that the substituent at C-5 might have an axial position (see the
Supporting Information), which is in line with an anti-attack of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Waldmann has recently proposed that such BIOS approach
could be improved if the new compounds are designed by
combining NP fragments which are selected according to
simple guidelines: the fragments should have complementary
heteroatomic patterns, should be embedded in NPs with
diverse bioactivities originating from different biosynthetic
pathways and, ideally, should be connected using synthetic
procedures that form new stereogenic centers. The resulting
compounds termed pseudo-NPs would resemble NPs not being
accessible via known biosynthetic pathways, thus spanning
areas of the biologically relevant chemical space not covered by
nature.[22,23] This proposal has been validated by synthesizing
chromopynones, novel glucose uptake inhibitors that unite
chromane and tetrahydropyrimidinone fragments in their
structures, or indotropanes embodying fragments characteristic
of indole-containing and tropane derived NPs. Moreover, these
libraries proved to be fundamentally different from BIOS
collections.[23,24]
However, the design of new chemotypes based solely on
natural-product-derived fragments could be leaving aside the
wealth of information that has been gathered during the last
century from medicinal chemistry research.[25] A sizeable set of
unnatural compounds also shows concrete and potent bio-
logical activities, although they have interesting pharmaco-
phores not present in nature. Thus, we propose to explore a
path in a further direction by combining the best of two worlds
by designing synthetic compounds that strategically join frag-
ments derived from NPs with others that are not found in
nature but are present in bioactive compounds. As several
chemoinformatic studies show that fully synthetic drug-like
compounds and NPs differ in many structural and physicochem-
ical features, our initial hypothesis is that the resulting
compounds could span across uncharted regions of biologically
relevant chemical space beyond that populated by compound
based solely on NPs.
In this paper, we present a preliminary exploration of this
hypothesis by designing a small set of compounds that
combine a steroid-derived fragment with unnatural polycyclic
scaffolds by using a synthetic sequence that comprises an Ugi
three-component reaction followed by a Pictet-Spengler cycliza-
tion. Moreover, the complex compounds thus obtained were
analyzed based on their structural features and physicochemical
properties by using chemoinformatic tools and evaluated for
their activities in selected biological systems.
ChemMedChem 2021, 16, 1–12
www.chemmedchem.org
2
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cmdc.202100052
ChemMedChem
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Scheme 1. Connecting fragments of natural and unnatural origin by an Ugi reaction/Pictet-Spengler reaction sequence alloys the synthesis of stereochemically
rich polycyclic compounds.
Scheme 2. Synthesis of complex compounds containing eight rings through the two-step cyclization sequence. Yields from the corresponding precursors are
in parentheses. Compounds 3a and 4a were obtained in a 4:1 diastereomeric ratio).
the isocyanide to the least hindered α-face of the steroidal ring
system. The suspected 5R configuration was confirmed in a
later stage, as described below.
The most noticeable differences between their 1H NMR
spectra were the signals corresponding to the methine hydro-
gen attached to the newly formed stereocenter (C-2’), which
showed a differential coupling pattern for each isomer, and the
chemical shifts of the indolic hydrogens, which differed in
1.5 ppm between the two compounds. Due to these com-
pounds’ complex structures, their absolute stereochemistry
assignment resulted in a challenge that we decided to face by
using the same NMR-based tools that are usually employed for
determining the structure of NPs of similar complexity.[32]
The next step involved an acid-promoted Pictet-Spengler
reaction of the Ugi adduct 2a, which, according to Wang et al.,
can be better performed in acidic media.[29] Thus, upon
dissolution in formic acid, compound 2a was kept at reflux until
its total disappearance. The reaction gave two main products
3a and 4a (in a 4:1 ratio) that were efficiently isolated by
column chromatography and proved to be diastereomers after
NMR and MS analysis (Scheme 2).
Firstly, the experimental 1H and 13C NMR signals of the
diastereomers were assigned entirely. As two new stereocenters
ChemMedChem 2021, 16, 1–12
www.chemmedchem.org
3
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cmdc.202100052
ChemMedChem
were formed, the first one during the Ugi-3CR (C-5) and the
other one during the Pictet-Spengler cyclization (C-2’), models
for each of the four possible stereoisomers were generated. The
compounds’ rigid polycyclic nature simplified this task because
they are expected to have few conformations in solution (see
the Supporting Information).
1
2
3
4
5
6
Subsequently, NMR shifts were predicted for every nucleus
in each structure using single point ab initio calculations. With
this information in hand, the stereochemistry was determined
using the DP4+ procedure developed by Grimblat et al., an
improved probabilistic approach that compares the predicted
and experimental chemical shifts.[33] Thus, DP4+ probabilities
were calculated for each diastereomer to assign the correct
configurations of the new stereocenters as (5R,2’S) and (5R,2’R)
for compounds 3a and 4a, respectively (further details in the
Supporting Information).
Once proved the synthetic approach‘s feasibility, the same
sequence was repeated but using in the first step alternatives
electron-rich aromatic isocyanides, such as 3,4-meth-
oxyphenylethyl and 4-methoxyphenylethyl isocyanide. The Ugi-
3CR proceeded efficiently in both cases, yielding the desired
products (2b and 2c shown in Scheme 3, respectively). But
when these intermediates were subjected to the same acidic
conditions, different outcomes were observed: although com-
pound 2b gave the expected mixture of diastereomers 3b and
4b, intermediate 2c generated only one product (3c) that by
comparison of its NMR coupling patterns and nOe’s with those
of the compounds mentioned above, proved to have the same
absolute configuration than 3a and 3b, that is 5R,2’R.
The absence of the corresponding diastereoisomer, which
reveals an almost complete stereoselectivity for the Pictet-
Spengler reaction, suggests kinetic control during the cycliza-
tion, probably due to the higher nucleophilicity of the 3,4-
dimethoxyphenyl ring compared to that of indole. Thus, some
preliminary calculations were performed to shed light on the
energetics of the cyclization process. As the final products’
stereochemistry is determined during the Pictet-Spengler
cyclization, only this step was studied. Starting from the most
stable conformation of the intermediate 2a, the reaction to
give the 5R,2’S product (3a) has an activation energy (ΔG�) of
10.42 kJ/mol, while the barrier to give the 5R,2’R product (4a)
was 33.18 kJ/mol. In both cases, the free energy of reaction
(ΔGr) of the processes were exothermic by 27.07 and 15.14 kJ/
mol, respectively, which is in line with its preferential formation
during the cyclization reaction. According to our study, the
steric interaction between the H-atom connected to the
aromatic ring involved in the cyclization and the hydrogen
atom at C-6 of the steroid might be responsible for the
observed product ratio.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Scheme 3. Additional polycyclic compounds obtained by the Ugi/Pictet-
Spengler sequence strategy. Yields from the corresponding precursors are in
parentheses. Compounds 3b–4b, 3e–4e, and 3f–4f were obtained in 4:1
diastereomeric ratios).
er, both results could explain why 3c is the only product
generated during the reaction.
Finally, and to further explore the generality of the synthetic
strategy, we expanded the same sequence by starting from
other steroids such as testosterone and stigmasterol, with
similar overall results. The resulting compounds, whose syn-
thesis and characterization are described in the Supporting
Information, and depicted in Scheme 3.
Geometry details can be found in the Supporting Informa-
tion. On the other hand, when the same calculations were
performed for the intermediate 2c, the barrier to give the
observed 5R,2’S product (3c) has an ΔG� of 19.62 kJ/mol, while
the barrier to give the hypothetic (5R,2’R) product was 39.62 kJ/
mol. Moreover, the ΔGr leading to 3c was exothermic by
0.56 kJ/mol when compared to that giving the diastereomeric
product, which was endothermic by 6.40 kJ/mol. Taken togeth-
Cheminformatic analysis
Once the desired compounds were obtained, some questions
arise: how do they compare to NPs or molecules from synthetic
origins? Do their structures encode the prevailing structural
ChemMedChem 2021, 16, 1–12
www.chemmedchem.org
4
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cmdc.202100052
ChemMedChem
features and physicochemical properties usually found in
bioactive molecules? We compared statistically the structural
and physicochemical properties found in each kind of com-
pound to gain further insight. We constructed a database of
687620 compounds: 228846 NPs included in the Universal
Natural Products Database (UNPD)[34] and 229989 synthetic
products picked up by simple random sampling from the All-
Clean ZINC database.[35] Likewise, a set of 228800 bioactive
compounds were also randomly selected from the ChEMBL
database.[36] Nineteen molecular descriptors known to describe
the topological, structural, and physicochemical properties were
calculated for every compound in the set, including the
synthetic molecules reported in this work. These descriptors,
which were previously validated by several researchers,[8,37] were
taken as a base for a principal component analysis (PCA). The
resulting first three principal components (PC1, PC2, and PC3)
retains 38.7, 18.5, and 13.8% of the total variance, thus
explaining 71% of the original dataset variance. (For further
information on the selected molecular descriptors and the PCA
results, please see the Supporting Information.)
As shown in Figure 1, the compounds described in this
work lie close to the region‘s border spanned by the
compounds from ZINC but are well inside the chemical space
populated by NPs. This analysis suggests that the new
compounds could be considered NPs mimics rather than fully
synthetic molecules, which, as discussed before, might be a
desirable feature when designing new chemical entities for
phenotypic assays. Interestingly, these compounds have a
Tanimoto similarity lower than 0.75 to the 429000 NPs collected
in the COCONUT database,[38] which suggests that, although
they globally resemble NPs, they are essentially different from
any known natural compound.
as percentages for the four cell lines. The most noticeable result
shown in Figure 2A is that compound 3c has a significant
antiproliferative action, which is more notorious in HepG2 cells.
The IC50 in this cell line (21 μM) is significantly lower than that
of 5-fluorouracil (302 μM) and comparable to cisplatin (19 μM),
two drugs commonly used in the clinical treatment of liver
cancer (Figure S4 in the Supporting Information). From this
preliminary screening, and although we have a small set of
compounds in hand, some interesting observations on the
structure-activity relationships could be advanced because
compound 3b, which has only one methoxy substituent at the
aromatic ring, lacks cytotoxicity in all tested cell lines at this
concentration. On the other hand, compound 3d, having the
same ring system as 3c but lacking the hydrocarbon side chain,
is also devoid of activity.
In contrast, the stigmasterol-derived compound 3g, which
has only minor structural differences at the side chain when
compared to 3c, also shows a selective, albeit moderate, activity
on HepG2 cells, suggesting that the presence of this aliphatic
chain could be a relevant structural feature for the observed
antiproliferative action.
Noteworthily, compound 2c, the acyclic precursor of 3c,
does not elicit any antiproliferative effect either, which hints
that the new bioactivity arises from the fusion of the steroidal
and pyrazinoisoquinoline scaffolds.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
We selected compound 3c to study further the molecular
mechanisms underlying its antiproliferative activity on HepG2
cells in the sight of these results. As shown in Figure 2B, the
compound 3c proved to decrease the viability in a dose-
dependent manner and was also able to reduce the number of
HepG2 colonies in a long-term in vitro colony formation assay
which might be regarded as an indirect estimation of the
antineoplastic effect of a compound[41] (Figure S5).
On the other hand, the Figure 1 shows that these new
compounds share structural and physicochemical properties
with the bioactive compounds picked from the ChEMBL data-
base.
Moreover, when observing the morphological features of
cell death by microscopy after 4’,6-diamidino-2-phenylindole
(DAPI) staining, we detected an increased number of nuclei
having chromatin condensation and fragmentation in cells
treated with 3c compared to untreated ones (Figure 2C),
suggesting that 3c elicits its antiproliferative action against
hepatocarcinoma cells through an apoptotic-like cell death
mechanism. It is a known feature of cancer cells to exhibit high
rates of lipid uptake and de novo lipid synthesis to support
rapid cell proliferation and tumor growth, and that these
processes are particularly crucial in liver cancer development
and progression.[42,43] Therefore, we speculated that compounds
3c might affect lipid homeostasis as part of its selective
cytotoxicity in HepG2 cells. Thus, the lipid content was
measured by Oil Red O (ORO) staining after treating HepG2 cells
with different concentrations of 3c in the presence or absence
of oleic acid (OA), a well-known enhancer of lipid accumulation.
As expected, the intracellular lipid of HepG2 content was
significantly increased in cells treated with OA in a dose-
response manner compared to untreated cells (Figure 3A).
However, when observing DAPI stained cells, we were
surprised to find that the number of nuclei exhibiting apoptotic
features decreased in cells treated with a combination of 3c
and OA, in comparison to those observed in cultures treated
Evaluation of the biological activity of the new compounds
The remaining question is whether the novel chemotypes’
possess any notable biological activity. As they were not
designed to satisfy a given bioactivity, they were screened,
along with their synthetic precursors, in an array of biological
assays, in a similar way that newly discovered NPs are
evaluated.[39] In particular, we examined their antiviral and
antiproliferative potency on different cell-based phenotypic
assays, as that this approach does not require any a priori
knowledge of a specific molecular mechanism of action.[40] The
antiviral activity against Herpes simplex 1 was assessed by the
cytopathogenic effect (CPE) reduction assay. Unfortunately,
none of the compounds showed significant activity.
The antiproliferative activity of the set of compounds was
evaluated in human colorectal (HT29 and Caco-2), hepatic
(HepG2), and prostatic (PC3) carcinoma cell lines by the well-
established MTT assay. Initially, all compounds were screened at
a single dose (50 μM), and the growth inhibition was quantified
ChemMedChem 2021, 16, 1–12
www.chemmedchem.org
5
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cmdc.202100052
ChemMedChem
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Figure 1. a)–c) PCA 2D scatter plots of the three principal components, ordered from highest to lowest percentage of explained variances: a) PC2 vs. PC1
(57.2%), b) PC3 vs. PC1 (52.5%), and c) PC3 vs. PC2 (32.3%). In (a)–(c), the dots represent the position of the compounds in the chemical space defined by the
PC and are colored according to compound families: green (NPs), blue (synthetic products), yellow (bioactive products), and red (synthesized compounds
described in this work). d), g), and j) 2D histograms of PC2 vs. PC1, e), h), and k) 2D histograms of PC3 vs. PC1, and f), i), and l) 2D histograms of PC3 vs. PC2,
for synthetic, natural and bioactive products, respectively.
ChemMedChem 2021, 16, 1–12
www.chemmedchem.org
6
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cmdc.202100052
ChemMedChem
shown that OA robustly alleviates additional cellular damages
1
2
3
4
5
6
7
8
9
caused by palmitic acid, such as endoplasmic reticulum (ER)
stress and mitochondrial dysfunction, both in HepG2 cells and
rat primary hepatocytes.[45] Hence, these earlier reports open
the possibility that cell death induced by 3c might occur due
either to ER stress or altered mitochondrial function. The first
mechanism was excluded because the antiproliferative activity
of 3c could not be abrogated by simultaneous treatment of the
cells with 4-phenylbutyric acid (4PBA), a chemical chaperone
known to alleviate ER stress (Figure 4A).[46]
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Then, and to characterize a possible mitochondria-mediated
cell death, we assessed cytochrome c (Cyt c) release from the
mitochondrial inter-membrane space to the cytosol by immu-
nofluorescence staining. This event can be considered a marker
of mitochondrial membrane permeabilization, independently of
cell death types.[47] As shown in Figure 4B, Cyt c-related
fluorescence became diffuse in the cytoplasm of HepG2 cells
treated with compound 3c for 24 h, suggesting the release of
Cyt c, and therefore the disruption of the mitochondrial
membrane potential. Moreover, a significant increment of ROS
production was observed in cells treated with 3c compared to
Figure 2. A) Heat map for the growth inhibition percentage (GIP) of
compounds 2a–4f against HT29, Caco-2, HepG2, and PC3 cell lines (72 h at
50 μM of the compounds). B) Dose-dependent effect of 3c on HepG2 cell
viability. The IC50 value was obtained as described in the Experimental
Section. C) Representative fluorescent micrographs of HepG2 cells treated or
not with 50 μM compound 3c for 24 h after DAPI staining (magnification
400x).
Figure 3. A) Quantification of ORO content in OA-induced HepG2 cells
treated or not with 25 μM of 3c. OD: optical density. B) Double ORO/DAPI
staining was performed to visualize the increase in intracellular lipid
droplets. HepG2 cell cultures were co-treated with 0.2 mM OA and 25 μM of
3c. C) Quantitative analysis of apoptotic cells. Results are presented as
mean�SD of three independent experiments (**p<0.01).
Figure 4. A) Cell death produced by 3c in the presence of 4PBA (1 mM), CsA
(10 μM), or Z-VAD-FMK (10 μM). Inhibitors were added 2 h before treatment
with 25 μM of 3c. The activity on HepG2 cell viability was determined by
MTT assay (three independent experiments, ***p<0.001). B) Immunostaining
with an anti-Cyt c antibody was performed to visualize Cyt c release to the
cytosol following treatment with 25 μM of 3c. Nuclei were labeled with
DAPI. C) ROS production displayed by microscopy with DCFDA (2’,7’-
dichlorofluorescein diacetate) as fluorescence probe after treatment with
25 μM compound 3c for 24 h. D) Identification of AIF nuclear translocation
by immunofluorescence assay. Co-staining with DAPI was used to detect the
nuclei, and the merged images clearly show the nuclear localization of AIF
induced by treatment with 25 μM of 3c for 24 h.
with 3c alone (Figure 3B). Thus, these results suggest not only
that lipid homeostasis does not seem to be involved in the
mechanism of cell death induced by 3c, but that oleic acid
alleviates the cytotoxic effects caused by this compound in
HepG2 cells (Figure 3C). The cytoprotective effect of oleic acid
has been previously described in the context of the apoptosis
triggered by palmitic acid in HepG2 cells.[44] Furthermore, it was
ChemMedChem 2021, 16, 1–12
www.chemmedchem.org
7
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cmdc.202100052
ChemMedChem
untreated HepG2 cells (Figure 4C), which is probably linked to Conclusion
1
2
3
4
5
6
7
8
9
mitochondrial dysfunction.[48,49]
Interestingly, the pharmacological blockade of the mito-
chondrial permeability transition pore (mPTPM) with cyclo-
sporine A (CsA), a well-known inhibitor that prevents pore
opening, did not reduce the antiproliferative effect of 3c in
HepG2 cells (Figure 4A), suggesting that the mPTPM would not
be participating in the mitochondrial failure induced by 3c.
Furthermore, since the release of Cyt c does not always lead to
caspase activation, we evaluated the effect of the highly
effective caspase inhibitor Z-VAD-FMK on 3c induced cytotox-
icity in HepG2 cells. As shown in Figure 4A, the cell death
promoted by 3c was not affected by this pan-caspase inhibitor,
hinting that 3c triggered a cellular death process in a caspase-
independent manner.
It is well known that mitochondria also contain several
caspase-independent death effectors, including apoptosis-in-
ducing factor (AIF), endonuclease G, and others.[50] Therefore, to
better characterize the mode of cell death elicited by 3c, the
mitochondrial release of AIF was investigated by immunofluor-
escence. As shown in Figure 4D, AIF presents a granular pattern
in the mitochondria of control cells, while after treatment with
3c, it translocates from mitochondria to the cytosol and the
nucleus. This result suggests that AIF release is involved in the
caspase-independent death pathway induced by 3c.
For 20 years now, it has been clear that to find new small
molecules endowed with unexpected biological properties,
thoughtful exploration of the chemical space is needed, which
is a unpromising task to undertake without some guiding
principles. As several strategies have been proposed, those
seeking to mimic the structures of NPs are supported by
theoretical considerations and validated by numerous exper-
imental results. In this sense, the recent concept of pseudo-NPs
introduced by Waldmann seems to be a very appealing
approach: they are designed by joining NP fragments with
diverse biosynthetic origins into new compounds that resemble
NPs but are not accessible through known biosynthetic path-
ways and proved to have novel biological properties. In this
work, we have preliminarily explored an extension of this
design criteria by fusing the polycyclic framework of steroids
-highly relevant from the biological point of view- with a
scaffold only found in unnatural, fully synthetic bioactive
compounds. The resulting compounds have complex hepta- or
octacyclic, stereochemically rich scaffolds that cannot be found
in nature because no known biosynthetic pathway can produce
them. Nevertheless, the resulting compounds share not only
global physicochemical properties with NPs but also interrogate
alternative regions of the chemical space.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Lastly, taking into account the evidence above that the
antiproliferative activity of 3c on HepG2 cells seems to be
related to the damage exerted on mitochondria, a possible
mechanism for the cytoprotective effect of OA arises, as it is
known that OA can protect hepatocytes against lipotoxicity by
inducing autophagy.[51] In this sense, we found that the
pharmacological inhibition of autophagy with concanamycin A
(ConA) enhanced the cytotoxic effect of 3c, as evidenced by an
increased number of cells having apoptotic-like nuclei (Fig-
ure S6). This finding suggests that OA-triggered autophagy may
account for the abrogation of 3c-induced cell death by
remotion of damaged mitochondria.
In summary, our results show that the new bioactive
chemotype induces HepG2 cell death in a programmed way in
which mitochondria play a crucial role. The cell death mecha-
nism presumably involves mitochondrial membrane permeabili-
zation, independently of mPTPM. In this sense, a recent chemo-
informatic study shows that molecules causing mitochondrial
toxicity can be characterized by having relatively restricted
values of physicochemical properties such as logP, molecular
weight, and surface area.[52]
Moreover, among the small set of new molecules, a
cytotoxic compound with antiproliferative activity on hepatic
HepG2 cells and a comparable in vitro potency to that of
antitumor drugs clinically used to treat hepatocarcinoma was
found, and with a mechanism of action that seems not to
involve caspase-triggered cell death, an exciting finding
because most of the currently used anticancer drugs, independ-
ent of their targets and mechanisms of action, elicit cell death
via caspase-dependent apoptosis.[53] However, cancer cells
usually became resistant to drugs during chemotherapy due to
apoptotic machinery‘s dysregulation. Thus, discovering new
drugs with alternative death mechanisms is a driving force to
start drug discovery programs aimed at the modulation of
tumor growth.
As a whole, this paper might be considered as a proof-of-
concept that paves the way to the idea that the connection of
pharmacophores not found in nature with scaffolds from NPs
using complexity-generating reactions could lead, efficiently, to
new compounds which may be endowed with unique bio-
logical activities.
Interestingly, the values for these properties for compound
3c are similar to that of other mitochondrial-disrupting Experimental Section
compounds, which reinforces our hypothesis on the underlying
mechanism of action. Thus, upon mitochondrial membrane
disruption, the consequent release of Cyt c and AIF to the
cytoplasm, as well as the accumulation of ROS, would lead to
cell death without caspase activation. Note that the antiprolifer-
ative effect may be modulated by combination with autophagy
inhibitors.
Synthesis of the compounds
General. For a full experimental description of the synthesis and
characterization of the new compounds, please see the Supporting
Information. The structures were confirmed by 1D and 2D NMR
spectroscopy from spectra recorded on a Bruker AM-500 (500 MHz
1
for H and 125 MHz for 13C). Chemical shifts (δ) are given in ppm
(downfield from TMS as the internal standard). Coupling constant
(J) values are in Hz. HRMS (ESI) analyses were performed on a
ChemMedChem 2021, 16, 1–12
www.chemmedchem.org
8
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cmdc.202100052
ChemMedChem
Bruker micrOTOF-Q II. Combustion analyses for the new com-
pounds were performed on an Exeter CE 440 Elemental Analyzer
and were within �0.4% of the theoretical values. Melting points
were determined on a Fisher Johns apparatus and are uncorrected.
As a representative example, the synthesis of compound 3c is
described below.
Studies on the stereoselectivity of the Pictet-Spengler cyclization. All
geometries optimizations were carried out at the M06-2X/6-31+
G(d,p) level of theory[55] using the “UltraFine” pruned integration
grid by Gaussian 09. The election of the DFT method was based on
similar examples found in the bibliography.[56] Vibrational frequen-
cies were performed to ensure that the geometries obtained
correspond to local minima (no imaginary vibration mode) or
transition structures (one imaginary vibration mode) in the
potential energy surfaces for the Pictet-Spengler cyclization reac-
tion. Moreover, these calculations provide the zero-point energy
(ZPE) corrections at the experimental temperature, 373 K. The effect
of the solvent, formic acid, was evaluated using the integral
equation formalism variant (IEFPCM) of the polarizable continuum
model (PCM) implemented in Gaussian09 rev. D.01 program
1
2
3
4
5
6
7
8
9
Synthesis of compound 3c. 500 mg (1.3 mmol) of the oxoacid 1a,
which was obtained from cholesterol following
a reported
procedure,[30] was dissolved in methanol and 1 equiv. of 2,2-
diethoxyethylamine was added. The mixture was stirred for 15 min
at room temperature after which
1 equivalent of 3,4-meth-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
oxyphenylethyl isocyanide was added. The reaction was kept at
room temperature until disappearance of the starting material
(72 h). The solvent was evaporated under reduced pressure and the
residue was taken in EtOAc and washed with NaOH (5% aq). The
crude product was purified by silica gel column chromatography
package.[57] GaussView
5 was employed to visualize all the
structures and atomic motion in the transition structures.
(hexane/EtOAc gradient) to give compound 2c with 57% yield.
Chemoinformatic analysis. Calculation of the physicochemical
descriptors for the new compounds and those contained in the
databases was performed with ChemAxon’s JChem for Excel
(release 20.11.0.644, 2020, ChemAxon; http://www.chemaxon.com).
Principal Component Analysis was carried out with R (version 3.6.1)
using factoextra, factoMineR, dendextend, and ggplot2 packages.
1
°
Melting point: 125 C. H NMR: 0.63 (3H, s), 0.72 (3H, s), 0.83 (3H, d,
J=2.3), 0.84 (3H, d, J=2.3), 0.87 (3H, d, J=6.5), 1.12 and 1.21 (6H, t,
J=7.0), 1.93 (1H, dt, J=12.5, 3.3), 2.07 (1H, t, J=12.1), 2.22 (1H,
ddd, J=16.4, 11.7, 7.2), 2.34 (1H, ddd, J=16.4, 9.6, 1.7), 2.52 (1H,
ddd, J=12.1, 10.3, 3.9), 2.64 and 2.70 (2H, m), 3.13 and 3.66 (2H, m),
3.20, 3.65, 3.66 and 3.71 (4H, m)„ 3.28 and 3.49 (2H, ddt, J=13.1,
13.0, 7.3), 3.82 (3H, s), 3.85 (3H, s), 4.82 (1H, dd, J=7.2, 3.3), 6.64
(1H, dd, J=8.1, 2.0), 6.67 (1H, d, J=1.9), 6.77 (1H, d, J=8.1), 7.64
(1H, s). 13C NMR: 12.2, 15.2, 15.6, 17.5, 18.7, 22.7, 22.8, 22.9, 23.9,
24.1, 27.1, 28.1, 28.4, 30.0, 30.9, 30.9, 35.1, 35.8, 35.9, 36.2, 39.6, 40.0,
41.1, 42.8, 43.2, 46.4, 47.7, 55.5, 55.9, 56.0, 56.2, 62.5, 64.1, 69.8,
100.4, 111.4, 111.9, 120.9, 131.7, 147.7, 149.0, 174.4, 174.8. HRMS
(ESI+): calcd for C43H71N2O6: 711.5307 [M+H+]; found: 711.5270.
Biological assays
Cell culture and reagents. HT29 cells, Caco-3 cells, PC3 cells, and
HepG2 cells were all obtained from ATCC (Rockville, USA) and
seeded in DMEM medium, whereas PC3 cells were grown in RPMI
1640 medium (Gibco, Grand Island, NY); both media were
supplemented with 10% fetal bovine serum (Gibco) and 50 μg/ml
gentamicin. Cultured cells were kept in a 5% CO2 atmosphere at
Compound 2c (50 mg, 0.07 mmol) was dissolved in 1 mL of formic
acid (97% aq.) and the resulting solution was refluxed. The reaction
was monitored by TLC, and after completion, the formic acid was
evaporated at reduced pressure. The residue was dissolved in
10 mL of EtOAc and washed with 20 mL of K2CO3 (s.s.) The
combined organic layers were dried over anhydrous sodium sulfate
and evaporated to give the crude product, which was purified by
°
37 C. The broad-spectrum caspase inhibitor z-VAD-mk was pur-
chased from Bio-Rad. 4-PBA, cyclosporine A, and concanamycin A
were purchased from Sigma-Aldrich.
Cell proliferation assay. The cell viability was determined by
colorimetric MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazo-
lium bromide) (Sigma-Aldrich) assay.[58] Cells were seeded in 96-well
plates at a density of 1x104 cells/well, and solutions containing
different concentrations of the tested compound were added. Each
determination was done in triplicate. After 72 h of incubation, MTT
solution (5 mg/mL in distilled water) was added to cells in the
silica gel column chromatography (hexane/EtOAc gradient) to give
1
°
compound 3c with 69% yield. Melting point: 167 C. H NMR: 0.69
(3H, s), 0.86 (3H, d, J=2.2), 0.87 (3H, d, J=2.3), 0.90 (3H, d, J=6.5),
1.07 (3H, s), 1.67 (1H, m), 1.70 (1H, m), 1.83 (1H, m), 1.86 (1H, m),
2.00 (1H, m), 2.05 (1H, td, J=12.6, 7.6), 2.32 (1H, m), 2.33 (1H, m),
2.46 (1H, m), 2.56 and 2.78 (2H, m), 2.70 (1H, m), 2.78 and 4.74 (2H,
m), 3.74 (1H, dd, J=13.3, 9.0), 3.84 (3H, s), 3.87 (3H, s), 4.73 (1H, m),
5.12 (1H, dd, J=13.5, 4.6), 6.52 (1H, s), 6.61 (1H, s). 13C NMR: 12.4,
15.6, 18.8, 21.6, 22.7, 23.0, 24.0, 24.1, 27.0, 27.1, 27.5, 28.1, 28.4, 28.5,
29.2, 34.6, 36.0, 36.3, 39.7, 39.9, 40.0, 40.1, 42.8, 43.1, 45.4, 53.6, 55.7,
56.0, 56.2, 56.3, 65.2, 108.4, 111.5, 126.6, 128.9, 148.1, 148.3, 171.6,
172.2. HRMS (ESI+): calcd for C39H59N2O4: 619.4469 [M+H+]; found
619.4442.
°
culture medium, and plates were incubated for 2 h at 37 C. Then,
the produced formazan was solubilized by the addition of 0.2 mL of
ethanol, and the absorbance values were immediately measured
using an ELISA plate reader (Eurogenetics MPR-A 4i) at a test
wavelength of 570 nm and a reference wavelength of 630 nm.
Results for each treated cell culture were normalized as
a
percentage of the absorbance to untreated controls. The drug
concentration needed to inhibit cell growth by 50% (IC50) was
found by fitting the concentration-response curves. In experiments
where two different treatments must be compared, the Mann-
Whitney test was performed to assess the confidence level of the
conclusions reported. The morphological changes of cells treated
with different amounts of compound 3c were seen using an
inverted microscope, and the images were recorded with a digital
camera.
Computational methods
Assessment of the absolute configuration of the new compounds.
Models for each of the four possible stereoisomers were generated
(without the side chain, for simplicity), and a conformational search
was performed. The corresponding structures having the lowest
energy were optimized to a B3LYP/6-31G(d) level. Chemical shift
tensors were calculated at the mPW1PW91/6-31+G(d,p) level with
a Polarizable Continuum Model (PCM) using chloroform as solvent.
Calculations were performed using Gaussian09 rev. D.01.[54] The
DP4+ probabilities were calculated by comparing the predicted
and observed chemical shifts (for 1H and 13C) according to the
protocol proposed by Grimblat et al.[33] Table S1 in the Supporting
Information contains the relevant information.
Analysis of nuclear morphology. Cell nuclear apoptotic morphology
was seen by staining with 4’,6-diamidino phenylindole (DAPI;
Sigma-Aldrich). HT29 cells grown on glass coverslips in 24-well
plates were incubated with compound 3c at different concen-
trations for 24 h. An additional sample was treated with 5-
fluorouracil 50 μM as a positive control. After washes with PBS, cells
°
were fixed with methanol for 10 min at À 20 C and stained with
DAPI (4 μg/mL in PBS) for 5 min at room temperature protected
ChemMedChem 2021, 16, 1–12
www.chemmedchem.org
9
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cmdc.202100052
ChemMedChem
from the light. Finally, coverslips were rinsed, mounted, and imaged
on a fluorescence microscope (Olympus BX51) with a DAPI filter.
Apoptosis was determined according to characteristic nuclear
morphology: the presence of apoptotic bodies, nuclear condensa-
tion, and fragmentation.[59]
Conflict of Interest
1
2
3
The authors declare no conflict of interest.
4
Keywords: antiproliferative · caspase-independent cell death ·
natural product mimics · Pictet-Spengler reaction · Ugi reaction
5
6
7
8
Clonogenic assay. The clonogenic assay stands for an indirect
estimation of neoplastic transformation. Briefly, HT29 cells were
seeded in 6-well plates at a density of 1x103 cells/well and allowed
to adhere for 12 h. Then, cells were treated with compound U2 at
different concentrations or 5-FU (XXμM) in complete medium. At
24 h, the culture medium was replaced with drug-free medium, and
cells were grown for an additional 11 days until visible colonies
appeared. Finally, cells were rinsed with PBS, fixed with 10%
formaldehyde for 10 min, and stained with 1% crystal violet (20%
ethanol in distilled water). Colonies showing at least 50 cells were
manually scored, and their images were recorded with a digital
camera.
9
[1] H. Chen, O. Engkvist, N. Blomberg, J. Li, MedChemComm 2012, 3, 312.
[2] R. R. Ramsay, M. R. Popovic-Nikolic, K. Nikolic, E. Uliassi, M. L. Bolognesi,
Clin. Transl. Med. 2018, 7, 3.
[3] F. Sams-Dodd, Drug Discovery Today 2013, 18, 211–217.
[4] F. Sams-Dodd, Drug Discovery Today 2005, 10, 139–147.
[5] M. K. Hellerstein, Metab. Eng. 2008, 10, 1–9.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
[6] J. A. Lee, E. L. Berg, J. Biomol. Screening 2013, 18, 1143–1155.
[7] M. Szabo, S. Svensson Akusjärvi, A. Saxena, J. Liu, G. Chandrasekar Ja-
nebjer, S. S. Kitambi, Drug Des. Dev. Ther. 2017, 11, 1957–1967.
[8] C. F. Stratton, D. J. Newman, D. S. Tan, Bioorg. Med. Chem. Lett. 2015, 25,
4802–4807.
[9] D. J. Newman, G. M. Cragg, J. Nat. Prod. 2016, 79, 629–661.
[10] B. David, J. L. Wolfender, D. A. Dias, Phytochem. Rev. 2015, 14, 299–315.
[11] C. A. Lipinski, Adv. Drug Delivery Rev. 2016, 101, 34–41.
[12] M. Feher, J. M. Schmidt, J. Chem. Inf. Comput. Sci. 2003, 43, 218–227.
[13] R. Bade, H. Chan, J. Reynisson, Eur. J. Med. Chem. 2010, 45, 5646–5652.
[14] M. Pascolutti, R. J. Quinn, Drug Discovery Today 2014, 19, 215–221.
[15] E. C. Barnes, R. Kumar, R. A. Davis, Nat. Prod. Rep. 2016, 33, 372–381.
[16] K. C. Morrison, P. J. Hergenrother, Nat. Prod. Rep. 2014, 31, 6–14.
[17] S. F. Martin, J. Org. Chem. 2017, 82, 10757–10794.
[18] D. Genis, M. Kirpichenok, R. Kombarov, Drug Discovery Today 2012, 17,
1170–1174.
Oil Red O Staining. HepG2 cells cultured in 96-well microplate or on
glass coverslips in 24-well plates were treated or not with OA at
different concentrations for 24 h, then undergoing ORO staining. In
brief, cells were fixed in 4% paraformaldehyde, and after washes
with PBS, the ORO working solution was added for 15 minutes at
room temperature. Then, ORO solution was discarded, and cells
were rinsed with distilled water until the solution became clear.
After drying, isopropanol (100%) was added to the microplate used
for ORO quantification, and absorbance was measured at 490 nm
by using an ELx808 microplate reader (Bio-Tek Instruments, Inc,
Winooski, VT). Dried coverslips were stained with DAPI and
mounted to visualize nuclear cell morphology and lipid droplets
distribution using an Olympus BX51 microscope. Magnification:
400x.
[19] H. van Hattum, H. Waldmann, J. Am. Chem. Soc. 2014, 136, 11853–
11859.
[20] R. S. Bon, H. Waldmann, Acc. Chem. Res. 2010, 43, 1103–1114.
[21] M. A. Koch, A. Schuffenhauer, M. Scheck, S. Wetzel, M. Casaulta, A.
Odermatt, P. Ertl, H. Waldmann, Proc. Natl. Acad. Sci. USA 2005, 102,
17272–17277.
[22] M. Grigalunas, A. Burhop, A. Christoforow, H. Waldmann, Curr. Opin.
Chem. Biol. 2020, 56, 111–118.
[23] G. Karageorgis, E. S. Reckzeh, J. Ceballos, M. Schwalfenberg, S. Sievers,
C. Ostermann, A. Pahl, S. Ziegler, H. Waldmann, Nat. Chem. 2018, 10,
1103–1111.
[24] T. Schneidewind, S. Kapoor, G. Garivet, G. Karageorgis, R. Narayan, G.
Vendrell-Navarro, A. P. Antonchick, S. Ziegler, H. Waldmann, Cell Chem.
Biol. 2019, 26, 512–523.e5.
[25] R. W. Brown, Future Med. Chem. 2021, 13, 211–224.
[26] Y. T. Yang, S. Du, S. Wang, X. Jia, X. Wang, X. Zhang, Steroids 2019, 151,
108465.
[27] P. Saikia, S. Gogoi, S. Gogoi, R. C. Boruah, Steroids 2014, 88, 1–6.
[28] F. Alonso, S. L. S. L. Acebedo, A. C. A. C. Bruttomesso, J. A. J. A. Ramírez,
Steroids 2008, 73, 1270–1276.
[29] W. Wang, S. Ollio, E. Herdtweck, A. Dömling, J. Org. Chem. 2011, 76,
637–644.
[30] J. T. Edward, D. Holder, W. H. Lunn, I. Puskas, Can. J. Chem. 1961, 39,
ROS detection. The cellular reactive oxygen species (ROS) generation
was detected as they convert the non-fluorescent reagent 2’,7’-
dichlorofluorescein diacetate (DCFDA) into DCF, a highly fluores-
cent compound. HepG2 cells grown on glass coverslips in 24-well
°
plates were stained with 25 μM of DCFDA for 30 min at 37 C
covered with aluminum foil. After washes with PBS, the appearance
of green fluorescence was examined under an Olympus BX51
microscope. Magnification: 400x.
Indirect immunofluorescence assay. HepG2 cells grown on glass
coverslips in 24-well plates were fixed with methanol and washed
three times with PBS. Then, cells were immunolabeled with anti-
cytochrome c or anti-AIF primaries antibodies (Alexis Biochemicals
Corporation). Next, cells were incubated with a FITC-conjugated
goat anti-mouse secondary antibody (Sigma-Aldrich). After staining
cells with DAPI to label the nucleus, coverslips were rinsed,
mounted, and examined under a fluorescence microscope (Olym-
pus BX51). Magnification: 400x.
599–600.
[31] L. Banfi, A. Basso, G. Guanti, R. Riva in Multicomponent Reactions (Eds.: J.
Zhu, H. Bienaymé), Wiley-VCH, Weinheim, 2005, pp. 1–32.
[32] M. W. Lodewyk, M. R. Siebert, D. J. Tantillo, Chem. Rev. 2012, 112, 1839–
1862.
Statistical analysis of the experiments was performed using Graph-
Pad Prism version 6.01 for Windows (GraphPad Software, La Jolla,
CA, USA; www.graphpad.com).
[33] N. Grimblat, M. M. Zanardi, A. M. Sarotti, J. Org. Chem. 2015, 80, 12526–
12534.
[34] J. Gu, Y. Gui, L. Chen, G. Yuan, H. Z. Lu, X. Xu, PLoS One 2013, 8, 1–10.
[35] T. Sterling, J. J. Irwin, J. Chem. Inf. Model. 2015, 55, 2324–2337.
[36] A. Gaulton, L. J. Bellis, A. P. Bento, J. Chambers, M. Davies, A. Hersey, Y.
Light, S. McGlinchey, D. Michalovich, B. Al-Lazikani, J. P. Overington,
Nucleic Acids Res. 2012, 40, D1100–D1107.
[37] T. A. Wenderski, C. F. Stratton, R. A. Bauer, F. Kopp, D. S. Tan, Methods
Mol. Biol. 2015, 1263, 225–242.
[38] M. Sorokina, C. Steinbeck, “COCONUT: the COlleCtion of Open NatUral
producTs. (Version 4) (Data set). Zenodo.,” can be found under http://
doi.org/10.5281/zenodo.3778405, 2020.
[39] Y. Futamura, K. Yamamoto, H. Osada, Biosci. Biotechnol. Biochem. 2017,
81, 28–31.
Acknowledgements
This work was supported by grants from Universidad de Buenos
Aires (UBACyT 20020170100679BA). We are also grateful to María
Marta Rancez for her assistance with English language editing.
ChemMedChem 2021, 16, 1–12
www.chemmedchem.org
10
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cmdc.202100052
ChemMedChem
[40] D. C. Swinney, J. Anthony, Nat. Rev. Drug Discovery 2011, 10, 507–519.
[41] N. A. P. Franken, H. M. Rodermond, J. Stap, J. Haveman, C. van Bree, Nat.
Protoc. 2006, 1, 2315–2319.
[42] D. F. Calvisi, C. Wang, C. Ho, S. Ladu, S. A. Lee, S. Mattu, G. Destefanis, S.
Delogu, A. Zimmermann, J. Ericsson, Gastroenterology 2011, 140, 1071–
1083.e5.
[43] N. Yahagi, H. Shimano, K. Hasegawa, K. Ohashi, T. Matsuzaka, Y. Najima,
M. Sekiya, S. Tomita, H. Okazaki, Y. Tamura, Y. Iizuka, K. Ohashi, R. Nagai,
S. Ishibashi, T. Kadowaki, M. Makuuchi, S. Ohnishi, J. I. Osuga, N.
Yamada, Eur. J. Cancer 2005, 41, 1316–1322.
[44] M. Ricchi, M. R. Odoardi, L. Carulli, C. Anzivino, S. Ballestri, A. Pinetti, L. I.
Fantoni, F. Marra, M. Bertolotti, S. Banni, A. Lonardo, N. Carulli, P. Loria,
J. Gastroenterol. Hepatol. 2009, 24, 830–840.
M. Caricato, A. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B.
Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, D.
Williams-Young, F. Ding, F. Lipparini, F. Egidi, J. Goings, B. Peng, A.
Petrone, T. Henderson, D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega,
G. Zheng, W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J.
Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
K. Throssell, J. A. M. Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E.
Brothers, K. N. Kudin, V. N. Staroverov, T. Keith, R. Kobayashi, J.
Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J.
Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo, R. Cammi, J. W.
Ochterski, R. L. Martin, K. Morokuma, O. Farkas, J. B. Foresman, D. J. Fox,
Gaussian 09, Gaussian Inc., 2013.
1
2
3
4
5
6
7
8
9
[55] Y. Zhao, D. G. Truhlar, Theor. Chem. Acc. 2008, 120, 215–241.
[56] L. J. McAllister, D. W. Bruce, P. B. Karadakov, Phys. Chem. Chem. Phys.
2014, 16, 2576.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
[45] X. Chen, L. Li, X. Liu, R. Luo, G. Liao, L. Li, J. Liu, J. Cheng, Y. Lu, Y. Chen,
Life Sci. 2018, 203, 291–304.
[46] U. Ozcan, E. Yilmaz, L. Ozcan, M. Furuhashi, E. Vaillancourt, R. O. Smith,
C. Z. Görgün, G. S. Hotamisligil, Science 2006, 313, 1137–40.
[47] S. Gupta, G. V. R. K. Prasad, A. Mukhopadhaya, J. Biol. Chem. 2015, 290,
31051–31068.
[48] B. Kim, Y. S. Song, Free Radical Res. 2016, 50, 1065–1070.
[49] N. Yao, Y. J. Li, Y. H. Lei, N. Hu, W. M. Chen, Z. Yao, M. Yu, J. S. Liu, W. C.
Ye, D. M. Zhang, J. Exp. Clin. Cancer Res. 2016, 35, 1–15.
[50] G. Kroemer, L. Galluzzi, C. Brenner, Physiol. Rev. 2007, 87, 99–163.
[51] S. Mei, H. M. Ni, S. Manley, A. Bockus, K. M. Kassel, J. P. Luyendyk, B. L.
Copple, W. X. Ding, J. Pharmacol. Exp. Ther. 2011, 339, 487–498.
[52] J. Hemmerich, F. Troger, B. Füzi, G. F. Ecker, Mol. Inf. 2020, 39, 2000005.
[53] S. Fulda, K. M. Debatin, Oncogene 2006, 25, 4798–4811.
[54] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.
Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H. Nakatsuji, X. Li,
[57] A. V. Marenich, C. J. Cramer, D. G. Truhlar, J. Phys. Chem. B 2009, 113,
6378–6396.
[58] F. Denizot, R. Lang, J. Immunol. Methods 1986, 89, 271–7.
[59] Y. Sawai, H. Murata, M. Horii, K. Koto, T. Matsui, N. Horie, Y. Tsuji, E.
Ashihara, T. Maekawa, T. Kubo, S. Fushiki, Oncol. Rep. 2013, 29, 941–5.
Manuscript received: January 21, 2021
Revised manuscript received: February 27, 2021
Accepted manuscript online: March 7, 2021
Version of record online: ■■■, ■■■■
ChemMedChem 2021, 16, 1–12
www.chemmedchem.org
11
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

FULL PAPERS
1
2
3
4
5
6
Dr. F. Alonso, A. Galilea, Dr. P. A.
Mañez, Dr. S. L. Acebedo, Dr. G. M.
Cabrera, Dr. M. Otero, Dr. A. A.
Barquero*, Dr. J. A. Ramírez*
7
8
9
1 – 12
Beyond Pseudo-natural Products:
Sequential Ugi/Pictet-Spengler
Reactions Leading to Steroidal
Pyrazinoisoquinolines That
Trigger Caspase-Independent
Death in HepG2 Cells
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Best of both worlds: Complex polycy-
clic stereochemically rich compounds
have been designed by connecting
fragments present in bioactive natural
products and bioactive synthetic
molecules. One of these natural
product mimics induces caspase-inde-
pendent death in human hepatocarci-
noma cells.