DMT1 Inhibitors Kill Cancer Stem Cells by Blocking Lysosomal Iron Translocation

Article abstract of DOI:10.1002/chem.202000159

Cancer stem cells (CSC) constitute a cell subpopulation in solid tumors that is responsible for resistance to conventional chemotherapy, metastasis and cancer relapse. The natural product Salinomycin can selectively target this cell niche by directly interacting with lysosomal iron, taking advantage of upregulated iron homeostasis in CSC. Here, inhibitors of the divalent metal transporter 1 (DMT1) have been identified that selectively target CSC by blocking lysosomal iron translocation. This leads to lysosomal iron accumulation, production of reactive oxygen species and cell death with features of ferroptosis. DMT1 inhibitors selectively target CSC in primary cancer cells and circulating tumor cells, demonstrating the physiological relevance of this strategy. Taken together, this opens up opportunities to tackle unmet needs in anti-cancer therapy.

Full text of DOI:10.1002/chem.202000159

A Journal of
Accepted Article
Title: DMT1 Inhibitors Kill Cancer Stem Cells by Blocking Lysosomal
Iron Translocation
Authors: Andreea L. Turcu, Antoine Versini, Nadjib Khene, Christine
Gaillet, Tatiana Cañeque, Sebastian Müller, and Raphaël
Rodriguez
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.202000159
Link to VoR: http://dx.doi.org/10.1002/chem.202000159
Supported by

10.1002/chem.202000159
Chemistry - A European Journal
DMT1 Inhibitors Kill Cancer Stem Cells by Blocking
Lysosomal Iron Translocation
Andreea L. Turcu,1† Antoine Versini,2,3,4† Nadjib Khene,2,3,4 Christine Gaillet,2,3,4 Tatiana
Cañeque,2,3,4 Sebastian Müller,2,3,4* Raphaël Rodriguez2,3,4*
1Laboratori de Química Farmacèutica (Unitat Associada al CSIC), Facultat de Farmàcia i
Ciències de l’Alimentació i Institut de Biomedicina, University of Barcelona, Av. Joan XXIII,
27-31, 08028, Barcelona, Spain.
2Institut Curie, 26 rue d’Ulm, 75248 Paris Cedex 05, France.
3PSL Université Paris, France.
4Chemical Biology of Cancer Laboratory, CNRS UMR 3666, INSERM U1143.
† These authors contributed equally to this work.
*Correspondence to: sebastian.muller@curie.fr, raphael.rodriguez@curie.fr
Abstract
Cancer stem cells (CSC) constitute a subpopulation of cells in solid tumors that is
responsible for resistance to conventional chemotherapy, metastasis and cancer relapse.
The natural product Salinomycin can selectively target this cell niche by directly
interacting with lysosomal iron, taking advantage of upregulated iron homeostasis in
CSC. Here, we identify inhibitors of the divalent metal transporter 1 (DMT1) that
selectively target CSC by blocking lysosomal iron translocation. This leads to lysosomal
iron accumulation, production of reactive oxygen species and cell death with features of
ferroptosis. DMT1 inhibitors selectively target CSC in primary cancer cells and
circulating tumor cells, demonstrating the physiological relevance of this strategy. Taken
together, this opens up opportunities to tackle unmet needs in anti-cancer therapy.
Tumors consist of heterogenous cell populations, among which a subset is responsible
for metastasis formation, cancer relapse and in many cases treatment resistance, constituting a
major drawback for the success of conventional anti-cancer therapy. These so-called ‘cancer
stem cells’ (CSC) are highly plastic and can transition reversibly between epithelial and
mesenchymal states[1] with distinct biochemical signatures, which opens up opportunities for
therapeutic intervention[2]. Their metastable mesenchymal state is characterized by an addiction
This article is protected by copyright. All rights reserved.

10.1002/chem.202000159
Chemistry - A European Journal
to iron[3], providing the means to specifically target and eradicate this cancer cell population.
One of the most promising small molecules that specifically targets CSC is the natural product
Salinomycin[4] (1) (Scheme 1). It selectively kills CSC by sequestering lysosomal iron, leading
to the production of reactive oxygen species (ROS) via Fenton chemistry, lysosomal membrane
permeabilization and subsequent cell death reminiscent of ferroptosis[3a]. Thus, strategies to
retain iron in lysosomes[3a, 5] represent an attractive therapeutic strategy to target and eradicate
CSC. A key protein in the regulation of iron homeostasis is the divalent metal transporter 1
(DMT1) located within the plasma membrane and membranes of late endosomes and
lysosomes, which enables the translocation of ferrous iron (Fe2+) to the cytosol following iron
endocytosis[6]. Thus, we hypothesized that inhibiting DMT1 could prevent iron translocation,
potentially leading to lysosomal iron overload, ROS production and cell death, providing a
novel therapeutic angle to eradicate iron-addicted CSC.
To this end, we synthesized a library of small molecules previously reported as DMT1
inhibitors, which can be classed into three groups: 1) ebselen (2)[7], 2) substituted pyrazoles
(3a-e)[8] and 3) benzylisothioureas (4a-e)[9] (Scheme 1, Supplementary Information).
Scheme 1. Molecular structures of small molecules reported as DMT1 inhibitors. A)
Salinomycin. B) Ebselen. C) Substituted pyrazoles. D) Benzylisothioureas.
This article is protected by copyright. All rights reserved.

10.1002/chem.202000159
Chemistry - A European Journal
Next, we evaluated the biological activity of this library of DMT1 inhibitors against
transformed human mammary HMLER[10] CD44high/CD24low cells, a well-established model of
breast CSC, and their HMLER CD44low/CD24high counterparts deprived of CSC properties,
using 1 as a positive control. In this assay, a small molecule selective for CSC preferentially
kills the HMLER CD44high/CD24low cell population, as reported for 1[3a]. We observed that
DMT1 inhibitors of group 2 (3a-e) exhibited IC50 concentrations of cell viability against
HMLER CD44high/CD24low in the low micromolar range (Table 1, Figure S1). In contrast, IC50
values of structurally distinct molecule 2 and group 3 (4a-e) were more elevated. Importantly,
almost all the molecules showed selectivity for HMLER CD44high/CD24low cells exhibiting
CSC properties, as defined by the ratio of IC50 HMLER CD44low/CD24high (Table 1).
Interestingly, while 1 exhibited a selectivity of 4.1 fold, we found that 3d exhibited a 2.8 fold
and 4a a 2.6 fold selectivity in this model.
IC50 HMLER IC50 HMLER
Ratio HMLER
CD44low
5.8
CD44high
1.4
CD44low/CD44high
1
2
4.1
1.1
1.2
1.3
1.5
2.8
1.3
2.6
1.3
1.6
1.8
0.9
19.4
3.1
18.0
2.5
3a
3b
3c
3d
3e
4a
4b
4c
4d
4e
0.4
0.3
2.1
1.4
1.1
0.4
16.4
13.9
3.9
12.6
5.4
3.0
16.7
37.2
101.3
10.2
20.5
108.8
Table 1. IC50 values of cell viability at 72 h measured for HMLER CD44low/CD24high and
HMLER CD44high/CD24low cells in cells treated with 1 and the DMT1 inhibitors 2-4e.
Selectivity ratios (CD44low/CD44high) are indicated, with values >1 indicating a selectivity for
cells with CSC properties.
These results indicated that DMT1 inhibitors show a preference for cells with features of CSC,
which prompted us to investigate their mechanism of action. To this end, we used the Fe2+-
specific turn-on fluorescent probe RhoNox-M, which can detect the presence of Fe2+ in
lysosomes[11], and monitored lysosomal Fe2+ by means of flow cytometry (Figure 1A).
Whereas previous studies employed an assay using calcein quenching to assess cellular iron,
using RhoNox-M allowed us to directly investigate lysosomal iron in relation to DMT1. Like
1, compounds 3a-e showed a pronounced increase of lysosomal Fe2+. In comparison, a
moderate increase of lysosomal Fe2+ was observed for 4a-d but not for 4e or 2. The moderate
This article is protected by copyright. All rights reserved.

10.1002/chem.202000159
Chemistry - A European Journal
efficacy of 4e may be assigned to a distinct configuration of the isothiourea moieties compared
to 4a-d, which may differentially affect target engagement. Although previously reported as a
DMT1 inhibitor, ebselen (2) was also reported as a ROS scavenger preventing ferroptotic cell
death[12]. Thus, reported effects on iron homeostasis might be linked to this effect as opposed
to a direct inhibition of DMT1 per se. Thus, the lack of increased lysosomal iron loading
together with the absence of selectivity for HMLER CD44high/CD24low cells makes 2 a suitable
negative control in our settings.
Figure 1. DMT1 inhibitors 3d and 4a increase lysosomal Fe2+ loading. A) Flow cytometry
analysis of RhoNox-M fluorescence in cells treated with 1, 2, 3a-e and 4a-e for 24 h. B)
This article is protected by copyright. All rights reserved.

10.1002/chem.202000159
Chemistry - A European Journal
Fluorescence microscopy images of RhoNox-1 and lysotracker in cells treated with 1, 2, 3d
and 4a for 24 h. Lysotracker deep red stains the lysosomes (green) and DAPI stains nuclear
DNA (blue). Scale bar, 10 µm. Quantification of 20 cells per condition. N = 3 biological
replicates. Colocalization analyses expressed as Pearson correlation coefficients. Kruskal-
Wallis test with Dunn’s correction. Bars and error bars, mean values +/- SD. * P ≤ 0.05, ** P ≤
0.01, *** P ≤ 0.001, **** P ≤ 0.0001. HMLER CD44high/CD24low cells were used throughout
the figure.
To further investigate the cellular mechanisms of action (MoA), we took forward the
most selective molecule of groups 2 and 3 of structurally distinct DMT1 inhibitors, namely 3d
and 4a, using 1 and 2 as positive and negative controls, respectively. We treated cells with each
molecule and monitored lysosomal iron loading using a related fluorescent Fe2+ selective turn-
on probe RhoNox-1[13] and lysotracker. In untreated (Unt.) cells and in cells treated with 2, we
observed a weak and diffuse staining of RhoNox-1, whereas cells treated with 1, 3d and 4a
exhibited an increase in fluorescence, specifically localized in lysosomal compartments as
defined by colocalization with lysotracker. This data indicated that these molecules block the
translocation of Fe2+ from the lumen of lysosomes to the cytosol, thereby leading to increased
lysosomal iron loading (Figure 1B).
Next, we investigated whether molecules 3d and 4a induced lysosomal ROS
production, leading on from their lysosomal Fe2+ loading properties. To this end, we monitored
by flow cytometry the levels of ROS using the CellROX probe, and found that 1, 3d and 4a
promoted the production of ROS. In contrast, levels of ROS in cells treated with compound 2
were lower and comparable to that found in untreated cells (Figure 2A). Consistent with the
accumulation of iron in lysosomes, fluorescence microscopy indicated the production of ROS
in this organelle in cells treated with 1, 3d and 4a, as assessed by colocalization analyses of
CellROX and lysotracker (Figure 2B). This, in turn, leads to lysosomal membrane
permeabilization (LMP), as defined by the release of Dextran-FITC (Figure S2). In agreement
with LMP, we observed increased levels of lipid ROS in cells treated with 3d and 4a (Figure
2C and Figure S3), suggesting a cell death mechanism sharing molecular features with
ferroptosis[3a, 14]
.
This article is protected by copyright. All rights reserved.

10.1002/chem.202000159
Chemistry - A European Journal
Figure 2. DMT1 inhibitors 3d and 4a promote the production of lysosomal ROS lipid
peroxidation. A) Flow cytometry analysis of fluorescence of CellROX in cells treated with 1,
2, 3d and 4a for 48 h. B) Fluorescence microscopy analysis of CellROX fluorescence and
Lysotracker. Cells were treated with 1, 2, 3d and 4a for 48 h. CellROX fluoresces upon ROS
induction (red), lysotracker stains the lysosomes (green) and DAPI stains nuclear DNA (blue).
Scale bar, 10 µm. Quantification of 20 cells per condition. Colocalization analyses expressed
as Pearson correlation coefficients. N = 3 biological replicates. Kruskal-Wallis test with
Dunn’s correction. Bars and error bars, mean values +/- SD. * P ≤ 0.05, ** P ≤ 0.01, *** P ≤
0.001, **** P ≤ 0.0001. C) Flow cytometry analysis of lipid ROS in cells treated with 1, 2, 3d
and 4a for 48 h using BODIPY 581/591 C11. HMLER CD44high/CD24low cells were used
throughout the figure.
We next investigated the capacity of 3d and 4a to kill CSC using physiologically
relevant models. To this end, we chose a) primary breast cancer cells, b) the triple-negative
human breast cancer cell line MDA-MB-468 and c) primary lung circulating tumor cells (CTC)
that recapitulate properties of disseminating cells susceptible to promote metastatic spread. We
promoted epithelial-to-mesenchymal transition (EMT)[1, 15] leading to mesenchymal CD44high
cell populations using cytokines, epithelial growth factor (EGF) or transforming growth factor
ß (TGF-ß). Induction of EMT was further confirmed by western blotting monitoring CD44 and
the mesenchymal proteins vimentin and fibronectin (Figure S4). We next assessed the IC50
values of cell viability for 1, 3d and 4a in these cells in the epithelial (CD44low) and stem-like
mesenchymal (CD44high) states.
This article is protected by copyright. All rights reserved.

10.1002/chem.202000159
Chemistry - A European Journal
1
3d
4a
IC50 primary breast CD44low in µM
IC50 primary breast CD44high in µM
21.6
5.1
4.0
1.3
3.1
26.1
5.9
Ratio primary breast
4.2
4.4
CD44low/CD44high
IC50 MDA-MB-468 CD44low in µM
6.5
2.7
2.4
2.4
0.6
4.0
4.6
1.1
4.2
IC50 MDA-MB-468 CD44high in µM
Ratio MDA-MB-468
CD44low/CD44high
IC50 lung CTC CD44low in µM
IC50 lung CTC CD44high in µM
Ratio lung CTC CD44low/CD44high
29.3
11.1
2.6
8.0
1.4
5.7
44.3
8.8
5.0
Table 2. IC50 values of cell viability measured for primary breast cancer cells, MDA-MB-468
cells and lung CTC treated with 1, 3d and 4a for 72 h. Selectivity ratios (CD44low/CD44high
)
are indicated, with values >1 indicating a selectivity for cells with CSC properties.
We found that 4a exhibited a selectivity for primary breast cancer cells in the mesenchymal
state comparable to that found for 1 and that both 3d and 4a showed significant selectivity for
CD44high MDA-MB-468 cells and CD44high lung CTC in the range of 4.0-5.7 fold, which was
superior to the selectivity observed for 1, ranging between 2.4-2.6 fold (Table 2, Figure S5).
Taken together, these data indicate that DMT1 inhibitors can selectively kill cancer cells with
stem-like properties by interfering with lysosomal iron regulation.
Here, we have generated a library of small molecule inhibitors of DMT1 function and
evaluated their capacity to specifically target cells with CSC properties. Some of these
molecules, notably 3d and 4a, show an efficiency and selectivity for these cells in the same
order of magnitude or better than 1, which remains a benchmark molecule in the context of
CSC targeting. We found that, like 1, these molecules promote increased iron loading in
lysosomes, leading to ROS production, lysosomal membrane permeabilization and cell death
reminiscent of ferroptosis. The MoA of these molecules may involve other complex
components given that DMT1 can also be found in the outer mitochondrial membrane and at
the plasma membrane [16]. Since mitochondrial metabolism is upregulated in certain subsets of
CSC[3c] and since mitochondrial iron and copper homeostasis play a central role in the
maintenance of these cells[17], inhibition of DMT1 activity may impact on mitochondrial iron
metabolism, which could potentially be exploited for therapeutic benefits. This study
establishes DMT1 function as a tractable strategy to specifically target the CSC niche. This
This article is protected by copyright. All rights reserved.

10.1002/chem.202000159
Chemistry - A European Journal
provides a basis for the development of innovative anti-cancer treatment modalities to avoid
cancer relapse and metastasis formation, which constitutes an unmet health need for cancer
patients.
Acknowledgements
We thank the CNRS, INSERM and Institut Curie for generous funding. R.R. laboratory is
supported by the European Research Council (grant agreement No [647973]), the Fondation
Charles Defforey-Institut de France and Ligue Contre le Cancer (Equipe Labellisée). Région
Ile de France is gratefully acknowledged for financial support of the 500 MHz NMR. A. V. is
funded by PSL Research University and A.L.T. by Fundació Pedro i Pons and Spanish
Ministerio de Educacion, Cultura y Deporte. We acknowledge the PICT-IBiSA@Pasteur
Imaging Facility of Institut Curie, member of the France-BioImaging national research
infrastructure. We thank P. Le Bacon for assistance with high-resolution microscopy. We
thank A. Puisieux for providing us with HMLER cell lines.
References
[1]
a) M. A. Nieto, R. Y. Huang, R. A. Jackson, J. P. Thiery, Cell 2016, 166, 21-45; b) T.
Brabletz, R. Kalluri, M. A. Nieto, R. A. Weinberg, Nat Rev Cancer 2018, 18, 128-134;
c) I. Pastushenko, A. Brisebarre, A. Sifrim, M. Fioramonti, T. Revenco, S. Boumahdi,
A. Van Keymeulen, D. Brown, V. Moers, S. Lemaire, S. De Clercq, E. Minguijon, C.
Balsat, Y. Sokolow, C. Dubois, F. De Cock, S. Scozzaro, F. Sopena, A. Lanas, N.
D'Haene, I. Salmon, J. C. Marine, T. Voet, P. A. Sotiropoulou, C. Blanpain, Nature
2018, 556, 463-468.
[2]
[3]
a) D. C. Voon, R. Y. Huang, R. A. Jackson, J. P. Thiery, Mol. Oncol. 2017, 11, 878-
891; b) S. Müller, T. Cañeque, V. Acevedo, R. Rodriguez, Isr. J. Chem. 2017, 57, 239-
250.
a) T. T. Mai, A. Hamai, A. Hienzsch, T. Cañeque, S. Müller, J. Wicinski, O. Cabaud,
C. Leroy, A. David, V. Acevedo, A. Ryo, C. Ginestier, D. Birnbaum, E. Charafe-
Jauffret, P. Codogno, M. Mehrpour, R. Rodriguez, Nat.Chem. 2017, 9, 1025-1033; b)
D. Basuli, L. Tesfay, Z. Deng, B. Paul, Y. Yamamoto, G. Ning, W. Xian, F. McKeon,
M. Lynch, C. P. Crum, P. Hegde, M. Brewer, X. Wang, L. D. Miller, N. Dyment, F. M.
Torti, S. V. Torti, Oncogene 2017, 36, 4089-4099; c) S. Müller, F. Sindikubwabo, T.
Cañeaue, A. Lafon, A. Versini, B. Lombard, D. Loew, T.-D. Wu, C. Genestier, E.
Charafe-Jauffret, A. Durand, C. Vallot, S. Baulande, S. Servant, R. Rodriguez,
BioRxvie 2019, doi.org/10.1101/693424; d) A. Hamai, T. Cañeque, S. Müller, T. T.
Mai, A. Hienzsch, C. Ginestier, E. Charafe-Jauffret, P. Codogno, M. Mehrpour, R.
Rodriguez, Autophagy 2017, 13, 1465-1466.
[4]
[5]
P. B. Gupta, T. T. Onder, G. Jiang, K. Tao, C. Kuperwasser, R. A. Weinberg, E. S.
Lander, Cell 2009, 138, 645-659.
H. Waldmann, L. Laraia, G. Garivet, D. J. Foley, N. Kaiser, S. Müller, S. Zinken, T.
Pinkert, J. Wilke, D. Corkery, A. Pahl, S. Sievers, P. Janning, C. Arenz, Y. Wu, R.
Rodriguez, Angew. Chem. Int. Ed. 2019.
[6]
M. Tabuchi, T. Yoshimori, K. Yamaguchi, T. Yoshida, F. Kishi, J. Biol. Chem. 2000,
275, 22220-22228.
This article is protected by copyright. All rights reserved.

10.1002/chem.202000159
Chemistry - A European Journal
[7]
[8]
[9]
a) H. A. Wetli, P. D. Buckett, M. Wessling-Resnick, Chem. Biol. 2006, 13, 965-972; b)
L. Xie, W. Zheng, N. Xin, J. W. Xie, T. Wang, Z. Y. Wang, Neurochem. Int. 2012, 61,
334-340.
J.-J. Cadieux, Z. Zhang, M. Mattice, A. Brownlie-Cutts, J. Fu, L. G. Ratkay, R. Kwan,
J. Thompson, J. Sanghara, J. Zhong, Y. P. Goldberg, Bioorg. Med. Chem. Lett. 2012,
22, 90-95.
Z. Zhang, V. Kodumuru, S. Sviridov, S. Liu, M. Chafeev, S. Chowdhury, N. Chakka, J.
Sun, S. J. Gauthier, M. Mattice, L. G. Ratkay, R. Kwan, J. Thompson, A. B. Cutts, J.
Fu, R. Kamboj, Y. P. Goldberg, J. A. Cadieux, Bioorg. Med. Chem. Lett. 2012, 22,
5108-5113.
[10] a) A. P. Morel, M. Lievre, C. Thomas, G. Hinkal, S. Ansieau, A. Puisieux, PLoS One
2008, 3, e2888; b) B. Elenbaas, L. Spirio, F. Koerner, M. D. Fleming, D. B. Zimonjic,
J. L. Donaher, N. C. Popescu, W. C. Hahn, R. A. Weinberg, Genes Dev. 2001, 15, 50-
65.
[11] M. Niwa, T. Hirayama, K. Okuda, H. Nagasawa, Org. Biomol. Chem. 2014, 12, 6590-
6597.
[12] S. J. Dixon, K. M. Lemberg, M. R. Lamprecht, R. Skouta, E. M. Zaitsev, C. E.
Gleason, D. N. Patel, A. J. Bauer, A. M. Cantley, W. S. Yang, B. Morrison, 3rd, B. R.
Stockwell, Cell 2012, 149, 1060-1072.
[13] T. Hirayama, K. Okuda, H. Nagasawa, Chem. Sci. 2013, 4, 1250-1256.
[14] M. Conrad, D. A. Pratt, Nat. Chem. Biol. 2019, 15, 1137-1147.
[15] I. Pastushenko, C. Blanpain, Trends Cell Biol. 2019, 29, 212-226.
[16] N. A. Wolff, A. J. Ghio, L. M. Garrick, M. D. Garrick, L. Zhao, R. A. Fenton, F.
Thevenod, FASEB J. 2014, 28, 2134-2145.
[17] a) K. Laws, G. Bineva-Todd, A. Eskandari, C. Lu, N. O'Reilly, K. Suntharalingam,
Angew. Chem. Int. Ed. 2018, 57, 287-291; b) S. Müller, A. Versini, F. Sindikubwabo,
G. Belthier, S. Niyomchon, J. Pannequin, L. Grimaud, T. Cañeque, R. Rodriguez, PLoS
One 2018, 13, e0206764.
This article is protected by copyright. All rights reserved.