1. Retro-1 Analogues Differentially Affect Oligonucleotide
Delivery and Toxin Trafficking
Bing Yang,[a]
Xin Ming,[a]
Hajer Abdelkafi,[b]
Valerie Pons,[b]
Aurelien Michau,[c]
Daniel Gillet,[c]
Jean-Christophe Cintrat,[b]
Julien Barbier,[b]
and Rudy Juliano*[a]
Retro-1 is a small molecule that displays two important biolog-
ical activities: First, it blocks the actions of certain toxins by al-
tering their intracellular trafficking. Second, it enhances the ac-
tivity of oligonucleotides by releasing them from entrapment
in endosomes. This raises the question of whether the two ac-
tions involve the same cellular target. Herein we report the ef-
fects of several Retro-1 analogues on both toxins and oligonu-
cleotides. We found analogues that affect toxins but not oligo-
nucleotides and vice-versa, while Retro-1 is the only compound
that affects both. This indicates that the molecular target(s) in-
volved in the two processes are distinct.
For more than two decades investigators have sought to har-
ness the precision implicit in the Watson–Crick base pairing of
antisense oligonucleotides, siRNA, and splice switching oligo-
nucleotides (SSOs) to the therapy of disease.[1]
However, oligo-
nucleotide therapeutics has thus far succeeded to only a limit-
ed degree despite massive investments in the chemistry and
formulation of these molecules and despite initial advances in
the clinic.[2]
A major limitation is the poor delivery of oligonu-
cleotides to their intracellular sites of action in the cytosol or
nucleus because of pharmacologically non-productive trapping
in endosomal compartments.[3]
Various lipid or polymeric nano-
carriers have been used to enhance oligonucleotide uptake
and to promote release from endosomes.[4]
However, the cat-
ionic nanocarriers typically used present problems in terms of
toxicity[5]
and biodistribution.[6]
Recently we have pursued an alternative approach that in-
volves the use of small molecules to modulate the processing
of oligonucleotides through the endosomal machinery. Thus
we have shown that the compound Retro-1 selectively releases
oligonucleotides from Rab 7/9 positive late endosomes leading
to a substantial enhancement of oligonucleotide pharmacolog-
ical effects both in cell culture and in vivo.[7]
Retro-1 is
a member of a group of compounds that were initially discov-
ered in a screen for inhibitors of bacterial and plant toxins.[8]
Retro-1 and some of its analogues inhibit the actions of certain
toxins such as Shiga and ricin by blocking their requisite trans-
port through the retrograde intracellular trafficking pathway.[9]
In addition to Retro-1, recent studies have identified other
classes of compounds that act to increase oligonucleotide ef-
fects by causing their release from endosomes,[10]
suggesting
that this approach may be an important general tool for oligo-
nucleotide pharmacology.
The fact that Retro-1 affects both toxin trafficking and oligo-
nucleotide release from endosomes raises the question of
whether the two effects involve the same or different targets
within the cell. The observation that the effect on oligonucleo-
tides requires a considerably higher concentration than the
effect on toxins suggests that more than one cellular target
may be involved. To address this issue we have examined
a number of Retro-1 analogues for their actions on both toxins
and oligonucleotides. We found compounds that affected
toxins but not oligonucleotides and vice-versa, with Retro-
1 being the only compound discovered thus far that could sig-
nificantly affect both. This suggests that the molecular target(s)
involved in enhancing oligonucleotide effectiveness are differ-
ent from those involved in blockade of toxin trafficking.
A series of Retro compound analogues were tested for their
ability to enhance the effect of a SSO in correcting splicing of
a luciferase reporter containing an aberrant intron.[7,11]
The ini-
tial screening for oligonucleotide-related effects involved ap-
proximately 50 compounds and was performed in 96-well
plates. Many of the tested compounds either failed to provide
significant induction of luciferase or clearly had substantial tox-
icity or both. From the initial set, six compounds were chosen
for further testing. The structures of these compounds are
shown in Figure 1. The compounds were tested for luciferase
induction with correction for cell protein using an assay in 24-
well plates, and were also tested for cytotoxicity.
As shown in Figure 2a the compounds Retro-1 and HA061
provided substantial enhancements of the effect of a SSO on
luciferase induction, while compounds VP173, VP174 and
VP184 provided modest but clear-cut enhancements. The com-
pounds were effective in the concentration range of ~60–
120 mm. In contrast, HA229 did not produce luciferase levels
that were significantly different from baseline. In several repeat
experiments HA061 was consistently more effective than
Retro-1 while HA229 was inactive. When tested for cytotoxicity,
all the compounds except VP184 displayed negligible toxicity
over the concentration range tested for luciferase induction
[a] B. Yang, Dr. X. Ming, Dr. R. Juliano
Division of Molecular Pharmaceutics, UNC Eshelman School of Pharmacy,
University of North Carolina, Chapel Hill, NC 27599 (USA)
E-mail: arjay@med.unc.edu
[b] H. Abdelkafi, V. Pons, Dr. J.-C. Cintrat, Dr. J. Barbier
Service de Chimie Bio-organique et Marquage (SCBM), IBITECS, CEA, LabEx
LERMIT, UniversitØ Paris-Saclay, 91191, Gif-sur-Yvette (France)
[c] A. Michau, Dr. D. Gillet
Service d’IngØnierie MolØculaire des ProtØines (SIMOPRO), IBITECS, CEA,
LabEx LERMIT, UniversitØ Paris-Saclay, 91191, Gif-sur-Yvette (France)
The ORCID identification number(s) for the author(s) of this article can
be found under http://dx.doi.org/10.1002/cmdc.201600463.
ChemMedChem 2016, 11, 1 – 6 2016 Wiley-VCH Verlag GmbH Co. KGaA, Weinheim1
These are not the final page numbers! ÞÞThese are not the final page numbers! ÞÞ
CommunicationsDOI: 10.1002/cmdc.201600463

2. (Figure 2b). However, VP184 displayed considerable toxicity in
the concentration range where luciferase induction effects
were observed.
The compounds were also tested for their ability to relieve
protein synthesis inhibition caused by Shiga toxin.[8a]
Com-
pounds HA229, VP184 and Retro-1 substantially diminished
toxin action, with HA229 being especially effective, while
VP173, VP174 and HA061 were without effect on toxins
(Figure 3).
Thus the Retro analogues include compounds that affect
toxins but not oligonucleotides, such as HA229, and other
compounds that affect oligonucleotides but not toxins, such
as HA061, VP173 and VP174. Retro-1 is the only compound we
have identified thus far that has both actions at nontoxic con-
centrations. While VP184 affects toxins it only affects oligonu-
cleotides at concentrations that manifest substantial cytotoxici-
ty. These results clearly indicate that toxin trafficking and en-
hancement of oligonucleotide effects involve distinct targets,
as there are molecules that exert one effect but not the other.
We had previously determined that Retro-1 enhances SSO
actions by allowing release of the oligonucleotide from non-
productive entrapment in endosomes thus providing access to
the cytosol and nucleus.[7]
Therefore, we expected that ana-
logues that enhanced oligonucleotide actions would cause oli-
gonucleotides to escape from endosomes and reach the nu-
cleus while those that failed to influence oligonucleotide ac-
tions would not. Using confocal microscopy, we examined the
effects of the compounds on subcellular distribution of a fluo-
rescent oligonucleotide (Figure 4a,b). In the control cells the
oligonucleotide was located within cytosolic vesicles, presuma-
bly endosomes. As expected, Retro-1 caused partial redistribu-
tion of the oligonucleotide from endosomes to the nucleus.
Surprisingly, however, there was less evidence of nuclear locali-
zation in cells treated with HA061; some faint nuclear fluores-
cence was observed in a few cells, but this was clearly less
than in the case of Retro-1.
Because the fluor-labeled oligonucleotide provided a relative-
ly weak signal, we chose to further study effects of the com-
pounds on endosomes by using a highly fluorescent
Alexa 488–dextran as an endosomal marker, as we have previ-
ously described.[10]
As shown in Figure 4c–j, in the control cells
the dextran was present in intracellular vesicles with no evi-
dence of accumulation in the cytosol or nucleus. In contrast,
Retro-1 clearly caused partial but substantial redistribution of
the dextran into the cytosol and nucleus. As expected there
was no release of dextran with the inactive compound HA229.
However, we also observed only a minor redistribution using
HA061, clearly less than that observed with Retro-1.
An alternative explanation for the strong effect of HA061 on
luciferase induction would be a direct action on RNA transcrip-
tion or splicing. However, when we delivered the SSO by elec-
troporation, thus bypassing all endocytotic mechanisms and al-
lowing the oligonucleotide to directly access the cytosol and
nucleus, there was no significant enhancement by HA061 over
the diluent control (Figure 5). This indicates that HA061 does
not directly affect transcription or splicing. These observations
indicate that certain compounds, such as HA061, can enhance
Figure 1. Structures of Retro-1 and derivatives.
Figure 2. Retro analogues: enhancing effects on oligonucleotides and cyto-
toxicity. a) Luciferase induction: HeLa Luc705 cells were preloaded overnight
with 100 nm SSO623 by incubation in complete medium. Cells were rinsed
and then exposed to various concentrations of the test compounds for 2 h
in DMEM+1% FBS. After removal of the compounds, cells were further in-
cubated for 4 h, and luciferase activity and cell protein content were deter-
mined. As a negative control a mismatched version of SSO623 (100 nm) was
also tested at 90 mm for each analogue tested. Values are the meanÆSEM of
n=3 replicates; results with mismatched oligonucleotide were at baseline
levels and are not shown. b) Cytotoxicity: Cells were exposed to compounds
under the same conditions as in panel (a). The incubation was continued to
24 h, and the viable cell number determined by Alamar Blue assay. Data nor-
malized to control untreated cells: 100%. Values are the meanÆSEM of
n=3 replicates. Differences in luciferase activity at 120 mm HA061, Retro-1,
VP174, and VP173 versus the inactive compound HA229 are significant at
the 95% level or greater, as determined by the Student t-test.
ChemMedChem 2016, 11, 1 – 6 www.chemmedchem.org 2016 Wiley-VCH Verlag GmbH Co. KGaA, Weinheim2
ÝÝ These are not the final page numbers!ÝÝ These are not the final page numbers!
Communications

3. delivery of oligonucleotide to nuclear splicesosomes without
causing an overall redistribution of oligonucleotide that is
readily detectable by fluorescence microscopy.
These studies demonstrate that the actions of Retro-type
compounds on toxins and on oligonucleotides involve differ-
ent molecular targets. At nontoxic concentrations, one subset
of compounds influences toxin trafficking but not the actions
of oligonucleotides, while another subset has the converse
effect. This distinction implies that optimization of these com-
pounds as toxin blockers and as oligonucleotide enhancers
should proceed along different synthetic chemistry pathways.
Because their effective concentrations are in the 100 mm range,
at present the Retro compounds are not sufficiently potent as
oligonucleotide enhancers to warrant their evaluation in thera-
peutic settings and thus additional structure–activity studies
will be needed to further develop these compounds. The cur-
rent studies indicate that such oligonucleotide enhancing
Retro analogues can be developed without impacting the vital
retrograde trafficking pathway.
A surprising sidelight to our studies concerns the apparent
differences in the actions of Retro-1 and HA061. The enhancing
effect of Retro-1 is paralleled by an obvious redistribution of
oligonucleotide from endosomes to the nucleus. Most likely
this is via initial release to the cytosol, as seen with the fluores-
cent dextran, followed by nuclear uptake. With HA061 a strong
oligonucleotide enhancing effect is seen with only subtle
changes in visible oligonucleotide redistribution. This indicates
the presence of an alternative trafficking pathway from endo-
somes to the nuclear splicing machinery that is quite efficient.
However, at this point we cannot suggest a definitive mecha-
nism for this pathway.
Experimental Section
Synthesis: Strategies for the synthesis of Retro compounds have
been described elsewhere.[8b]
Cellular assays of effect and toxicity: Effects of SSOs were tested
using HeLa cells stably transfected with a reporter cassette com-
prised of the coding sequence of firefly luciferase interrupted by
an abnormal intron (HeLa Luc705). Delivery of an appropriate SSO
to the cell nucleus results in corrected splicing and increased ex-
pression of the reporter.[7,11]
All oligonucleotide assays were done
in the absence of any transfection agents other than the Retro ana-
logues being tested. Cytotoxicity of the Retro compounds was
monitored using the Alamar Blue assay.[12]
Cell protein determina-
tions were carried out with the BCA assay (Pierce) as per the sup-
plier’s directions. The various Retro analogues were dissolved in
DMSO at 30 mm and added directly to the cells in culture. Assays
were done in 96- or 24-well formats. The SSO used was termed
SSO623 and was a 2’-O-methylphosphorothioate, whose sequence
has been described elsewhere.[7]
Intoxication assays with Shiga toxin (Stx): HeLa cells were main-
tained at 378C under 5% CO2 in Dulbecco’s modified Eagle’s
medium (DMEM, Invitrogen) supplemented with 10% fetal bovine
serum (FBS), 4.5 gLÀ1
glucose, 100 UmLÀ1
penicillin, 100 mgmLÀ1
streptomycin, 4 mm glutamine, 5 mm pyruvate. The cells were
plated at a density of 50000 cells per well in 96-well Cytostar-TTM
scintillating microplates (PerkinElmer) with scintillator incorporated
into the polystyrene plastic. After incubation with either 30 mm
Retro compound (or 0.1% DMSO) for 4 h at 378C, cells were chal-
lenged with increasing doses of Stx in the continued presence of
compounds. After incubation for 20 h, the medium was removed
and replaced with DMEM without leucine (Eurobio) containing
10% FBS, 2 mm l-glutamine, 0.1 mm non-essential amino acids,
1% penicillin/streptomycin supplemented by 0.5 mCimLÀ1
[14
C]leucine. The cells were grown for an additional 7 h at 378C in
Figure 3. Evaluation of protective activity toward Shiga toxin (Stx): HeLa cells were incubated for 4 h with test compound (30 mm, *) or carrier only (DMSO,
*) before the addition of Stx at the indicated concentrations for 20 min. The media were removed and replaced with DMEM containing [14
C]leucine at
0.5 mCimLÀ1
for 7 h before counting. Data points represent the meanÆSEM of duplicates of a representative experiment.
ChemMedChem 2016, 11, 1 – 6 www.chemmedchem.org 2016 Wiley-VCH Verlag GmbH Co. KGaA, Weinheim3
These are not the final page numbers! ÞÞThese are not the final page numbers! ÞÞ
Communications

4. an atmosphere of 5% CO2 and 95% air. Protein biosynthesis was
then determined by measuring the incorporation of radiolabeled
leucine into cells using a Wallac 1450 MicroBeta liquid scintillation
counter (PerkinElmer). The mean percentage of protein biosynthe-
sis was determined and normalized from duplicate wells. All values
are expressed as meansÆSEM, and data were fitted with Prism
ver. 5 software (GraphPad Inc., San Diego, CA, USA).
Confocal microscopy: Visualization of effects on endosomes in-
volved the use of Alexa 488 labeled 10000 Da dextran (Life Tech-
nologies) or a TAMRA-labeled version of SSO623.[7]
Cells were
imaged on an Olympus FV1000 MPE laser scanning confocal micro-
scope with environmental chamber to maintain 378C, 40% humidi-
ty and 5% CO2; images were collected with a 60” oil immersion
lens. Nuclear to cytoplasmic ratios of fluorescence were quantitat-
ed using NIH Image J.
Acknowledgements
This work was supported by US National Institutes of Health
(NIH) grant R01A151964 to R.L.J. and by the Joint Ministerial Pro-
gram of RD against CBRNE Risks, Agence Nationale de la Re-
Figure 4. a) Effects on the subcellular distribution of oligonucleotide: Hela Luc705 cells (50000) were seeded into glass-bottom dishes and then incubated
overnight with 1 mm TAMRA-623 oligonucleotide. After removal of the oligonucleotide, cells were treated for 2 h in medium+10% FBS with 120 mm Retro-
1 or HA061 and then rinsed. Live cells were observed using a confocal microscope with environmental stage. Yellow arrows indicate cells with distinct nuclear
fluorescence. b) Nuclear/cytosolic intensity ratios of the TAMRA oligonucleotide. Values are the meanÆSEM of n=6 replicates. c)–j) Effects on endosome per-
meability: Hela cells (50000) were seeded into glass-bottom dishes and briefly incubated at 378C in DMEM+10% FBS for attachment. Alexa 488–dextran
(10 kDa, 200 mgmLÀ1
) was added to the medium, and the cells were incubated for 24 h and then rinsed. Cells were then placed in medium+10% FBS and
treated with 140 mm test compounds for 2 h (or left as untreated controls). The compounds were removed, the cells were further incubated for 24 h in
DMEM+10% FBS, and live cells imaged with an Olympus FV1200 confocal microscope at 378C. Both fluorescence and DIC/fluorescence overlap images are
shown. (c,d) untreated control, (e,f) Retro-1 treatment, (g,h) HA061 treatment, (i,j) HA229 treatment. Results shown are typical of three independent assays.
Figure 5. Electroporation of the SSO. HeLa Luc705 cells were electroporated
with various amounts of SSO623 or with mismatched oligonucleotide
(MM623). The cells were cultured for 3 h to allow cell attachment. Cells were
then treated for 2 h with 120 mm Retro-1 or HA061 or with DMSO diluent.
After removal of the compounds, cells were incubated for an additional 4 h
and then luciferase activity and cell protein content were determined.
ChemMedChem 2016, 11, 1 – 6 www.chemmedchem.org 2016 Wiley-VCH Verlag GmbH Co. KGaA, Weinheim4
ÝÝ These are not the final page numbers!ÝÝ These are not the final page numbers!
Communications

5. cherche (ANR, France) (grant Anti-HUS ANR-14-CE16-0004), LabEx
LERMIT (grant R3 RetroLeishma), the Conseil RØgional d’ le de
France (grant from the DIM Malinf initiative 140101 and CEA) to
D.G., J.C.C., and J.B.
Keywords: endocytosis · oligonucleotides · Retro-1 · toxins ·
trafficking
[1] a) C. F. Bennett, E. E. Swayze, Annu. Rev. Pharmacol. Toxicol. 2013, 50,
259–293; b) J. K. Watts, D. R. Corey, J. Pathol. 2011, 226, 365–379;
c) J. C. Burnett, J. J. Rossi, Chem. Biol. 2012, 19, 60–71; d) R. L. Juliano,
Nucleic Acids Res. 2016, 44, 6518–6548.
[2] M. T. Tse, Nat. Rev. Drug Discovery 2013, 12, 179.
[3] a) J. Gilleron, W. Querbes, A. Zeigerer, A. Borodovsky, G. Marsico, U.
Schubert, K. Manygoats, S. Seifert, C. Andree, M. Stoter, H. Epstein-
Barash, L. Zhang, V. Koteliansky, K. Fitzgerald, E. Fava, M. Bickle, Y. Ka-
laidzidis, A. Akinc, M. Maier, M. Zerial, Nat. Biotechnol. 2013, 31, 638–
646; b) A. K. Varkouhi, M. Scholte, G. Storm, H. J. Haisma, J. Controlled
Release 2011, 151, 220–228.
[4] a) R. Kanasty, J. R. Dorkin, A. Vegas, D. Anderson, Nat. Mater. 2013, 12,
967–977; b) J. Li, Y. Wang, Y. Zhu, D. Oupicky, J. Controlled Release 2013,
172, 589–600.
[5] S. Akhtar, Expert Opin. Drug Metab. Toxicol. 2010, 6, 1347–1362.
[6] R. Juliano, J. Bauman, H. Kang, X. Ming, Mol. Pharmaceutics 2009, 6,
686–695.
[7] X. Ming, K. Carver, M. Fisher, R. Noel, J. C. Cintrat, D. Gillet, J. Barbier, C.
Cao, J. Bauman, R. L. Juliano, Nucleic Acids Res. 2013, 41, 3673–3687.
[8] a) B. Stechmann, S. K. Bai, E. Gobbo, R. Lopez, G. Merer, S. Pinchard, L.
Panigai, D. Tenza, G. Raposo, B. Beaumelle, D. Sauvaire, D. Gillet, L. Jo-
hannes, J. Barbier, Cell 2010, 141, 231–242; b) H. Abdelkafi, A. Michau,
A. Clerget, D. A. Buisson, L. Johannes, D. Gillet, J. Barbier, J. C. Cintrat,
ChemMedChem 2015, 10, 1153–1156.
[9] L. Johannes, C. Wunder, Traffic 2011, 12, 956–962.
[10] B. Yang, X. Ming, C. Cao, B. Laing, A. Yuan, M. A. Porter, E. A. Hull-Ryde,
J. Maddry, M. Suto, W. P. Janzen, R. L. Juliano, Nucleic Acids Res. 2015,
43, 1987–1996.
[11] M. R. Alam, V. Dixit, H. Kang, Z. B. Li, X. Chen, J. Trejo, M. Fisher, R. L. Ju-
liano, Nucleic Acids Res. 2008, 36, 2764–2776.
[12] X. Ming, W. Ju, H. Wu, R. R. Tidwell, J. E. Hall, D. R. Thakker, Drug Metab.
Dispos. 2009, 37, 424–430.
Received: September 9, 2016
Revised: October 12, 2016
Published online on , 0000
ChemMedChem 2016, 11, 1 – 6 www.chemmedchem.org 2016 Wiley-VCH Verlag GmbH Co. KGaA, Weinheim5
These are not the final page numbers! ÞÞThese are not the final page numbers! ÞÞ
Communications

6. COMMUNICATIONS
B. Yang, X. Ming, H. Abdelkafi, V. Pons,
A. Michau, D. Gillet, J.-C. Cintrat,
J. Barbier, R. Juliano*
–
Retro-1 Analogues Differentially Affect
Oligonucleotide Delivery and Toxin
Trafficking
Clarity in Retro-spect: Both oligonucle-
otides (red ribbons) and toxins (orange
squares) enter cells by endocytosis and
traffic to early endosomes (EE). Oligos
traffic to late endosomes (LE) and
thence to lysosomes (LY), where they
are degraded. Toxins traffic via the Ret-
romer complex (R) to the trans Golgi
(TG) to become active. Retro-1 blocks
toxin trafficking and also releases oligos
from late endosomes.
ChemMedChem 2016, 11, 1 – 6 www.chemmedchem.org 2016 Wiley-VCH Verlag GmbH Co. KGaA, Weinheim6
ÝÝ These are not the final page numbers!ÝÝ These are not the final page numbers!