Structure-Activity Relationships of JMV4463, a Vectorized Cathepsin D Inhibitor with Antiproliferative Properties: The Unique Role of the AMPA-Based Vector

Article abstract of DOI:10.1002/cmdc.201500457

Cathepsin D (CathD) is overexpressed and secreted by several solid tumors and stimulates their growth, the mechanism of which is still not understood. In this context, the pepstatin bioconjugate JMV4463 [Ac-arg-O₂Oc-(Val)₃-Sta-Ala-St

Full text of DOI:10.1002/cmdc.201500457

DOI: 10.1002/cmdc.201500457
Full Papers
Structure–Activity Relationships of JMV4463, a Vectorized
Cathepsin D Inhibitor with Antiproliferative Properties:
The Unique Role of the AMPA-Based Vector
Lubomir L. Vezenkov, ClØment A. Sanchez, Virginie Bellet, Vincent Martin, Marie Maynadier,
Nadir Bettache, Vincent Lisowski, Jean Martinez, Marcel Garcia, Muriel Amblard, and Jean-
FranÅois Hernandez*[a]
Cathepsin D (CathD) is overexpressed and secreted by several
solid tumors and stimulates their growth, the mechanism of
which is still not understood. In this context, the pepstatin bio-
conjugate JMV4463 [Ac-arg-O2Oc-(Val)3-Sta-Ala-Sta-(AMPA)4-
NH2; O2Oc=8-amino-3,6-dioxaoctanoyl, Sta=statine, AMPA=
ortho-aminomethylphenylacetyl], containing a new kind of
cell-penetrating vector, was previously shown to exhibit potent
antiproliferative effects in vitro and to delay the onset of
tumors in vivo. In this study, we performed a structure–activity
relationship analysis to evaluate the significance of the inhibi-
tor and vector moieties of JMV4463. By modifying both statine
residues of pepstatin we found that the antiproliferative activi-
ty is correlated with CathD inhibition, supporting a major role
of the catalytic activity of intracellular CathD in cancer cell pro-
liferation. Replacing the vector composed of four AMPA units
with other vectors was found to abolish cytotoxicity, although
all of the conjugates enabled pepstatin transport into cells. In
addition, the AMPA4 vector must be localized at the C terminus
of the bioconjugate. The unexpected importance of the vector
structure and position for cytotoxic action suggests that
AMPA4 enables pepstatin to inhibit the proteolysis of critical
CathD substrates involved in cell proliferation via a unique
mechanism of action.
Introduction
Cathepsin D (CathD) is overexpressed and secreted by
a number of solid tumors (breast, prostate, melanoma, colorec-
tal cancer, glioma, etc.) and was shown to be involved in their
progression and development of metastasis.[1–6] In the clinic,
elevated CathD levels are related to a poor prognosis in vari-
ous cancers.[7–11] At first recognized for its role in the degrada-
tion of proteins in lysosomes, CathD was also proposed to be
responsible for more specific functions that control tissue re-
generation[12] and that regulate essential cellular processes[13]
such as the activation or regulation of various growth factors
and/or the degradation of growth inhibitors,[14–17] and the inac-
tivation of chemokines involved in the antitumor immune re-
sponse.[18] CathD also contributes to the activation of anti-
apoptotic proteins such as Aven and the induction of autopha-
gy.[19–24] Therefore, CathD constitutes a pertinent target in
cancer; inhibition of its intracellular or extracellular proteolytic
activity could be an efficient way to help control or impede
tumor growth.[15,25–29] We previously reported that a cell-pene-
trating CathD inhibitor (JMV4463) exhibits potent anti-prolifer-
ative effects in vitro and delays tumor emergence and growth
in tests with mice xenografted with breast cancer cells.[30] Al-
though CathD is present in all cells, the absence of observable
toxicity toward normal human fibroblasts and in in vivo experi-
ments indicates that targeting the overexpression of CathD in
tumor cells can be considered as a selective strategy.
JMV4463 is composed of three main regions: 1) pepstatin,
a potent CathD inhibitor (KD ~0.5 nm),[31,32] which only poorly
penetrates cells by itself; 2) a tetramer of ortho-aminomethyl-
phenylacetyl residues (AMPA4), which belongs to a new class
of vectors based on short oligomers of constrained dipeptide
mimetics called cell-penetrating non-peptides (CPNP), which
target the endolysosomal pathway;[33] and 3) a hydrophilic
moiety that improves aqueous solubility (Figure 1).
[a] Dr. L. L. Vezenkov,+ C. A. Sanchez,+ Dr. V. Bellet,+ Dr. V. Martin,
Dr. M. Maynadier, Dr. N. Bettache, Prof. V. Lisowski, Prof. J. Martinez,
Dr. M. Garcia, Dr. M. Amblard, Dr. J.-F. Hernandez
To determine the structural features required for efficient in-
ternalization and inhibition of tumor proliferation, we synthe-
sized several analogues of JMV4463 (Table 1). In particular, we
replaced the AMPA4 vector with various cell-penetrating pep-
tides (CPPs) and CPNPs. We also assessed the importance of
the positions of the various components in the conjugate and
that of the two statine residues of the pepstatin moiety.
Institut des BiomolØcules Max Mousseron (IBMM)
UMR5247 CNRS, UniversitØ de Montpellier, ENSCM, FacultØ de Pharmacie
15 avenue Charles Flahault, 34093 Montpellier Cedex 5 (France)
E-mail: jean-francois.hernandez@univ-montp1.fr
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201500457: characterization (yields,
HPLC, MS data) of conjugates 1–19, synthesis of Fmoc-LBD-OH 25, and
1H NMR spectra of 25 and synthetic intermediates 23 and 24.
ChemMedChem 2016, 11, 302 – 308
302
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 1. Structure of JMV4463, highlighting the hydrophobic, pepstatin, and AMPA4 regions.
Results and Discussion
Design of bioconjugates 1–19
For solubility reasons, all bioconjugates except 12 and 13 con-
tain a hydrophilic moiety composed of a d-arginine (arg) resi-
due and a short poly(ethylene glycol) segment, 8-amino-3,6-di-
oxaoctanoyl (O2Oc). For most of the conjugates this moiety
was attached at the N terminus, but it was also placed at the
C terminus in the reverse order (O2Oc-arg, compound 9) or
partially moved (in compound 11) (Table 1). Several analogues
that differ from JMV4463 in the number of AMPA units (com-
pounds 2–4 with 2, 3, and 5 units, respectively) were prepared
in order to assess the optimal length required for efficient
translocation and anti-proliferative activity. We also explored
the significance of the position of the AMPA tetramer (AMPA4)
vector within the conjugate. A fully reversed conjugate (AMPA4
at the N terminus, compound 9) was prepared as well as a con-
jugate in which AMPA4 was placed between the hydrophilic
moiety and the inhibitor (compound 10). The O2Oc residue
was moved as a spacer between pepstatin and the vector
(compound 11). Substitution of AMPA4 with other vectors was
also evaluated, including the highly charged and hydrophilic
CPPs penetratin[34] (12) and octaarginine (Arg8)[35] (13) as well
as other uncharged CPNPs composed of d-benzothiazepine
(DBT4, compound 14)[33] or l-benzodiazepine (LBD4, compound
15). Finally, we explored the significance of the ortho-AMPA
structure by preparing the corresponding analogues with the
meta-AMPA (compound 16) and para-AMPA (compound 17)
isomers. In the case of the CPP-containing bioconjugates 12
and 13, the hydrophilic moiety was omitted because their
vector moieties (penetratin, Arg8, respectively) are themselves
highly hydrophilic.
Figure 2. Structures of special residues contained in bioconjugates 1–19.
The pepstatin moiety was also modified (Figure 2). The hy-
droxy group of the statine residue is known to be crucial for
the inhibition of aspartyl proteases like CathD. To evaluate the
significance of CathD inhibition in the biological activity of
JMV4463, the two statine residues of the pesptatin moiety
were replaced with g-leucine (compound 5), a modification re-
ported to lead to a less potent inhibitor,[36] or with g-alanine
residues (compound 6), which should further decrease the in-
hibitory activity. Such compounds should confirm the correla-
tion between antiproliferative activity and CathD inhibition.
The importance of each statine residue was also evaluated by
replacing them with a g-alanine residue (7 and 8). Finally, neg-
ative controls were included, which were previously shown to
be non-cytotoxic.[30] Compound 18 is composed of the hydro-
philic and pepstatin moieties, but is unable to efficiently pene-
trate cells. Compound 19 is composed of the hydrophilic and
AMPA4 moieties and does not inhibit CathD.
Table 1. Structures of bioconjugates 1–19.
Compd
Structure
Compd
Structure
1[a]
2
3
4
5
6
7
8
Ac-arg-O2Oc-(Val)3-Sta-Ala-Sta-(AMPA)4-NH2
Ac-arg-O2Oc-(Val)3-Sta-Ala-Sta-(AMPA)2-NH2
Ac-arg-O2Oc-(Val)3-Sta-Ala-Sta-(AMPA)3-NH2
Ac-arg-O2Oc-(Val)3-Sta-Ala-Sta-(AMPA)5-NH2
Ac-arg-O2Oc-(Val)3-gLeu-Ala-gLeu-(AMPA)4-NH2
Ac-arg-O2Oc-(Val)3-gAla-Ala-gAla-(AMPA)4-NH2
Ac-arg-O2Oc-(Val)3-Sta-Ala-gAla-(AMPA)4-NH2
Ac-arg-O2Oc-(Val)3-gAla-Ala-Sta-(AMPA)4-NH2
Ac-(AMPA)4-(Val)3-Sta-Ala-Sta-O2Oc-arg-NH2
Ac-arg-O2Oc-(AMPA)4-(Val)3-Sta-Ala-Sta-NH2
11
12
13
14
15
16
17
18
19
Ac-arg-(Val)3-Sta-Ala-Sta-O2Oc-(AMPA)4-NH2
Iva-(Val)2-Sta-Ala-Sta-penetratin-NH2
Iva-(Val)2-Sta-Ala-Sta-(Arg)8-NH2
Ac-arg-O2Oc-(Val)3-Sta-Ala-Sta-(DBT)4-NH2
Ac-arg-O2Oc-(Val)3-Sta-Ala-Sta-(LBD)4-NH2
Ac-arg-O2Oc-(Val)3-Sta-Ala-Sta-(mAMPA)4-NH2
Ac-arg-O2Oc-(Val)3-Sta-Ala-Sta-(pAMPA)4-NH2
Ac-arg-O2Oc-(Val)3-Sta-Ala-Sta-NH2
9
10
Ac-arg-O2Oc-(AMPA)4-NH2
[a] Bioconjugate 1=JMV4463.
ChemMedChem 2016, 11, 302 – 308
www.chemmedchem.org
303
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
All pepstatin conjugates containing a CPNP vector should
be stable toward CathD, and no cleavage of either the vector
or the hydrophilic moiety is expected before interaction of
pepstatin with CathD. Indeed, the amide links between pepsta-
tin and each surrounding moiety involve at least one non-a-
amino acid residue. These conjugates should also be highly
stable in cells toward any peptidase, although the Val3 seg-
ment of pepstatin moiety could be considered as a potential
cleavage site. All compounds were synthesized by standard
solid-phase synthesis following the Fmoc strategy and using
HBTU/DIEA for coupling, as previously described.[30]
cine residues in place of the two statines showed a lower in-
hibitory activity (~63% at 10 mm), while compound 6, with two
g-alanine residues, exhibited almost no inhibition at the same
concentration. Replacing only one statine residue by g-alanine
led to compounds 7 and 8 with inhibitory potencies similar to
that of pepstatin A when tested at 10 mm. However, only com-
pound 7 retained high inhibitory potency at 10 nm, similar to
all pesptatin-containing compounds. These results indicate the
important role of the first statine residue in the sequence of
pepstatin for potent CathD inhibition. They are in agreement
with previously published data concerning pepstatin A.[37] Fi-
nally, IC50 values were determined for representative com-
pounds (1, 5–7, 9, and 18) and were found to correlate with
the values at 10 mm and 10 nm.
Inhibitory potency against purified CathD
We first assessed the inhibitory potency of the bioconjugates
against purified CathD (Table 2). All compounds containing un-
modified pepstatin (1–4, 9–18) showed inhibitory activity simi-
lar to that of pepstatin A, as they exhibited comparable inhibi-
tion percentages at 10 mm (>90%) and 10 nm (>85%), indicat-
ing that the presence of the vectors and the hydrophilic
moiety do not interfere with pepstatin recognition by the
enzyme. As hypothesized, compound 5 containing two g-leu-
Capacity to penetrate cells: measurement of intracellular
CathD activity
The ability of the compounds to penetrate cells was assessed
with MDA-MB-231 tumor cells. After 24 h incubation at a con-
centration of 10 mm, the cells were washed, lysed, and super-
natants were acidified and submitted to enzymatic tests to
measure the levels of CathD inhibition (Table 2). The reported
inhibition is only related to the ability of conjugates to pene-
trate cells, but does not indicate in which cell compartment
they are localized. When comparing compounds 1–4, which
differ in the number of AMPA units present, only compounds
1 and 4 were found to potently inhibit CathD proteolytic activ-
ity. This indicates that a minimum of four AMPA units is re-
quired to enable cell penetration. As expected, compound 5,
with two g-leucine residues, showed low intracellular CathD in-
hibitory activity (~28%) in accordance with its low activity
against isolated CathD, whereas compound 6, with g-alanine
residues was found to be totally inactive. Compounds 7 and 8,
containing only one statine residue, were equally potent to
1 in inhibiting intracellular CathD activity. It is important to
note that compound 8 showed >90% inhibition on living
cells, although it was less potent than compounds 1 and 7
toward purified CathD at 10 nm. This suggests that the biocon-
jugate efficiently penetrates, allowing sufficiently high intracel-
lular concentrations to be reached.
Table 2. Inhibition of proteolytic CathD activity by bioconjugates 1–19.[a]
Compd
IC50 [nm][b]
Purified CathD
Inhib. [%][b]
Cell CathD
Inhib. [%][c]
10 mm
10 nm
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
2.30Æ0.09
95Æ2*
89Æ7*
84Æ5*
81Æ3*
63Æ2*
13Æ5
93Æ1*
89Æ1*
90Æ1*
86Æ7*
22Æ2
7Æ4
94Æ4*
9Æ19
2Æ9
–
–
–
85Æ8*
28Æ11
1Æ9
41.00Æ1.76
>10000
4.40Æ0.10
94Æ2*
92Æ1*
91Æ1*
94Æ5*
91Æ1*
90Æ2*
93Æ2*
92Æ2*
94Æ1*
92Æ1*
92Æ1*
92Æ3*
1Æ6
91Æ2*
15Æ8
89Æ1*
81Æ7*
88Æ1*
85Æ4*
80Æ2*
88Æ1*
90Æ1*
90Æ1*
86Æ1*
87Æ1*
1Æ4
91Æ4*
96Æ4*
97Æ2*
8Æ8
–
1.10Æ0.03
–
–
–
–
–
–
–
–
95Æ3*
91Æ7*
85Æ2*
98Æ1*
78Æ2*
98Æ3*
93Æ4*
8Æ14
6Æ13
10Æ10*
0Æ5
Overall, it is notable that the efficiency of AMPA4 to deliver
pepstatin was extremely high, as it allows a near total inhibi-
tion of the elevated CathD concentration (ꢀ50 pmolmgÀ1 cel-
lular proteins) present in these cancer cells. Compounds 9–11,
which differ from 1 by the order of their functional moieties
(hydrophilic moiety, pepstatin, AMPA4) gave disparate results.
The presence of the vector at the N terminus (9) or at the
C terminus (11) induced intracellular CathD inhibition similar to
that of 1. In the case of compound 11, this result shows that it
is possible to introduce a spacer between the vector and the
enzyme inhibitor while maintaining intracellular delivery. How-
ever, insertion of the vector between the hydrophilic moiety
and pepstatin (in 10) abolished the activity, suggesting the in-
ability of this bioconjugate to penetrate cells. We have shown
that JMV4463 (1) penetrates cells via the endolysosomal path-
way,[30] but its mode of interaction with membranes and of in-
17
18
19
2.80Æ0.16
–
–
–
pepstatin A[d]
DMSO
95Æ1*
0Æ9
–
0Æ8
[a] The enzymatic activity of CathD was determined by FRET using a pep-
tide substrate coupled to fluorophore (EDANS) and quencher
a
a
(DABCYL), measured at l 355 nm in a microtiter plate reader after 90 min
incubation at 378C with 5 mm substrate at pH 3.5. Percent CathD inhibi-
tion is reported for bioconjugates 1–17 together with 18, 19, and 1%
DMSO as negative controls. [b] Inhibition of CathD activity determined by
incubation of isolated CathD (250 ng) with test compounds at either
10 mm or 10 nm final concentration in cell-free conditions; a 50–1 nm con-
centration range was used for IC50 determination. [c] Inhibition of intracel-
lular CathD: after incubation of MDA-MB-231 cells for 24 h with test com-
pounds at 10 mm, cellular CathD activity was determined in 50 mg cellular
proteins. Values are the meanÆSD of three separate experiments; *p<
0.05 relative to DMSO control (Student t-test). [d] Data from Ref. [30].
ChemMedChem 2016, 11, 302 – 308
www.chemmedchem.org
304
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
ternalization has not yet been determined. It can be hypothe-
sized that the highly hydrophobic AMPA4 vector requires loca-
tion at one end of the bioconjugate for efficient internalization,
whereas the presence of the two neighboring moieties on
both sides in compound 10 would prevent cell membrane in-
teraction.
enter cells (Table 2), the results highlight that their internaliza-
tion does not necessarily lead to cytotoxicity.
Conjugates containing unmodified pepstatin or pepstatin
with only one statine residue together with an AMPAn (n=4,5)
vector located at their C terminus (4, 7, 8, and 11) exhibited cy-
totoxicity similar to that of JMV4463 (compound 1). It is inter-
esting to note that it was possible to introduce a spacer be-
tween the CathD inhibitor and the vector (compound 11) with-
out loss of activity and that only one statine residue (com-
pounds 7 and 8) was sufficient to induce cell toxicity. The
latter result was further confirmed by comparing the activities
of compounds 7 and 8 with that of 1 in a concentration–re-
sponse experiment, which showed a very good superimposi-
tion of the three cytotoxicity curves (Figure 4).
Finally, the ability of other vectors to translocate pepstatin
into cells was evaluated. As expected, the CPPs penetratin (12)
and polyarginine Arg8 (13) enabled pepstatin to enter cells as
indicated by the strong inhibition of intracellular CathD mea-
sured after cell lysis. In addition, the two CPNP vectors DBT4
(14) and LBD4 (15) also led to intracellular inhibition. The latter
was found to be slightly less efficient (~78% inhibition for 15
instead of >94% for compounds 1 and 14), suggesting that
replacing the sulfur atom (DBT) with an NH group (LBD) led to
lower penetration efficiency. Strikingly, bioconjugates 16 and
17 were found to be as efficient as compound 1 in inhibiting
intracellular CathD, indicating that the meta- and para-AMPA
tetramers (mAMPA4 and pAMPA4, respectively) are able to
translocate pepstatin into cells. Interestingly, the vectors LBD4,
mAMPA4 and pAMPA4 have never been reported before, and
the present results suggest that the family of CPNPs can be
further broadened.
However, changing the AMPA-based vector to a CPP (12, 13)
or a different CPNP (14, 15) fully abolished cytotoxicity. More
strikingly, isomers 16 and 17 were found to be non-cytotoxic,
showing that the ortho arrangement of the AMPA aminometh-
yl and acetyl groups is crucial. These results might be ex-
Toxicity toward cancer cells
The cellular toxicity of analogues at 10 mm was evaluated with
MDA-MB-231 breast cancer cells after an incubation period of
three days (Figure 3). As expected, bioconjugates 2, 3, 6, 10
and the negative controls 18 and 19, which were unable to in-
hibit intracellular CathD, were found to be non-cytotoxic. Con-
sistent with its intermediate CathD inhibitory activity, com-
pound 5 showed intermediate cytotoxicity. Therefore, the re-
sults obtained for compounds 1, 5 (two g-leucine residues),
and 6 (two g-alanine residues) showed that the antiprolifera-
tive effect of compound 1 (JMV4463) is related to intracellular
CathD inhibition. For all other inhibitor conjugates shown to
Figure 4. Viability of MDA-MB-231 breast cancer cells. Cells were treated
with pepstatin (control) and with compounds 1, 7, and 8 for three days, all
at 1, 3, 6, and 10 mm. Growth was evaluated by MTT assay as described in
the Experimental Section. Values are expressed as a percentage of control
cells treated with vehicle alone (1% DMSO) and are the meanÆSD of three
independent experiments.
Figure 3. Cytotoxicity of compounds 1–19 toward MDA-MB-231 breast cancer cells. Cells were treated with each compound at 10 mm for three days, and
growth was evaluated by MTT assay as described in the Experimental Section. All values are expressed as a percentage of control cells treated with vehicle
alone (1% DMSO) and are the meanÆSD of three independent experiments. The different types of structural variations are indicated at top (AMPAn: variable
AMPA oligomer length, Pepstatin: substitution of statine residues, Order: order of the three functional moieties, Vector: vectors other than AMPA4).
ChemMedChem 2016, 11, 302 – 308
www.chemmedchem.org
305
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
plained by an insufficient addressing of these conjugates to
the CathD-containing compartment and/or to a rapid externali-
zation from endosomes following endocytosis. In the case of
the DBT4-containing bioconjugate 14, its inactivity on cell
growth was unexpected, as the DBT vector was shown to very
efficiently penetrate cells via the endolysosomal pathway.[33]
The reasons for this behavior are unclear, but it is clear that
AMPA4 possesses specific properties not shared by other vec-
tors. Compound 9, which presents a reverse composition rela-
tive to 1, was shown to penetrate cells (Table 2), but is not cy-
totoxic. This result indicates that the position of the AMPA4
vector is crucial, as already highlighted by the results obtained
with the non-cell-penetrating compound 10. Altogether, these
results suggest that a restrictive and specific mode of interac-
tion between CathD and its inhibitor occurs inside the cell
compartment, which could be achieved only by bioconjugates
containing the AMPA4 vector at the C terminus.
Experimental Section
Synthesis of compounds 1–19
Fmoc-protected amino acid derivatives, HBTU and Fmoc Rink
amide polystyrene resin (100–200 mesh, 0.46 mmolgÀ1) were pur-
chased from Iris Biotech. Fmoc-(3S,4S)-4-amino-3-hydroxy-5-cyclo-
hexylpentanoic acid (Fmoc-statine), Fmoc-gLeu-OH and Fmoc-gAla-
OH were purchased from NeoMPS (Strasbourg, France). Other re-
agents and solvents were purchased from Riedel-de-Han, Carlo
Erba or Acros organics and used without purification. Solvents
used for HPLC and LC–MS were of HPLC grade. All final com-
pounds were purified by reversed-phase HPLC, and the purity as-
sessed by analytical reversed-phase HPLC was found to be >95%.
Fmoc-AMPA-OH and Fmoc-DBT-OH were prepared as previously
described.[38,39] The synthesis of Fmoc-LBD-OH (compound 25) is
described in the Supporting Information.
Anchoring on Rink amide PS resin. Fmoc-Rink amide resin was
conditioned for 30 min in DMF and submitted to the standard de-
protection cycle, using DMF/piperidine (80:20 v/v) solution for
30 min. After washing steps, the first residue was loaded onto the
resin through a standard coupling cycle.
Conclusions
This structure–activity study points to a major role of the cata-
lytic activity of intracellular CathD in the proliferation of breast
cancer cells and to the critical importance of the AMPA4 vector.
In addition, we found it possible to remove one of the two sta-
tine residues of the pepstatin moiety, allowing a more eco-
nomical chemical synthesis.
Coupling. The coupling reaction was carried out manually in plas-
tic syringes equipped with frits. HBTU (3 equiv) as coupling agent,
DIEA (3 equiv) as base, and Fmoc-AA-OH (3 equiv) (according to
resin loading) were dissolved in DMF and added to the resin. The
reaction was stirred for 3 h at room temperature. The reaction was
monitored by the standard TNBS test.[40]
Deprotection. Fmoc group removal of Fmoc-protected amino
acids, except Fmoc-AMPA-OH, was carried out using DMF/piperi-
dine (80:20 v/v) solution twice for 30 min. Fmoc group removal of
Fmoc-AMPA-OH was carried out using a solution of DMF/piperi-
dine/DBU (92:4:4 v/v/v) twice for 30 min.
Concerning the pepstatin moiety of the conjugate, its rele-
vance in antiproliferative activity is clearly evidenced by the
loss of activity resulting from substitution of both statine resi-
dues with non-hydroxylated g-amino acids. This lends signifi-
cant support to previous studies, indicating that the mitogenic
activity of CathD is at least partially due to its catalytic activity
either by activation of growth factor precursors or by inactiva-
tion of growth inhibitors.[14–16] JMV4463 is an inhibitor of CathD
catalytic activity with high clinical potential and is also a highly
valuable tool to explore the mechanisms of cancer cell prolifer-
ation involving intracellular CathD.
Washing steps were performed after each coupling and deprotec-
tion step. Once with MeOH, once with CH2Cl2, three times with
DMF, and three times again with CH2Cl2.
Cleavage of Rink amide PS resin. Bioconjugates were simultane-
ously cleaved from resin and deprotected for 2 h in a mixture of
TFA, TIS, and H2O (95:2.5:2.5 v/v/v). After removal of the resin by
filtration, the TFA was concentrated in vacuo. Compounds were
precipitated by the addition of Et2O and filtered. They were dis-
solved in a solution of CH3CN/H2O (50:50 v/v) containing 0.1% TFA
and freeze dried.
The translocation of pepstatin bioconjugates into cells was
enabled by all selected vectors, but their antiproliferative activ-
ity was strictly dependent on the structure of the vector and
its position. It is clear that AMPA4 possesses a unique mode of
action, the mechanism of which is not yet understood. This
finding suggests that the structure of AMPA4 imparts the bio-
conjugate with the capacity to prevent the proteolysis of criti-
cal substrates involved in cell proliferation. CathD is considered
a multifunctional enzyme, and its precursor pro-CathD under-
goes several steps of maturation during cell trafficking, and
several binding partners have also been identified.[13] CathD
could perform various functions during the cell trafficking. In
addition, active but not fully matured forms of CathD, such as
the 48 kDa form, are present in endosomes and they probably
contribute to these functions. We therefore hypothesize that
thanks to the AMPA4 vector, JMV4463 can more efficiently rec-
ognize immature forms of CathD and/or localize more efficient-
ly in the CathD-containing compartment involved in the ob-
served effect. These hypotheses are now under investigation
to understand the specific role of the vector.
Purification. All bioconjugates were purified by preparative HPLC
(Waters 4000 apparatus) carried out on a C18 reversed-phase
column (C18 Deltapak column, 100 mm”40 mm, 15 mm, 100 Š)
with a mixture of H2O+0.1% TFA and CH3CN+0.1% TFA in gradi-
ent mode at a flow rate of 50 mLminÀ1 with UV detection at
l 214 nm. HPLC and MS analytical data of the bioconjugates 1–19
are given in Supporting Information Table 1.
Biological evaluation
Cell culture and viability assay. Human breast MDA-MB-231 cancer
cell lines were purchased from ATCC (American Type Culture Col-
lection, Manassas, VA, USA) and were allowed to grow in a humidi-
fied atmosphere at 378C under 5% CO2. MDA-MB-231 cells were
cultured in Dulbecco’s modified Eagle’s medium (DMEM) supple-
mented with 10% fetal bovine serum (FBS). Cell viability was as-
sessed in 96-well plates by the MTT method after treatment of ad-
ChemMedChem 2016, 11, 302 – 308
www.chemmedchem.org
306
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[3] E. Liaudet-Coopman, M. Beaujouin, D. Derocq, M. Garcia, M. Glondu-
Lassis, V. Laurent-Matha, C. PrØbois, H. Rochefort, F. Vignon, Cancer Lett.
2006, 237, 167–179.
[4] C. Kirana, H. Shi, E. Laing, K. Hood, R. Miller, P. Bethwaite, J. Keating,
T. W. Jordan, M. Hayes, R. Stubbs, Int. J. Proteomics 2012, 2012, 245819.
[5] P. Benes, V. Vetvicka, M. Fusek, Crit. Rev. Oncol. Hematol. 2008, 68, 12–
28.
herent cells with compounds or with 1% DMSO (vehicle) for three
days. A solution of MTT (0.5 mgmLÀ1) was added to each well and
incubated for 4 h. The MTT-formazan product was dissolved in
EtOH/DMSO (50:50) and estimated by measuring absorbance at
l 540 nm in a Multiskan FC plate reader (Fisher Scientific, Illkirch,
France).
[6] F. L. Pruitt, Y. He, O. E. Franco, M. Jiang, J. M. Cates, S. W. Hayward, Pros-
tate 2013, 73, 476–488.
[7] G. Ferrandina, G. Scambia, F. Bardelli, P. Benedetti Panici, S. Mancuso, A.
Messori, Br. J. Cancer 1997, 76, 661–666.
[8] J. A. Foekens, M. P. Look, J. Bolt de Vries, M. E. Meijer van Gelder, W. L.
van Putten, J. G. Klijn, Br. J. Cancer 1999, 79, 300–307.
[9] M. E. Fukuda, Y. Iwadate, T. Machida, T. Hiwasa, Y. Nimura, Y. Nagai, M.
Takiguchi, H. Tanzawa, A. Yamaura, N. Seki, Cancer Res. 2005, 65, 5190–
5194.
[10] T. Mimae, K. Tsuta, A. M. Maeshima, M. Okada, H. Asamura, T. Kondo, H.
Tsuda, Pathol. Res. Pract. 2012, 208, 534–540.
[11] Y. Chai, W. Wu, C. Zhou, J. Zhou, Arch. Gynecol. Obstet. 2012, 286, 465–
471.
[12] P. Saftig, M. Hetman, W. Schmahl, K. Weber, L. Heine, H. Mossmann, A.
Kçster, B. Hess, M. Evers, K. von Figura, EMBO J. 1995, 14, 3599–3608.
[13] Z. Khalkhali-Ellis, M. J. C. Hendrix, Biol. Med. 2014, 6, 206.
[14] P. NirdØ, D. Derocq, M. Maynadier, M. Chambon, I. Basile, M. Gary-Bobo,
M. Garcia, Oncogene 2010, 29, 117–127.
[15] V. Laurent-Matha, P. F. Huesgen, O. Masson, D. Derocq, C. PrØbois, M.
Gary-Bobo, F. Lecaille, B. Rebi›re, G. Meurice, C. OrØar, R. E. Hollings-
worth, M. Abrahamson, G. Lalmanach, C. M. Overall, E. Liaudet-Coop-
man, FASEB J. 2012, 26, 5172–5181.
CathD proteolytic activity. To detect CathD protease activity, Fçrster
resonance energy transfer (FRET) was used. A fluorophore (EDANS)
and a quencher (DABCYL) dye were coupled to the C- and N-termi-
nal ends of a peptide substrate highly selective for CathD. On
intact peptides, the emission energy of the fluorophore was cap-
tured by the quencher. Following cleavage of the substrate, the
quencher is no longer able to absorb the fluorescent energy of the
fluorophore, and this increase in fluorescence was measured. The
FRET substrate for CathD (JMV3672, DABCYL-bAla-Arg-Pro-Ile-Leu-
Phe-Phe-Arg-Leu-Asp(EDANS)-NH2) was not commercially available
and was synthesized according to standard procedures. The pep-
tide sequence used was reported to be cleaved by CathD, but not
by other cathepsins.[41] In the first case, purified recombinant
human CathD (250 ng) was incubated in 0.1m citrate buffer
(pH 3.5) with 10 mm and 10 nm compounds as described.[41] In the
second case, the compounds were incubated for 24 h with living
cells, then cells were washed with phosphate-buffered saline (PBS),
lysed and supernatants used for FRET analysis. The substrate (final
concentration 5 mm) was then incubated at 378C with 20 mL of cell
lysate supernatants at pH 3.5. After 90 min, the plate was read in
a PerkinElmer 1420 Victor 2 fluorescent plate reader at lex =
355 nm and lem =538 nm. The background fluorescence of the
FRET substrate was later subtracted out. Results were presented as
percent of inhibition of CathD activity.
[16] T. Braulke, M. Claussen, P. Saftig, M. Wendland, K. Neifer, B. Schmidt, J.
Zapf, K. von Figura, C. Peters, Prog. Growth Factor Res. 1995, 6, 265–
271.
[17] C. A. Conover, D. D. De Leon, J. Biol. Chem. 1994, 269, 7076–7080.
[18] L. Hasan, L. Mazzucchelli, M. Liebi, M. Lis, R. E. Hunger, A. Tester, C. M.
Overall, M. Wolf, J. Immunol. 2006, 176, 6512–6522.
Abbreviations. AMPA: ortho-aminomethylphenylacetyl, CathD: ca-
thepsin D, CPNP: cell-penetrating non-peptide, CPP: cell-penetrat-
ing peptide, DABCYL: 4-([4-(dimethylamino)phenyl]azo)benzoyl,
DBT: d-benzothiazepine, DBU: 1,8-diazabicyclo[5.4.0]undec-7-ene,
DIEA: N,N-diisopropylethylamine, DMF: N,N-dimethylformamide,
DMSO: dimethyl sulfoxide, EDANS: 5-([2-aminoethyl]amino)naph-
thalene-1-sulfonyl, Fmoc: 9-fluorenylmethyloxycarbonyl, HBTU: N-
[(1H-benzotriazol-1-yloxy)(dimethylamino)methylene]-N-methylme-
thanaminium hexafluorophosphate, LBD: l-benzodiazepine, MTT:
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, O2Oc:
8-amino-3,6-dioxaoctanoyl, PS: polystyrene, TFA: trifluoroacetic
acid, TIS: triisopropylsilane, TNBS: trinitrobenzenesulfonic acid.
[19] T. Berg, T. Gjøen, O. Bakke, Biochem. J. 1995, 307, 313–326.
[20] M. Garcia, N. Platet, E. Liaudet, V. Laurent, D. Derocq, J. P. Brouillet, H.
Rochefort, Stem Cells 1996, 14, 642–650.
[21] I. M. Melzer, S. B. Fernµndez, S. Bçsser, K. Lohrig, U. Lewandrowski, D.
Wolters, S. Kehrloesser, M. L. Brezniceanu, A. C. Theos, P. M. Irusta, F.
Impens, K. Gevaert, M. Zçrnig, Cell Death Differ. 2012, 19, 1435–1445.
[22] Y.-S. Hah, H. S. Noh, J. H. Ha, J. S. Ahn, J. R. Hahm, H. Y. Cho, D. R. Kim,
Cancer Lett. 2012, 323, 208–214.
[23] P. Montcourrier, P. H. Mangeat, G. Salazar, M. Morisset, A. Sahuquet, H.
Rochefort, Cancer Res. 1990, 50, 6045–6054.
[24] H. Wallin, M. Abrahamson, U. Ekstrçm, J. Biol. Chem. 2013, 288, 17019–
17029.
[25] M. Glondu, E. Liaudet-Coopman, D. Derocq, N. Platet, H. Rochefort, M.
Garcia, Oncogene 2002, 21, 5127–5134.
[26] O. Masson, A. S. Bach, D. Derocq, C. PrØbois, V. Laurent-Matha, S. Pattin-
gre, E. Liaudet-Coopman, Biochimie 2010, 92, 1635–1643.
[27] Y. Liu, Y. Zhou, K. Zhu, PLoS ONE 2012, 7, e45910.
[28] M. Maynadier, R. Farnoud, P. J. Lamy, V. Laurent-Matha, M. Garcia, H. Ro-
chefort, Int. J. Oncol. 2013, 43, 1683–1690.
[29] O. Achour, N. Bridiau, M. Kacem, R. Delatouche, S. Bordenave-Juchereau,
F. Sannier, V. ThiØry, J. M. Piot, T. Maugard, I. Arnaudin, Biochimie 2013,
95, 2010–2017.
Acknowledgements
This work was supported by the CNRS, Montpellier University, the
Fondation pour la Recherche MØdicale (FRM DCM20121225739),
and by the RØgion Languedoc-Roussillon (“Chercheur d’avenir”
2011 grant to V.L.). We also thank Pierre Sanchez (Institut des Bio-
molØcules Max Mousseron, Montpellier) for mass spectrometry
analyses.
[30] M. Maynadier, L. L. Vezenkov, M. Amblard, V. Martin, C. Gandreuil, O.
Vaillant, M. Gary-Bobo, I. Basile, J.-F. Hernandez, M. Garcia, J. Martinez, J.
Controlled Release 2013, 171, 251–257.
[31] H. Umezawa, T. Aoyagi, H. Morishima, M. Matsuzaki, M. Hamada, J. Anti-
biot. 1970, 23, 259–262.
Keywords: antiproliferation · cathepsin D · cell-penetrating
[32] C. G. Knight, A. J. Barrett, Biochem. J. 1976, 155, 117–125.
[33] L. L. Vezenkov, M. Maynadier, J.-F. Hernandez, M.-C. Averlant-Petit, O.
Fabre, E. Benedetti, M. Garcia, J. Martinez, M. Amblard, Bioconjugate
Chem. 2010, 21, 1850–1854.
non-peptides
relationships
·
pepstatin delivery
·
structure–activity
[34] A. Joliot, A. Prochiantz, Nat. Cell Biol. 2004, 6, 189–196.
[35] P. A. Wender, D. J. Mitchell, K. Pattabiraman, E. T. Pelkey, L. Steinman,
J. B. Rothbard, Proc. Natl. Acad. Sci. USA 2000, 97, 13003–13008.
[36] D. H. Rich, E. Sun, J. Singh, Biochem. Biophys. Res. Commun. 1977, 74,
762–767.
[1] H. Rochefort, F. Capony, M. Garcia, V. Cavaill›s, G. Freiss, M. Chambon,
M. Morisset, F. Vignon, J. Cell. Biochem. 1987, 35, 17–29.
[2] I. Bartenjev, Z. Rudolf, B. Stabuc, I. Vrhovec, T. Perkovic, A. Kansky, Int. J.
Dermatol. 2000, 39, 599–602.
ChemMedChem 2016, 11, 302 – 308
www.chemmedchem.org
307
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[37] P. Majer, J. R. Collins, S. V. Gulnik, J. W. Erickson, Protein Sci. 1997, 6,
1458–1466.
[41] P. E. Scarborough, K. Guruprasad, C. Topham, G. R. Richo, G. E. Conner,
T. L. Blundell, B. M. Dunn, Protein Sci. 1993, 2, 264–276.
[38] M. Amblard, M. Calm›s, V. Roques, S. Tabet, A. Loffet, J. Martinez, Org.
Prep. Proced. Int. 2002, 34, 405–415.
[39] N. Raynal, M.-C. Averlant-Petit, G. Berge, C. Didierjean, M. Marraud, C.
Duru, J. Martinez, M. Amblard, Tetrahedron Lett. 2007, 48, 1787–1790.
[40] W. S. Hancock, J. E. Battersby, D. R. K. Harding, Anal. Biochem. 1975, 69,
497–503.
Received: October 6, 2015
Revised: November 19, 2015
Published online on December 7, 2015
ChemMedChem 2016, 11, 302 – 308
www.chemmedchem.org
308
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim