Tumor-Targeted Delivery of an EGFR Inhibitor Prodrug via Site-Specific Albumin Conjugation.

Introduction
A
major challenge in anticancer therapy is the limited tumor specificity
of many drugs, which often leads to severe adverse effects. Human serum albumin (HSA) has long been identified
as a very promising tool for improved drug delivery and tuning of
pharmacokinetic properties. With a half-life of ∼19 days, a
plasma concentration of 35–50 g/L (∼640 μM), and
several hydrophobic binding pockets, HSA serves as a natural transporter
of diverse endogenous ligands and drugs, significantly prolonging
their circulation.
−

The remarkably long half-life of HSA results from
binding to the neonatal fragment crystallizable (Fc) receptor after
endocytosis, effectively protecting it from lysosomal degradation. Moreover, albumin accumulates in solid tumors
through both active and passive targeting mechanisms. First, it is
actively taken up by cancer cells via endocytosis, serving as a nutrient
source. Second, albumin passively targets
tumors through the enhanced permeability and retention (EPR) effect,
entering malignant tissue through leaky vasculature and accumulating
due to impaired lymphatic drainage.
−

Emphasizing this, our
recent study demonstrated albumin accumulation in tumors using a radioactive
[89Zr]­Zr-DFO* complex covalently conjugated to HSA via
a maleimide linker. The use of a rather
long-lived radioisotope in form of a highly stable HSA-conjugate yielded
exceptionally high tumor-to-tissue ratios, visualizing the remarkable
efficiency of albumin-mediated tumor targeting. Some oncological drugs
already make use of albumin for drug delivery, including US Food and
Drug Administration (FDA)-approved Abraxane, a nanoparticle albumin-bound
formulation of paclitaxel.
,
Another example is
the experimental drug Aldoxorubicin, a maleimide-functionalized derivative
of doxorubicin that has advanced to phase 3 clinical trials (ClinicalTrials.gov number,
NCT02049905). Following intravenous administration, it selectively
forms a covalent bond with the Cys34 residue of endogenous
albumin.
,,

Tyrosine
kinase inhibitors (TKIs) are targeted therapeutics and
represent one of the most prolific areas in the development of novel
anticancer drugs. TKIs bind to the adenosine triphosphate (ATP)-binding
pocket of receptor tyrosine kinases (RTKs), disrupting intracellular
downstream signaling and inducing apoptosis.
,
As of 2025, the FDA has approved 75 protein kinase inhibitors to
treat neoplasms, including 45 receptor TKIs. As TKIs are directed toward molecular targets involved in cancer
cell proliferation, an increased selectivity for cancerous over healthy
cells is frequently observed.
,
However, as RTK-induced
signaling is upregulated but not confined to malignant cells, severe adverse effects remain an issue. Some
of the most common ones are cardiovascular events (e.g., hypertension,
heart failure, etc.), gastrointestinal symptoms (e.g., nausea, diarrhea,
etc.) and skin rash.
,

Many approved TKIs are
well-known to electrostatically bind to
albumin, often reaching >90% protein binding in serum.
−

Yet, to the best of our knowledge, there is no direct evidence that
this interaction contributes to tumor targeting. In contrast, several
studies even suggest that due to this extensive protein binding, the
low levels of unbound drug in serum can limit their clinical activity.
−

This widely accepted principle, known as the Free Drug Hypothesis,
states that only the unbound or free fraction of a drug is responsible
for the activity at the therapeutic target site.
,
Of note, also TKI toxicity is influenced by their interaction with
albumin. Numerous studies demonstrated that patients with hypoalbuminemia
show significantly more adverse events, leading to early treatment
discontinuation. This implies that elevated
levels of free drug can cause severe toxicity and that albumin binding
acts like a protective reservoir at the expense of decreased drug
activity.
Therefore, we envisioned that the activity of TKIs
could be improved
if the drug is bound to albumin covalently via a maleimide linker
and can be selectively released upon activation in the tumor tissue
(Figure
).
While the benefit of albumin binding is well established for small
molecular chemotherapeutic agents, to the best of our knowledge, such
approaches have not yet been reported for TKIs. To enable this tumor-specific
drug release, we selected the well-established cathepsin B-cleavable
dipeptide valine-citrulline (ValCit) with a para-aminobenzyl
carbonyl (PABC) spacer,
,
which is part of several
approved antibody drug conjugates (ADCs) for anticancer therapy. Cathepsin B is a lysosomal cysteine protease
that is overexpressed in a variety of cancers. Moreover, malignant cells frequently secrete the enzyme
into the extracellular space, generating a proteolytic tumor microenvironment.
,
As a model TKI, we selected a derivative of the third generation
epidermal growth factor receptor (EGFR) inhibitor osimertinib, which
is approved for the first-line therapy of EGFR mutation-positive non-small
cell lung cancer (NSCLC) (Figure
A).

Altogether, the system comprises (1) a maleimide
moiety (Mal) for
binding to endogenous albumin after intravenous administration, (2)
a piperazine (Pip) linker motif for enhanced aqueous solubility, (3)
a ValCit dipeptide fragment as the cathepsin B-cleavable linker, (4)
the PABC spacer to enable enzyme accessibility and (5) the EGFR inhibitor,
connected via a carbamate (Figure
B). Cleavage of the dipeptide C-terminus finally initiates
the self-immolation of the PABC fragment, releasing CO2 and the free osimertinib derivative (Figure
B). As a first
step, we synthesized and evaluated several derivatives of osimertinib
with a modified dimethylamino group moiety (NMe2) to enable
conjugation of the EGFR inhibitor. Based on docking experiments, the
desired target compound Mal-Pip-ValCit (Figure
B) was synthesized together
with a noncleavable reference containing a glycine-glycine (GlyGly)
dipeptide motif. For both drugs, kinase screening and cell culture
experiments confirmed the prodrug-nature of the new compounds. Subsequently,
cathepsin B cleavage and albumin-binding studies were performed. Finally,
we demonstrated promising in vivo anticancer activity
in a cathepsin B-overexpressing, EGFR-dependent xenograft mouse model.

Results
and Discussion

Synthesis of Osimertinib Derivatives
To enable conjugation
of osimertinib to the maleimide-ValCit fragment, three derivatives
harboring an NH amine group were synthesized according to an adapted
literature procedure (Figure
A and Scheme S1). In the first step, a nucleophilic aromatic substitution
of N-(4-fluoro-2-methoxy-5-nitrophenyl)-4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-amine with the respective amine
substituents was performed. Next, the aromatic nitro group was reduced
to an aniline using Fe and NH4Cl in EtOH/H2O,
creating only a mild pH at which the Boc-groups were stable. The aniline
was then acylated using acryloyl chloride in dichloromethane (DCM),
whereby slow addition and low temperatures (−80 °C) were
essential to avoid diacylation. Finally, acidic deprotection of the
Boc group using trifluoroacetic acid (TFA) afforded OsiNHMe, OsiNH

2
and OsiPropNHMe.

EGFR Inhibition Profile and Anticancer Activity of Osimertinib
Derivatives in Vitro and in Vivo
To investigate the impact of the modifications of osimertinib
on the EGFR-inhibitory potential, cell-free kinase inhibition assays
were performed on double-mutant EGFR (L858R/T790M) by a commercial
provider (Figure S1). The data revealed
that OsiNHMe (IC50 = 1.45 nM) and OsiNH

2
(IC50 = 1.56 nM) had similar
activity to osimertinib, which was the most active compound with an
IC50 of 1.11 nM. Interestingly, OsiPropNHMe was distinctly less active at an IC50 of 10.8 nM. To
assess whether these cell-free effects translate into biological systems,
viability assays using a panel of EGFR-dependent cell lines with known
sensitivity to osimertinib were performed (Table
and Figure S2). In more detail, the experiments were conducted in the NSCLC cell
line H1975, a well-characterized osimertinib-sensitive model carrying
the EGFR L858R/T790M double mutation; the NSCLC cell line H1650, which
harbors an EGFR exon 19 deletion (E746-A750del); the NSCLC cell line
HCC827, which carries the same EGFR exon 19 deletion and is known
for high EGFR overexpression; and the epidermoid carcinoma cell line
A431, which overexpresses wild-type EGFR. Cells were exposed to increasing
concentrations of the drugs for 72 h and their viability was measured
by MTT assay. In these experiments, all compounds showed efficacy
similar or even slightly greater than that of osimertinib. Of note,
our generated IC50 values of osimertinib in H1975 cells
differed from literature values. However,
when looking at the exact shape of the dose response curves, in both
studies an extended plateau exactly in the IC50 range can
be observed. Thus, we hypothesize that already small variations in
the assay protocols (e.g., used cell number, assay incubation time
or calculation method) could distinctly impact the IC50 values. Consequently, we repeated the viability experiments in this
cell line with MTT measurements after different development times.
Indeed, incubation with the MTT solution only for 30 min resulted
in a ∼10-fold drop of the IC50 values: osimertinib
(0.3 μM), OsiNHMe (0.2 μM) and OsiNH

2
(0.2 μM) (Figure S3). Only in case of OsiPropNHMe the values remained
largely unchanged. Overall, this effect can probably be associated
with the rather cytostatic than cytotoxic activity of EGFR inhibition.
Consequently, we decided to check, how the in vitro antitumor activity translates into the in vivo situation.
To this end, therapy experiments on H1975 xenografts in C.B.-17SCID/SCID mice were performed (Figure
). All drugs were administered via oral gavage
once daily, five times a week, for 2 weeks at doses equimolar to 29.8
mg/kg osimertinib mesylate, which corresponds to 25 mg/kg osimertinib,
the reported maximum tolerated dose (MTD).

Treatment revealed that the tumor
growth-inhibiting activity of
both OsiNHMe and OsiNH

2
was similar to osimertinib, while OsiPropNHMe was distinctly less effective. These data of OsiNHMe are in good agreement with the literature, reporting that this metabolite
of osimertinib retains a similar potency and selectivity profile. Moreover, these experiments revealed that both OsiNHMe and OsiNH

2
could
be potential candidates for prodrug strategies, while OsiPropNHMe does not have sufficient EGFR-targeting properties. This is very
surprising, considering that only one additional CH2 unit
was introduced in the region extending into the solvent front, while
the moiety interacting with the binding site remained unchanged. Thus,
we selected OsiNHMe as the structurally closest osimertinib
derivative for the synthesis of the prodrugs (see below).

Molecular Docking
and Dynamics Studies
The intended
prodrug system for OsiNHMe comprises the PABC spacer,
the ValCit dipeptide as the trigger unit for cathepsin B and the maleimide
moiety for albumin binding (Figure
B) with a noncleavable GlyGly derivative as a reference.
To evaluate the viability of the prodrug concept, in a first step
in silico molecular docking studies of Mal-Pip-ValCit and Mal-Pip-GlyGly with the EGFR were performed. Building
on our previous investigation of an EGFR-inhibiting oxaliplatin­(IV)
complex, the double mutant L858R/T790M
EGFR kinase domain (PDB ID: 5CAS) was used. As a reference,
osimertinib was docked into the EGFR-binding site, where it adopted
a pose consistent with that of the cocrystallized ligand (Figure S4A), and achieved a comparable docking
score (Table
). Also, OsiNHMe fitted very well into the EGFR-binding pocket, with
its 1-methyl-3-pyrimidin-indole pharmacophore engaging the active
site similarly to osimertinib. Notably, the third-best pose, as well
as most of the top ten ranked poses, closely superimpose with that
of osimertinib (Figure
A). Although the top two do not perfectly overlap, they remain within
the binding pocket and show comparable docking scores (Table
) making them equally plausible
binding modes. Conjugation of osimertinib to the peptide-PABC moiety
significantly altered the binding behavior. The docking results clearly
showed that both Mal-Pip-ValCit and Mal-Pip-GlyGly (Figure
B,C) prodrugs
failed to position the 1-methyl-3-pyrimidin-indole scaffold within
the EGFR-binding pocket. Instead, these bulky conjugates adopted poses
outside the active site, and their docking scores are markedly lower
than those of osimertinib and OsiNHMe, indicating a substantial
reduction in binding affinity (Table
). Among the top ten poses, only one for Mal-Pip-ValCit (2nd-best pose) and two for Mal-Pip-GlyGly (7th and
8th-best poses) position the pharmacophore within the pocket, and
even in these cases, the docking scores are low. Molecular dynamics
(MD) simulations were performed to evaluate the stability of the docked
conformations of the prodrugs relative to OsiNHMe. The
root-mean-square deviation (RMSD, Å) profiles (Figure S4B–D) indicate that, over the 100 ns simulation, OsiNHMe retains its docking pose with minimal fluctuation
(<0.8 Å), consistent with a stable ligand–receptor
interaction. In contrast, Mal-Pip-ValCit and Mal-Pip-GlyGly exhibit larger positional deviations (approximately 4 Å), consistent
with weaker binding affinity. These compounds remain with the pharmacophore
largely displaced from the active site, instead sampling the outer
protein surface. Overall, these findings suggest that the prodrug
strategy could effectively mask their ability to bind to the EGFR.

Synthesis of the Albumin-Binding, Cathepsin B-Cleavable EGFR-Inhibitor
Prodrugs
The target compounds Mal-Pip-ValCit and Mal-Pip-GlyGly were synthesized in nine steps via
a universal precursor with a para-nitrophenol (PNP)
active ester moiety, which enabled late-stage incorporation of amine-functionalized
molecules (Scheme
). To prevent Michael addition side reactions, the maleimide was
protected through a Diels-Alder reaction with 2,5-dimethylfuran, which
can conveniently be removed by heating at ∼90 °C. For enhanced aqueous solubility, a protonatable
piperazine moiety was introduced as a linker between the maleimide
and ValCit dipeptide, as recently reported by our group. Briefly, after 2,5-dimethylfuran protection
of maleimide, alkylation with 2-bromoethyl ether to yield pMal-O-Br
was performed. Next, piperazine was introduced to generate pMal-O-Pip.
Subsequent coupling with tert-butyl 3-bromopropionate
gave pMal-O-Pip-COO
t
Bu. Finally,
t
Bu-deprotection yielded pMal-O-Pip-COOH in
28% yield over five steps (Scheme
). Notably, heating at 50 °C during the second
step reduced the endo isomer content from 16 % (observed immediately
after the Diels-Alder reaction) to 2 %, due to maleimide deprotection.
This highlights the importance of a high excess of the exoisomer (thermodynamic
product) to prevent product loss from cycloreversion of the endoisomer
(kinetic product). Next, a coupling reaction of pMal-Pip-COOH with
H2N-GlyGly-PAB-OH (synthesized in 2 steps) or commercially
available H2N-ValCit-PAB-OH was performed using the coupling
agent 1-ethyl-3-(3-dimethylaminopropyl)­carbodiimide hydrochloride
(EDC·HCl) with N-hydroxybenzotriazole (HOBt).
The benzyl alcohols were then activated using a slight excess of bis­(4-nitrophenyl)
carbonate (3 equiv) to generate the respective PNP-active esters,
which were purified via column chromatography using a MeOH/DCM mixture
as the eluent. Interestingly, during solvent removal after chromatographic
purification, the initially colorless solutions turned bright yellow,
indicating the release of PNP. Mass spectrometry confirmed the formation
of the respective methyl carbonate esters, formed by nucleophilic
substitution of the MeOH from the eluent upon DCM evaporation (Figure S5). The reaction was likely promoted
by the piperazine amines which create basic conditions similar to
those induced by other tertiary amines. However, adding high-boiling
toluene to the mixture during solvent removal prevented the concentration
of MeOH and thus suppressed the substitution reaction (Figure S6). Both PNP-derivatives were obtained
in excellent yields (∼90%) and subsequently reacted with OsiNHMe (Figure
A). Finally, maleimide deprotection was achieved by heating in DMSO
at 90 °C for 3 h, followed by purification via preparative high-performance
liquid chromatography (HPLC). The desired compounds were obtained
in good yields over the last four steps (Mal-Pip-GlyGly: 40%, Mal-Pip-ValCit: 35%).
After successful synthesis, the solubility of the target compounds
was tested in aqueous isotonic media (e.g., 0.9% NaCl or 5% glucose),
as required for intravenous (i.v.) administration in later in vivo studies. As mentioned above, in mice, the MTD of
osimertinib mesylate is 29.8 mg/kg. Therefore, the target compounds
needed to be soluble at ∼10 mM to enable equimolar dosing of
the prodrugs. Complete dissolution was easily achieved using 20 vol
% propylene glycol in a 5% glucose solution, with no precipitation
observed for >24 h.
Of note, the target compounds were also
synthesized harboring a
PEG6-linker (Scheme S2). Unfortunately,
neither compound was soluble at the required concentration of 10 mM,
even in the presence of various solubilizing additives at their maximum
tolerable concentration (20% propylene glycol or 30% PEG400). Thus, Mal-PEG-GlyGly and Mal-PEG-ValCit were omitted
from further investigations.

Stability, Kinase Inhibition and Anticancer
Activity of Mal-Pip-ValCit
and Mal-Pip-GlyGly in Vitro
First the stability
of the Mal-Pip-ValCit and Mal-Pip-GlyGly prodrugs (10 μM) was tested in 10 mM phosphate buffer (PB)
at pH 7.4 and 37 °C using analytical HPLC. Apart from the expected
maleimide hydrolysis (Figure S7), both
compounds remained stable for >25 h (Figure S8). Next, we investigated whether the new compounds are indeed
prodrugs,
unable to inhibit the EGFR (without cathepsin B activation). To this
end, the EGFR-inhibitory potential of Mal-Pip-GlyGly and Mal-Pip-ValCit was evaluated in a cell-free kinase inhibition
assay (Figure S9). The data revealed a
complete loss of EGFR-inhibitory activity with IC50 values
>10 μM, in contrast to the potent OsiNHMe activity
of 1.45 nM (Figure S1). In good agreement,
also in the viability assays using EGFR mutant H1975 cells the observed
IC50 values shifted from 3.92 μM and 2.48 μM
for free osimertinib and OsiNHMe, respectively, to >10
μM for both prodrugs after 72 h incubation (Figure
). To evaluate the effects
of the new drugs in healthy cells, we performed viability assays with
murine fibroblasts (Figure
). Noteworthy, in contrast to osimertinib, both prodrugs had
no relevant impact on the viability of these cells. Together these
data confirm that the new compounds are indeed stable prodrugs, without
any premature release of OsiNHMe, which might prevent
off-target toxicity of osimertinib in healthy tissue.

Cathepsin B Cleavage Studies
Cathepsin B cleavage studies
with Mal-Pip-ValCit and Mal-Pip-GlyGly were
performed to demonstrate enzyme accessibility and prodrug activation.
Therefore, cathepsin B from human liver was activated in a solution
of DL-dithiothreitol (DTT) and ethylenediaminetetraacetic acid (EDTA)
and subsequently incubated with the compounds in sodium acetate buffer
at pH 5 (Figure S10). Of note, the added DTT quickly reacted with the maleimide
moiety of Mal-Pip-ValCit and Mal-Pip-GlyGly in a Michael addition reaction, bridging two molecules via the DTT-thiol
groups (Figure S11). However, this process
did not influence the catalytic reaction, as after only 30 min ca.
half of Mal-Pip-ValCit was cleaved and OsiNHMe released (Figure S12). After 1.5 h the
cleavage was complete. In contrast, Mal-Pip-GlyGly was
completely stable in the presence of cathepsin B and no release of OsiNHMe was observed (Figure S13A). As additional control samples, Mal-Pip-ValCit and Mal-Pip-GlyGly were incubated with DTT and EDTA without cathepsin
B, revealing no release of OsiNHMe (Figure S13B,C). In order to evaluate the stability of the
new drugs in cell culture conditions, we tested their activity in
the cell model Caki-1 (Figure S14), which
is characterized by very high levels of lysosomal cathepsin B (Figure S15). In
contrast to osimertinib, both prodrugs had no significant activity
for up to 72 h indicating their distinct intracellular stability despite
the high lysosomal cathepsin B expression of the cells. This can presumably
be explained by the inability of the prodrugs to reach the lysosomes
due to their albumin-mediated drug uptake via endocytosis.

Albumin-Binding
Studies
To investigate the albumin-binding
properties of Mal-Pip-ValCit and Mal-Pip-GlyGly, HSA (∼300 μM) was incubated with the drugs in a 6:1
ratio in PB at pH 7.4 and 37 °C, and analyzed via HPLC. For both Mal-Pip-ValCit and Mal-Pip-GlyGly (Figure S16A,B), already after 10 min >95%
were
albumin-bound, accompanied by a significant increase in the albumin
peak area. A very small additional peak appeared, likely corresponding
to the hydrolyzed maleimide derivatives. The albumin conjugates remained
stable for >24 h (data not shown). As TKIs like osimertinib are
well-known
to bind to albumin also through electrostatic interaction,
,
an additional experiment was performed, blocking the maleimide of Mal-Pip-ValCit with N-acetylcysteine (NAC)
prior to incubation with albumin (Figure S16C). No albumin binding was detected for >2 h. This confirmed that
the rapid binding observed with the free maleimide (Figure S16A,B) is not electrostatic, since such adducts would
dissociate under HPLC conditions and remain detectable as a separate
peak.
Next, we explored the cathepsin B-mediated release of
the EGFR inhibitor from the albumin conjugate (Figure S17). Therefore, we incubated Mal-Pip-ValCit with HSA (∼300 μM) and added cathepsin B (NaOAc buffer,
pH 5) similarly to the cathepsin B cleavage assay above. To follow
the reaction, we selected 375 nm as the detection wavelength, which
is characteristic for OsiNHMe. At the 0 h time point,
∼85% of Mal-Pip-ValCit was bound to albumin and
∼10% of hydrolyzed Mal-Pip-ValCit were observed.
After 4 h of incubation with cathepsin B, ∼75% of the EGFR
inhibitor was released while ∼20% remained associated with
the albumin peak. This experiment clearly confirmed efficient drug
release from the albumin-drug conjugate.
The Cys34-conjugation of maleimide was investigated
by incubation of Mal-Pip-ValCit, Mal-Pip-GlyGly and OsiNHMe with albumin at 37 °C and the remaining
free thiol content was determined via UV–vis spectrophotometry
using 2,2′-dithiodipyridine (DTDP). The initial levels of free Cys34 in the used HSA solution
was 33 ± 3%, which was subsequently normalized to 100%. Coincubation
of Mal-Pip-GlyGly or Mal-Pip-ValCit gradually
reduced the free thiol content (Figure
). In contrast, OsiNHMe, without a maleimide
moiety, did not affect the quantity of free Cys. However, the binding at Cys34 was not quantitative
with 61% Mal-Pip-GlyGly and 53% Mal-Pip-ValCit bound at a 1:1 ratio. In general, the maleimide-Cys34 interaction is considered rather fast. Therefore, the binding kinetics of Mal-Pip-GlyGly to
Cys34 was separately investigated at a 1:1 ratio, revealing
that the molecule was already bound to Cys34 after 20 min
(60%) and did not increase considerably (62%) after a further 70 min
of incubation (Figure S18).
Given the discrepancy between the efficient and
covalent albumin
binding of the prodrug (Figure S16) and
the high amount of free thiol (∼40%) of albumin after equimolar
incubation with Mal-Pip-ValCit (Figure
), most likely nucleophilic
groups other than Cys34 are involved. We tested this theory
by repeating the DTDP assay (Figure
) with Mal-Pip-ValCit at pH 6 instead
of pH 7.4, which further protonates lysine residues and prevents them
from engaging in aza-Michael addition reactions (with maleimide).
Indeed, ∼20% higher binding to the albumin-SH was observed,
indicating an involvement of lysines in covalent maleimide binding
to albumin at pH 7.4 (data not shown). Of note, Cys34 is
situated in a cleft in subdomain IA of HSA, which could hinder its
fast interaction with large molecules. However, HSA harbors several
hydrophobic binding pockets. Thus, the
compound may partially bind electrostatically to one of the other
albumin pockets, where it can react with lysine residues rather than
reaching the Cys34 pocket.

Serum Extraction of Osimertinib
and OsiNHMe
To determine
the stability of the prodrug in circulation, the common strategy is
the quantification of the free TKI in serum. Therefore, both osimertinib
(in accordance with existing protocols) and OsiNHMe were spiked into mouse serum, precipitated
with an excess of acetonitrile and after centrifugation the supernatant
was analyzed via HPLC. The first time point immediately after sample
preparation revealed high recovery of the EGFR inhibitors (>95%).
However, extended incubation times for up to 24 h at 37 °C resulted
in a distinct progressive decrease of the extracted concentrations
of both osimertinib and OsiNHMe (Figure
A,B). Also, when using a pure HSA solution
(∼600 μM in PB) a similar behavior was observed, indicating
fast covalent albumin binding (Figure
A,B). Notably, covalent conjugation of OsiNHMe to albumin was not detected under cathepsin B cleavage conditions
at ∼pH 5 (data not shown). In mouse serum the recovery of both
osimertinib and OsiNHMe was even less than 50% after
2 h of incubation and only ∼10% after 24 h. To confirm the
nature of this interaction, 50 μM osimertinib was incubated
with albumin (∼1.5 mM) at 37 °C for 24 h, purified using
size-exclusion chromatography and subsequently analyzed by mass spectrometry
(Figure
C). The data
clearly revealed covalent conjugation of osimertinib to HSA.
These data are well in agreement with recent literature reports,
showing that osimertinib covalently conjugates to different lysine
residues of albumin and that this reaction
is irreversible. Unfortunately, this
prevented quantitative serum stability studies or detailed pharmacokinetics
of the prodrugs Mal-Pip-ValCit and Mal-Pip-GlyGly, as the released inhibitor binds covalently to albumin.

Tolerability
and anticancer activity of Mal-Pip-ValCit and Mal-Pip-GlyGly in Vivo
To select an appropriate cancer model for in vivo activity studies, we first assessed the tissue cathepsin
B expression of different xenograft tumors (H1975, H1650 and HCT116)
by immunohistochemistry. Analysis of stained sections using QuPath
0.5.1 software revealed that H1650 xenografts exhibited the strongest
cathepsin B levels (Figure
A,C). In parallel to the cathepsin B stains, also albumin
uptake into the tumor tissue was confirmed by immunohistochemistry
(Figure
B).
Based on their high cathepsin
B expression, H1650 cells were selected
as model for the evaluation of the new prodrugs in vivo (Figure
D,E). All
drugs were administered at doses equimolar to osimertinib mesylate
(29.8 mg/kg) twice per week. Indeed, treatment with all three TKIs
resulted in sustained and significant tumor growth inhibition. Mal-Pip-ValCit was more effective than Mal-Pip-GlyGly and even osimertinib alone, which supports the expected albumin-mediated
accumulation of the prodrug in the tumor. Accordingly, Mal-Pip-ValCit treatment resulted in a significant survival benefit (>100%)
compared
to the control. While animals in the control, Mal-Pip-GlyGly, and osimertinib groups reached average survival relatively fast
(within 39 days for the control group, 70 days for Mal-Pip-GlyGly, and 55 days for osimertinib), those receiving Mal-Pip-ValCit survived considerably longer (average 90 days), remaining viable
for up to 122 days. The clear survival gap between Mal-Pip-ValCit and Mal-Pip-GlyGly underscores the biological relevance
of linker activation in vivo, where cathepsin B activity
enables the selective release of the cytotoxic payload. In contrast, Mal-Pip-GlyGly, which cannot be enzymatically cleaved by cathepsin
B, is significantly less active in the tumor environment despite its
albumin-mediated delivery to the tumor. Notably, Mal-Pip-GlyGly still showed significant antitumor activity compared to the control,
which may result from activation during physiological degradation
of albumin. Importantly, no significant
weight loss (Figure S19) or signs of systemic
toxicity (Figure S20) were observed during
the course of treatment, indicating good tolerability. Together, this
indicates that incorporation of TKIs in albumin-targeted cathepsin
B-cleavable prodrug systems, are a valuable strategy to improve their
antitumor activity.

Conclusion
The exceptional pharmacokinetic
properties of albumin, including
its high plasma concentration, long circulatory half-life and ability
to accumulate in tumor tissue, highlight the protein as an important
vehicle for the targeted delivery of anticancer therapeutics. Many
endogenous ligands and drugs, including TKIs, are known to electrostatically
bind to albumin. However, in contrast to covalently bound therapeutics,
there is no direct evidence that electrostatic albumin association
of TKIs also leads to enhanced tumor targeting, beyond the well-known
effect of prolonged systemic circulation. Moreover, according to the
free drug hypothesis, only the unbound drug fraction in the body is
able to interact with its pharmacological target. Together, this suggests
that the activity of TKIs in vivo may even be reduced
because of their strong electrostatic binding to albumin. Notably,
for osimertinib, both our data and recent literature reports additionally
indicate extensive covalent binding to albumin lysines and an inability to reverse this reaction. This further reduces the availability of osimertinib
for target inhibition. Consequently, this study introduces a drug
delivery system for TKIs that allows their covalent attachment to
the Cys34 residue of endogenous albumin via a maleimide
moiety after intravenous administration. To enable tumor-specific
release, a cathepsin B-cleavable ValCit dipeptide coupled to a PABC
spacer was employed. For the selection of the active component, three
derivatives of the EGFR inhibitor osimertinib were synthesized, harboring
an NH group for attachment to the ValCit-PABC fragment, and the most
promising candidate OsiNHMe was identified after evaluation in vitro and in vivo. The target compound Mal-Pip-ValCit efficiently released OsiNHMe in
the presence of cathepsin B and demonstrated excellent stability in
the absence of the enzyme. As expected, the reference compound Mal-Pip-GlyGly was unaffected in either condition. The prodrug
nature of Mal-Pip-ValCit and Mal-Pip-GlyGly was demonstrated in a cell-free kinase inhibition assay against
double mutant EGFR (L858R/T790M), as no IC50 value could
be determined within the tested concentration range (up to 10 μM).
This finding was further supported by a cell-based MTT assay. Binding
studies revealed efficient covalent conjugation of both prodrugs to
albumin. Finally, in vivo studies were performed
in EGFR-mutant H1650-bearing mice, strongly overexpressing cathepsin
B. Indeed, Mal-Pip-ValCit demonstrated the strongest
tumor regression as well as the longest overall survival when directly
compared to Mal-Pip-GlyGly and osimertinib. These data
underscore the increased antitumor activity of a TKI when covalently
conjugated to albumin over electrostatic binding in blood, guiding
future strategies for improved therapeutics.

Experimental
Part

Chemicals
pMal-O-Pip-COOH·2HCl was synthesized
according to literature. The synthetic
strategies of OsiNHMe, OsiNH

2
, OsiPropNHMe
and
pMal-PEG6-COOH were adapted
from literature. Milli-Q water (18.2 MΩ cm, Merck Milli-Q Advantage,
Darmstadt, Germany) was used for synthesis and for analytical and
preparative RP-HPLC. Anhydrous solvents (DCM and DMF) over molecular
sieves were purchased from Thermo Scientific. All other chemicals
and solvents were purchased from commercial suppliers (Aaron Chemicals,
abcr, Acros Organics, Alfa Aesar, Ambeed, BLD Pharm, Fisher Scientific,
Fluka, Sigma-Aldrich and TCI Europe, VWR) and used without purification.
Silica gel (particle size 40–63 μm) for column chromatography
was purchased from VWR. No unexpected or unusually high safety hazards
were encountered. Electronspray ionization (ESI) mass spectra for
intermediate compounds were recorded on a Bruker amaZon speed ETD
mass spectrometer and high-resolution mass spectra of final compounds
were measured on an Orbitrap Exploris 120 mass spectrometer in positive
mode at the Mass Spectrometry Centre of the University of Vienna. 1H NMR spectra of intermediate compounds were recorded in deuterated
dimethyl sulfoxide (DMSO-d
6, 99.8% D,
purchased from Eurisotop) at 25 °C on a Bruker FT-NMR spectrometer
AV NEO 500 at 500.10 MHz. For the characterization of the final compounds OsiNHMe, OsiNH

2
, OsiPropNHMe, Mal-PEG-ValCit, Mal-PEG-GlyGly, Mal-Pip-ValCit and Mal-Pip-GlyGly dry
DMSO-d
6 from glass ampules (99.8% D, purchased
from Eurisotop) was used. One- (1H and 13C)
and two-dimensional (COSY, HSQC and HMBC) NMR spectra were recorded
at 25 °C using a Bruker FT-NMR spectrometer AV NEO 500 MHz. 1H NMR spectra were measured at 500.32 MHz and 13C NMR spectra at 125.82 MHz. All NMR spectra were recorded at the
NMR Centre of the University of Vienna. Chemical shifts (ppm) were
referenced internally to the residual solvent peaks. For the description
of the spin multiplicities the following abbreviations were used:
s = singlet, d = doublet, t = triplet, q = quartet, bs = broad singlet,
dd = doublet of doublets, dt = doublet of triplets, quint = quintet,
sxt = sextet, m = multiplet. The 1H, 13C, COSY,
HSQC and HMBC NMR spectra of the final compounds are depicted in Figures S21–S27 and the HPLC runs in Figures S28 and S29. All compounds are >95%
pure
by HPLC. Purification by preparative HPLC was performed on an XBridge
BEH C18 OBD Prep Column (19 mm × 250 mm) on an Agilent 1260 Infinity
II system. Flow rates of 17 mL/min were constant for each run at 20
°C. Elemental analysis measurements were performed on a Eurovector
EA 3000 CHNS–O Elemental Analyzer at the Microanalytical Laboratory
of the University of Vienna and are within ± 0.4%, confirming
>95% purity.

Synthesis

Tert-Butyl
(2-((5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)-2-nitrophenyl)­(methyl)­amino)­ethyl)­(methyl)­carbamate
(1a)

N-(4-Fluoro-2-methoxy-5-nitrophenyl)-4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-amine (2.00 g, 5.08 mmol, 1.0 equiv), tert-butyl methyl­(2-(methylamino)­ethyl)­carbamate (1.47 g,
7.63 mmol, 1.5 equiv) and K2CO3 (1.42 g, 10.17
mmol, 2.0 equiv) were dissolved in 35 mL DMF and heated at 100 °C
for 15 h. After cooling to room temperature, the mixture was poured
into 80 mL H2O, the precipitate filtered off, washed 3×
with H2O and dried in vacuo at 60 °C. 1a was obtained as a bright orange solid in 95% yield (2.72 g, 4.84
mmol). 1H NMR (500 MHz, DMSO-d
6): δ 8.68 (d, J = 21.7 Hz, 1H), 8.37 (d, J = 7.8 Hz, 1H), 8.33 (s, 1H), 8.32 (d, J = 5.4 Hz, 1H), 8.10 (s, 1H), 7.52 (d, J = 8.2 Hz,
1H), 7.25 (t, J = 8.0 Hz, 1H), 7.22 (d, J = 5.4 Hz, 1H), 7.12 (t, J = 7.5 Hz, 1H), 6.83 (d, J = 43.1 Hz, 1H), 3.97 (s, 3H), 3.88 (s, 3H), 3.41 (t, J = 6.0 Hz, 2H), 3.32–3.26 (m, 2H, overlap with water
signal), 2.87 (d, J = 5.6 Hz, 3H), 2.77 (d, J = 5.6 Hz, 3H), 1.35 (d, J = 13.3 Hz,
9H) ppm. MS (m/z) calcd C29H35N7O5: (M + H)+, 562.28;
found, 562.47.

Tert-Butyl (2-((2-amino-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­ethyl)­(methyl)­carbamate
(2a)
1a (2.63 g, 4.68 mmol, 1.0 equiv) was suspended
in 95 mL EtOH and 32 mL H2O, then Fe (1.59 g, 28.10 mmol,
6.0 equiv) and NH4Cl (0.19 g, 3.51 mmol, 0.75 equiv) were
added. The mixture was heated at 100 °C for 18 h and afterward
concentrated in vacuo. The residue was taken up in 40 mL DCM, filtered
and the filter washed with another 40 mL DCM. The filtrate was concentrated
in vacuo to give the crude product as a light-brown foam, which was
purified by silica column chromatography using 2.5% MeOH in DCM as
the eluent. 2a was obtained as a light-yellow foam in
91% yield (2.30 g, 4.28 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 8.42 (d, J = 8.0
Hz, 1H), 8.30 (s, 1H), 8.27 (d, J = 5.3 Hz, 1H),
7.78 (s, 1H), 7.55–7.47 (m, 2H), 7.24 (t, J = 7.2 Hz, 1H), 7.19–7.13 (m, 2H), 6.77 (s, 1H), 4.41 (s,
2H), 3.88 (s, 3H), 3.75 (s, 3H), 2.94 (t, J = 6.6
Hz, 2H), 2.79 (d, J = 18.4 Hz, 3H), 2.63 (s, 3H),
1.40 (s, 9H) ppm. 2H below the water signal. MS (m/z) calcd C29H37N7O3: (M + H)+, 532.30; found, 532.42.

Tert-Butyl (2-((2-Acrylamido-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­ethyl)­(methyl)­carbamate
(3a)
2a (2.30 g, 4.28 mmol, 1.0 equiv) and DIPEA
(1.51 mL, 8.55 mmol, 2.0 equiv) were dissolved in 50 mL dry DCM and
cooled to ca. −80 °C using liquid N2/acetone.
A solution of acryloyl chloride (0.36 mL, 4.28 mmol, 1.0 equiv) in
13 mL dry DCM was added under Ar atmosphere with an automated syringe
pump at a flow rate of 40 μL/min. After the addition, the mixture
was warmed up very slowly by allowing the cold bath to reach room
temperature overnight. The reaction mixture was diluted with DCM to
a total volume of 100 mL and washed with 80 mL saturated NaHCO3 solution. The organic phase was dried over Na2SO4 and reduced in vacuo to give the crude product as
a yellow foam, which was purified by silica column chromatography
using 10% acetone in DCM as the eluent. 3a was obtained
as a pale-yellow foam in 74% yield (1.87 g, 3.16 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 9.23–9.04
(m, 1H), 8.96 (s, 1H), 8.61 (s, 1H), 8.32 (d, J =
5.3 Hz, 1H), 8.26 (d, J = 8.0 Hz, 1H), 7.89 (s, 1H),
7.52 (d, J = 8.2 Hz, 1H), 7.27–7.20 (m, 2H),
7.16 (t, J = 7.3 Hz, 1H), 6.98 (s, 1H), 6.67 (dd, J = 16.7, 10.1 Hz, 1H), 6.26 (d, J = 16.9
Hz, 1H), 5.75 (d, J = 11.5 Hz, 1H), 3.90 (s, 3H),
3.87 (s, 3H), 2.99 (t, J = 6.0 Hz, 2H), 2.82–2.73
(m, 3H), 2.70 (s, 3H), 1.37 (s, 9H) ppm. 2H below the water signal.
MS (m/z) calcd C32H39N7O4: (M + H)+, 586.31;
found, 586.41.

N-(4-Methoxy-2-(methyl­(2-(methylamino)­ethyl)­amino)-5-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­acrylamide (OsiNHMe)
3a (1.87 g, 3.16 mmol, 1.0 equiv) was dissolved
in 40 mL DCM and TFA (8.61 mL, 110.63 mmol, 35.0 equiv) was added.
The mixture was stirred at room temperature for 1 h and the solvent
removed in vacuo. The residue was dissolved in Milli-Q-H2O, lyophilized and additionally dried in vacuo at 40 °C, affording
the crude product as a bright-yellow, solid TFA salt (2.43 g, including
solvent residues). A total of 101 mg were purified by preparative
HPLC on an XBridge BEH C18 OBD Prep Column (19 mm × 250 mm) using
a mixture of 44% MeOH and 56% Milli-Q water with 0.1% TFA as the eluent.
The product fractions were combined, MeOH removed in vacuo and the
aqueous phase lyophilized. OsiNHMe was obtained in 69%
yield (67 mg, 0.090 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 9.39 (s, 1H, H22), 9.03 (bs, 1H,
H14), 8.70 (s, 1H, H1), 8.67–8.51 (m, 1H, H20 + 2H, H30 due
to protonation), 8.39–8.10 (m, 2H, H7 + H12), 7.56 (d, J = 8.2 Hz, 1H, H4), 7.33 (d, J = 6.1 Hz,
1H, H11), 7.27 (t, J = 7.6 Hz, 1H, H5), 7.15 (t, J = 7.4 Hz, 1H, H6), 7.04 (s, 1H, H17), 6.72 (dd, J = 16.9, 10.2 Hz, 1H, H24), 6.28 (dd, J = 17.0, 1.7 Hz, 1H, H25), 5.78 (dd, J = 10.3, 1.5
Hz, 1H, H25), 3.91 (s, 3H, H2), 3.85 (s, 3H, H21), 3.25 (t, J = 5.7 Hz, 2H, H27), 3.16 (quint, J =
5.6 Hz, 2H, H28), 2.66–2.60 (m, 6H, H26 + H29) ppm. 13C NMR (126 MHz, DMSO-d
6): δ 164.35
(only in 2D, C10), 163.39 (C23), 158.38 (q, J = 32.9
Hz, Cq-TFA), 156.42 (only in 2D, C13), 151.36 (only in
2D, C12), 148.47 (only in 2D, C16), 140.64 (only in 2D, C18), 137.96
(C3), 136.22 (C1), 132.10 (C24), 126.79 (C25), 125.67 (C19), 125.41
(C8), 122.75 (C5), 122.54 (only in 2D, C15), 121.98 (C7), 121.74 (C6),
118.09 (only in 2D, C20), 116.70 (q, J = 296.9 Hz,
CF3-TFA), 112.02 (C9), 110.90 (C4), 106.60 (C11), 105.23
(C17), 56.11 (C21), 50.73 (C27), 45.57 (C28), 42.53 (C26), 33.37 (C2),
32.46 (C29) ppm. HRMS (m/z) calcd
C27H31N7O2: (M + H)+, 486.2612; found, 486.2602. Elemental analysis (%) Calcd
for C27H31N7O2*2TFA*1.5H2O: C, 50.27; H, 4.90: N, 13.24. Found: C, 50.38; H, 4.57;
N, 13.10.

Tert-Butyl (2-((5-Methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)-2-nitrophenyl)­(methyl)­amino)­ethyl)­carbamate
(1b)

N-(4-Fluoro-2-methoxy-5-nitrophenyl)-4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-amine (0.62 g, 1.58 mmol), tert-butyl (2-(methylamino)­ethyl)­carbamate (0.42 g, 2.37
mmol) and K2CO3 (0.44 g, 3.15 mmol) were dissolved
in 18 mL DMF and heated at 100 °C for 9 h. After cooling to room
temperature, the mixture was poured into 100 mL H2O, the
precipitate filtered off and dried in vacuo at 40 °C. 1b was obtained as a red solid in quantitative yield (0.88 g, 1.58
mmol). 1H NMR (500 MHz, DMSO-d
6): δ 8.66 (s, 1H), 8.36 (d, J = 7.6 Hz, 1H),
8.33 (s, 1H), 8.31 (d, J = 5.4 Hz, 1H), 8.09 (s,
1H), 7.52 (d, J = 8.2 Hz, 1H), 7.24 (t, J = 7.7 Hz, 1H), 7.21 (d, J = 5.4 Hz, 1H), 7.12 (t, J = 7.4 Hz, 1H), 6.86 (s, 1H), 6.84 (t, J = 5.2 Hz, 1H), 3.97 (s, 3H), 3.88 (s, 3H), 3.25–3.15 (m,
4H), 2.84 (s, 3H), 1.33 (s, 9H) ppm. MS (m/z) calcd C28H33N7O5: (M + H)+, 548.26; found, 548.30.

Tert-Butyl (2-((2-Amino-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­ethyl)­carbamate
(2b)
1b (0.87 g, 1.56 mmol) was suspended in
33 mL EtOH and 11 mL H2O, then Fe (0.53 g, 9.35 mmol) and
NH4Cl (0.063 g, 1.17 mmol) were added. The mixture was
heated at 100 °C for 3.5 h and afterward concentrated in vacuo.
The residue was taken up in 60 mL DCM containing 10% MeOH, filtered
and the filter washed with another 60 mL DCM containing 10% MeOH.
The filtrate was washed 2× with 60 mL brine, dried over Na2SO4 and concentrated in vacuo to give the crude
product as a light-brown foam, which was purified by silica column
chromatography using 2% MeOH in DCM as the eluent. 2b was obtained as a yellow foam in 92% yield (0.76 g, 1.43 mmol). 1H NMR (500 MHz, DMSO-d
6): δ
8.42 (d, J = 8.0 Hz, 1H), 8.31 (s, 1H), 8.26 (d, J = 5.3 Hz, 1H), 7.79 (s, 1H), 7.51 (d, J = 8.2 Hz, 1H), 7.49 (s, 1H), 7.24 (t, J = 7.6 Hz,
1H), 7.19–7.13 (m, 2H), 6.89 (t, J = 5.7 Hz,
1H), 6.74 (s, 1H), 4.50 (s, 2H), 3.88 (s, 3H), 3.74 (s, 3H), 3.08
(q, J = 6.3 Hz, 2H), 2.85 (t, J =
6.6 Hz, 2H), 2.58 (s, 3H), 1.38 (s, 9H) ppm. MS (m/z) calcd C28H35N7O3: (M + H)+, 518.29; found, 518.32.

Tert-Butyl (2-((2-Acrylamido-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­ethyl)­carbamate
(3b)
2b (0.69 g, 1.29 mmol) and DIPEA (0.46
mL, 2.59 mmol) were dissolved in 15 mL dry DCM and cooled to ca. −80
°C using liquid N2/acetone. A solution of acryloyl
chloride (0.11 mL, 1.29 mmol) in 4.2 mL dry DCM was added under Ar
atmosphere with an automated syringe pump at a flow rate of 40 μL/min.
After the addition, the mixture was warmed up very slowly by allowing
the cold bath to reach −15 °C. Stirring was then continued
in an ice bath for 10 min, then at room temperature for 40 min. The
reaction mixture was diluted with 70 mL DCM containing 10% MeOH and
washed 3× with 70 mL saturated NaHCO3 solution. The
organic phase was dried over Na2SO4 and reduced
in vacuo to give the crude product as a yellow foam, which was purified
by silica column chromatography using 10% acetone in DCM as the eluent. 3b was obtained as a yellow foam in 87% yield (0.68 g, 1.13
mmol). 1H NMR (500 MHz, DMSO-d
6): δ 9.23 (s, 1H), 9.02 (s, 1H), 8.63 (s, 1H), 8.32 (d, J = 5.3 Hz, 1H), 8.25 (d, J = 7.9 Hz, 1H),
7.91 (s, 1H), 7.52 (d, J = 8.2 Hz, 1H), 7.26–7.21
(m, 2H), 7.15 (t, J = 7.5 Hz, 1H), 7.04 (t, J = 5.3 Hz, 1H), 6.98 (s, 1H), 6.72 (dd, J = 16.9, 10.2 Hz, 1H), 6.27 (dd, J = 16.8, 1.1 Hz,
1H), 5.75 (dd, J = 10.5, 1.6 Hz, 1H), 3.91 (s, 3H),
3.86 (s, 3H), 3.10 (q, J = 6.0 Hz, 2H), 2.90 (t, J = 6.2 Hz, 2H), 2.64 (s, 3H), 1.37 (s, 9H) ppm. MS (m/z) calcd C31H37N7O4: (M + H)+, 572.30; found, 572.35.

N-(2-((2-Aminoethyl)­(methyl)­amino)-4-methoxy-5-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­acrylamide (OsiNH

2
)
3b (0.67 g, 1.11
mmol) was dissolved in 20 mL DCM and TFA (4.55 mL, 58.42 mmol) was
added. The mixture was stirred at room temperature for 40 min and
the solvent removed in vacuo. The residue was dissolved in Milli-Q
water, lyophilized and additionally dried in vacuo at 40 °C,
affording the crude product as a bright-yellow, solid TFA salt (0.83
g, including solvent residues). A total of 100 mg were purified by
preparative HPLC on an XBridge BEH C18 OBD Prep Column (19 mm ×
250 mm) using a mixture of 44% MeOH and 56% Milli-Q water with 0.1%
TFA as the eluent. The product fractions were combined, MeOH removed
in vacuo and the aqueous phase lyophilized. OsiNH

2
was obtained in 75% yield (73 mg, 0.10 mmol). 1H NMR (500 MHz, DMSO-d
6): δ
9.34 (s, 1H, H22), 9.04 (bs, 1H, H14), 8.70 (s, 1H, H1), 8.62 (s,
1H, H20), 8.38–8.10 (m, 2H, H7 + H12), 8.01–7.80 (m,
3H, H29 due to protonation), 7.56 (d, J = 8.2 Hz,
1H, H4), 7.33 (d, J = 6.1 Hz, 1H, H11), 7.27 (t, J = 7.6 Hz, 1H, H5), 7.16 (t, J = 7.4 Hz,
1H, H6), 7.02 (s, 1H, H17), 6.72 (dd, J = 17.0, 10.2
Hz, 1H, H24), 6.26 (dd, J = 17.0, 1.7 Hz, 1H, H25),
5.78 (dd, J = 10.0, 1.5 Hz, 1H, H25), 3.91 (s, 3H,
H2), 3.84 (s, 3H, H21), 3.21 (t, J = 5.7 Hz, 2H,
H27), 3.10–3.01 (m, 2H, H28), 2.61 (s, 3H, H26) ppm. 13C NMR (126 MHz, DMSO-d
6): δ 164.34
(only in 2D, C10), 163.27 (C23), 158.31 (q, J = 32.9
Hz, Cq-TFA), 156.15 (only in 2D, C13), 151.65 (only in
2D, C12), 148.48 (only in 2D, C16), 140.84 (only in 2D, C18), 137.96
(C3), 136.31 (C1), 132.18 (C24), 126.67 (C25), 125.70 (C19), 125.41
(C8), 122.77 (C5), 122.68 (only in 2D, C15), 121.99 (C7), 121.79 (C6),
118.02 (only in 2D, C20), 116.72 (q, J = 297.0 Hz,
CF3-TFA), 112.01 (C9), 110.90 (C4), 106.55 (C11), 105.10
(C17), 56.10 (C21), 52.13 (C27), 42.36 (C26), 36.41 (C28), 33.38 (C2)
ppm. HRMS (m/z) calcd C26H29N7O2: (M + H)+, 472.2455;
found, 472.2450. Elemental analysis (%) Calcd for C26H29N7O2*2TFA*1.5H2O: C, 49.59;
H, 4.72; N, 13.49. Found: C, 49.66; H, 4.35; N, 13.30.

Tert-Butyl Methyl­(3-(methylamino)­propyl)­carbamate

N,N′-Dimethylpropane-1,3-diamine
(1.80 mL, 14.33 mmol) was dissolved in 40 mL dry DCM and cooled to
0 °C. A solution of di-tert-butyl dicarbonate
(1.07 g, 4.81 mmol) in 20 mL dry DCM was added under Ar atmosphere
with an automated syringe pump at a flow rate of 50 μL/min.
The mixture was warmed to room temperature and stirred for 45 min.
Then, the mixture was washed with 70 mL Milli-Q water and the aqueous
phase extracted 3× with EtOAc. The combined organic phases were
washed 2× with brine, dried over MgSO4 and concentrated
in vacuo. Tert-butyl methyl­(3-(methylamino)­propyl)­carbamate
was obtained as a clear oil in 89% yield (0.94 g containing 7% EtOAc
residue, 4.30 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 3.17 (t, J = 7.1 Hz, 2H),
3.15–3.07 (m, 1H), 2.75 (bs, 3H), 2.40 (t, J = 6.7 Hz, 2H), 2.25 (s, 3H), 1.57 (quint, J = 6.8
Hz, 2H), 1.38 (s, 9H) ppm. MS (m/z) calcd C10H22N2O2: (M
+ H)+, 203.18; found, 203.12.

Tert-Butyl
(3-((5-Methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)-2-nitrophenyl)­(methyl)­amino)­propyl)­(methyl)­carbamate
(1c)

N-(4-Fluoro-2-methoxy-5-nitrophenyl)-4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-amine (0.50 g, 1.27 mmol), OsiPropNHMe-1
(0.42 mg, 1.91 mmol) and K2CO3 (0.36 g, 2.54
mmol) were dissolved in 8 mL DMF and heated at 100 °C for 9 h.
After cooling to room temperature, the mixture was poured into 80
mL H2O, the precipitate filtered off and dried in vacuo
at 40 °C. 1c was obtained as a bright orange solid
in 95% yield (0.69 g, 1.21 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 8.66 (s, 1H), 8.42–8.28 (m,
3H), 8.10 (s, 1H), 7.52 (d, J = 8.2 Hz, 1H), 7.24
(t, J = 7.8 Hz, 1H), 7.22 (d, J =
5.4 Hz, 1H), 7.11 (t, J = 7.6 Hz, 1H), 6.80 (s, 1H),
3.96 (s, 3H), 3.88 (s, 3H), 3.22–3.09 (m, 4H), 2.82 (s, 3H),
2.75 (s, 3H), 1.78 (quint, J = 6.9 Hz, 2H), 1.37
(s, 9H) ppm. MS (m/z) calcd C30H37N7O5: (M + Na)+, 598.27; found, 598.33.

Tert-Butyl (3-((2-Amino-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­propyl)­(methyl)­carbamate
(2c)
1c (0.69 g, 1.20 mmol) was suspended in
25 mL EtOH and 8.4 mL H2O, then Fe (0.41 g, 7.21 mmol)
and NH4Cl (0.049 g, 0.90 mmol) were added. The mixture
was heated at 100 °C for 5 h and afterward concentrated in vacuo.
The residue was taken up in 60 mL DCM containing 10% MeOH, filtered
and the filter washed with another 60 mL DCM containing 10% MeOH.
The filtrate was washed with 60 mL brine, dried over MgSO4 and concentrated in vacuo to give the crude product as a light-brown
foam, which was purified by silica column chromatography using a 1–2%
MeOH gradient in DCM as the eluent. 2c was obtained as
a yellow foam in 89% yield (0.59 g, 1.07 mmol). 1H NMR
(500 MHz, DMSO-d
6): δ 8.42 (d, J = 8.0 Hz, 1H), 8.30 (s, 1H), 8.27 (d, J = 5.3 Hz, 1H), 7.78 (s, 1H), 7.52 (s, 1H), 7.51 (d, J = 8.6 Hz, 1H), 7.27–7.22 (m, 1H), 7.18–7.13 (m, 2H),
6.74 (s, 1H), 4.45 (s, 2H), 3.88 (s, 3H), 3.74 (s, 3H), 3.22 (t, J = 7.1 Hz, 2H), 2.82 (t, J = 5.9 Hz, 2H),
2.75 (bs, 3H), 2.57 (s, 3H), 1.65 (bs, 2H), 1.38 (s, 9H) ppm. MS (m/z) calcd C30H39N7O3: (M + H)+, 546.32; found, 546.36.

Tert-Butyl (3-((2-Acrylamido-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­propyl)­(methyl)­carbamate
(3c)
2c (0.30 g, 0.54 mmol) and DIPEA (0.19
mL, 1.070 mmol) were dissolved in 6 mL dry DCM and cooled to ca. −80
°C using liquid N2/acetone. A solution of acryloyl
chloride (0.047 mL, 0.56 mmol) in 1.8 mL dry DCM was added under Ar
atmosphere with an automated syringe pump at a flow rate of 40 μL/min.
After the addition, the mixture was stirred at −70 °C
for 10 min, then stirring was then continued at room temperature for
45 min. The reaction mixture was diluted with 25 mL of DCM containing
10% MeOH and washed 3× with 25 mL saturated NaHCO3 solution. The organic phase was dried over Na2SO4 and reduced in vacuo to give the crude product as a yellow
foam, which was purified by silica column chromatography using a 6–10%
acetone gradient in DCM as the eluent. 3c was obtained
as a yellow foam in 84% yield (0.27 g, 0.45 mmol). 1H NMR
(500 MHz, DMSO-d
6): δ 9.51–9.11
(m, 1H), 9.03–8.85 (m, 1H), 8.65–8.55 (m, 1H), 8.32
(d, J = 5.3 Hz, 1H), 8.25 (d, J =
7.9 Hz, 1H), 7.89 (s, 1H), 7.52 (d, J = 8.2, 1H),
7.26–7.21 (m, 2H), 7.15 (t, J = 7.3 Hz, 1H),
6.93 (s, 1H), 6.87–6.64 (m, 1H), 6.26 (d, J = 17.0 Hz, 1H), 5.73 (d, J = 10.6 Hz, 1H), 3.90
(s, 3H), 3.86 (s, 3H), 3.21 (bs, 2H), 2.85 (t, J =
6.5 Hz, 2H), 2.78–2.70 (m, 3H), 2.67–2.55 (m, 3H), 1.64
(bs, 2H), 1.43–1.34 (m, 9H) ppm. MS (m/z) calcd C33H41N7O4: (M + H)+, 600.33; found, 600.37.

N-(4-Methoxy-2-(methyl­(3-(methylamino)­propyl)­amino)-5-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­acrylamide (OsiPropNHMe)
3c (0.27 g, 0.45 mmol) was dissolved
in 8 mL DCM and TFA (1.22 mL, 15.70 mmol) was added. The mixture was
stirred at room temperature for 35 min and the solvent removed in
vacuo. The residue was dissolved in Milli-Q water, lyophilized and
additionally dried in vacuo at 40 °C, affording the crude product
as a bright-yellow, solid TFA salt (0.36 g, including solvent residues).
A total of 170 mg were purified by preparative HPLC on an XBridge
BEH C18 OBD Prep Column (19 mm × 250 mm) using a mixture of 45%
MeOH and 55% Milli-Q water with 0.1% TFA as the eluent. The product
fractions were combined, MeOH removed in vacuo and the aqueous phase
lyophilized. OsiPropNHMe was obtained in 64% yield (116
mg, 0.14 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 10.06–8.83 (m, 2H, H14 + H22), 8.74
(s, 1H, H1), 8.66–8.35 (m, 1H, H20 + 2H, H30 due to protonation),
8.35–8.06 (m, 2H, H7 + H12), 7.57 (d, J =
8.2 Hz, 1H, H4), 7.35 (d, J = 6.3 Hz, 1H, H11), 7.29
(t, J = 7.6 Hz, 1H, H5), 7.17 (t, J = 7.3 Hz, 1H, H6), 7.01 (s, 1H, H17), 6.72 (dd, J = 16.9, 10.2 Hz, 1H, H24), 6.23 (dd, J = 17.0,
1.6 Hz, 1H, H25), 5.74 (dd, J = 10.3, 1.4 Hz, 1H,
H25), 3.92 (s, 3H, H2), 3.84 (s, 3H, H21), 3.00 (t, J = 6.2 Hz, 2H, H27), 2.97–2.89 (m, 2H, H29), 2.70 (s, 3H,
H26), 2.54 (t, J = 5.4 Hz, 3H, H31), 1.79 (quint, J = 7.3 Hz, 2H, H28) ppm. 13C NMR (126 MHz, DMSO-d
6): δ 165.11 (only in 2D, C10), 163.23
(C23), 158.36 (q, J = 33.4 Hz, Cq-TFA),
155.16 (only in 2D, C13), 149.20 (only in 2D, C12), 148.85 (only in
2D, C16), 141.40 (only in 2D, C18), 138.04 (C3), 136.91 (C1), 132.24
(C24), 126.53 (C25), 125.60 (C19), 125.41 (C8), 122.94 (C5), 121.99
(C6 + C7 and only in 2D, C15), 118.70 (only in 2D, C20), 116.53 (q, J = 296.1 Hz, CF3-TFA), 111.91 (C9), 111.00 (C4),
106.38 (C11), 104.95 (C17), 56.04 (C21), 52.53 (C27), 46.43 (C29),
42.03 (C26), 33.45 (C2), 32.59 (C31), 23.34 (C28) ppm. HRMS (m/z) calcd C28H33N7O2: (M + H)+, 500.2768; found,
500.2765. Elemental analysis (%) Calcd for C28H33N7O2*3TFA*H2O: C, 47.50; H, 4.46;
N, 11.40. Found: C, 47.61; H, 4.42; N, 11.47.

(9H-Fluoren-9-yl)­methyl (2-((2-((4-(hydroxymethyl)­phenyl)­amino)-2-oxoethyl)­amino)-2-oxoethyl)­carbamate
(Fmoc-GlyGly-PAB-OH)
(((9H-Fluoren-9-yl)­methoxy)­carbonyl)­glycylglycine
(1.00 g, 2.77 mmol, 1.0 equiv) was suspended in 30 mL DCM and 4-aminobenzyl
alcohol (0.70 g, 5.53 mmol, 2.0 equiv) was added. 10 mL MeOH were
added and the mixture sonicated until a clear light-brown solution
was obtained. EEDQ (1.38 g, 5.53 mmol, 2.0 equiv) was added and the
mixture stirred for 19 h, whereby a white solid precipitated. The
solid was filtered off, washed with a mixture of 25% MeOH in DCM and
dried in vacuo. Fmoc-GlyGly-PAB-OH was obtained as a white solid in
87% yield (1.20 g containing 8% DCM residue, 2.41 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 9.81
(s, 1H), 8.22 (t, J = 5.6 Hz, 1H), 7.89 (d, J = 7.5 Hz, 2H), 7.72 (d, J = 7.5 Hz, 2H),
7.64 (t, J = 6.0 Hz, 1H), 7.54 (d, J = 8.4 Hz, 2H), 7.42 (t, J = 7.4 Hz, 2H), 7.33 (t, J = 7.2 Hz, 2H), 7.24 (d, J = 8.4 Hz, 2H),
5.09 (t, J = 5.7 Hz, 1H), 4.43 (d, J = 5.7 Hz, 2H), 4.30 (d, J = 7.0 Hz, 2H), 4.24 (t, J = 6.9 Hz, 1H), 3.89 (d, J = 5.7 Hz, 2H),
3.68 (d, J = 6.0 Hz, 2H) ppm. MS (m/z) calcd C26H25N3O5: (M + Na)+, 482.17; found, 482.17.

2-Amino-N-(2-((4-(hydroxymethyl)­phenyl)­amino)-2-oxoethyl)­acetamide
(H

2

N-GlyGly-PAB-OH)
Fmoc-GlyGly-PAB-OH (1.19 g, 2.38 mmol, 1.0 equiv) was
dissolved in 20 mL DMF and piperidine (8 mL, 79.99 mmol, 33.6 equiv)
was added. The mixture was stirred for 1 h at room temperature, then
all volatiles were removed in vacuo. The residue was triturated in
40 mL DCM and the product filtered off and washed with DCM. H2N-GlyGly-PAB-OH was obtained as a white solid in 80% yield
(0.45 g, 1.91 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 9.92 (s, 1H), 8.20 (s, 1H), 7.53 (d, J = 8.2 Hz, 2H), 7.25 (d, J = 8.2 Hz, 2H),
5.10 (s, 1H), 4.44 (s, 2H), 3.93 (s, 2H), 3.16 (s, 2H), 2.04 (s, 2H)
ppm. MS (m/z) calcd C11H15N3O3: (M + H)+, 260.10;
found, 260.10.

3-(4-(2-(2-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)­ethoxy)­ethyl)­piperazin-1-yl)-N-(2-((2-((4-(hydroxymethyl)­phenyl)­amino)-2-oxoethyl)­amino)-2-oxoethyl)­propenamide
(pMal-O-Pip-GlyGly-PAB-OH)
pMal-O-Pip-COOH·2HCl
(333 mg, 0.67 mmol, 1.0 equiv) in a dry flask was dissolved in 13
mL dry DMF at room temperature under Ar atmosphere and NEt3 (0.94 mL, 6.74 mmol, 10.0 equiv) and EDC·HCl (198 mg, 1.01
mmol, 1.5 equiv) were added. The mixture was stirred for 10 min, then
HOBt·H2O (158 mg, 1.01 mmol, 1.5 equiv) was added
and stirring was continued for another 10 min. Finally, H2N-GlyGly-PAB-OH (240 mg, 1.01 mmol, 1.5 equiv) was added and the
mixture was stirred for 17 h. All volatiles were removed in vacuo
and the residue purified by preparative HPLC on an XBridge BEH C18
OBD Prep Column (19 mm × 250 mm) using a mixture of 13% CH3CN and 87% Milli-Q water with 0.1% TFA as the eluent. The
product fractions were combined, CH3CN removed in vacuo
and the aqueous phase lyophilized twice. pMal-O-Pip-GlyGly-PAB-OH
was obtained as a white solid TFA salt in 76% yield (444 mg, 0.51
mmol). 1H NMR (500 MHz, DMSO-d
6): δ 9.84 (s, 1H), 8.47–8.36 (m, 1H), 8.26 (t, J = 5.0 Hz, 1H), 7.54 (d, J = 8.4 Hz, 2H),
7.24 (d, J = 8.5 Hz, 2H), 6.38 (s, 2H), 4.43 (s,
2H), 3.90 (d, J = 5.8 Hz, 2H), 3.79 (d, J = 5.5 Hz, 2H), 2.90 (s, 2H), 1.54 (s, 6H) ppm. The remaining 20
aliphatic H-signals of the pMal-O-Pip linker and the alcohol signal
overlap with the broad water peak (∼3.70–2.70 ppm).
MS (m/z) calcd C32H44N6O8: (M + H)+, 641.33;
found, 641.30.

4-(2-(2-(3-(4-(2-(2-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)­ethoxy)­ethyl)­piperazin-1-yl)­propanamido)­acetamido)­acetamido)­benzyl
(4-nitrophenyl) carbonate (pMal-O-Pip-GlyGly-PAB-PNP)
pMal-O-Pip-GlyGly-PAB-OH (250 mg, 0.29 mmol, 1.0 equiv) in a dry
flask was dissolved in 10 mL dry DMF at room temperature under Ar
atmosphere and bis­(4-nitrophenyl) carbonate (268 mg, 0.86 mmol, 3.0
equiv) and DIPEA (0.76 mL, 4.32 mmol, 15.0 equiv) were added. The
mixture was stirred for 7 h, then all volatiles were removed in vacuo.
The residue was purified by silica column chromatography using a gradient
of 10–30% MeOH in DCM as the eluent. The product-containing
fractions were combined and 30 mL toluene were added before concentrating
them in vacuo. pMal-O-Pip-GlyGly-PAB-PNP was obtained as an off-white
solid in 97% yield (262 mg containing 9% toluene and 5% DIPEA·TFA
salt, 0.28 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 9.94 (bs, 1H), 8.51–8.34 (m, 1H), 8.34–8.28
(m, 2H), 8.22 (bs, 1H), 7.66 (d, J = 8.2 Hz, 2H),
7.60–7.53 (m, 2H), 7.42 (d, J = 8.6 Hz, 2H),
6.37 (s, 2H), 5.25 (s, 2H), 3.91 (d, J = 5.8 Hz,
2H), 3.76 (s, 2H), 3.57–3.43 (m, 6H), 3.07–2.81 (m,
2H, overlap with CH
exo signal), 2.89 (s,
2H), 1.53 (s, 6H) ppm. The remaining 12 aliphatic H-signals of the
pMal-O-Pip linker overlap with the DMSO peak (∼2.70–2.20
ppm). MS (m/z) calcd C39H47N7O12: (M + H)+, 806.34;
found, 806.31.

4-(2-(2-(3-(4-(2-(2-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)­ethoxy)­ethyl)­piperazin-1-yl)­propanamido)­acetamido)­acetamido)­benzyl
(2-((2-acrylamido-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­ethyl)­(methyl)­carbamate
(pMal-O-Pip-GlyGly-PAB-OsiNHMe)
pMal-O-Pip-GlyGly-PAB-PNP
(260 mg, 0.28 mmol, 1.0 equiv) and OsiNHMe (293 mg, 0.38 mmol, 1.4
equiv) in a dry flask were dissolved in 10 mL dry DMF at room temperature
under Ar atmosphere and DIPEA (0.26 mL, 1.47 mmol, 5.3 equiv) was
added. The mixture was stirred for 4 h, then all volatiles were removed
in vacuo. The residue was purified by silica column chromatography
using a gradient of 10–25% MeOH in DCM as the eluent. pMal-O-Pip-GlyGly-PAB-OsiNHMe
was obtained as a pale-yellow solid in 85% yield (307 mg containing
6% MeOH, 0.25 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 9.81 (s, 1H), 9.11 (d, J =
24.5 Hz, 1H), 8.95 (s, 1H), 8.61 (s, 1H), 8.48–8.36 (m, 1H),
8.32 (d, J = 5.3 Hz, 1H), 8.26 (d, J = 7.8 Hz, 1H), 8.20–8.12 (m, 1H), 7.87 (s, 1H), 7.65–7.55
(m, 2H), 7.52 (d, J = 8.2 Hz, 1H), 7.33–7.26
(m, 2H), 7.26–7.20 (m, 2H), 7.16 (t, J = 7.3
Hz, 1H), 6.94 (d, J = 22.9 Hz, 1H), 6.64 (dd, J = 16.8, 10.3 Hz, 1H), 6.35 (s, 2H), 6.24 (d, J = 17.7 Hz, 1H), 5.76–5.66 (m, 1H), 4.99 (s, 2H), 3.90 (s,
3H), 3.90–3.88 (m, 2H), 3.85 (d, J = 17.0
Hz, 3H), 3.74 (d, J = 5.5 Hz, 2H), 3.49 (t, J = 5.6 Hz, 2H), 3.46–3.37 (m, 6H), 3.08–2.98
(m, 2H), 2.87 (s, 2H), 2.84–2.78 (m, 3H), 2.71–2.60
(m, 3H, overlap with DMSO 13C satellite signal), 1.52 (s,
6H) ppm. The remaining 14 aliphatic H-signals of the pMal-O-Pip linker
overlap with the DMSO peak (∼2.60–2.10 ppm). MS (m/z) calcd C60H73N13O11: (M + H)+, 1152.56; found,
1152.53.

4-(2-(2-(3-(4-(2-(2-(2,5-Dioxo-2,5-dihydro-1H-pyrrol-1-yl)­ethoxy)­ethyl)­piperazin-1-yl)­propanamido)­acetamido)­acetamido)­benzyl
(2-((2-Acrylamido-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­ethyl)­(methyl)­carbamate
(Mal-Pip-GlyGly)
pMal-O-Pip-GlyGly-PAB-OsiNHMe
(293 mg, 0.24 mmol) was dissolved in 6 mL DMSO and stirred at 90 °C
for 3 h. All volatiles were removed in vacuo and the residue was purified
by preparative HPLC on an XBridge BEH C18 OBD Prep Column (19 mm ×
250 mm) using a mixture of 30% CH3CN and 70% Milli-Q water
with 0.1% TFA as the eluent. The product fractions were combined,
CH3CN removed in vacuo and the aqueous phase lyophilized
twice. Mal-Pip-GlyGly was obtained as a yellow solid in 64% yield
(221 mg, 0.15 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 9.98 (s, 1H, H36), 9.55 (bs, 1H, H14 or H22),
9.22 (d, J = 17.7 Hz, 1H, H14 or H22), 8.76 (s, 1H,
H1), 8.48 (t, J = 5.5 Hz, 1H, H42), 8.44 (bs, 1H,
H20), 8.31 (t, J = 5.6 Hz, 1H, H39), 8.29–8.03
(m, 2H, H7 + H12), 7.67–7.49 (m, 3H, H4 + H34), 7.37 (d, J = 6.4 Hz, 1H, H11), 7.33–7.23 (m, 3H, H5 + H33),
7.17 (t, J = 6.9 Hz, 1H, H6), 7.04 (s, 2H, H49),
7.03–6.94 (m, 1H, H17), 6.65 (dd, J = 16.8,
10.3 Hz, 1H, H24), 6.21 (d, J = 17.1 Hz, 1H, H25),
5.80–5.60 (m, 1H, H25), 4.99 (s, 2H, H31), 3.92 (s, 3H, H2),
3.90 (d, J = 5.8 Hz, 2H, H38), 3.86–3.76 (m,
5H, H21 + H41), 3.66 (t, J = 4.5 Hz, 2H, H45), 3.59
(t, J = 5.3 Hz, 2H, H47), 3.53 (t, J = 5.2 Hz, 2H, H46), 3.50–2.86 (m, 16H, HPip‑linker + H27 + H28), 2.81 (d, J = 16.0 Hz, 3H, H29), 2.74
(d, J = 26.8 Hz, 3H, H26), 2.59 (t, J = 6.3 Hz, 2H, H44) ppm. 13C NMR (126 MHz, DMSO-d
6): δ 171.05 (C48), 169.94 (C43), 169.26
(C40), 167.66 (C37), 165.59 (only in 2D, C10), 163.05 (C23), 158.45
(q, J = 33.8 Hz, Cq-TFA), 155.69 + 155.45
(C30 + C30′), 149.15 (only in 2D, C16), 147.70 (only in 2D,
C12), 141.93 (only in 2D, C18), 138.51 (C3 or C35), 138.08 (C3 or
C35), 137.46 (C1), 134.63 (C49), 132.10 (C24), 131.71 (C32), 128.42
(C33), 126.55 (C25), 125.42 (C8), 125.12 + 125.03 (C19 + C19′),
123.06 (C5), 122.14 (C6 + C7), 119.05 (C20 + C34), 116.47 (q, J = 295.6 Hz, CF3-TFA), 111.84 (C9), 111.04 (C4),
106.17 (C11), 104.90 + 104.67 (C17 + C17′), 67.34 (C46), 66.06
(C31), 64.99 (C45), 55.93 (C21), 55.07 (CPip‑linker), 53.28 + 52.78 (C27 + C27′), 51.95 (CPip‑linker), 46.55 + 45.89 (C28 + C28′), 42.59 (C38), 42.09 (C41), 41.83
+ 41.50 (C26 + C26′), 36.59 (C47), 34.46 + 34.00 (C29 + C29′),
33.52 (C2), 30.29 (C44) ppm. C13, C15 and four CPip‑linker signals could not be observed. HRMS (m/z) calcd C54H65N13O10: (M + H)+, 1056.5050; found, 1056.5031. Elemental
analysis (%) Calcd for C54H65N13O10*3.5TFA: C, 50.35; H, 4.74; N, 12.51. Found: C, 50.25; H,
4.71; N, 12.30.

(2S)-2-((2S)-2-(3-(4-(2-(2-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)­ethoxy)­ethyl)­piperazin-1-yl)­propanamido)-3-methylbutanamido)-N-(4-(hydroxymethyl)­phenyl)-5-ureidopentanamide (pMal-O-Pip-ValCit-PAB-OH)
pMal-O-Pip-COOH·2HCl (333 mg, 0.67 mmol, 1.0 equiv)
in a dry flask was dissolved in 13 mL dry DMF at room temperature
under Ar atmosphere and NEt3 (0.94 mL, 6.74 mmol, 10.0
equiv) and EDC·HCl (198 mg, 1.01 mmol, 1.5 equiv) were added.
The mixture was stirred for 10 min, then HOBt·H2O
(158 mg, 1.01 mmol, 1.5 equiv) was added and stirring was continued
for another 10 min. Finally, H2N-ValCit-PAB-OH (388 mg,
1.01 mmol, 1.5 equiv) was added and the mixture was stirred for 17
h. All volatiles were removed in vacuo and the residue purified by
preparative HPLC on an XBridge BEH C18 OBD Prep Column (19 mm ×
250 mm) using a mixture of 16% CH3CN and 84% Milli-Q water
with 0.1% TFA as the eluent. The product fractions were combined,
CH3CN removed in vacuo and the aqueous phase lyophilized
twice. pMal-O-Pip-ValCit-PAB-OH was obtained as a white solid TFA
salt in 66% yield (452 mg, 0.45 mmol). 1H NMR (500 MHz,
DMSO-d
6): δ 9.93 (s, 1H), 8.26–7.99
(m, 2H), 7.53 (d, J = 7.8 Hz, 2H), 7.23 (d, J = 7.4 Hz, 2H), 6.38 (s, 2H), 6.01 (s, 1H), 5.43 (bs, 2H),
4.48–4.35 (m, 3H), 4.26 (t, J = 7.0 Hz, 1H),
2.91 (s, 2H), 2.04–1.94 (m, 1H), 1.74–1.49 (m, 8H),
1.48–1.30 (m, 2H), 0.87 (d, J = 6.4 Hz, 3H),
0.84 (d, J = 6.3 Hz, 3H) ppm. The remaining 20 aliphatic
H-signals of the pMal-O-Pip linker and the alcohol signal overlap
with the broad water peak (∼3.70–2.70 ppm). Two H-signals
of the citrulline amino acid residue overlap with the DMSO signal
(∼2.60–2.40 ppm). MS (m/z) calcd C39H58N8O9: (M
+ H)+, 783.44; found, 783.44.

4-((2S)-2-((2S)-2-(3-(4-(2-(2-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)­ethoxy)­ethyl)­piperazin-1-yl)­propanamido)-3-methylbutanamido)-5-ureidopentanamido)­benzyl
(4-nitrophenyl) Carbonate (pMal-O-Pip-ValCit-PAB-PNP)
pMal-O-Pip-ValCit-PAB-OH (290 mg, 0.29 mmol, 1.0 equiv) in a dry
flask was dissolved in 10 mL dry DMF at room temperature under Ar
atmosphere and bis­(4-nitrophenyl) carbonate (267 mg, 0.86 mmol, 3.0
equiv) and DIPEA (0.76 mL, 4.30 mmol, 15.0 equiv) were added. The
mixture was stirred for 7 h, then all volatiles were removed in vacuo.
The residue was purified by silica column chromatography using a gradient
of 10–30% MeOH in DCM as the eluent. The product-containing
fractions were combined and 30 mL toluene were added before concentrating
them in vacuo. pMal-O-Pip-ValCit-PAB-PNP was obtained as an off-white
solid in 88% yield (252 mg containing 5% toluene, 0.25 mmol). 1H NMR (500 MHz, DMSO-d
6): δ
10.07 (s, 1H), 8.35–8.29 (m, 2H), 8.22 (bs, 1H), 8.17–8.08
(m, 1H), 7.65 (d, J = 8.6 Hz, 2H), 7.59–7.54
(m, 2H), 7.41 (d, J = 8.6 Hz, 2H), 6.36 (s, 2H),
5.98 (t, J = 5.5 Hz, 1H), 5.42 (s, 2H), 5.24 (s,
2H), 4.39 (dt, J = 8.0, 5.2 Hz, 1H), 4.25 (dd, J = 8.4, 6.5 Hz, 1H), 3.54–3.38 (m, 6H), 3.07–2.91
(m, 2H), 2.88 (s, 2H), 1.98 (sxt, J = 6.8 Hz, 1H),
1.75–1.55 (m, 2H), 1.53 (s, 6H), 1.50–1.32 (m, 2H),
0.87 (d, J = 6.8 Hz, 3H), 0.84 (d, J = 6.8 Hz, 3H) ppm. The remaining 14 aliphatic H-signals of the pMal-O-Pip
linker overlap with the DMSO peak (∼2.70–2.10 ppm).
MS (m/z) calcd C46H61N9O13: (M + H)+, 948.45;
found, 948.42.

4-((2S)-2-((2S)-2-(3-(4-(2-(2-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)­ethoxy)­ethyl)­piperazin-1-yl)­propanamido)-3-methylbutanamido)-5-ureidopentanamido)­benzyl
(2-((2-acrylamido-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­ethyl)­(methyl)­carbamate
(pMal-O-Pip-ValCit-PAB-OsiNHMe)
pMal-O-Pip-ValCit-PAB-PNP
(250 mg, 0.25 mmol, 1.0 equiv) and OsiNHMe (250 mg, 0.33 mmol, 1.3
equiv) in a dry flask were dissolved in 8.5 mL dry DMF at room temperature
under Ar atmosphere and DIPEA (0.22 mL, 1.25 mmol, 5.0 equiv) was
added. The mixture was stirred for 4 h, then all volatiles were removed
in vacuo. The residue was purified by silica column chromatography
using a gradient of 10–30% MeOH in DCM as the eluent. pMal-O-Pip-ValCit-PAB-OsiNHMe
was obtained as a pale-yellow solid in 92% yield (327 mg containing
9% MeOH, 0.23 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 9.98 (s, 1H), 9.11 (d, J =
20.8 Hz, 1H), 8.94 (s, 1H), 8.60 (s, 1H), 8.32 (d, J = 5.3 Hz, 1H), 8.26 (d, J = 8.0 Hz, 1H), 8.25–8.18
(m, 1H), 8.11 (d, J = 7.0 Hz, 1H), 7.87 (s, 1H),
7.62–7.54 (m, 2H), 7.52 (d, J = 8.2 Hz, 1H),
7.33–7.26 (m, 2H), 7.26–7.20 (m, 2H), 7.16 (t, J = 7.2 Hz, 1H), 6.94 (d, J = 22.3 Hz,
1H), 6.64 (dd, J = 16.8, 10.2 Hz, 1H), 6.35 (s, 2H),
6.24 (d, J = 18.3 Hz, 1H), 5.97 (t, J = 5.6 Hz, 1H), 5.75–5.68 (m, 1H), 5.40 (s, 2H), 4.99 (s,
2H), 4.43–4.33 (m, 1H), 4.23 (dd, J = 8.4,
6.5 Hz, 1H), 3.90 (s, 3H), 3.85 (d, J = 16.3 Hz,
3H), 3.49 (t, J = 5.5 Hz, 2H), 3.47–3.37 (m,
6H), 3.08–2.90 (m, 4H), 2.87 (s, 2H), 2.84–2.76 (m,
3H), 2.72–2.61 (m, 3H, overlap with DMSO 13C satellite
signal), 1.97 (sxt, J = 6.7 Hz, 1H), 1.74–1.55
(m, 2H), 1.52 (s, 6H), 1.48–1.31 (m, 2H), 0.86 (d, J = 6.8 Hz, 3H), 0.82 (d, J = 6.8 Hz, 3H)
ppm. The remaining 14 aliphatic H-signals of the pMal-O-Pip linker
overlap with the DMSO peak (∼2.60–2.10 ppm). MS (m/z) calcd C67H87N15O12: (M + H)+, 1294.67; found,
1294.61.

4-((S)-2-((S)-2-(3-(4-(2-(2-(2,5-Dioxo-2,5-dihydro-1H-pyrrol-1-yl)­ethoxy)­ethyl)­piperazin-1-yl)­propanamido)-3-methylbutanamido)-5-ureidopentanamido)­benzyl
(2-((2-acrylamido-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­ethyl)­(methyl)­carbamate
(Mal-Pip-ValCit)
pMal-O-Pip-ValCit-PAB-OsiNHMe
(311 mg, 0.22 mmol) was dissolved in 5.5 mL DMSO and stirred at 90
°C for 3 h. All volatiles were removed in vacuo and the residue
was purified by preparative HPLC on an XBridge BEH C18 OBD Prep Column
(19 mm × 250 mm) using a mixture of 32% CH3CN and
68% Milli-Q water with 0.1% TFA as the eluent. The product fractions
were combined, CH3CN removed in vacuo and the aqueous phase
lyophilized twice. Mal-Pip-ValCit was obtained as a yellow solid in
65% yield (219 mg, 0.14 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 10.06 (s, 1H, H36), 9.91–8.97
(m, 2H, H14 + H22), 8.75 (s, 1H, H1), 8.46 (s, 1H, H20), 8.38–7.95
(m, 4H, H7 + H12 + H39 + H42), 7.66–7.48 (m, 3H, H4 + H34),
7.36 (d, J = 6.2 Hz, 1H, H11), 7.28 (t, J = 7.2 Hz, 3H, H5 + H33), 7.21–7.11 (m, 1H, H6), 7.04 (s,
2H, H49), 7.03–6.95 (m, 1H, H17), 6.65 (dd, J = 16.7, 10.4 Hz, 1H, H24), 6.21 (d, J = 17.0 Hz,
1H, H25), 6.08 (s, 1H, H53), 5.93–5.08 (m, 3H, H25 + H55),
4.99 (s, 2H, H31), 4.45–4.34 (m, 1H, H38), 4.26 (t, J = 7.5 Hz, 1H, H41), 3.92 (s, 3H, H2), 3.82 (d, J = 16.1 Hz, 3H, H21), 3.70–3.62 (m, 2H, H45), 3.59
(t, J = 5.1 Hz, 2H, H47), 3.53 (t, J = 5.1 Hz, 2H, H46), 3.50–2.84 (m, 18H, HPip‑linker + H27 + H28 + H52), 2.80 (d, J = 14.5 Hz, 3H, H29),
2.74 (d, J = 22.2 Hz, 3H, H26), 2.67–2.55
(m, 2H, H44), 1.98 (sxt, J = 6.5 Hz, 1H, H56), 1.74–1.52
(m, 2H, H50), 1.50–1.29 (m, 2H, H51), 0.87 (d, J = 6.5 Hz, 3H, H57), 0.84 (d, J = 6.6 Hz, 3H, H57)
ppm. 13C NMR (126 MHz, DMSO-d
6): δ 171.05 (C40 + C48), 170.65 (C37), 169.46 (C43), 165.38
(only in 2D, C10), 163.04 (C23), 159.02 (C54), 158.43 (q, J = 33.4 Hz, Cq-TFA), 155.69 + 155.43 (C30 +
C30′), 148.80 (only in 2D, C12), 148.96 (only in 2D, C16),
142.10 (only in 2D, C18), 138.66 (C3 or C35), 138.06 (C3 or C35),
137.22 (C1), 134.64 (C49), 132.12 (C24), 131.67 (C32), 128.48 (C33),
126.54 (C25), 125.42 (C8), 125.14 + 125.00 (C19 + C19′), 123.00
(C5), 122.07 (C6 + C7), 119.02 (C20 + C34), 116.58 (q, J = 296.3 Hz, CF3-TFA), 111.88 (C9), 111.01 (C4), 106.24
(C11), 104.84 + 104.62 (C17 + C17′), 67.32 (C46), 66.09 (C31),
65.12 (C45), 57.54 (C41), 55.94 (C21), 55.14 (CPip‑linker), 53.29 + 52.78 (C27 + C27′), 53.13 (C38), 52.14 (CPip‑linker), 46.55 + 45.91 (C28 + C28′), 41.86 + 41.53 (C26 + C26′),
38.56 (C52), 36.61 (C47), 34.47 + 33.99 (C29 + C29′), 33.48
(C2), 30.72 (C56), 30.36 (C44), 29.35 (C50), 26.91 (C51), 19.24 (C57),
18.19 (C57) ppm. C13, C15 and four CPip‑linker signals
could not be observed. HRMS (m/z) calcd C61H79N15O11:
(M + H)+, 1198.6156; found, 1198.6159. Elemental analysis
(%) Calcd for C61H79N15O11*3TFA: C, 52.24; H, 5.37; N, 13.64. Found: C, 51.93; H, 5.33; N,
13.34.

Tert-Butyl 1-(2,5-Dioxo-2,5-dihydro-1H-pyrrol-1-yl)-3,6,9,12,15,18-hexaoxahenicosan-21-oate (Mal-PEG

6

-COO

t

Bu)

Tert-butyl
1-amino-3,6,9,12,15,18-hexaoxahenicosan-21-oate (1.00 g, 2.39 mmol,
1.0 equiv) was dissolved in 15 mL CHCl3 and stirred in
an ice bath for 10 min. N-Methoxycarbonylmaleimide
(0.76 g, 4.79 mmol, 2.0 equiv), tetrabutylammonium hydrogen sulfate
(0.73 g, 2.15 mmol, 0.9 equiv) and NEt3 (0.44 mL. 3.11
mmol, 1.3 equiv) were added in this order and stirred in the cold
for 15 min. Thirty mL of saturated NaHCO3 solution were
added and the mixture was stirred vigorously for 10 min in the cold,
before allowing the mixture to warm up to room temperature. Stirring
was continued for 20 h, then the phases were separated and the aqueous
layer extracted 2× with 20 mL CHCl3. The combined
organic phases were concentrated in vacuo to give the crude product
as a red oil, which was purified by silica column chromatography using
10% MeOH in EtOAc as the eluent. Mal-PEG6-COO
t
Bu was obtained as a clear oil in 89% yield (1.04
g, 2.13 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 7.03 (s, 2H), 3.60–3.54 (m, 4H), 3.53–3.44
(m, 22H), 2.41 (t, J = 6.2 Hz, 2H), 1.39 (s, 9H)
ppm. MS (m/z) calcd C23H39NO10: (M + Na)+, 512.25; found,
512.28.

Tert-Butyl 1-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)-3,6,9,12,15,18-hexaoxahenicosan-21-oate
(pMal-PEG

6

-COO

t

Bu)
Mal-PEG6-COO
t
Bu (1.03 g, 2.10 mmol, 1.0 equiv)
was dissolved in 35 mL CH3CN and 2,5-dimethylfuran (2.3
mL, 21.04 mmol, 10.0 equiv) was added. The mixture was heated at 60
°C for 24 h, then all volatiles were removed in vacuo. pMal-PEG6-COO
t
Bu was obtained as a yellow
oil in 93% yield (1.19 g, 1.96 mmol). The molar ratio of exo/endo
was 1.00:0.28, according to NMR. 1H NMR (500 MHz, DMSO-d
6): δ 6.37 (s, 1.47H, CH
mal‑exo), 6.22 (s, 0.40H, CH
mal‑endo), 3.58 (t, J = 6.2 Hz, 2H),
3.53–3.42 (m, 24H), 3.27 (s, 0.43H, CH
endo), 2.89 (s, 1.44H, CH
exo),
2.41 (t, J = 6.2 Hz, 2H), 1.63 (s, 1.22H, CH
3,endo), 1.53 (s, 4.50H, CH
3,exo), 1.39 (s, 9H) ppm. Ca. 4% of deprotected maleimide
species could be observed at 7.03 ppm. MS (m/z) calcd C29H47NO11: (M
+ Na)+, 608.30; found, 608.32.

1-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)-3,6,9,12,15,18-hexaoxahenicosan-21-oic
Acid (pMal-PEG

6

-COOH)
pMal-PEG6-COO
t
Bu
(1.18 g, 1.94 mmol, 1.0 equiv) was dissolved in 10 mL DCM and TFA
(5 mL, 64.25 mmol, 33.2 equiv) was added. The mixture was stirred
at room temperature for 1.5 h, then all volatiles were removed and
the residue purified by silica column chromatography using a gradient
of 2–10% MeOH in DCM as the eluent. The product-containing
fractions were combined and concentrated. The resulting brown oil
was dissolved in Milli-Q water, causing a phase separation of the
product in water and a dark brown liquid. The aqueous phase was separated
and lyophilized. pMal-PEG6-COOH was obtained as a light-orange
oil in 83% yield (0.92 g, 1.61 mmol). The molar ratio of exo/endo
was 1.00:0.22, according to NMR. 1H NMR (500 MHz, DMSO-d
6): δ 12.14 (bs, 1H), 6.37 (s, 1.52H,
CH
mal‑exo), 6.22 (s, 0.34H, CH
mal‑endo), 3.59 (t, J = 6.4 Hz, 2H), 3.52–3.43 (m, 24H), 3.27 (s, 0.36H, CH
endo), 2.89 (s, 1.53H, CH
exo), 2.43 (t, J = 6.4 Hz, 2H), 1.63 (s, 1.05H,
CH
3,endo), 1.53 (s, 4.72H, CH
3,exo) ppm. Ca. 7% of deprotected maleimide species could
be observed at 7.03 ppm. MS (m/z) calcd C25H39NO11: (M + Na)+, 552.24; found, 552.27.

1-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)-N-(2-((2-((4-(hydroxymethyl)­phenyl)­amino)-2-oxoethyl)­amino)-2-oxoethyl)-3,6,9,12,15,18-hexaoxahenicosan-21-amide
(pMal-PEG

6

-GlyGly-PAB-OH)
pMal-PEG6-COOH (100 mg, 0.18 mmol, 1.0 equiv)
in a dry flask was dissolved in 3.5 mL dry DMF at room temperature
under Ar atmosphere and NEt3 (0.25 mL, 1.76 mmol, 10.0
equiv) and EDC·HCl (52 mg, 0.26 mmol, 1.5 equiv) were added.
The mixture was stirred for 10 min, then HOBt·H2O
(41 mg, 0.26 mmol, 1.5 equiv) was added and stirring was continued
for another 10 min. Finally, H2N-GlyGly-PAB-OH (63 mg,
0.26 mmol, 1.5 equiv, see above) was added and the mixture was stirred
for 20 h. All volatiles were removed in vacuo and the residue purified
by preparative HPLC on an XBridge BEH C18 OBD Prep Column (19 mm ×
250 mm) using a mixture of 22% CH3CN and 78% Milli-Q water
with 0.1% TFA as the eluent. The product fractions were combined,
CH3CN removed in vacuo and the aqueous phase lyophilized
twice. pMal-PEG6-GlyGly-PAB-OH was obtained as a white
solid in 81% yield (106 mg, 0.14 mmol). The molar ratio of exo/endo
was 1.00:0.16, according to NMR. 1H NMR (500 MHz, DMSO-d
6): δ 9.73 (s, 1H), 8.26 (t, J = 5.6 Hz, 1H), 8.17 (t, J = 5.7 Hz, 1H), 7.56 (d, J = 8.5 Hz, 2H), 7.24 (d, J = 8.5 Hz, 2H),
6.36 (s, 1.69H, CH
mal‑exo), 6.22
(s, 0.26H, CH
mal‑endo), 5.06 (bs,
1H), 4.43 (s, 2H), 3.88 (d, J = 5.9 Hz, 2H), 3.74
(d, J = 5.7 Hz, 2H), 3.62 (t, J =
6.5 Hz, 2H), 3.53–3.43 (m, 24H), 3.27 (s, 0.32H, CH
endo), 2.89 (s, 1.69H, CH
exo), 2.42 (t, J = 6.5 Hz, 2H), 1.63 (s, 0.80H, CH
3,endo), 1.53 (s, 5.19H, CH
3,exo) ppm. Ca. 2% of deprotected maleimide species could
be observed at 7.02 ppm. MS (m/z) calcd C36H52N4O13:
(M + Na)+, 771.34; found, 771.36.

4-(1-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)-21,24-dioxo-3,6,9,12,15,18-hexaoxa-22,25-diazaheptacosan-27-amido)­benzyl
(4-nitrophenyl) Carbonate (pMal-PEG

6

-GlyGly-PAB-PNP)
pMal-PEG6-GlyGly-PAB-OH (97 mg, 0.13 mmol, 1.0 equiv) in a dry flask was dissolved
in 4.5 mL dry DMF at room temperature under Ar atmosphere and bis­(4-nitrophenyl)
carbonate (121 mg, 0.39 mmol, 3.0 equiv) and DIPEA (0.34 mL, 1.94
mmol, 15.0 equiv) were added. The mixture was stirred for 3 h, then
all volatiles were removed in vacuo. The residue was purified by silica
column chromatography using 8% MeOH in DCM as the eluent. The product-containing
fractions were combined and 10 mL toluene were added before concentrating
them in vacuo. pMal-PEG6-GlyGly-PAB-PNP was obtained as
a white waxy solid in 89% yield (105 mg, 0.11 mmol). The molar ratio
of exo/endo was 1.00:0.13, according to NMR. 1H NMR (500
MHz, DMSO-d
6): δ 9.89 (s, 1H), 8.34–8.29
(m, 2H), 8.27 (t, J = 5.6 Hz, 1H), 8.19 (t, J = 5.8 Hz, 1H), 7.67 (d, J = 8.5 Hz, 2H),
7.59–7.54 (m, 2H), 7.42 (d, J = 8.5 Hz, 2H),
6.36 (s, 1.69H, CH
mal‑exo), 6.22
(s, 0.22H, CH
mal‑endo), 5.25 (s,
2H), 3.90 (d, J = 5.8 Hz, 2H), 3.75 (d, J = 5.7 Hz, 2H), 3.62 (t, J = 6.5 Hz, 2H), 3.52–3.43
(m, 24H), 3.27 (s, 0.24H, CH
endo), 2.88
(s, 1.70H, CH
exo), 2.42 (t, J = 6.5 Hz, 2H), 1.62 (s, 0.67H, CH
3,endo), 1.53 (s, 5.17H, CH
3,exo). Ca. 4% of
deprotected maleimide species could be observed at 7.02 ppm. MS (m/z) calcd C43H55N5O17: (M + Na)+, 936.35; found,
936.35.

4-(1-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)-21,24-dioxo-3,6,9,12,15,18-hexaoxa-22,25-diazaheptacosan-27-amido)­benzyl
(2-((2-acrylamido-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­ethyl)­(methyl)­carbamate
(pMal-PEG

6

-GlyGly-PAB-OsiNHMe)
pMal-PEG6-GlyGly-PAB-PNP (97 mg, 0.11 mmol,
1.0 equiv) and OsiNHMe (106 mg, 0.14 mmol, 1.3 equiv) in a dry flask
were dissolved in 5 mL dry DMF at room temperature under Ar atmosphere
and DIPEA (94 μL, 0.53 mmol, 5.0 equiv) was added. The mixture
was stirred for 2 h, then all volatiles were removed in vacuo. The
residue was purified by silica column chromatography using 7% MeOH
in DCM as the eluent. pMal-PEG6-GlyGly-PAB-OsiNHMe was
obtained as a yellow sticky foam in 95% yield (152 mg containing 16%
DIPEA·TFA salt, 0.10 mmol). The molar ratio of exo/endo was 1.00:0.11,
according to NMR. 1H NMR (500 MHz, DMSO-d
6): δ 9.81 (s, 1H), 9.11 (d, J =
23.7 Hz, 1H), 8.94 (s, 1H), 8.61 (s, 1H), 8.32 (d, J = 5.3 Hz, 1H), 8.30–8.21 (m, 2H, overlap with DIPEA·TFA
signal), 8.17 (t, J = 5.7 Hz, 1H), 7.88 (s, 1H),
7.59 (s, 2H), 7.52 (d, J = 8.2 Hz, 1H), 7.33–7.26
(m, 2H), 7.26–7.21 (m, 2H), 7.16 (t, J = 7.5
Hz, 1H), 6.94 (d, J = 22.9 Hz, 1H), 6.64 (dd, J = 16.7, 10.3 Hz, 1H), 6.35 (s, 1.73H, CH
mal‑exo), 6.24 (d, J = 17.5 Hz,
1H), 6.21 (s, 0.20H, CH
mal‑endo), 5.75–5.66 (m, 1H), 4.99 (s, 2H), 3.90 (s, 3H), 3.88 (d, J = 5.9 Hz, 2H), 3.87–3.80 (m, 3H), 3.74 (d, J = 5.6 Hz, 2H), 3.61 (t, J = 6.3 Hz, 2H,
overlap with DIPEA·TFA signal), 3.51–3.43 (m, 24H), 3.40
(t, J = 6.5 Hz, 2H), 3.26 (s, 0.29H, CH
endo), 3.09–2.97 (m, 2H), 2.88 (s, 1.77H, CH
exo), 2.85–2.77 (m, 3H), 2.66 (d, J = 26.1 Hz, 3H), 2.41 (t, J = 6.4 Hz,
2H), 1.62 (s, 0.62H, CH
3,endo), 1.52 (s,
5.32H, CH
3,exo) ppm. Ca. 6% of deprotected
maleimide species could be observed at 7.01 ppm. MS (m/z) calcd C64H81N11O16: (M + Na)+, 1282.58; found, 1282.58.

4-(1-(2,5-Dioxo-2,5-dihydro-1H-pyrrol-1-yl)-21,24-dioxo-3,6,9,12,15,18-hexaoxa-22,25-diazaheptacosan-27-amido)­benzyl
(2-((2-acrylamido-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­ethyl)­(methyl)­carbamate
(Mal-PEG-GlyGly)
pMal-PEG6-GlyGly-PAB-OsiNHMe
(135 mg, 0.090 mmol) was dissolved in 2.3 mL DMSO and stirred at 90
°C for 3 h. All volatiles were removed in vacuo and the residue
was purified by preparative HPLC on an XBridge BEH C18 OBD Prep Column
(19 mm × 250 mm) using a mixture of 35% CH3CN and
65% Milli-Q water with 0.1% TFA as the eluent. The product fractions
were combined, CH3CN removed in vacuo and the aqueous phase
lyophilized twice. Mal-PEG-GlyGly was obtained as a yellow solid in
66% yield (77 mg, 0.059 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 9.84 (s, 1H, H36), 9.74–8.85
(m, 2H, H14 + H22), 8.75 (s, 1H, H1), 8.48 (bs, 1H, H20), 8.35–8.10
(m, 4H, H7 + H12 + H39 + H42), 7.65–7.50 (m, 3H, H4 + H34),
7.35 (d, J = 6.3 Hz, 1H, H11), 7.32–7.23 (m,
3H, H5 + H33), 7.22–7.12 (m, 1H, H6), 7.07–6.90 (m,
3H, H17 + H48), 6.65 (dd, J = 16.8, 10.3 Hz, 1H,
H24), 6.21 (d, J = 17.0 Hz, 1H, H25), 5.78–5.62
(m, 1H, H25), 4.99 (s, 2H, H31), 3.92 (s, 3H, H2), 3.88 (d, J = 5.6 Hz, 2H, H38), 3.82 (d, J = 18.4
Hz, 3H, H21), 3.74 (d, J = 5.6 Hz, 2H, H41), 3.61
(t, J = 6.5 Hz, 2H, H45), 3.57–3.53 (m, 2H,
H46), 3.52–3.41 (m, 24H, H28 + HPEG‑linker), 3.18–3.00 (m, 2H, H27), 2.81 (d, J = 15.4
Hz, 3H, H29), 2.73 (d, J = 27.4 Hz, 3H, H26), 2.41
(t, J = 6.4 Hz, 2H, H44) ppm. 13C NMR
(126 MHz, DMSO-d
6): δ 171.02 (C43
or C47), 170.96 (C43 or C47), 169.46 (C40), 167.70 (C37), 165.52 (only
in 2D, C10), 163.03 (C23), 158.05 (q, J = 33.5 Hz,
Cq-TFA), 155.69 + 155.44 (C30 + C30′), 148.95 (only
in 2D, C16), 138.49 (C3 or C35), 138.07 (C3 or C35), 137.22 (C1),
134.58 (C48), 132.11 (C24), 131.70 (C32), 128.39 (C33), 126.55 (C25),
125.41 (C8), 125.15 + 125.04 (C19 + C19′), 123.00 (C5), 122.08
(C6 + C7), 119.06 (C20 + C34), 111.87 (C9), 111.02 (C4), 106.24 (C11),
104.85 + 104.65 (C17 + C17′), 69.80 + 69.77 + 69.65 + 69.56
+ 69.41 (s) + 66.95 (CPEG‑linker), 66.70 (C45),
66.07 (C31), 55.95 (C21), 53.28 + 52.77 (C27 + C27′), 46.57
+ 45.89 (C28 + C28′), 42.65 (C38), 42.29 (C41), 41.85 + 41.54
(C26 + C26′), 36.80 (C46), 35.91 (C44), 34.45 + 34.00 (C29
+ C29′), 33.48 (C2) ppm. The CF3-TFA, C12, C13,
C15 and C18 signals could not be observed. HRMS (m/z) calcd C58H73N11O15: (M + H)+, 1164.5360; found, 1164.5353.
Elemental analysis (%) Calcd for C58H73N11O15*TFA*H2O: C, 55.59; H, 5.91; N,
11.89. Found: C, 55.59; H, 5.82; N, 11.62.

1-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)-N-((S)-1-(((S)-1-((4-(hydroxymethyl)­phenyl)­amino)-1-oxo-5-ureidopentan-2-yl)­amino)-3-methyl-1-oxobutan-2-yl)-3,6,9,12,15,18-hexaoxahenicosan-21-amide
(pMal-PEG

6

-ValCit-PAB-OH)
pMal-PEG6-COOH (100 mg, 0.18 mmol, 1.0 equiv)
in a dry flask was dissolved in 3.5 mL dry DMF at room temperature
under Ar atmosphere and NEt3 (0.25 mL, 1.76 mmol, 10.0
equiv) and EDC·HCl (52 mg, 0.26 mmol, 1.5 equiv) were added.
The mixture was stirred for 10 min, then HOBt·H2O
(41 mg, 0.26 mmol, 1.5 equiv) was added and stirring was continued
for another 10 min. Finally, H2N-ValCit-PAB-OH (101 mg,
0.26 mmol, 1.5 equiv) was added and the mixture was stirred for 20
h. All volatiles were removed in vacuo and the residue purified by
preparative HPLC on an XBridge BEH C18 OBD Prep Column (19 mm ×
250 mm) using a mixture of 25% CH3CN and 75% Milli-Q water
with 0.1% TFA as the eluent. The product fractions were combined,
CH3CN removed in vacuo and the aqueous phase lyophilized
twice. pMal-PEG6-ValCit-PAB-OH was obtained as a white
solid in 87% yield (136 mg, 0.15 mmol). The molar ratio of exo/endo
was 1.00:0.16, according to NMR. 1H NMR (500 MHz, DMSO-d
6): δ 9.88 (s, 1H), 8.08 (d, J = 7.6 Hz, 1H), 7.87 (d, J = 8.6 Hz, 1H), 7.54 (d, J = 8.5 Hz, 2H), 7.23 (d, J = 8.5 Hz, 2H),
6.36 (s, 1.69H, CH
mal‑exo), 6.22
(s, 0.27H, CH
mal‑endo), 5.98 (s,
1H), 5.40 (bs, 2H), 4.42 (s, 2H), 4.38 (dt, J = 7.9,
5.1 Hz, 1H), 4.22 (dd, J = 8.4, 6.8 Hz, 1H), 3.64–3.55
(m, 2H, overlap with water signal), 3.53–3.43 (m, 24H, overlap
with water signal), 3.27 (s, 0.30H, CH
endo), 3.06–2.90 (m, 2H), 2.89 (s, 1.75H, CH
exo), 2.46 (t, J = 7.2 Hz, 1H, overlap with
DMSO signal), 2.39 (t, J = 6.3 Hz, 1H), 1.96 (sxt, J = 6.8 Hz, 1H), 1.74–1.55 (m, 2H), 1.63 (s, 0.91H,
CH
3,exo), 1.53 (s, 5.17H, CH
3,exo), 1.49–1.31 (m, 2H), 0.86 (d, J = 6.8 Hz, 3H), 0.83 (d, J = 6.8 Hz, 3H) ppm. The
OH proton could not be observed. Ca. 3% of deprotected maleimide species
could be observed at 7.02 ppm. MS (m/z) calcd C43H66N6O14:
(M + Na)+, 913.45; found, 913.47.

4-((23S,26S)-1-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)-23-isopropyl-21,24-dioxo-26-(3-ureidopropyl)-3,6,9,12,15,18-hexaoxa-22,25-diazaheptacosan-27-amido)­benzyl
(4-Nitrophenyl) carbonate (pMal-PEG

6

-ValCit-PAB-PNP)
pMal-PEG6-ValCit-PAB-OH (125 mg, 0.14 mmol, 1.0 equiv) in a dry flask was
dissolved in 4.9 mL dry DMF at room temperature under Ar atmosphere
and bis­(4-nitrophenyl) carbonate (131 mg, 0.42 mmol, 3.0 equiv) and
DIPEA (0.37 mL, 2.10 mmol, 15.0 equiv) were added. The mixture was
stirred for 3 h, then all volatiles were removed in vacuo. The residue
was purified by silica column chromatography using 10% MeOH in DCM
as the eluent. The product-containing fractions were combined and
15 mL toluene were added before concentrating them in vacuo. pMal-PEG6-ValCit-PAB-PNP was obtained as a white waxy solid in 79%
yield (123 mg containing 5% toluene residue, 0.11 mmol). The molar
ratio of exo/endo was 1.00:0.14, according to NMR. 1H NMR
(500 MHz, DMSO-d
6): δ 10.05 (s,
1H), 8.34–8.29 (m, 2H), 8.13 (d, J = 7.4 Hz,
1H), 7.87 (d, J = 8.6 Hz, 1H), 7.65 (d, J = 8.5 Hz, 2H), 7.59–7.54 (m, 2H), 7.41 (d, J = 8.6 Hz, 2H), 6.36 (s, 1.61H, CH
mal‑exo), 6.22 (s, 0.23H, CH
mal‑endo),
5.98 (t, J = 5.7 Hz, 1H), 5.41 (s, 2H), 5.24 (s,
2H), 4.38 (dt, J = 7.8, 5.6 Hz, 1H), 4.23 (dd, J = 8.4, 6.9 Hz, 1H), 3.64–3.56 (m, 2H), 3.53–3.42
(m, 24H), 3.27 (s, 0.27H, CH
endo), 3.08–2.90
(m, 2H), 2.88 (s, 1.62H, CH
exo), 2.47
(t, J = 7.3 Hz, 1H, overlap with DMSO signal), 2.39
(t, J = 6.3 Hz, 1H), 1.96 (sxt, J = 6.8 Hz, 1H), 1.76–1.55 (m, 2H), 1.63 (s, 0.75H, CH
3,endo), 1.53 (s, 4.98H, CH
3,exo), 1.50–1.32 (m, 2H), 0.86 (d, J = 6.8 Hz, 3H), 0.83 (d, J = 6.8 Hz, 3H) ppm. Ca.
5% of deprotected maleimide species could be observed at 7.02 ppm.
MS (m/z) calcd C50H69N7O18: (M + Na)+, 1078.46;
found, 1078.45.

4-((23S,26S)-1-(4,7-Dimethyl-1,3-dioxo-1,3,3a,4,7,7a-hexahydro-2H-4,7-epoxyisoindol-2-yl)-23-isopropyl-21,24-dioxo-26-(3-ureidopropyl)-3,6,9,12,15,18-hexaoxa-22,25-diazaheptacosan-27-amido)­benzyl
(2-((2-Acrylamido-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­ethyl)­(methyl)­carbamate
(pMal-PEG

6

-ValCit-PAB-OsiNHMe)
pMal-PEG6-ValCit-PAB-PNP (115 mg, 0.10 mmol,
1.0 equiv) and OsiNHMe (103 mg, 0.13 mmol, 1.3 equiv) in a dry flask
were dissolved in 5 mL dry DMF at room temperature under Ar atmosphere
and DIPEA (91 μL, 0.52 mmol, 5.0 equiv) was added. The mixture
was stirred for 2 h, then all volatiles were removed in vacuo. The
residue was purified by silica column chromatography using 10% MeOH
in DCM as the eluent. The product-containing fractions were combined,
concentrated in vacuo and the residue triturated in a mixture of Et2O and EtOAc (1:3). The supernatant was decanted and the residue
dried in vacuo. pMal-PEG6-ValCit-PAB-OsiNHMe was obtained
as a yellow sticky solid in 71% yield (109 mg containing 5% solvent
residue from DCM and EtOAc, 0.074 mmol). The molar ratio of exo/endo
was 1.00:0.12, according to NMR. 1H NMR (500 MHz, DMSO-d
6): δ 9.98 (s, 1H), 9.12 (d, J = 19.7 Hz, 1H), 8.93 (s, 1H), 8.61 (s, 1H), 8.31 (d, J = 5.3 Hz, 1H), 8.26 (d, J = 7.6 Hz, 1H), 8.11 (d, J = 7.4 Hz, 1H), 8.00–7.87 (m, 1H), 7.86 (d, J = 8.6 Hz, 1H), 7.62–7.54 (m, 2H), 7.52 (d, J = 8.2 Hz, 1H), 7.32–7.25 (m, 2H), 7.26–7.21
(m, 2H), 7.16 (t, J = 7.5 Hz, 1H), 6.95 (d, J = 22.4 Hz, 1H), 6.64 (dd, J = 16.8, 10.3
Hz, 1H), 6.35 (s, 1.64H, CH
mal‑exo), 6.24 (d, J = 18.0 Hz, 1H), 6.21 (s, 0.19H, CH
mal‑endo), 5.97 (t, J = 5.6 Hz, 1H), 5.75–5.68 (m, 1H), 5.40 (s, 2H), 4.99 (s,
2H), 4.41–4.34 (m, 1H), 4.22 (dd, J = 8.4,
6.9 Hz, 1H), 3.90 (s, 3H), 3.85 (d, J = 16.4 Hz,
3H), 3.62–3.56 (m, 2H), 3.52–3.43 (m, 24H), 3.40 (t, J = 6.4 Hz, 2H), 3.27 (s, 0.27H, CH
endo), 3.08–2.90 (m, 4H), 2.88 (s, 1.65H, CH
exo), 2.85–2.76 (m, 3H), 2.67 (d, J = 24.8 Hz, 3H), 2.46 (t, J = 7.2 Hz, 1H, overlap
with DMSO signal), 2.38 (t, J = 6.5 Hz, 1H), 1.99–1.92
(m, 1H, overlap with EtOAc signal), 1.74–1.55 (m, 2H), 1.62
(s, 0.65H, CH
3,endo), 1.53 (s, 5.11H,
CH
3,exo), 1.48–1.30 (m, 2H), 0.85
(d, J = 6.7 Hz, 3H), 0.82 (d, J =
6.8 Hz, 3H) ppm. Ca. 7% of deprotected maleimide species could be
observed at 7.02 ppm. MS (m/z) calcd
C71H95N13O17: (M + Na)+, 1424.69; found, 1424.69.

4-((23S,26S)-1-(2,5-Dioxo-2,5-dihydro-1H-pyrrol-1-yl)-23-isopropyl-21,24-dioxo-26-(3-ureidopropyl)-3,6,9,12,15,18-hexaoxa-22,25-diazaheptacosan-27-amido)­benzyl
(2-((2-Acrylamido-5-methoxy-4-((4-(1-methyl-1H-indol-3-yl)­pyrimidin-2-yl)­amino)­phenyl)­(methyl)­amino)­ethyl)­(methyl)­carbamate
(Mal-PEG-ValCit)
pMal-PEG6-ValCit-PAB-OsiNHMe
(97 mg, 0.066 mmol) was dissolved in 1.7 mL DMSO and stirred at 90
°C for 3 h. All volatiles were removed in vacuo and the residue
was purified by preparative HPLC on an XBridge BEH C18 OBD Prep Column
(19 mm × 250 mm) using a mixture of 37% CH3CN and
63% Milli-Q water with 0.1% TFA as the eluent. The product fractions
were combined, CH3CN removed in vacuo and the aqueous phase
lyophilized twice. Mal-PEG-ValCit was obtained as a yellow solid in
64% yield (62 mg, 0.042 mmol). 1H NMR (500 MHz, DMSO-d
6): δ 10.00 (s, 1H, H36), 9.91–8.89
(m, 2H, H14 + H22), 8.76 (s, 1H, H1), 8.45 (s, 1H, H20), 8.36–8.16
(m, 2H, H7 + H12), 8.13 (d, J = 7.4 Hz, 1H, H39),
7.89 (d, J = 8.6 Hz, 1H, H42), 7.65–7.49 (m,
3H, H4 + H34), 7.36 (d, J = 6.4 Hz, 1H, H11), 7.32–7.24
(m, 3H, H5 + H33), 7.22–7.14 (m, 1H, H6), 7.02 (s, 2H, H48),
7.07–6.92 (m, 1H, H17), 6.66 (dd, J = 16.9,
10.3 Hz, 1H, H24), 6.21 (d, J = 17.1 Hz, 1H, H25),
6.02 (s, 1H, H52), 5.71 (d, J = 9.7 Hz, 1H, H25),
5.44 (bs, 2H, H54), 4.98 (s, 2H, H31), 4.41–4.34 (m, 1H, H38),
4.25–4.20 (m, 1H, H41), 3.92 (s, 3H, H2), 3.82 (d, J = 18.2 Hz, 3H, H21), 3.63–3.53 (m, 4H, H45 + H46),
3.52–3.40 (m, 24H, H28 + HPEG‑linker), 3.18–3.05
(m, 2H, H27), 3.05–2.90 (m, 2H, H51), 2.80 (d, J = 14.7 Hz, 3H, H29), 2.74 (d, J = 22.9 Hz, 3H,
H26), 2.48–2.34 (m, 2H, H44), 1.96 (sxt, J = 6.7 Hz, 1H, H55), 1.74–1.53 (m, 2H, H49), 1.49–1.30
(m, 2H, H50), 0.85 (d, J = 6.7 Hz, 3H, H56), 0.82
(d, J = 6.7 Hz, 3H, H56) ppm. 13C NMR
(126 MHz, DMSO-d
6): δ 171.21 (C40),
170.96 (C47), 170.64 (C37), 170.35 (C43), 165.50 (only in 2D, C10),
163.07 (C23), 158.95 (C53), 158.22 (q, J = 33.8 Hz,
Cq-TFA), 155.69 + 155.42 (C30 + C30′), 148.98 (only
in 2D, C16), 147.17 (only in 2D, C12), 138.68 (C35), 138.09 (C3),
137.46 (C1), 134.59 (C48), 132.09 (C24), 131.63 (C32), 128.46 (C33),
126.58 (C25), 125.42 (C8), 125.13 + 124.99 (C19 + C19′), 123.07
(C5), 122.16 (C6 + C7), 119.01 (C20 + C34), 116.44 (q, J = 295.8 Hz, CF3-TFA), 111.84 (C9), 111.05 (C4), 106.20
(C11), 104.87 + 104.65 (C17 + C17′), 69.79 + 69.71 + 69.65
+ 69.50 + 69.42 (CPEG‑linker), 66.95 (C45 + CPEG‑linker), 66,10 (C31), 57.51 (C41), 55.95 (C21),
53.25 + 52.73 (C27 + C27′), 53.15 (C38), 46.55 + 45.90 (C28
+ C28′), 41.84 + 41.51 (C26 + C26′), 38.59 (C51), 36.81
(C46), 35.93 (C44), 34.47 + 34.00 (C29 + C29′), 33.52 (C2),
30.63 (C55), 29.27 (C49), 26.88 (C50), 19.21 (C56), 18.15 (C56) ppm.
The CF3-TFA, C13, C15 and C18 signals could not be observed.
HRMS (m/z) calcd C65H87N13O16: (M + H)+, 1306.6466;
found, 1306.6464. Elemental analysis (%) Calcd for C65H87N13O16*1.5TFA: C, 55.28; H, 6.04; N,
12.32. Found: C, 55.15; H, 6.07; N, 12.12.

HPLC/LC–MS
Stability Measurements
200 μM
stock solutions of Mal-Pip-ValCit and Mal-Pip-GlyGly in DMSO were diluted 1:20 (v/v) with 10 mM PB at pH 7.4, yielding
a final compound concentration of 10 μM and a total content
of 5% (v/v) DMSO. The samples were incubated at 37 °C and measured
over 25 h. HPLC runs were performed using a Waters Acquity UPLC BEH
C18 column (130Å, 1.7 μm, 3 mm × 50 mm) on a Dionex
Thermo Scientific UltiMate 3000 HPLC system, equipped with a SRD-3400
degasser, an HPG-3400RS binary pump, a WPS-3000TBRS autosampler, a
TCC-3000SD column oven and a DAD-3000 UV–vis detector. Chromatograms
were evaluated at a wavelength of 250 nm. Milli-Q water (mobile phase
A) and acetonitrile (mobile phase B), both containing 0.1% TFA, were
used as eluents. The flow rate was consistent at 0.6 mL/min for all
measurements using the following gradient: 0–0.5 min A 75:25
B, 0.5–5.5 min linear gradient to A 45:55 B, 5.5–6.5
min A 5:95 B, 6.5–9.2 min A 75:25 B. LC–MS analyses
of the same samples at the 4 h time points were conducted on a Waters
Acquity UPLC BEH C18 column (130 Å, 1.7 μm, 3 mm ×
50 mm) on an Agilent 1260 Infinity II system equipped with a Flexible
pump, a 1260 VWD UV–vis detector and the LC-MSD system. Chromatograms
were evaluated at a wavelength of 250 nm. Milli-Q water (mobile phase
A) and acetonitrile (mobile phase B), both containing 0.1% formic
acid, were used as eluents. The flow rate was consistent at 0.6 mL/min
for all measurements using the following gradient: 0–0.5 min
A 75:25 B, 0.5–5.5 min linear gradient to A 45:55 B, 5.5–5.6
min linear gradient to A 5:95 B, 5.6–7.5 min A 5:95 B, 7.5–8.0
min linear gradient to A 75:25 B.

Cathepsin B Cleavage Assay
Cathepsin B form human liver
was purchased from Calbiochem (cat: 219362). The cleavage of OsiNHMe was monitored via HPLC on the same instrument and
column used for HPLC stability measurements (see above). Chromatograms
were evaluated at a wavelength of 250 nm. Milli-Q water (mobile phase
A) and acetonitrile (mobile phase B), both containing 0.1% TFA, were
used as eluents. The flow rate was consistent at 0.6 mL/min for all
measurements using the following gradient: 0–0.5 min A 95:5
B, 0.5–5.5 min linear gradient to A 5:95 B, 5.5–6.5
min A 5:95 B, 6.5–9.2 min A 95:5 B. LC–MS analyses of
enzyme-free samples were conducted on the same instrument and column
used for HPLC stability measurements (see above). Chromatograms were
evaluated at a wavelength of 220 nm. Milli-Q water (mobile phase A)
and acetonitrile (mobile phase B), both containing 0.1% formic acid,
were used as eluents. The flow rate was consistent at 0.6 mL/min for
all measurements using the following gradient: 0–0.5 min A
95:5 B, 0.5–6–0 min linear gradient to A 5:95 B, 6.0–7.0
min A 5:95 B, 7.0–7.1 min linear gradient to A 95:5 B. In the
case of preincubation of Mal-Pip-ValCit with HSA, the
same instrument, column, eluent and gradient as for “HPLC albumin
binding studies” was used (see below).
The following
solutions were freshly prepared. Activation buffer: 30 mM DTT and
15 mM EDTA-Na2 in Milli-Q water. Procedure without albumin
preincubation: 5 μL of the cathepsin B solution were mixed with
10 μL of activation buffer and incubated at room temperature
for 15 min. 2 min before the first HPLC-injection, the enzyme mixture
was diluted with 1.5 mL 25 mM NaOAc buffer containing 1 mM EDTA. Then
7 μL of 10 mM Mal-Pip-ValCit or Mal-Pip-GlyGly in Milli-Q water were added, yielding a final compound concentration
of 46 μM, and the samples incubated in the autosampler at 37
°C. Enzyme-free controls were prepared using 5 μL H2O instead of cathepsin B. Procedure with albumin preincubation:
5 μL of 20 mM Mal-Pip-ValCit in DMSO was mixed
with 15 μL H2O, then 980 μL of ∼300
μM HSA (200 g/L Albunorm diluted 1:10 with 10 mM PB, pH 7.4)
were added and the sample incubated at 37 °C for 3.5 h. Albumin
binding was monitored via HPLC. 5 μL of the cathepsin B solution
were mixed with 10 μL of activation buffer and incubated at
room temperature for 15 min. 2 min before the first HPLC-injection,
the enzyme mixture was diluted with 807 μL 100 mM NaOAc buffer
containing 2 mM EDTA. Then 700 μL of the HSA-Mal-Pip-ValCit
solution were added, yielding a final compound concentration of 46
μM, and the samples incubated in the autosampler at 37 °C.

HPLC Albumin-Binding Studies
Incubation with albumin:
2.5 mM stock solutions of Mal-Pip-ValCit and Mal-Pip-GlyGly in Milli-Q water with 25% DMSO were added (1:50 v/v) to ∼300 μM HSA (200 g/L albunorm
diluted 1:10 with 150 mM PB, pH 7.4) and incubated at 37 °C.
Preincubation with NAC: 10 μL of 5 mM N-acetylcysteine
(NAC) in 150 mM PB (pH 7.4) were combined with 100 μL of 0.5
mM Mal-Pip-ValCit in Milli-Q water containing 5% DMSO
and incubated at room temperature for ∼10 min. Then, 11 μL
of this solution were combined with 89 μL of ∼300 μM
HSA (200 g/L Albunorm diluted 1:10 with 150 mM PB, pH 7.4) and incubated
at 37 °C. All samples had a final compound concentration of 50
μM with 0.5% v/v DMSO. The samples were analyzed via HPLC. RP-HPLC
runs were performed using a Waters Acquity UPLC Peptide BEH C18 column
(300 Å, 1.7 μm, 2.1 mm × 100 mm) on a Dionex Thermo
Scientific UltiMate 3000 HPLC system, equipped with a SRD-3400 degasser,
an HPG-3400RS binary pump, a WPS-3000TRS autosampler, a TCC-3000RS
column oven and a DAD-3000RS UV–vis detector. Chromatograms
were evaluated at a wavelength of 300 nm. Milli-Q water (mobile phase
A) and acetonitrile (mobile phase B), both containing 0.1% TFA, were
used as eluents. The flow rate was consistent at 0.4 mL/min for all
measurements using the following gradient: 0–0.5 min A 68:32
B, 0.5–6.5 min linear gradient to A 5:95 B, 6.5–7.6
min A 5:95 B, 7.6–7.7 min linear gradient to A 68:32 B, 7.7–9.2
min A 68:32 B.

UV–Vis Albumin-Binding Studies
To remove N-acetyl-dl-tryptophan, caprylic
acid and NaCl
from Albunorm, 0.5 mL of a 200 g/L albunorm solution were purified
via size-exclusion chromatography (Sephadex G25, purchased from Cytiva)
using Milli-Q water as the eluent and finally lyophilized. The interaction
of the compounds Mal-Pip-GlyGly, Mal-Pip-ValCit and OsiNHMe with the Cys34 residue of HSA
was investigated via the DTDP method described in our former work. The available Cys thiol content in HSA was determined
to be 33 ± 3%. Compound binding was tested according to the following
setup: ca. 6.6 μM Cys34 thiol (20 μM HSA) and
various equivalents of compound (0, 0.25, 0.5, 0.75, 1.0, 1.5 equiv)
were incubated for 20 min at pH 7.40 (100 mM PB) and 37 °C and
the UV–vis spectra were recorded (a), then 33 μM DTDP
(from 4 mM acidic stock solution) was added and the UV–vis
spectra were measured after another 2 min waiting time (b). Spectrum
(a) was subtracted from spectrum (b) and the obtained difference spectrum
was deconvoluted to the sum of the colored reaction product 2-thiopyridone
(2-TP) and the excess DTDP with the solver add-in of MS Excel in the
wavelength range 310–400 nm. The calculated concentration of
2-TP (ε342nm = 7870 M–1×cm–1, own data obtained from reaction with reduced glutathione)
corresponds to the concentration of free Cys34 thiols in
the sample. Time dependence of the maleimide–Cys34 interaction was investigated for Mal-Pip-GlyGly: HSA
and the compound were mixed in a 1:1 ratio (ca. 6 μM) and DTDP
was added after 5, 10, 20, 40, 60, and 90 min waiting time. Afterward,
the experiment proceeded as described above. UV–Vis spectra
were recorded between 200 and 600 nm on an Agilent Cary 3500 spectrophotometer
equipped with an eight-position multicell holder unit.

Osimertinib
and OsiNHMe Recovery
2 mM DMSO stock solutions
of osimertinib or OsiNHMe in were added (1:10 v/v) to
either mouse serum, a solution of ∼600 μM HSA (200 g/L
albunorm diluted 1:5 with 150 mM PB, pH 7.4) or Milli-Q water and
incubated at 37 °C. At 0, 2, 4 and 24h, 20 μL aliquots
were sampled and 60 μL acetonitrile were added. The samples
were vortexed for 5 s, kept at 4 °C for 10 min and then centrifuged
at 4 °C for 10 min at 10,000 rpm. 40 μL of the supernatant
were diluted 1:1 v/v with 0.1% TFA
and the samples analyzed by HPLC on the same instrument and column
used for HPLC stability measurements (see above). Chromatograms were
evaluated at a wavelength of 250 nm. Milli-Q water (mobile phase A)
and acetonitrile (mobile phase B), both containing 0.1% TFA, were
used as eluents. The flow rate was consistent at 0.6 mL/min for all
measurements using the following gradient: 0–0.5 min A 95:5
B, 0.5–5.5 min linear gradient to A 5:95 B, 5.5–6.5
min A 5:95 B, 6.5–8.0 min A 95:5 B.

Analysis of Covalent HSA–Osimertinib
Conjugate
A 0.5 mM DMSO stock solution of osimertinib was
added (1:10 v/v) to a solution of
∼1.5 mM HSA
(200 g/L Albunorm diluted 1:1 with 100 mM PB, pH 7.4) and incubated
at 37 °C for 24 h. 0.5 mL of the albumin solution were purified
using Sephadex G25 Fine, according to the same protocol used for Albunorm
purification (see above). A small aliquote of the solid albumin sample
was dissolved in 10 μL of 2% acetonitrile in H2O
and 1 μL injected onto a Waters BioResolve Reversed-Phase Monoclonal
Antibody Polyphenyl analytical column (2.1 × 150 mm, 2.7 μm,
450 Å) using a Vanquish Horizon UHPLC system coupled to an Orbitrap
QExactive mass spectrometer (Thermo Fisher Scientific) at the Mass
Spectrometry Centre of the University of Vienna. The column oven temperature
was set to 40 °C. H2O (mobile phase A) and acetonitrile/H2O 9:1 (mobile phase B), both containing 0.1% FA, were used
as eluents. The flow rate was consistent at 0.4 mL/min using the following
gradient: 0–20.0 min linear gradient from A 85:15 B to A 40:60
B, 20.0–20.3 min linear gradient to A 10:90 B, 20.3–21.3
min A 10:90 B, 21.3–21.6 min linear gradient to A 85:15 B,
21.6–26.0 min A 85:15 B. The high-resolution MS spectra were
recorded in positive ion mode in the range m/z 700–3000 at a resolution of 17,500 (fwhm at 200m/z). The following HESI ion source settings
were applied: electrospray voltage: 3.5 kV, ion transfer capillary
temperature: 275 °C. In-source collision induced dissociation
(isCID 40 eV) was used to remove noncovalently bound solvents and
salt adducts. Charge state determination and deconvolution of ESI
mass-to-charge ratio spectra and determination of the molecular mass
of the sample was performed using the MagTran software program.

Cell-Free Kinase Screening
The EGFR
kinase-inhibitory
potential of OsiNHMe, OsiNH

2
, OsiPropNHMe and osimertinib was evaluated against
the double mutant EGFR T790M/L858R kinase using the SelectScreen Biochemical
Kinase Profiling Service at Life Technologies (ThermoFisher Scientific,
Madison, WI). The compounds were screened in duplicate using the Z′-LYTE
Assay with a final DMSO concentration of 1%, in the presence of ATP
levels at the apparent ATP-Km of the respective kinase.
The EGFR kinase-inhibitory potential of Mal-Pip-ValCit and Mal-Pip-GlyGly was evaluated against the double-mutant
EGFR T790M/L858R kinase using the Kinase Screening Assay Services
at Reaction Biology (Freiburg, Germany). The compounds were screened
in singlicate using the 33PanQinase Activity Assay with
a final DMSO concentration of 1%, in the presence of ATP levels at
the apparent ATP-Km of the respective kinase.

Molecular
Docking
Molecular docking studies were performed
using AutoDock 4.2. The crystal structure
of the EGFR kinase domain mutant “TMLR” (PDB ID: 5CAS) was retrieved from
the Protein Data Bank and used as the receptor model. Preparation
of both the receptor and ligands was carried out using Molecular Graphics
Laboratory (MGL) of The Scripps Research Institute (version 1.5.7).
Ligand structures were energy-minimized and optimized using LigPrep
(Schrödinger Release 2025–2: LigPrep, Schrödinger,
LLC, New York, NY, 2025). A grid box was defined to fully encompass
the binding pocket of the protein and to allow for exploration of
potential ligand-receptor interactions. The grid dimensions were set
to 100 × 100 × 100 points with a spacing of 0.375 Å.
Grid centers were based on the center of mass of the cocrystallized
ligand, with coordinates: x = −51.907, y = −0.508, and z = −23.173
(PDB ID: 5CAS). Docking calculations were carried out using the Lamarckian Genetic
Algorithm. The resulting binding affinities were expressed as estimated
free energies of binding in kcal/mol. To validate the docking protocol,
the original cocrystallized ligand in 5CAS was removed and redocked
into the binding site. The resulting pose showed excellent overlap
with the experimental structure, confirming the reliability of the
docking setup. Molecular graphics were performed with UCSF ChimeraX,
developed by the Resource for Biocomputing, Visualization, and Informatics
at the University of California, San Francisco, with support from
National Institutes of Health R01-GM129325 and the Office of Cyber
Infrastructure and Computational Biology, National Institute of Allergy
and Infectious Diseases.

Molecular
dynamics (MD) simulations were performed using Desmond, as implemented in the Schrödinger suite. All the investigated systems were solvated in
a cubic water box with dimensions of 15 Å × 10 Å ×
10 Å using the TIP3P water model. First, the total charge of
the system was neutralized, and then additional Na+ Cl– ions were added to achieve a physiological salt concentration
of 0.150 M. Simulations were carried out under NPT ensemble conditions at a temperature of 300 K and a pressure of
1 atm. Prior to the production run, the system was relaxed using Desmond’s
default relaxation protocol to remove unfavorable contacts and equilibrate
the system. The production MD simulation was run for a total of 100
ns, with trajectory frames recorded every 100 ps, resulting in approximately
1000 frames. The simulations were performed using the OPLS4 force
field, as implemented in Desmond. Structural
stability and conformational changes of the protein-ligand complexes
were evaluated through trajectory analysis, including root-mean-square
deviation (RMSD) calculations.

Cell Culture
Cell-based
experiments were conducted
using the human NSCLC cell lines H1975 (CRL-5908), H1650 (CRL-5883),
HCC827 (CRL-2868), human epidermoid carcinoma cell line A431 (CRL-1555),
human colorectal carcinoma cell line HCT116 (CCL-247), human clear
cell renal carcinoma cell line Caki-1 (HTB-46), and healthy murine
fibroblasts (derived from C57BL/6 mouse in-house). H1975, H1650 and
HCC827 cells were cultured in RPMI-1640 media, A431 cells were cultured
in Dulbecco’s modified eagle’s medium (DMEM), all supplemented
with 10% fetal bovine serum (FBS). HCT116 and Caki-1 cells were cultured
in McCoy’s 5A media supplemented with 10% FBS and 1% l-glutamine. Healthy murine fibroblasts were cultured in DMEM supplemented
with 10% FBS and 1% l-glutamine. All cells apart from murine
fibroblasts were obtained from the American Type Culture Collection
(ATCC, Manassas, VA, USA) and maintained at 37 °C in a humidified
incubator with 5% CO2. Mycoplasma contamination was routinely tested for, and all cell lines were
used within ten passages from thawing to ensure consistency and reliability
of experimental outcomes.

MTT Viability Assay
Cells were seeded
in 96-well plates
at a density of 3–4 × 104 cells per well in
100 μL of medium and allowed to adhere overnight at 37 °C
in a humidified atmosphere containing 5% CO2. Test compounds
were initially dissolved in DMSO to prepare 10 mM stock solutions
and subsequently diluted in growth medium, ensuring a final DMSO concentration
below 1%. After 24 h, cells were exposed to 100 μL of compound
dilutions in triplicate, at final concentrations ranging from 0 to
10 μM. Following a 72 h incubation at 37 °C and 5% CO2, cell viability was assessed using the MTT assay (EZ4U, Biomedica,
Vienna, Austria), according to the manufacturer’s instructions.

Western Blotting
Western blot analysis was used to
evaluate the expression levels of Cathepsin B in H1975, H1650 and
Caki-1 cell lines as described previously.

Immunohistochemistry
Formalin-fixed paraffin-embedded
tissue sections were incubated at 65 °C for 10 min, deparaffinized,
and rehydrated. For cathepsin B immunohistochemistry, antigen retrieval
was performed by boiling the sections in 10 mM citrate buffer (pH
6.0) for 30 min. Sections were then incubated overnight at 4 °C
in a humidified chamber with the primary antibody (Cathepsin B (D1C7Y)
XP Rabbit mAb or Albumin antibody PA5–85166, Thermo Fisher).
Detection of antibody binding was carried out using the UltraVision
LP Large Volume Detection System HRP Polymer (Lab Vision), following
the manufacturer’s protocol (Thermo Fisher Scientific Inc.).
Signal development was performed using liquid DAB and substrate chromogen
system (DAB; Dako). For hematoxylin and eosin (H/E) staining, rehydrated
sections were subjected to nuclear staining with Harris’ hematoxylin
(Merck) for 2 min. Slides were then blued in Scott’s solution
(Morphisto) for 45 s. Cytoplasmic counterstaining was performed with
Eosin Y (Sigma-Aldrich) for 1 min. Slides were subsequently dehydrated
through graded ethanol, cleared in n-butyl acetate,
and mounted with Entellan mounting medium (Merck). All stained slides
were imaged using standard laboratory imaging systems. Images were
prepared for presentation using SlideViewer (version 2.9.0.229983),
and analyses were performed in QuPath (version 0.6.0).

Animals
A total of 57 C.B.17SCID/SCID mice
(8–16 weeks old purchased from Javier) were utilized for the in vivo experiments. Animals were housed under specific
pathogen-free (SPF) conditions, and all procedures were conducted
within a laminar airflow cabinet to maintain sterility. All animal
experiments were carried out in accordance with the guidelines and
approvals of the Ethics Committee for the Care and Use of Laboratory
Animals at the Medical University of Vienna (Ethic number: 2022–0.770.291).
Animal well-being was carefully monitored throughout the study, with
daily assessments of body weight and clinical signs of distress.

Therapy Experiments
For the xenograft experiments,
the cancer cells were injected subcutaneously into the right flank
of mice at the concentration of 1 × 106 per mouse.
Therapy started at a mean tumor volume of 100–150 mm3. For the H1975 experiment, osimertinib mesylate (29.8 mg/kg), OsiNHMe (37.0 mg/kg), OsiNH

2
(36.4 mg/kg) and OsiPropNHMe (43.0 mg/kg) were
administered per os, all formulated in 0.5% hydroxypropyl methyl cellulose.
Control group received 0.5% hydroxypropyl methyl cellulose alone.
All groups received drugs five times a week for 2 weeks.
In
the H1650 experiment, Mal-Pip-ValCit and Mal-Pip-GlyGly were administered intravenously at a dose of 83.7 mg/kg and 75.7
mg/kg respectively, while osimertinib mesylate was administered per
os at equimolar doses of 29.8 mg/kg. Both Mal-Pip-ValCit and Mal-Pip-GlyGly were formulated in a vehicle consisting
of 20% propylene glycol (PG), 45% of a 5% glucose and 35% murine blood
serum, while osimertinib was formulated in 0.5% hydroxypropyl methyl
cellulose. Control animals received the 20% PG dissolved in 5% glucose
vehicle solution (i.v.). Drugs and control solution were administered
two times per week for 2 weeks. Tumor dimensions were measured daily
using calipers. Subsequently, the tumor volume was estimated using
the established modified ellipsoidal formula: V =
(1/2) × L × W
2, where L is the tumor length and W the width. At the conclusion of the experiment, animals were sacrificed
by cervical dislocation, and tumors and organs were collected, fixed
in 4% formaldehyde (Carl Roth) for 24 h, and embedded in paraffin
using a KOS tissue processor (Milestone).

Supplementary Material