Unraveling the Interaction Between Carboxylesterase 1C and the Antibody- Drug Conjugate SYD985: Improved Translational PKPD by Using Ces1c Knockout Mice

Author Manuscript Published OnlineFirst on August 9, 2018; DOI: 10.1158/1535-7163.MCT-18-0329 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Unraveling the interaction between carboxylesterase 1c and the antibody- drug conjugate SYD985: improved translational PKPD by using Ces1c knockout mice

Ruud Ubink(1), Eef H.C. Dirksen(1), Myrthe Rouwette(1), Ebo S. Bos(1), Ingrid Janssen(1),

David F. Egging(1), Eline M. Loosveld(1), Tanja A. van Achterberg(1), Kim Berentsen(1),

Miranda M.C. van der Lee(1), Francis Bichat(2), Olivier Raguin(2), Monique A.J. van der

Vleuten(1), Patrick G. Groothuis(1), Wim H.A. Dokter(1).

Authors’ Affiliations: (1)Department of Preclinical, Synthon Biopharmaceuticals BV,

Nijmegen, the Netherlands. (2)Oncodesign, Dijon, France.

Running title: CES1c knockout mice for ADC development

Corresponding Author: Ruud Ubink, Synthon Biopharmaceuticals BV, Microweg 22, PO Box

7071, 6503 GN Nijmegen, the Netherlands. Tel: +31 (0)24 372 7700, email: [email protected]

Disclosure of potential conflict of interest: All authors are employees of Synthon

Biopharmaceuticals BV except for Francis Bichat and Olivier Raguin, who are employees of

Oncodesign

Financial support for this study was provided by Synthon Biopharmaceuticals BV.

Downloaded from mct.aacrjournals.org on September 29, 2021. © 2018 American Association for Cancer Research. Author Manuscript Published OnlineFirst on August 9, 2018; DOI: 10.1158/1535-7163.MCT-18-0329 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Abstract

Carboxylesterase 1c (CES1c) is responsible for linker-drug instability and poor pharmacokinetics

(PK) of several antibody-drug conjugates (ADCs) in mouse, but not in monkey or human.

Preclinical development of these ADCs could be improved if the PK in mice would more closely resemble that of human and is not affected by an enzyme that is irrelevant for humans. SYD985, a HER2-targeting ADC based on trastuzumab and linker-drug vc-seco-DUBA is also sensitive to

CES1c. In the present studies, we first focused on the interaction between CES1c and SYD985

by size- exclusion chromatography, Western blotting and LC-MS/MS analysis, using

recombinant CES1c and plasma samples. Intriguingly, CES1c activity not only results in release

of the active toxin DUBA but also in formation of a covalent bond between CES1c and the linker

of vc-seco-DUBA. Mass spectrometric studies enabled identification of the CES1c cleavage site on the linker-drug and the structure of the CES1c adduct. To assess the in vivo impact, CES1c -/-

SCID mice were generated which showed stable PK for SYD985, comparable to that in monkey

and human. Patient-derived xenograft (PDX) studies in these mice showed enhanced efficacy

compared to PDX studies in CES1c +/+ mice and provided a more accurate prediction of clinical

efficacy of SYD985, hence delivered better quality data. It seems reasonable to assume that

CES1c -/- SCID mice can increase quality in ADC development much broader for all ADCs that

carry linker-drugs susceptible to CES1c, without the need of chemically modifying the linker-

drug to specifically increase PK in mice.

Introduction

Stability of the linker-drug in plasma has been, and still is, a major challenge in the development of antibody-drug conjugates (ADCs). On the one hand, linkers should be stable enough to prevent release of the active toxin in plasma, while on the other hand, linkers should allow for an efficient release of active toxin once the ADC targets a tumor or tumor cell. The difficulty of developing optimal linker characteristics is illustrated by the limited amount of linker-drug technologies in ADCs that have been studied in the clinic (1,2), although novel linker-drug strategies are emerging (for recent reviews see 3, 4, 5, 6). For Mylotarg ®, the first approved

ADC based on a pH-labile hydrazone linker, chemical instability of the linker-drug in the blood circulation probably contributed to the temporary withdrawal from the market (7, 8). Besides early cleavage, other possible causes of instability in plasma include maleimide exchange of the cysteine-conjugated linker-drugs to other thiol-containing molecules circulating in plasma, such as serum albumin (9) and enzymatic instability by circulating enzymes in plasma (10, 11), which can cleave functional groups like esters, amides, lactones, lactams, carbamides, sulfonamides and peptide mimetics (12).

SYD985, a HER2-targeting ADC based on trastuzumab and vc-seco-DUBA, a cleavable linker-duocarmycin payload, was found to be stable in human and monkey plasma, but unstable in mouse and rat plasma (13, 14). Despite its instability in mouse plasma, and its subsequent poor exposure in mice, SYD985 was found to be very active in HER2 3+, 2+ and 1+ breast cancer patient-derived xenograft (PDX) models, whereas T-DM1 only showed significant antitumor activity in HER2 3+ models (15). Although good anti-tumor activity of SYD985 was observed in these mouse models, it was still hypothesized that the poor stability yields an underestimation of the expected anti-tumor activity in humans. Therefore, in parallel to the

clinical development of SYD985, subsequent preclinical studies aimed at i) identifying the cause

of instability of SYD985 in mouse and rat plasma, ii) testing the hypothesis that anti-tumor

activity in mice is improved when plasma stability of SYD985 improves, and iii) delivering a

mouse model that more closely resembles the monkey and human PK to allow for reliable

assessment of the minimal effective dose in humans. In addition, and most ideally, such a mouse

model could be used for other ADC programs as well and deliver more relevant PK/PD data in

preclinical studies right away. As reported earlier, the rodent-specific carboxylesterase 1c

(CES1c) was identified as the crucial enzyme that cleaves vc-seco-DUBA-based ADCs, leading

to the poor PK of SYD985 in mice and rats compared to that observed in human and monkey

(14, 15). After these reports, it was published that CES1c is also responsible for cleavage of

other valine-citrulline-p-aminobenzyloxycarbonyl (vc-PABC) containing linker-drugs in mouse

plasma and that the extent of CES1c-mediated cleavage depends on the site of conjugation (16).

Several additional papers report instability in mouse and rat plasma of vc-PABC or carbamate-

containing linker-drugs (10, 17), indicating CES1c activity might be a more widespread cause of

linker-drug instability in these species, hampering the development of ADCs.

In this paper, we will first present the data that led to the identification of CES1c as the

enzyme responsible for the instability of SYD985, causing release of active toxin. In order to be

complete and set the context right, we needed to recapture a limited amount of data that we

published before, mainly as supplementary data. We will subsequently identify the cleavage site

of CES1c on the linker-drug vc-seco-DUBA, being different from the cleavage site identified by

(16). Furthermore, it was found that CES1c activity not only cleaves the linker-drug, but also

leads to the formation of a covalent bond between CES1c and the remaining linker. To

circumvent the translational issues caused by the CES1c activity, we developed a mouse model

which closely mimics the human PK and is suitable for xenograft studies, by cross-breeding

CES1c -/- mice with SCID mice. Using this mouse model we demonstrate that improved PK

indeed leads to improved efficacy in PDX studies and improved human dose predictions for

SYD985. Potentially, this model would also increase PK/PD relationships of other ADCs that

contain linker-drugs that are susceptible to CES1c cleavage.

Materials and methods

ADCs and related materials

SYD985 was prepared as previously described (15). [3H]-SYD985 was prepared at Pharmaron,

Cardiff, UK, using the same procedure and a linker-drug in which two tritium labels were

incorporated in the hydroxybenzoyl moiety of the toxin. N-acetyl cysteine-quenched linker-drug

(NAc-Cys-linker-drug) was prepared by mixing linker-drug and N-acetyl cysteine in a 1:10

molar ratio in an acetonitrile/water mixture for 4 hours followed by purification by preparative

HPLC. The head-to-head studies comparing T-DM1 with SYD985 were conducted with two

batches of T-DM1 from Roche, EU batch N0001B02 (in nude mice) and N1037B19 (in CES1c -

/- SCID mice).

In vitro plasma stability and cleavage by CES1c

The antibody-drug conjugate SYD985 was spiked into pooled female mouse (BALB/c), rat

(Sprague Dawley), monkey (Macaca fascicularis) and human K2-EDTA plasma, at a

concentration of 100 μg/mL and incubated at 37°C. After 0, 1, 6, 24, 48 and 96 hours of

incubation, plasma samples were snap-frozen and stored at -80°C until bioanalysis. Recombinant

mouse CES1c (rCES1c, Cusabio Biotech, CSB-MP338557MO) was spiked in human K2-EDTA

plasma at 0, 10, 100, 200, and 400 μg/mL) together with 100 μg/mL SYD985. After 96 hours of

incubation at 37°C, plasma samples were snap-frozen in liquid nitrogen and stored at -80°C until bioanalysis.

PK studies and bioanalytical assays

Adult female CES1c -/- (B6-Ces1ctm1.1Loc/J, Charles River) mice, CES1c -/- SCID (B6-

Ces1ctm1.1Loc/J x B6.CB17-Prkdcscid/SzJ, Charles River) mice and MAXF1162 tumor-bearing

CES1c -/- SCID mice, rats (Wistar, Charles River) and monkeys (Macaca fascicularis,

Mauritian) were dosed intravenously with 0.3, 1, 3, 8 and/or 10 mg/kg SYD985. Blood samples were taken at multiple time points after dosing, cooled on ice water, and processed to K2-EDTA

plasma as soon as possible. Plasma samples were snap-frozen in liquid nitrogen and stored at -

80°C until bioanalysis. SYD985 plasma levels were quantified using ELISA-based methods with

anti-idiotype capture and reporting for total antibody, and anti-toxin capture and anti-idiotype

reporting for conjugated antibody, as previously described (15). Based on the reported plasma

levels, PK parameters were calculated in WinNonlin version 6.3.

Release of DUBA in mouse and rat plasma

Human tumor cell line SW-620 (ATCC® Number: CCL-227™, Lot Number: 58483168, P: 86

received on 19SEP2012) was obtained from and characterized by the American Type Culture

Collection. No further cell-line authentication was conducted. Mycoplasma contamination in cell

cultures was tested at Minerva Biolabs GmbH (Berlin, Germany) using the Venor®GeM Prime

test. No detectable levels of mycoplasma were found in the working cell bank p:86+14 tested on

the 11th of April in 2013. The number of passages between collection and use in the described

experiments is 5 and 6. SW-620 cells (90µL/well; 4,000 cells/well) were plated in 96-wells

plates in RPMI-1640 medium (Lonza), containing FBS, heat-inactivated (HI; Gibco-Life

Technologies) and 80U/mL Pen/Strep and incubated at 37°C, 5% CO₂ overnight. After an

overnight incubation 10 µL SYD985 or DUBA was added to each well of the 96-wells plate, containing 10% mouse (BALB/c), CES1c -/- mouse, rat (Wistar), monkey (Macaca fascicularis) or human plasma. Serial dilutions were made in culture medium with plasma, to reach a final

concentration range of 200 nM (10 µg/mL) to 6.3 pM (0.316 ng/mL) for SYD985 and 10 nM to

0.316 pM for DUBA and 1% plasma. The cell viability was measured after 6 days using the

CTG assay kit (Promega, G7572).

Affinity extraction of SYD985-protein complexes

SYD985 was spiked at 100 µg/mL in mouse (BALB/c) and CES1c -/- mouse plasma and

incubated up to 48 or 96 hours at 37°C, followed by storage at -80°C until extraction. For the

T=0 samples, SYD985 was spiked in plasma and immediately frozen. SYD985 and associated

protein complexes were extracted from plasma using an anti-idiotype mini-antibody (Biorad,

AbD15916) coupled via amino coupling to NHS-activated sepharose 4 fast flow (GE

Healthcare). Prior to incubation of the SYD985 plasma samples with the anti-idiotype coupled

sepharose, a 50% ammonium sulfate precipitation step was performed to improve the purity of

the extracted samples be selectively precipitating the IgG fraction, including SYD985. Following

incubation with the anti-idiotype coupled sepharose, the SYD985 sample was transferred to a

disposable column (Thermo Scientific, cat# 29920). After removal of the flow-through,

sequential washes (1 M NaCl in PBS, 10% acetonitrile in PBS and PBS) were performed to

remove non-specifically bound proteins. SYD985 and associated protein complexes were eluted

with 10 mM glycine-HCl, pH 2.5, followed by immediate neutralization with 0.1 M Tris-HCl,

pH 8.5.

SYD985 analysis by SEC

Radiolabeled [3H]-SYD985 was incubated at 100 µg/mL in plasma from female mouse

(BALB/c), CES1c -/- mouse, rat (Wistar), monkey (Macaca fascicularis) and human for 6 hours

at 37°C. The resulting plasma samples were analyzed directly by size exclusion chromatography

(SEC) using a Waters BEH200 SEC column (4.6 x 150 mm, 1.7 µm particles) and isocratic

elution with 100 mM NaPO4, pH 6.8 + 0.3 M NaCl. Eluting compounds were detected using a

Model 4 β-Ram Radiodetector (LabLogic, Sheffield, Yorkshire, UK).

For ‘cold’ (i.e. not radiolabeled) SYD985 that was incubated in plasma of mouse under the

conditions mentioned above, anti-idiotype affinity-extracted material was evaporated to dryness

and reconstituted in 100 mM NaPO4, pH 6.8 + 0.3 M NaCl before analyses by SEC using the

same analytical column and solvent as described for radiolabeled SYD985, but using UV

absorbance detection at 214 nm and 330 nm. Eluting compounds, including the high molecular

weight species, were collected, buffer-exchanged and concentrated into 50 mM Tris-HCl pH 8.0

using 10K centrifugal filters (Amicon). 8M urea was added to denature the proteins, followed by

reduction and alkylation with dithiothreitol (DTT; 4.5 mM final concentration in 50 mM

Tris.HCl, pH 8.5, 37°C, 1 hour) and iodoacetamide (IAM; 9 mM final concentration in 50 mM

Tris.HCl, pH 8.5, 37°C, 1 hour in the dark), respectively. Following an overnight (o/n) precipitation in EtOH, the resulting pellets were digested using trypsin (4 hours, 100 mM

Tris.HCl, pH 8.5).

Coomassie blue staining and Western blot analysis

Samples obtained by affinity extraction were evaporated to dryness, reconstituted in Milli-Q water and electrophoretically separated under non-reducing conditions on a 3-8% Tris-Acetate gel (Thermo scientific) for total protein staining, or on a 4 -12% Bis-Tris gel (Thermo Scientific) for Western blot analysis. Total protein staining was performed with Bio-safe Coomassie stain

(Bio-Rad) according to manufacturer’s instructions. For Western blot analysis, proteins were transferred to a PVDF membrane (GE Healthcare) with a Mini Trans-Blot® Electrophoretic

Transfer Cell (Biorad). The PDVF membrane was blocked with Western blocker solution

(Sigma-Aldrich). SYD985 and associated protein complexes were detected with an HRP-labeled anti-idiotype mini-antibody (Biorad). CES1c was detected with a biotinylated polyclonal rabbit anti-CES1c antibody (Abcam) in combination with streptavidin-HRP (R&D systems). SYD985 and associated protein complexes and CES1c were visualized by using the Metal Enhanced DAB

Substrate Kit (Thermo Scientific).

In-gel protein digestion

Protein bands of interest were excised from the Coomassie-stained SDS-PAGE and washed with

Milli-Q water and acetonitril, to remove Coomassie and SDS before proceeding to the next step.

Following in-gel reduction and alkylation with DTT (10 mM final concentration in 10 mM

Tris.HCl, pH 8.5, 37°C, 30 min) and IAM (20 mM final concentration in 10 mM Tris.HCl, pH

8.5, RT, 45 min in the dark), respectively, proteins were in-gel digested using trypsin (o/n , 10

mM Tris.HCl, pH 8.5).

Incubation of rCES1c with NAc-Cys-linker-drug and SYD985 and subsequent proteolytic

digestion

Recombinant CES1c was incubated with NAc-Cys-linker-drug or SYD985 in phosphate- buffered saline (PBS), at a concentration of 100 µg/mL for 96 hours at 37°C under gentle shaking. The resulting sample was acidified to pH 3.5 using glycine.HCl and reduced and

alkylated using TCEP (10 mM final concentration, 37°C, 30 min.) and N-ethyl maleimide

(NEM; 20 mM final concentration, RT, 45 min. in the dark), respectively. Considering that a

potential (serine) ester of CES1c with NAc-Cys-linker-drug is probably base-labile, all incubations and sample preparations of the resulting CES1c – NAc-Cys-linker-drug complex

were performed at low pH (~ 3). For protein digestion, pepsin was added and the mixture was

incubated at 37°C (pH 3.0, o/n).

LC-MS/MS analysis of proteolytic digests

Proteolytic peptide mixtures were separated by reversed phase liquid chromatography (Dionex

Ultimate 3000) using a Waters BEH-300 C18 column (1.0 x 150 mm, 1.7 µm particles) and a 40’

5-95% linear gradient of acetonitile + 0.1% formic acid (FA) in Milli-Q + 0.1% FA at a flow rate

of 0.1 mL/min and a column temperature of 30 °C. Eluting peptides were detected using

electrospray ionization mass spectrometry (Thermo Orbitrap Fusion, heated electrospray

ionization (HESI) at a spray voltage of 3.8 kV) with MS detection in the Orbitrap at 120K resolution over the scan range m/z 400-1600 and further analyzed using data-dependent tandem

mass spectrometry (MS/MS) in the quadrupole, using HCD (30%), on the top-3 most intense

precursor ions detected per survey scan to enable protein identification. Fragments were detected in the ion trap. Proteins were identified against the Swissprot database using Mascot search engine.

In addition, MS and MS/MS fragmentation data were interpreted manually using Qualbrowser

(Thermo XCalibur 3.0.63) by comparing proteolytic digests of rCES1c with and without NAc-

Cys-linker-drug and proteolytic digests of rCES1c and SYD985 individually to that of the incubated mixtures to investigate and characterize cross-linked peptide candidates.

Patient-derived xenograft studies

The in vivo antitumor activity of SYD985 versus T-DM1 was tested as single-dose therapy in

MAXF1162 breast cancer PDX model (Charles River) in both CES1c +/+ NMRI-Foxn1nu mice

as well as CES1c -/- SCID mice. Initial tumor volumes at the day of randomization and treatment

ranged from 46 to 252 mm3. The HER2 FISH and IHC status of the tumor from the PDX model

as determined by the CROs was independently confirmed. Studies were conducted as previously

described (15). All in vivo studies and protocols were approved by the local animal care and use

committees according to established guidelines.

Results

CES1c expressed in rodent plasma cleaves vc-seco-DUBA

In vitro plasma stability studies, as well as in vivo PK studies, showed that SYD985 conjugated

antibody levels rapidly decrease in mouse and rat plasma, but are stable in monkey and human

plasma (Figure 1A, 1B and S1), similar to what was previously found for closely related ADCs

(13, 14). A literature study pointed towards mouse and rat-specific CES1c as a potential

hydrolyzing enzyme. To assess its role in SYD985 instability, increasing amounts of

recombinant mouse CES1c (rCES1c) were spiked in human plasma in combination with 100

µg/mL SYD985 and incubated at 37°C up to 96 hours. SYD985 conjugated antibody levels

decreased with increasing CES1c concentration (Figure 1C, previously published as

supplementary data; 15). Significant cleavage was observed as of 100 µg/mL CES1c which are

physiologically relevant CES1c concentrations, comparable to the in vivo CES1c concentration

in mice (80 µg/mL)(18). A PK study with SYD985 in CES1c homozygous -/- mice,

heterozygous CES1c +/- and homozygous CES1c +/+ wild-type littermates, showed that the PK

in homozygous CES1c -/- mice was similar to the PK in monkey and human (Figure 1D and

Table S1, previously published as supplementary data in 15).

Mechanism of action of CES1c-mediated vc-seco-DUBA cleavage

To study if CES1c activity results in release of the toxin DUBA, SYD985 was incubated

with HER2-negative cells in the presence or absence of 1% wild type mouse, CES1c -/- mouse,

rat, monkey or human plasma (Figure 1E, Table S2 and (15)). Compared to the medium control,

presence of 1% plasma caused a clear shift in the IC50 for rat (8.08 nM) and especially mouse

(0.82 nM) plasma, but not with monkey, human, or CES1c -/- mouse plasma, all showing an

IC50 > 100 nM.

As part of a metabolism study, radiolabeled SYD985 was analyzed by size exclusion

chromatography (SEC) after incubation in mouse, rat, CES1c -/- mouse, monkey and human

plasma. In SEC profiles of mouse and rat plasma a new peak appeared (Figure 2 B, D), eluting at

an earlier retention time (7.15 minutes) compared to the SYD985 peak (RT = 7.8 - 8.02 minutes).

This new peak was hardly observed in samples taken from CES1c -/- mouse, monkey and human

plasma (Figure 2 A, C and E), suggesting that a SYD985-containing compound with a higher

molecular weight than SYD985 alone was formed in rat and mouse plasma. To identify the

compound(s) eluting in this mouse and rat-specific peak, mouse plasma was incubated with

SYD985 up to 48 hours and SYD985 and SYD985-associated material was subsequently

extracted from plasma using anti-idiotype extraction. SEC profiles of the affinity-extracted

material confirmed the appearance of a SYD985-containing peak, at an earlier retention time as compared to SYD985 itself (Figure 2F). The extract was also analyzed using non-reducing SDS-

PAGE and Coomassie staining, which revealed a unique band pattern around 200 - 250 kDa

(Figure 2G). Although SYD985 has a molecular weight of ~150 kDa, bands were also observed

at lower molecular weight in the extracted samples. This is due to the fact that as a result of

conjugation on interchain cysteines, used to manufacture SYD985, its structural integrity

(partially) relies on non-covalent interactions that are disrupted during SDS-PAGE analysis. The

200 – 250 kDa bands were cut out of the gel, trypsin-digested and analyzed by LC-MS/MS. Next to trastuzumab, CES1c (Uniprot entry P23953) was confidently identified based on 6 unique peptides (9% sequence coverage, Mascot score 337) in the lower migrating band (around 200 kDa), while CES1c was also identified in the higher migrating band (around 250 kDa) based on

12 unique peptides (25% sequence coverage, Mascot score 745). CES1c was also observed by

LC-MS/MS analysis of a direct (i.e. in solution) digest of affinity-extracted material from wild- type mouse plasma following incubation of SYD985 at 37oC for 96 hours. Western blot studies showed the presence of both SYD985 and CES1c in the 200 - 250 kDa bands, since it stained positive for SYD985 using an anti-idiotypic antibody, and for CES1c using an anti-CES1c antibody (Figure 2 H, I), respectively.

Next, it was attempted to identify the CES1c cleavage site(s) on the linker-drug in

SYD985 and the position of the covalent bond between CES1c and SYD985. Covalent interactions between carbamates and esterases have been described for human CES1 and CES2

(19). Based on that work, it was hypothesized that the interaction potentially occurred via the active site serine (Ser221) in CES1c and one of the carbamates present in the linker-drug (Figure

3A). Due to the structural orientation of the carbamates, the covalent bond formation with the

ADC would most likely proceed via the carbamate connecting the linker to the DNA-alkylating moiety of the toxin, resulting in concomitant release of the toxin.

To investigate this hypothesis, rCES1c was first incubated with N-acetylcysteine- quenched linker-drug (NAc-Cys-linker-drug), for 72 hours at 37oC in PBS. Following reduction, alkylation and proteolytic digestion with pepsin (all at pH ~ 3), LC-MS/MS analysis of the generated peptide mixture resulted in the identification of five covalently modified CES1c peptides (Figure 3B, C, D, E and Table 1), all including the active site serine S221. The proposed

covalent interaction was confirmed by the accurate masses of these modified peptides and the

concomitant mass increase of 968.1 Da with respect to the corresponding original (unmodified)

peptide mass as a result of the binding of NAc-Cys-linker-drug (minus the released toxin

DUBA). Furthermore, the characteristic fragment ions observed in the corresponding MS/MS

data, such as the neutral loss of 779.3 Da, corresponding to the majority of the NAc-Cys-linker-

drug modification; the loss of the complete NAc-Cys-linker-drug modification (985.4 Da)

resulting in the formation of a dehydroalanine moiety from the serine residue, and several NAc-

Cys-linker-drug specific fragment ions (Figure 3E) supported the hypothesized structure of the

CES1c-linker-drug cross-link. In addition, free seco-DUBA, the pro-toxin that is released upon linker-drug cleavage and is stable at acidic pH, was also identified in the incubated digest at pH

< 3, confirming the release of the toxin.

Next, SYD985 was incubated with rCES1c in PBS during 72 hours at 37oC. Following the same, low pH, sample work up as described for the NAc-Cys-linker-drug incubated sample, the generated peptide mixture was analyzed using LC-MS/MS. This dataset revealed several candidate cross-linked peptide molecules. These were confirmed using a combination of their

accurate mass and the distinctive fragment ions that were also observed when rCES1c was

incubated with NAc-Cys-linker-drug, corresponding to the rCES1c part of a cross-linked

peptide. Subtracting the mass of the rCES1c sequence (with part of the linker-drug on it), as well

as the remainder of the linker-drug part that was conjugated to the cysteine residue originating

from SYD985, from the precursor mass, revealed one of the cross-linked peptide candidates to

contain the trastuzumab sequence KSFNRGEC (Figure S2), originating from a conjugated

trastuzumab light chain. This confirmed the presence of a SYD985-rCES1c cross-link through

the linker-drug side chain via the proposed mechanism. Finally, this conclusion was underscored

by the identification of a cross-link between a rCES1c peptide and the SYD985 hinge region- containing heavy chain peptide THTC(NEM)PPCPAPELLGGPSVF (Figure S2), that was originally conjugated to one linker-drug. The other cysteine in the sequence was found to be alkylated with N-ethylmaleimide as a result of the sample preparation. Overall, four different rCES1c sequences (all containing the active site serine, S221) were found to be cross-linked to three different peptides originating from the light chain, and to one peptide originating from the hinge region of the heavy chain (Table 2).

Efficacy studies in CES1c knockout mice

Since in vivo CES1c activity could not be completely blocked by carboxylesterase inhibitors

(Figure S3), efficacy studies were performed in CES1c -/- SCID mice. Before efficacy studies were initiated, it was first verified whether the SYD985 PK in CES1c -/- SCID mice was identical to that in the original CES1c -/- strain. As shown in Figure 4A, the PK profiles indeed overlapped. Pilot efficacy studies with CES1c -/- SCID mice using different PDX models (15), showed tumor take and growth rates similar to those observed in the nude mice. In these models, efficacy was indeed improved at 3 mg/kg SYD985 (Figure S4). To confirm the data obtained at a single dose level of 3 mg/kg, and to get an accurate estimate of the fold change in efficacy, an experiment comprising of different dose levels was performed in the HER2 3+ MAXF1162 PDX model. The efficacy of different SYD985 dosages was compared head-to-head against T-DM1, similar to a previous study performed in nude mice (15), to control for potential strain differences between the nude and SCID background. A non-binding isotype control ADC, bearing the same linker-drug and with the same drug-antibody ratio, was included to study the

effect of improved PK on efficacy for a non-target-binding ADC. The efficacy of SYD985 was

improved almost threefold in CES1c -/- SCID mice showing tumor-regression at 3 mg/kg

SYD985 in these animals whereas 10 mg/kg SYD985 induced tumor-stasis in CES1c +/+ nude mice. The isotype control also showed efficacy at 3 mg/kg, in contrast to the lack of efficacy in

nude mice. This difference is most likely also caused by an increased exposure of intact isotype

control ADC. Extracellular cleavage or pinocytosis of the isotype control could explain the

efficacy, which is not uncommon for the isotype control (15). As expected, since T-DM1 is not

susceptible to CES1c cleavage, T-DM1 showed similar efficacy in nude mice versus CES1c -/-

SCID mice, with prolonged tumor stasis at 10 mg/kg.

Recently, phase I clinical data were published describing the efficacy of SYD985 at the

recommended phase 2 dose of 1.2 mg/kg in HER2 positive metastatic breast cancer patients (20,

21). The 1.2 mg/kg dose was shown to be efficacious with an overall response rate (ORR) of

33% (16 out of 48 patients) and a progression-free survival (PFS) of 9.4 months (95% CI: 4.5 -

12.4). The human PK of SYD985 at 1.2 mg/kg (AUCinf = 1725 h.µg/mL) almost overlaps with

the PK observed in MAXF1162 tumor bearing CES1c -/- SCID mice at the 1 mg/kg dose that

shows tumor stasis (AUCinf = 1332 h.µg/mL, Figure 4D and Table S3).

Discussion

Until recently, all xenograft studies with vc-seco-DUBA-based ADCs, including SYD985, were

performed in immune-compromised mice expressing CES1c. Despite the low exposure levels for

conjugated antibody in these mice, SYD985 showed good efficacy, which was superior to that of

T-DM1 in numerous xenograft models with different levels of HER2 target expression (15).

However, since the SYD985 PK profiles in the mouse were dramatically different from those in

monkey and the anticipated profile in human based on in vitro plasma stability data, an accurate

estimation of an effective human dose based on efficacy in PDX models was not possible. Based

on extrapolation of available PK data for SYD985 in CES1c +/+ mice (either nude or SCID)

AUCs of conjugated antibody at tumor static dosages are estimated to be between 150-260

h.µg/mL. This is based on data generated in the MAXF1162 PDX model and the BT474 model

(15) and in normal (non-immunedeficient) mice. This AUC is more than tenfold below the

observed AUC in the clinic at the recommended phase 2 dose of 1.2 mg/kg that showed efficacy

(21). Thus, using the AUC at tumor static dose in CES1c +/+ mice to predict the effective human

dose would lead to a significant under-prediction of this dose. It was anticipated that the

translational value of PDX models would improve if the PK profile in mice was more in line

with the anticipated human PK profile. As will be discussed later, the AUC of conjugated

SYD985 at a tumor static dose of 1 mg/kg in CES1c -/- mice in the MAXF1162 PDX model very

nicely correlated with that in humans at the recommended phase 2 dose at 1.2 mg/kg which

turned out to be efficacious.

Based on the in vivo PK data it was reasoned that the instability of SYD985 is most likely

caused by cleavage or release of the linker-drug, rather than clearance of the entire ADC, since

total antibody levels were similar in all tested species. It was hypothesized that the instability is

the result of hydrolysis of one of the carbamate moieties in the linker-drug since these are known substrates for esterase-like activity (22, 23). Preliminary studies with human esterases suggested

that neither human CES1 and CES2 nor the plasma esterase BCHE seem to be able to cleave the

linker-drug in SYD985 (data not shown). Based on a literature study focused on species differences in esterase activity in plasma, the mouse- and rat-specific CES1c was identified as a

candidate hydrolyzing enzyme (18, 24, 25). Incubation of SYD985 in human plasma spiked with

rCES1c indeed revealed that CES1c caused instability of SYD985. The final evidence was

provided by a PK study in CES1c -/- mice, showing a PK profile similar to the PK in monkey and human. The identification of CES1c triggered additional studies to reveal it’s mode of

action. The increase in potency observed after in vitro incubation of SYD985 with HER-2

negative cells and 1% mouse or rat plasma, compared to medium, human, monkey and CES1c -/-

plasma suggested active toxin DUBA is released by CES1c activity. LC-MS/MS analysis

confirmed the presence of significant amounts of DUBA in plasma samples from mice and rats

dosed with SYD985.

Most surprisingly, CES1c activity not only results in release of DUBA: here it is shown

that it also results in the formation of high molecular weight (HMW) species, as identified by

SEC. Using SDS-PAGE, Western blotting and LC-MS/MS, these HMW species were found to

contain both trastuzumab and CES1c, which suggests that at least one CES1c molecule is

capable of covalently binding to SYD985. Next, attempts to identify the site of interaction

between CES1c and the linker-drug disclosed it to be the carbamate group connecting the linker

and toxin. Other carbamates in the linker-drug seem to remain unaffected. CES1c activity results

in release of DUBA and formation of a covalent bond between SYD985 and CES1c, most likely

via the serine (Ser221) in its active site, although it cannot be excluded that the modification occurs on serine-222, adjacent to the active site serine. CES1c-SYD985 cross-links were

observed both in the hinge region of SYD985 heavy chain and SYD985 light chain. All together,

these interaction studies clarified the structure of the covalent bond between CES1c and SYD985

that is –unexpectedly- formed upon CES1c cleavage of the vc-seco-DUBA linker-drug at a

specific carbamate.

Clearly, CES1c activity is a more general concern in the development of ADCs. In a

recent publication, Dorywalska et al. (15) showed that CES1c also cleaves other linker-drugs.

Their vc-PABC containing linker-drug was postulated to be cleaved at the carbamate C-terminal

to the citrulline moiety. Cleavage by, or covalent interaction with, CES1c would not result in a complex between the ADC and the esterase due to the structural orientation of the carbamate in

the linker-drug (19). Even though this linker-drug and vc-seco-DUBA share the vc-PABC

structure, we found no evidence for CES1c-mediated cleavage C-terminal to citrulline, or for

hydrolysis at the other carbamate positions in vc-seco-DUBA, underlining the complexity of the

structure-activity relationship of CES1c. Several other publications describe carbamate-

containing linker-drugs that are unstable in studies with mice and rat and are therefore potentially

sensitive to cleavage by CES1c (17, 26). So far, modifying the structure of the linker-drug is the

general mitigation to enhance ADC stability in mouse plasma, even though the linker-drug is

perfectly stable in human plasma and from that perspective does not require further optimization.

In order to circumvent linker-drug optimization specifically for the mouse, it was attempted to

block mouse-specific CES1c activity in vivo. Initial experiments with esterase inhibitors failed to

completely block CES1c activity. Thus, immune-compromised CES1c -/- mice were needed to

evaluate the full impact of the absence of CES1c activity on PK and in vivo efficacy in PDX

models. Cross-breeding of CES1c -/- mice with SCID and nude mice was setup in parallel. The

cross-breeding of CES1c -/- mice with nude mice showed poor reproductive performance,

whereas the SCID cross-breeding was successful. Pilot PDX studies in CES1c -/- SCID mice

suggested improved efficacy of SYD985, compared to previous studies in CES1c +/+ nude mice.

The full PKPD study in the MAXF1162 PDX model with T-DM1 to control for potential strain

differences confirmed improved efficacy of SYD985 in CES1c -/- SCID mice showing tumor

stasis at 1 mg/kg. At this dose, the PK profile and AUC are very similar to those at the

recommended phase 2 dose in human. This shows that use of CES1c -/- mice is a good and for us

the preferred alternative to unnecessary structural optimization of the linker-drug that might even

compromise on human-relevant parameters.

Conclusions

Even though it is known that the rodent-specific esterase CES1c is able to cleave linker-drugs on

ADCs at distinct sites, the impact of CES1c activity on the predictive value of PK, efficacy and safety studies with ADCs in mice and rats should not be underestimated. Here it is shown that

cleavage of vc-seco-DUBA results in the formation of a covalent bond between CES1c and the

ADC, which indicates that the structure-activity relationship of CES1c is more complex than

originally thought. It not only depends on the structure of the linker-drug, but also on the site of

conjugation in the ADC. Consequently, developing linker-drugs that are stable in both mice and

humans, remains challenging. The use of CES1c -/- mice widens the spectrum of linker-drugs

suitable for the development of ADCs, thereby increasing the chance that ADCs are successfully

developed.

References

1 Lambert J.M. (2013) Drug-conjugated antibodies for the treatment of cancer. Br J Clin Pharmacol. 76:248-62.

2 Jain N., Smith S.W., Ghone S., Tomczuk B. (2015) Current ADC Linker Chemistry. Pharm Res. 32:3526-3540.

3 Lu J., Jiang F., Lu A., Zhang G. (2016) Linkers Having a Crucial Role in Antibody-Drug Conjugates. Int J Mol Sci. 17:561.

4 Beck A., Goetsch L., Dumontet Cl., Corvaïa N. (2017) Strategies and challenges for the next generation of antibody-drug conjugates. Nat Rev Drug Discov. 16:315-337.

5 Drake P.M., Rabuka D. (2017) Recent Developments in ADC Technology: Preclinical Studies Signal Future Clinical Trends. BioDrugs. 31:521-531.

6 Frigerio M., Kyle A.F. (2017) The Chemical Design and Synthesis of Linkers Used in Antibody Drug Conjugates. Curr Top Med Chem. 17:3393-3424.

7 Ducry L., Stump B. (2010) Antibody-drug conjugates: linking cytotoxic payloads to monoclonal antibodies. Bioconjug Chem. 21:5-13.

8 Van der Velden V.H., te Marvelde J.G., Hoogeveen P.G., Bernstein I.D., Houtsmuller A.B., Berger M.S., et al. (2001) Targeting of the CD33-calicheamicin immunoconjugate Mylotarg (CMA-676) in acute myeloid leukemia: in vivo and in vitro saturation and internalization by leukemic and normal myeloid cells. Blood 97:3197-3204.

9 Shen B.Q., Xu K., Liu L., Raab H., Bhakta S., Kenrick M., et al. (2012) Conjugation site modulates the in vivo stability and therapeutic activity of antibody-drug conjugates. Nat Biotechnol. 22:184-189.

10 Durbin K.R., Nottoli M.S., Catron N.D., Richwine N., Jenkins G.J. (2017) High-Throughput, Multispecies, Parallelized Plasma Stability Assay for the Determination and Characterization of Antibody−Drug Conjugate Aggregation and Drug Release. ACS Omega 2:4207-4215.

11 Shadid M., Bowlin S., Bolleddula J. (2017) Catabolism of antibody drug conjugates and characterization methods. Bioorg Med Chem. 25:2933-2945.

12 Di L., Kerns E.H., Hong Y., Chen H. (2005) Development and application of high throughput plasma stability assay for drug discovery. Int J Pharm. 297:110-119.

13 Dokter W., Ubink R., van der Lee M., van der Vleuten M., van Achterberg T., Jacobs D., Loosveld E., et al. (2014) Preclinical profile of the HER2-targeting ADC SYD983/SYD985:

introduction of a new duocarmycin-based linker-drug platform. Mol Cancer Ther. 13:2618- 2629.

14 Elgersma RC, Coumans RG, Huijbregts T, Menge WM, Joosten JA, Spijker HJ, et al. (2015) Design, Synthesis, and Evaluation of Linker-Duocarmycin Payloads: Toward Selection of HER2-Targeting Antibody-Drug Conjugate SYD985. Mol Pharm. 12:1813-1835.

15 Van der Lee M.M., Groothuis P.G., Ubink R., van der Vleuten M.A., van Achterberg T.A., Loosveld E.M., et al. (2015) The Preclinical Profile of the Duocarmycin-Based HER2- Targeting ADC SYD985 Predicts for Clinical Benefit in Low HER2-Expressing Breast Cancers. Mol Cancer Ther. 14:692-703.

16 Dorywalska M., Dushin R., Moine L., Farias S.E., Zhou D., Navaratnam T., et al. (2016) Molecular Basis of Valine-Citrulline-PABC Linker Instability in Site-Specific ADCs and Its Mitigation by Linker Design. Mol Cancer Ther. 15:958-970.

17 Gupta N., Kancharla J., Kaushik S., Ansari A., Hossain S., Goyal R., et al. (2017) Development of a facile antibody-drug conjugate platform for increased stability and homogeneity. Chem Sci. 8:2387-2395.

18 Li B., Sedlacek M., Manoharan I., Boopathy R., Duysen E.G., Masson P., et al. (2005) Butyrylcholinesterase, paraoxonase, and albumin esterase, but not carboxylesterase, are present in human plasma. Biochem Pharmacol. 70:1673-84.

19 Crow J.A., Bittles V., Borazjani A., Potter P.M., Ross M.K. (2012) Covalent inhibition of recombinant human carboxylesterase 1 and 2 and monoacylglycerol lipase by the carbamates JZL184 and URB597. Biochem Pharmacol. 84:1215-1222.

20 Van Herpen C.M.L., Banerji U., Mommers E.C.M., Koper N.P., Goedings P., Lopez J., et al. (2015) Phase I dose-escalation trial with the DNA-alkylating anti-HER2 antibody-drug conjugate SYD985. ESMO 2015 poster presentation

21 Saura C., Thristlethwaite F., Banerji U., Lord S., Moreno V., MacPherson I., et al (2018) A phase I expansion cohorts study of SYD985 in heavily pretreated patients with HER2-positive or HER2-low metastatic breast cancer. ASCO 2018 poster presentation

22 Yang Y., Aloysius H., Inoyama D.,Chen Y., Hun L. (2011) Enzyme-mediated hydrolytic activation of prodrugs. Acta Pharmaceutica Sinica B 1:143–159.

23 Laizure S.C., Herring V., Hu Z., Witbrodt K., Parker R.B. (2013) The role of human carboxylesterases in drug metabolism: have we overlooked their importance? Pharmacotherapy. 33:210-222.

24 Berry L.M., Wollenberg L., Zhao Z. (2009) Esterase Activities in the Blood, Liver and Intestine of Several Preclinical Species and Humans. Drug Met. Lett. 3:70-77.

25 Duysen E.G., Cashman J.R., Schopfer L.M., Nachon F., Masson P., Lockridge O. (2012) Differential sensitivity of plasma carboxylesterase-null mice to parathion, chlorpyrifos and chlorpyrifos oxon, but not to diazinon, dichlorvos, diisopropylfluorophosphate, cresyl saligenin phosphate, cyclosarin thiocholine, tabun thiocholine, and carbofuran. Chem Biol Interact. 195:189-98.

26 Dorywalska M., Strop P., Melton-Witt J.A., Hasa-Moreno A., Farias S.E., Galindo Casas M., et al. (2015) Site-Dependent Degradation of a Non-Cleavable Auristatin-Based Linker- Payload in Rodent Plasma and Its Effect on ADC Efficacy. PLoS One. 10:e0132282.

Table 1. Characteristics of the peptides containing the CES1c active site serine (Ser221) that were found to be conjugated to NAc-Cys-linker-drug.

Peptide sequence Theoretical Observed m/z (z) Observed m/z (z) Most predominant Peptide mass (Da) without conjugate with conjugate fragment ion m/z GES221SGG 492.19 493.19 (1) 730.79 (2) 681.27 GES221SGGIS 691.37 692.37 (1) 830.85 (2) 881.44 GES221SGGISV 790.37 791.37 (1) 880.39 (2) 980.50 GES221SGGISVS 877.40 878.40 (1) 923.90 (2) 1067.50 IFGES221SGGIS 951.45 952.45 (1) 960.93 (2) 1141.62

Table 2. Characteristics of the peptides containing the CES1c active site serine (Ser221) that were found to be conjugated to SYD985.

Precursor CES1c Linker Residual mass SYD985 sequence (1+) fragment* mass Light chain 828.04 (3+) 1140.6 618.28 723.25 FNRGEC 2482.13 773.74 (3+) 979.5 618.28 723.30 FNRGEC 2319.23 741.32 (3+) 880.4 618.28 723.30 FNRGEC 2221.98

813.03 (3+) 880.4 618.28 938.43 KSFNRGEC 2437.11 684.32 (4+) 880.4 618.28 1235.60 PVTKSFNRGEC 2734.28 Heavy chain 836.39 (4+) 680.3 618.28 2045.99 THTC(NEM)PPCPAPELLGGPSVF 3342.55 891.18 (4+) 880.4 618.28 2064.02 THTC(NEM)PPCPAPELLGGPSVF 3561.70 +H O 2 1181.89 (3+) 880.4 618.28 2046.01 THTC(NEM)PPCPAPELLGGPSVF 3543.69 *CES1c peptide fragments (uncharged): 680.3 = GESSGG; 880.4 = GESSGGIS; 979.5 = GESSGGISV; 1140.6 =

IFGESSGGIS. ** ‘+H2O’ refers to hydrolysis (ring opening) of the maleimide moiety in the linker drug.

Figure 1: In vitro and in vivo kinetics of SYD985. A, Conjugated antibody concentration during

a 96-hour incubation of 100 µg/mL SYD985 at 37°C in human, monkey, rat and mouse plasma

(mean +/- SEM, n=3). B, Total (TAb) and conjugated antibody (Conj. Ab) concentration (mean

+/- SEM, n = 2 for rat, 3 for mouse and 5 for monkey) after single-dose IV administration of 10 mg/kg SYD985 to mice, rats (dose normalized from 8 mg/kg) or monkeys. C, Conjugated

antibody concentrations of 100 µg/ml SYD985 after 96 hrs in vitro incubation at 37°C in human

plasma with increasing concentrations of mouse recombinant carboxylesterase 1c (rCES1c )

(mean ± SEM, n = 2). D, Conjugated antibody concentrations in plasma in CES1c knock-out (-/-

), CES1c heterozygous (+/-) and CES1c wild-type (+/+) mice, after a single IV bolus injection of

5 mg/kg SYD985 (mean +/- SEM, n = 3). E, Cytotoxic activity of released active toxin on

HER2-negative SW-620 cells after 6 days incubation of the toxin DUBA in plasma free medium

(positive control) or SYD985 in the presence of 1% mouse, CES1c -/- mouse, rat, monkey or

human plasma in the culture medium. Data show the percentage of survival ± SD of two

experiments performed in duplicate.

Figure 2: A – E, SEC-HPLC radiochromatograms of [3H]-SYD985 after 6 hours incubation in plasma of indicated species at 100 µg/mL and 37oC. F – I, analysis of high molecular weight

(HMW) species in affinity-extracted material obtained from BALB/c mouse plasma incubated

with SYD985. F, Confirmatory SEC-HPLC UV chromatogram of non-radiolabeled SYD985 and

associated material affinity-extracted after 48 hours of incubation in wild-type mouse plasma at

37°C. G, Coomassie staining of affinity-extracted samples from BALB/c mouse plasma using

non-reducing SDS-PAGE revealed an increase in the formation of high molecular weight species

over time. H, a Western blot with the anti-idiotype antibody directed against SYD985 showing

the presence of SYD985 in the high molecular weight species. I, a Western blot with an anti-

CES1c antibody showing the presence of CES1c in the high molecular weight species.

Figure 3. A, Structure of NAc-quenched linker-drug with carbamate moieties indicated by

circles. The blue-circled group was hypothesized to be the one resulting in the covalent

interaction upon binding of CES1c. B, proposed structure of the linker-drug-modified CES1c active site-containing peptide GESSGGISV. C, D, proposed structures of the corresponding

predominant fragment ions formed from the linker-drug-modified CES1c active site-containing

peptide GESSGGISV upon collision-induced dissociation: one that loses part of the linker

moiety (C) and one that loses the complete side chain modification, resulting in the formation of

a dehydroalanine (dhA) residue (D). Please refer to the middle spectrum in panel E for the

MS/MS data of this modified peptide. E, ion trap MS/MS fragmentation spectra (HCD @30 eV)

of the five linker-drug modified CES1c peptides that were found after incubation of the

carboxylesterase with the NAc-Cys--linker-drug. The predominant fragment resulting from the

loss of most of the conjugated NAc-Cys-linker-drug moiety (780.3 Da) from the precursor ion is

clearly visible (note: all precursors are doubly-charged), just like the presence of the

dehydroalanine-containing peptide fragments (at m/z ‘precursor ion’ – 987.3 Da) and linker-

specific fragments, such as those at m/z 163.1, 207.2, 240.3 and 349.1.

Figure 4: Pharmacokinetics of SYD985 in A, CES1c -/- mice versus CES1c -/- SCID mice after

an IV a dose of 3 mg/kg (mean +/- SEM, n = 3) and D, after dosing 0.3, 1 and 3 mg/kg in

MAXF1162 tumor bearing CES1c -/- SCID mice. Antitumor activity of SYD985 compared to T-

DM1 and isotype control ADC in the HER-2 3+ MAXF1162 breast cancer PDX model in CES1c

+/+ Nude mice (B, C) or in CES1c -/- SCID mice (E, F) (mean +/- SEM, n = 6-8).

Figure 1

A B C T A b m o n k e y C o n j A b M o n k e y 1 5 0 1 0 0 0 1 0 0 0 T A b R a t C o n j A b R a t

M o n k e y R a t T A b M o u s e C o n j A b M o u s e

)

) )

M o u s e L H u m a n L

L 1 0 0

g g g 1 0 0

1 0 0 µ

(

1 0

t t

t 1

e e

e 1 0 5 0

o o

o 0 . 1

C C

1 0 . 0 1 0 0 5 0 1 0 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 0 1 0 0 2 0 0 3 0 0 4 0 0

T i m e ( h ) T i m e ( h ) r C E S 1 c ( µ g / m L )

D E 1 0 0 0 1 2 0 C E S 1 c - / -

C E S 1 c + / - ) 1 0 0

L 1 0 0 C E S 1 c + / +

/ g

l 8 0

µ M e d iu m

(

1 0 v i

n H u m a n

r i

t 6 0

u M o n k e y

t 1 R a t

n %

e 4 0 M o u s e c

n C e s 1 c - / - o 0 . 1 C 2 0 m o u s e

D U B A 0 . 0 1 0 0 1 0 0 2 0 0 3 0 0 - 1 3 - 1 2 - 1 1 - 1 0 - 9 - 8 - 7 - 6

T i m e ( h ) L o g c o n c e n t r a t io n ( M )

Downloaded from mct.aacrjournals.org on September 29, 2021. © 2018 American Association for Cancer Research. CPM CPM CPM 10000.0 12000.0 14000.0 16000.0 10000.0 12000.0 14000.0 10000.0 12000.0 14000.0 2000.0 4000.0 6000.0 8000.0 2000.0 4000.0 6000.0 8000.0 2000.0 4000.0 6000.0 8000.0 0.0 0.0 0.0 G G SDS- 0.00 0.00 0.00 0 0.22 mins C E A PAGE 1 2.17 2.17 mins Downloaded from 6

Author manuscriptshavebeenpeerreviewedandacceptedforpublicationbutnotyetedited. 24 Author ManuscriptPublishedOnlineFirstonAugust9,2018;DOI:10.1158/1535-7163.MCT-18-0329 6.07 mins 6.07 mins 6.07

96 HMW 7.15 mins HMW 7.15 mins HMW 7.15 HMW SYD9857.80 8SYD985.02 mins SYD9858.02 min s 8.67 8.88 mins 10.00 10.00 9.53 mins 9.53 mins 10.00 mct.aacrjournals.org kDa kDa 100 150 250 75

14.08 mins H CES1c -/- mouseCES1c -/- a 0 Id

AU 1000.0 2000.0 3000.0 4000.0 5000.0 1000.0 2000.0 3000.0 4000.0 5000.0 6000.0 0.00 0.10 0.20 0.30 0.40 0.50 0.60 24 0.0 0.0

0.00 0.00

5.50 0.22 mins F B D 96

6.00

6.50 kDa kDa 100 150 250

75 7.00 6.07 mins 6.07 mins HMW 6.50 mins 7.50 7HMW.15 mins 7HMW.15 mins I I 7.80 mins a SYD985 SYD9857.80 mins 0 CES1c

Minutes 8.67 mins 8.00 SYD985 SYD985 8.88 mins 10.00

9.32 mins 10.00

1 10.18 mins 8.50

9.00 6

Figure2 14.08 mins

9.50 14.08 mins 24

10.00

96 Mouse 10.50 Mouse Mouse 17.98 mins 17.98 mins Rat Rat 20.00 20.00 19.07 mins 19.07 mins 11.00 mins kDa kDa mins 100 150 250

75 Author Manuscript Published OnlineFirst on August 9, 2018; DOI: 10.1158/1535-7163.MCT-18-0329 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Figure 3

A C V

G G E S G S

V Exact mass: m/z 980.49 (1+) V B S S I D I G G E S G S G G E S G dhA

Exact mass: m/z 880.39 (2+) Exact mass: m/z 774.36 (1+)

681.2722 NL: 2.57E4 GESSGG E 100 Full ms2 [email protected]

50 349.0958 207.1686 588.2110 730.3232 1035.2388 1210.5594 1401.1116 0 NL: 1.49E5 GESSGGIS 881.4395 100 Full ms2 [email protected]

50 349.2315 675.3375 240.2588 987.4726 1251.5571 1390.4729 1567.6655 0 980.5031 NL: 1.36E5 GESSGGISV 100 Full ms2 [email protected]

50 349.2002 774.4719 240.2942 588.3439 1085.4099 1350.6631 1512.6050 1717.5129

0 Relative Abundance Relative 1067.4961 NL: 2.21E4 GESSGGISVS 100 Full ms2 [email protected]

50 349.1088 861.3724 240.1815 588.2286 1142.3851 1399.5941 1607.2213 1808.9518 0 1141.6213 NL: 8.60E4 IFGESSGGIS 100 Full ms2 [email protected]

50 349.1663 935.4435 240.2594 588.2765 830.3983 1226.6826 1530.7034 1717.8667 0 200 400 600 800 1000 1200 1400 1600 1800 m/z

Figure 4

A B 2 0 0 0 C 2 0 0 0 1 0 0 V e h ic le I V V e h ic le I V C E S 1 c - / - I s o t y p e c o n t r o l 3 m g / k g T - D M 1 1 m g / k g

S Y D 9 8 5 1 m g / k g

C E S 1 c - / - S C I D T - D M 1 3 m g / k g )

S Y D 9 8 5 3 m g / k g )

) T - D M 1 1 0 m g / k g

3 1 5 0 0 3 1 5 0 0

L S Y D 9 8 5 1 0 m g / k g

(

m o

u 1 0 0 0 u

i 1 0 0 0

1 0 l

u u

o 5 0 0 5 0 0

T C

1 0 0 0 2 0 0 4 0 0 6 0 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0

T i m e ( h ) T i m e ( d ) T i m e ( d )

1 5 0 0 D 1 0 0 E V e h ic le I V F 1 5 0 0 V e h ic le I V 3 m g / k g I s o t y p e c o n t r o l 3 m g / k g 1 m g / k g T - D M 1 1 m g / k g S Y D 9 8 5 0 . 3 m g / k g

0 . 3 m g / k g T - D M 1 3 m g / k g

) )

S Y D 9 8 5 1 m g / k g ) 3

L T - D M 1 1 0 m g / k g

1 0 3

m m

S Y D 9 8 5 3 m g / k g m

/ 1 0 0 0 m

g 1 0 0 0

(

l u

t 1

o e 5 0 0 o

c 5 0 0

u o

0 . 1 T

T C

0 . 0 1 0 0 0 1 0 0 2 0 0 3 0 0 4 0 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0

T i m e ( h ) T i m e ( d ) T i m e ( d )

Unraveling the interaction between carboxylesterase 1c and the antibody-drug conjugate SYD985: improved translational PKPD by using CES1c knockout mice

Ruud Ubink, Eef HC Dirksen, Myrthe Rouwette, et al.

Mol Cancer Ther Published OnlineFirst August 9, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-18-0329

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2018/08/09/1535-7163.MCT-18-0329.DC1

Author Author manuscripts have been peer reviewed and accepted for publication but have not yet been Manuscript edited.

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/early/2018/08/09/1535-7163.MCT-18-0329. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.