Microtubule and tubulin binding and regulation of microtubule dynamics by the antibody drug conjugate (ADC) payload, monomethyl auristatin E (MMAE): Mechanistic insights into MMAE ADC peripheral neuropathy
Rebecca L. Best a, 1, Nichole E. LaPointe a, 1, Olga Azarenko a, 1, Herb Miller a,
Christine Genualdi b, Stephen Chih a, 2, Ben-Quan Shen c, Mary Ann Jordan a, Leslie Wilson a,
Stuart C. Feinstein a,*, Nicola J. Stagg b,*
a Neuroscience Research Institute and Department of Molecular, Cellular and Developmental Biology, University of California, Santa Barbara, CA 93106, USA
b Safety Assessment, Genentech Inc., 1 DNA Way, South San Francisco, CA 94080, USA
c Preclinical and Translational PK, Genentech Inc., 1 DNA Way, South San Francisco, CA 94080, USA

A R T I C L E I N F O

Editor: Lawrence Lash

Keywords:
Peripheral neuropathy Antibody drug conjugate Auristatin
Anti-cancer
Microtubule-targeting agent Peripheral nervous system
A B S T R A C T

Monomethyl auristatin E (MMAE) is a potent anti-cancer microtubule-targeting agent (MTA) used as a payload in three approved MMAE-containing antibody drug conjugates (ADCs) and multiple ADCs in clinical development to treat different types of cancers. Unfortunately, MMAE-ADCs can induce peripheral neuropathy, a frequent adverse event leading to treatment dose reduction or discontinuation and subsequent clinical termination of many MMAE-ADCs. MMAE-ADC-induced peripheral neuropathy is attributed to non-specific uptake of the ADC in peripheral nerves and release of MMAE, disrupting microtubules (MTs) and causing neurodegeneration. However, molecular mechanisms underlying MMAE and MMAE-ADC effects on MTs remain unclear. Here, we characterized MMAE-tubulin/MT interactions in reconstituted in vitro soluble tubulin or MT systems and evaluated MMAE and vcMMAE-ADCs in cultured human MCF7 cells. MMAE bound to soluble tubulin hetero- dimers with a maximum stoichiometry of ~1:1, bound abundantly along the length of pre-assembled MTs and with high affinity at MT ends, introduced structural defects, suppressed MT dynamics, and reduced the kinetics and extent of MT assembly while promoting tubulin ring formation. In cells, MMAE and MMAE-ADC (via nonspecific uptake) suppressed proliferation, mitosis and MT dynamics, and disrupted the MT network. Comparing MMAE action to other MTAs supports the hypothesis that peripheral neuropathy severity is deter- mined by the precise mechanism(s) of each individual drug-MT interaction (location of binding, affinity, effects on morphology and dynamics). This work demonstrates that MMAE binds extensively to tubulin and MTs and causes severe MT dysregulation, providing convincing evidence that MMAE-mediated inhibition of MT- dependent axonal transport leads to severe peripheral neuropathy.

⦁ Introduction
Microtubule targeting agents (MTAs) are important chemothera- peutic drugs used to combat many types of cancer (Argyriou et al., 2012; Argyriou et al., 2011; Carlson and Ocean, 2011; Windebank and Grisold, 2008). A few examples include vincristine, vinblastine, vinorelbine, vinflunine, eribulin, paclitaxel and ixabepilone. Mechanistically, MTA action derives from their ability to alter microtubule (MT) dynamics
and/or regulatory mechanisms controlling MT dynamics and MT-based transport, which can, in turn, lead to tumor cell death via cell cycle arrest (Argyriou et al., 2012; Field et al., 2014; Jordan and Wilson, 2004; Poruchynsky et al., 2015).
Another class of MTAs are the auristatins, analogues of naturally occurring peptides called dolastatins that were first isolated from the sea hare Dolabella auricularia (Pettit et al., 1987). Bai et al. (1990) demon- strated that the potent cytotoxicity and inhibition of proliferation

* Corresponding authors.
E-mail addresses: [email protected] (S.C. Feinstein), [email protected] (N.J. Stagg).
1 Contributed equally.
2 Present address: Penn State College of Medicine, 700 HMC Crescent Road, Hershey, PA 17033.

https://doi.org/10.1016/j.taap.2021.115534

Received 9 February 2021; Received in revised form 9 April 2021; Accepted 9 April 2021
Available online 20 April 2021
0041-008X/Published by Elsevier Inc.

exhibited by dolastatin 10 (in the picomolar range) correlated with its ability to inhibit MT assembly and tubulin-dependent GTPase activity. The net effect was cell cycle arrest and apoptosis. Initial clinical trials of dolastatin 10, however, were problematic because of toxic effects (Margolin et al., 2001; Mirsalis et al., 1999; Newman et al., 1994; Pitot et al., 1999). Monomethyl auristatin E (MMAE) is a more recently developed member of the auristatin class of MTAs that also shows very potent cytotoxicity, although, like dolostatin 10, this potency precludes its therapeutic use as a free drug due to its extensive cytotoxicity to normal (non-tumor cells).
While MTA chemotherapeutics exploit the critical importance of MTs in tumor cells, MTs are also essential for many functions in non-target tissues, including the peripheral nervous system. MTs are important for the maintenance of highly elongated neuronal morphologies and axonal transport, the rapid movement of cargo between neuronal cell bodies and distal nerve endings (Morfini et al., 2009). As a result, the peripheral nervous system, which lacks the protection conferred upon the central nervous system by the blood-brain barrier, is highly sus- ceptible to deleterious MTA-induced effects. Among the most frequent and serious side effects of MTA treatment is chemotherapy-induced peripheral neuropathy (CIPN), which is characterized by a disruption of peripheral nerve signaling, manifesting with clinical symptoms in humans ranging from numbness and tingling to hypersensitivity and severe neuropathic pain and causing significant axon loss, myelin morphological changes, and reduction of nerve conduction velocity in mouse models (Benbow et al., 2017). Clinically, CIPN symptoms generally exhibit a “stocking/glove” pattern, beginning at the most distal extremities, such as the fingertips and toes, and progressing proximally toward the trunk, suggesting that the longest axons are the most vulnerable (Argyriou et al., 2012; Carlson and Ocean, 2011; Windebank and Grisold, 2008). The symptoms of CIPN can be suffi- ciently debilitating to limit treatment, and in some cases, can be life
threatening. The only current strategies to address MTA-induced CIPN
are dose delay, dose reduction, or discontinuation of treatment. Though MTAs may have a broad application in anti-cancer strategies, CIPN poses a major obstacle to successful clinical anti-cancer efforts.
One strategy to manage the systemic toxicity of these potent drugs is to covalently couple them to an antibody directed to a specific antigen expressed on the surface of target tumor cells. These antibody-drug complexes are known as antibody-drug conjugates (ADCs) including the MMAE-ADCs that are FDA-approved for different cancers: Adcetris® (bretuximab vedotin, BV) for Hodgkin’s lymphoma and anaplastic large cell lymphoma, Polivy™ (polatuzumab vedotin) for diffuse large B-cell lymphoma, and Padcev™ (enfortumab vedotin-ejfv) for bladder cancer. Upon internalization into the tumor target cells, the MMAE drug payload is released from ADCs through proteolytic cleavage of the dipeptide linker (valine-citrulline) by cathepsin B. However, despite their targeted nature, many ADCs continue to exhibit a surprisingly high incidence of adverse effects. A recent meta-analysis found that these adverse effects are specific to the conjugated drug, i.e. the “payload” (Masters et al., 2017). In the case of an ADC with MMAE as the payload, peripheral neuropathy is a major adverse effect that can be both severe and dose limiting, and further, it has been shown to be independent of the target of the antibody (Saber and Leighton, 2015). Additionally, the significant clinical incidence of peripheral neuropathy with MMAE- conjugated ADCs was not predicted in earlier non-clinical toxicology
studies (Stagg et al., 2016). However, through investigative assessments,
the peripheral neuropathy observed in patients has been attributed primarily to nonspecific uptake of the ADC by peripheral nerves and subsequent release of its MT-targeted payload, (i.e., MMAE), leading to nerve degeneration (Stagg et al., 2016). Release of free MMAE system- ically has also been observed in patients and in nonclinical studies and could potentially be taken up by peripheral nerves, but the levels of free circulating drug are very low and unlikely to have much of an effect.
As there is still quite substantial exposure of the peripheral nerves to the MMAE payload as a result of the nonspecific uptake of the ADC,
understanding how MMAE interacts with MTs leading to peripheral neuropathy is of great interest. One often-stated hypothesis for MTA- induced CIPN posits that the drugs interfere with normal MT- dependent axonal transport, which in turn initiates subsequent neuro- degeneration. Interestingly, despite MTAs sharing the same molecular target (namely MTs), some of the newer MTAs exhibit reduced in- cidences and severities of CIPN. Investigations into this observation suggested that these differential frequencies of severe CIPN might derive from the different binding sites MTAs bind to on MTs and/or soluble tubulin and their resultant differential effects on MT dynamics (Benbow et al., 2017). It has become increasingly clear that gaining a more mechanistic understanding of MTA binding and regulation of MT dy- namics is key to understanding both normal MT action and the resultant cellular and clinical consequences of misregulation by MTAs. Genualdi et al., 2019 analyzed published data on in vitro microtubule (MT) properties for different MTAs that cause varying levels of CIPN in pa- tients and discussed possible mechanisms related to interactions with tubulin and MTs for the reduced CIPN with some MTAs. Eribulin, vinorelbine and vinfluinine, which all have less severe CIPN than the vinca alkaloids or taxanes (Carlson and Ocean, 2011; Gradishar, 2011), have unique MT properties consisting of reduced affinity and limited binding to MTs (i.e. bind only to the ends and not along the length).
Binding more potently to tubulin in the absence of neuronal βIII tubulin
was also observed with eribulin (Wilson et al., 2015) and may suggest specificity for tumor tubulin over neuronal tubulin.
The current study was designed to investigate the binding of MMAE as a free drug or as an ADC to MTs and free tubulin subunits, and its effects upon MT dynamics and MT morphology as well as cell prolifer- ation, cell cycle regulation and the generation of mitotic spindle ab- normalities in cultured human cells. In combination with comparisons made to other MTAs, our data provide further insights into the molec- ular mechanisms underlying normal MMAE action as well as those governing MMAE-ADC-induced peripheral neuropathy.
⦁ Materials and methods
⦁ Tubulin preparation
MAP (MT-associated protein)-rich tubulin (~70% tubulin, 30% MAPs) was obtained from crude bovine brain extract by three poly- merization/depolymerization cycles (Miller and Wilson, 2010). MAP- free tubulin was further purified with a phosphocellulose ion ex- change column (Miller and Wilson, 2010). Tubulin was stored in PEM50 (50 mM PIPES pH 6.8, 1 mM MgSO4, 1 mM EGTA) with 0.1–0.2 mM GTP
—
at 7–12 mg/mL, as measured by Bradford versus a BSA standard. Tubulin was drop-frozen in liquid nitrogen and stored at 80 ◦C. Rhodamine-labeled tubulin was prepared from MAP-free tubulin as
described (Hyman et al., 1991). Biotinylated tubulin was purchased from Cytoskeleton. GMPCPP was purchased from Jena Bioscience.
⦁ Drug stocks
Unlabeled MMAE, tritium-labeled MMAE ([3H]-MMAE), and MMAE ADC (CNJ2985-vc-MMAE) were provided by Genentech. [3H]-MMAE
(specific activity 863 mCi per mmole) was stored as a 1 mM stock in DMSO (Sigma). Unlabeled drug was stored as a 10 mM stock in DMSO. The MMAE ADC was stored in water with 5–10% sucrose, and molarity calculations were based on the molecular weight of the antibody.
⦁ MMAE binding to soluble tubulin in vitro
MAP-free tubulin (0.25 mg/mL, 2.3 μM) was prepared in PEM100 buffer (100 mM PIPES pH 6.8, 1 mM MgSO4, 1 mM EGTA) with 100 μM GTP and then incubated with different concentrations of [3H]-MMAE (0.1 μM to 20 μM). After 12–15 min at 30o C, the mixture was subjected to size exclusion chromatography to separate free drug from drug/

tubulin/GTP complexes using Zeba Spin (2 mL) desalting columns pre-
pared according to the manufacturer’s instructions (Thermo Scientific) and centrifuged at 1000 xg for 4 min. The eluate was analyzed for pro- tein content by Bradford and [3H]-MMAE content by liquid scintillation
counting. Data shown are mean and SEM from 3 to 4 independent ex- periments per MMAE concentration.

⦁ MMAE binding to MAP-rich MTs in vitro
MAP-rich tubulin (33 μM) was polymerized to steady-state in PEM100 with 1–2 mM GTP at 30 ◦C. Resulting MTs were sheared 6 times
to increase the MT number concentration and to tighten the MT length distribution, and then incubated an additional 15 min to re-establish steady-state. Transmission electron microscopy (TEM) samples were taken to determine MT length distributions (measured using a Zeiss
MOP III morphometric digitizing pad), and then MTs were aliquoted and incubated with [3H]-MMAE (1 μM to 20 μM). After 6 min, free drug was separated from MT-bound drug by centrifugation through 30% glycerol/
10% DMSO cushions at 40,000 X g, 30 ◦C for 1 h. Pellets were washed extensively and solubilized in water at 4 ◦C for 18 h, then analyzed for protein and [3H]-MMAE content as described above. Molecules of bound drug per MT was calculated using a value of 1690 tubulin dimers per μm of MT length and a median MT length of 3.63 μM. Data shown are mean and SEM of 2–4 independent experiments per concentration for 0.1–15 μM MMAE, and a single experiment for 20 μM MMAE.

⦁ MMAE-induced MT disassembly in vitro
MTs were assembled from MAP-rich tubulin as for binding experi- ments. After recovery to steady state, MMAE or MMAE ADC was added and light scattering at 350 nm (A350) was monitored. Samples for TEM were removed after 6 min, paralleling the drug exposure time in the MT- binding experiments, and then the A350 signal was measured until it plateaued.

⦁ Transmission electron microscopy (TEM)
MTs were fixed with glutaraldehyde (final concentration, 0.1–0.2%) and then placed on 200-mesh formvar-coated copper grids (Ted Pella). After 1.5 min, sample droplets were displaced with cytochrome c (1 mg/ mL) to improve staining, and then the grid was rinsed with deionized water and stained for 30 s with 1% uranyl acetate. MTs were imaged on a JEOL 1230 electron microscope at 80 kV with AMT image capture software.

⦁ In vitro MT assembly and MT end morphology assay
Solutions of MAP-rich tubulin (~2.4 mg/mL total protein; 16 μM tubulin) and MMAE or MMAE ADC were prepared on ice in PEM100, and then MT assembly at 30 ◦C was monitored by A350. After 1 h, samples were removed and the remaining volume was centrifuged at 40,000 X g, 30 ◦C for 1 h. Pellets were solubilized, and protein con- centration was determined as above. Data are the result of 2–4 inde-
±
pendent experiments per drug concentration and are shown as mean SEM. TEM images shown are from a single representative experiment. MT lengths were measured using Fiji (ImageJ, NIH).
≥
Morphology of MT ends was determined from 5 TEM images per condition. Images from 3 independent experiments were pooled and anonymized, and then each MT end was categorized as “blunt”, “splayed” (slightly flared), or “frayed” (having a projection at least as long as the MT width). Rare ends that did not fall into any of these categories were excluded from analysis. The proportion of frayed ends in each image was plotted as a function of MMAE concentration and analyzed by logistic regression in JMP Pro 13 (SAS).
⦁ In vitro MT dynamics assay
MTs were grown on coverslips from two types of seeds (no difference was observed between the types): (i) axonemes prepared from sea urchin sperm (Yenjerla et al., 2010) and attached directly to the coverslip; (ii) GMPCPP MT seeds prepared (Gell et al., 2010) from a mixture of MAP- free rhodamine-labeled tubulin, unlabeled tubulin, and biotinylated tubulin and attached via PEG-poly-L-lysine-biotin (Surface Solutions) and NeutrAvidin (Thermo Fisher). Coverslips were blocked with casein (4 mg/mL). MAP-free tubulin (10–11 μM 14% rhodamine-labeled) was diluted into PMEM buffer (87 mM Pipes, 36 mM MES, 1 mM EGTA, 1.4 mM MgSO4, pH 6.8) with 1 mM GTP, 10 mM DTT, an oxygen scavenging mixture (119 nM catalase, 214 U/mL glucose oxidase, and 40 mM D- glucose; Sigma C9322, Sigma G2133, and EM Sciences DX0145-3, respectively), and 100 nM of drug or vehicle (0.1% final concentration
DMSO in all mixtures). Samples were incubated at 32–34 ◦C for 30 min
×
to reach steady state and then imaged on a temperature-controlled Leica SP8 resonant scanning confocal microscope (63 oil objective), 1 frame every 2 s (10 min max per field of view, 45 min max per sample). Ky- mographs were generated and analyzed in Fiji (ImageJ, NIH) with KymographBuilder and then every second frame was averaged on the “time” axis to improve visibility, resulting in a final interval of 4 s (matching the frame rate used to collect the cellular data described below). Data were extracted from kymographs using the Velocity Mea- surement Tool (http://dev.mri.cnrs.fr/projects/imagej-macros/wiki/Ve locity_Measurement_Tool). Definitions and criteria for all parameters
were essentially identical to those described for cellular data, except that the criterion for growth and shortening events were ≥ 0.32 μm vs. ≥ 0.3 μm in cells. Data were pooled from 5 experiments.

⦁ Statistical analysis of MT dynamics data
Data were processed and formatted in R (v.3.3.3; (Team, R, 2016a, 2016b)), and then bootstrap analysis (106 iterations) was performed in Statistics101 (v.4.6; statistics101.net) to determine whether or not
observed differences between the means for control and MMAE were statistically significant. Graphs were generated in R with ggplot2 (H. Wickham, 2009) and then detailed in Adobe Illustrator. Boxplots show median (horizontal line) and 25th–75th percentile; a white square in- dicates the mean. Whiskers extend 1.5 times the interquartile range, and more extreme data points are shown individually. Sample sizes are indicated on the x-axis.
⦁ Cell culture
MCF7 human breast adenocarcinoma cells stably transfected with EGFP-α-tubulin cDNA (Azarenko et al., 2014) were cultured in DMEM with 10% fetal bovine serum (Atlanta Biologicals, Inc.), 1% final con- centration of nonessential amino acids, 100 units/mL penicillin (Sigma), and 100 μg/mL streptomycin (Sigma) at 37 ◦C, 5.5% CO2. These cells
were indistinguishable from unmodified MCF7 cells except for their fluorescent MTs and their doubling time of 35 h, which was 20% slower than unmodified MCF7 cells. They constitutively express EGFP- α-tubulin and are referred to in this manuscript as simply MCF7 cells.
⦁ MMAE cellular uptake assay
×
To determine intracellular drug concentrations (Okouneva et al., 2008), MCF7 cells were seeded into poly-L-lysine-treated scintillation vials (1 105 cells, 2 mL). After 24 h, media was replaced with fresh
media containing 10 nM [3H] MMAE or unlabeled drug. Media was
removed 30 min to 20 h after drug addition, vials rinsed twice with 1 mL PBS, and Ready Protein+(Beckman Coulter, Inc.) added for scintillation
counting (LS 6500 Multi-Purpose Scintillation Counter; Beckman Coulter, Inc.). Intracellular drug concentration = (moles of intracellular drug bound/mean cell volume) × (the number of cells per vial). Mean

cell volume was calculated based upon the mean diameter of cells rounded by trypsinization (4.57 × 10—12 L; n = 30).
⦁ Cell proliferation assay
Proliferation was measured by a modified sulforhodamine B (SRB) assay (Oroudjev et al., 2010). Cells were seeded (5000 cells/200 μL) in 96-well plates and incubated 24 h. Fresh medium with or without the MMAE was added, and incubation continued for 96 additional hr. Cells were fixed, stained with SRB reagent, and optical density determined
(490 nm; Victor3V Wallac 1420 Spectrophotometer, Perkin-Elmer).
± ≥
Triplicates of each condition were tested in each experiment. Results are mean SEM of 4 experiments. The concentrations that inhibited cell proliferation by 50% were calculated using Prism 4.0 software (GraphPad Software, Inc).
⦁ Cell viability and cell cycle analysis
×
± ≥
≥
Cells were seeded (6 104 cells/2 mL) in 6-well plates (24 h), and then incubated with drug for 24, 48, or 72 h. For viability, all cells were harvested and stained with ViaCount DNA binding dyes (5 min; EMD Millipore) and analyzed by flow cytometry (EasyCyte flow cytometer; Guava Technologies, Inc.) to distinguish live from apoptotic/dead cells. For cell-cycle analysis, floating and adherent cells were collected, permeabilized with ice-cold 70% ethanol (Sigma), washed with PBS, and stained with cell-cycle reagent (EMD Millipore). The DNA content of 5000 cells was measured by flow cytometry and analyzed with ModFit LT software (Verity Software House). Results are mean SEM of 3
independent experiments.

⦁ Mitotic arrest assay
×
MCF7 cells were seeded as for cell viability and incubated with drug for 24 h; both floating and attached cells were collected and fixed in 3.7% formaldehyde (30 min), followed by cold methanol (10 min; Sigma; (Okouneva et al., 2008)). Fixed cells were mounted on glass slides with antifade agent Prolong Gold-DAPI (Life Technologies, Inc.) to visualize DNA and examined by fluorescence microscopy (Nikon Eclipse E800, 40 objective). Mitotic cells were rounded cells with condensed chromosomes and no nuclear envelope. The percentage of mitotic cells
was determined by counting ≥500 cells for each condition. Results are mean ± SEM, ≥3 experiments.
⦁ Cell immunofluorescence assay
Cells were seeded as for cell viability, incubated with drug for 24 h, fixed with 3.7% formaldehyde followed by cold methanol (Okouneva et al., 2008), and stained with mouse monoclonal α-tubulin antibody (1:1000 DM1A, Sigma) and FITC-conjugated goat anti-mouse antibody (1:1000, Cappel MP Biochemicals). Centrosomes were stained with rabbit polyclonal anti-pericentrin (1:500, AB4448; Abcam) and rhodamine-conjugated goat anti-rabbit (1:500, Cappel). Cells were
×
mounted as above and imaged using a spectral confocal Olympus Fluoview1000 microscope (FLV1000S, 60 oil, N.A. 1.4 objective; Olympus).
⦁ Cellular MT dynamics assay
×
± ×
Cells were seeded on coverslips in 6-well plates (3 104 cells/mL, 2 mL/well) for 24 h, and then media was replaced with fresh media containing 0.5 nM MMAE. Time-lapse images (38 frames/4-s intervals/ 1 h at 37 1 ◦C) were taken with a 100 objective on the same mi-
croscope used for in vitro MT dynamics. Changes in the length of indi- vidual MTs were graphed versus time as “life history” plots. MT dynamics parameters were analyzed with IGOR Pro 6.0 (Oroudjev et al., 2010). Changes in length ≥ 0.3 μm were defined as growth or shortening
events, <0.3 μm were pause (or attenuation). The time-based catastro- phe frequency is the number of catastrophes (transitions from growth or pause to shortening) divided by total time spent growing and paused.
≥
Rescue frequency is the number of transitions from shortening to growth or pause divided by total time spent shortening. Dynamicity is total length grown and shortened divided by total duration of imaging a MT. Data were pooled from 3 experiments and then imported into R for statistical analysis and graphing (described above).
⦁ Results

⦁ MMAE action on purified tubulin heterodimer subunits and microtubules
⦁ MMAE binds to soluble tubulin heterodimers with a maximum stoichiometry of ~1:1
To determine the binding stoichiometry of MMAE to free tubulin subunits, soluble MAP-free tubulin heterodimers were incubated with 3H-MMAE for 12 min at 30 ◦C, and then tubulin-bound drug was sepa-
rated from free drug by size exclusion chromatography. As shown in Fig. 1, MMAE binding saturated at ~1.3 MMAE molecules/tubulin heterodimer and had an apparent Kd of ~1.6 μM, suggesting that MMAE binds to tubulin heterodimer subunits at a ratio of approximately 1:1.
⦁ MMAE binds with high affinity to MT ends and lower affinity along the MT length
Next, the interaction between MMAE and pre-assembled MTs was investigated. MAP-rich tubulin was polymerized to steady-state at 30 ◦C and then incubated with 3H-MMAE for 6 min, after which MT-bound
drug was separated from free drug by centrifugation. Control experi- ments demonstrated that the brief 6 min incubation with MMAE did not affect the mass of MTs recovered after centrifugation although it did result in a modest drop in A350 signal (Supplemental Fig. 1), indicating that this brief treatment of MTs with MMAE causes a structural alter- ation that affects light scattering without affecting the ability to pellet. As shown in Fig. 2, even at the highest concentration of MMAE tested (20 μM), MMAE binding did not even begin to saturate. In fact, even at this high concentration, the drug bound to only ~20% of the total tubulin present in the MT polymers. These data indicate a low affinity, high stoichiometry binding of MMAE along the length of the MTs. Additionally, careful examination of the data from the low MMAE concentration part of the curve (inset) suggests a limited number of high-affinity (sub-micromolar) MMAE binding sites, most likely corre- sponding to the MT ends. This MMAE binding pattern is quite similar to that of vinblastine binding to MTs (Wilson et al., 1982).
⦁ MMAE binding induces structural defects at both MT ends and along the MT length that may expose additional binding sites
The high abundance of MT binding sites for MMAE prompted us to examine the structural effects of MMAE upon pre-assembled MTs. MAP- rich MTs were prepared and incubated with MMAE as above and then examined by transmission electron microscopy (TEM). As shown in Fig. 3, MMAE-treated MTs exhibited both peeling of protofilamentous structures at their ends and marked “bubble-like” defects along their lengths – the latter structures appear to be protofilaments that have loosened from the MT body in an unusual way and bulge or protrude out in a bubble or partial loop that is connected linearly (but not laterally) to the other seemingly normal protofilaments. This is especially clear in the 5 μM MMAE images (Fig. 3). These observations are consistent with recent work indicating that MMAE can interfere with lateral tubulin- tubulin contacts in MTs (Waight et al., 2016). To the best of our knowledge, the observed bubble-like defects along the MT lengths are unique among all other MTAs. Additionally, many MTs appeared “wavy” as opposed to the usual rigid rod morphology. Finally, some MTs appeared severely affected while others in the same field appeared relatively normal, raising the possibility that MMAE action along the MT

Fig. 1. MMAE binding to soluble tubulin heterodimers. [3H]-MMAE (0.1–20 μM) was incubated with soluble, MAP-free tubulin for 12–15 min at 30 ◦C, and then tubulin and bound MMAE was separated from unbound MMAE by size exclusion chromatography. MMAE bound with an apparent Kd = 1.6 μM and a maximum stoichiometry of ~1.3 drug molecules per tubulin heterodimer. Each data point is the average from 3 to 4 experiments.

Fig. 2. MMAE binding to microtubules. [3H]-MMAE (0.1–20 μM) was incubated with microtubules made from MAP-rich tubulin for 6 min at 30 ◦C, and microtubule- bound MMAE was separated from unbound MMAE by centrifugation. The ratio of 3H-MMAE to tubulin in the pellet was determined. Each point is the average of 2–4 experiments (inset) or 1–4 experiments (overall), and the line shows a best fit. See Methods for additional experimental details.

length may be cooperative, i.e., that MMAE binding may cause struc- tural effects that expose additional MMAE binding sites. This would be consistent with the lack of binding saturation observed in Fig. 2, even at the highest MMAE concentrations tested.
⦁ MMAE binding potently suppresses MT dynamics
The fundamental action of most MTAs is to modulate (usually to suppress) the dynamic behaviors of MTs, although different MTAs reportedly affect different individual parameters underlying this behavior (reviewed in (Jordan and Wilson, 2004)). Therefore, we evaluated the effects of MMAE upon the regulation of MT dynamics
using MTs composed of purified, MAP-free tubulin. MTs were assembled to equilibrium (32 ◦C, 30 min) from seeds attached to glass coverslips, in the presence or absence of 100 nM MMAE. Under these conditions,
tubulin was present in vast excess over drug. Key parameters of MT dynamics were quantified, including the (i) rate of growth and short- ening, (ii) duration of growth and shortening events, (iii) fraction of time spent growing, shortening or attenuated, (iv) frequency of rescues (transitions from shortening to attenuation or growth) and catastrophes (transitions from growth or attenuation to shortening), and (v) overall “dynamicity”, which is the total length grown and shortened by a MT divided by the total observation time of that MT.

Fig. 3. Microtubules exposed to MMAE post-assembly show structural defects at their ends and along their lengths. Microtubules exposed to MMAE for 6 min post- assembly have peeling ends (white arrows) and defects along their lengths (black arrows). The microtubule preparation and incubation time are identical to those used in in the microtubule binding experiments shown in Fig. 2. See Supp. Fig. 1 for corresponding light scattering data.

As shown in Table 1 (right side of table), the most marked effects of MMAE on bovine MT dynamics were to reduce the growth and short- ening lengths per event. MMAE also reduced the percentage of time MTs spent growing, increased the percentage of time that MTs spent in an attenuated state, increased the duration of attenuation events, and greatly reduced overall dynamicity. Box plot presentations of these data are presented in Supplemental Fig. 2. Finally, although MMAE trended toward decreasing shortening rate, growth rate and catastrophe fre- quency, the differences did not reach statistical significance.
⦁ MMAE potently reduces the kinetics and extent of MT assembly
In order to gain additional insights into the mechanisms of MMAE action, we next asked what effects MMAE might have upon the kinetics and steady-state products of MT assembly when it is present during MT
assembly. MAP-rich tubulin was mixed on ice with varying concentra- tions of MMAE, and then transferred to 30 ◦C to initiate MT assembly. As shown in Fig. 4A, the presence of MMAE led to a concentration- dependent reduction in light scattering (A350), with an IC50 of ~3 μM.
These same data are plotted in a normalized manner in Supplemental Fig. 3, to better define the half time (t1/2) for maximal assembly. A similar concentration-dependence was observed on the mass of

Table 1
Effects of MMAE on dynamic instability of microtubules in MCF7 cells or in vitro microtubules from MAP-free bovine brain tubulin. Control (vehicle) and MMAE columns show mean ± SEM, and MMAE columns also indicate percent change from control and its direction. Statistical analysis of the difference between the means was determined, and the 95% CI for the difference and the p-value are indicated (see Methods).
Parameter MCF7 cells (MAP-rich) MAP-free bovine brain tubulin

Control MMAE, 0.5 nM (% change)
Difference in means (95% CI)
p-Value Control MMAE, 100 nM (%
change)
Difference in means (95% CI)
p- Value

Growth rate (μm/min) 9.7 ± 0.20 7.5 ± 0.15 (—23%) 2.19 (1.71, 2.68) <0.0001 1.1 ±
0.05
Growth length (μm) 1.8 ± 0.07 1.2 ± 0.05 (—30%) 0.52 (0.35, 0.79) <0.0001 2.9 ±
0.31
0.97 ± 0.07 (—12%) 0.13 (—0.05, 0.30) NS
1.6 ± 0.13 (—45%) 1.30 (0.68, 1.98) 0.0002

Growth duration (min) 0.19 ±
0.007
0.18 ± 0.007
(—6%)
0.01 (—0.01, 0.03) NS 2.8 ±
0.30
1.9 ± 0.18 (—32%) 0.89 (0.24, 1.58) 0.009

Shortening rate (μm) 14.8 ± 0.5 11.0 ± 0.3 (—26%) 3.81 (2.65, 5.00) 0.0001 12.3 ±
1.9
Shortening length (μm) 1.8 ± 0.11 1.1 ± 0.06 (—38%) 0.69 (0.46, 0.93) <0.0001 3.8 ±
0.56
9.1 ± 1.3 (—26%) 3.20 (—1.17, 7.96) NS
2.2 ± 0.31 (—43%) 1.59 (0.40, 2.87) 0.012

Catastrophe frequency (#/min/MT)
3.0 ± 0.16 2.4 ± 0.15 (—20%) 0.61 (0.19, 1.03) 0.004 0.31 ±
0.06
0.24 ± 0.04 (—22%) 0.70 (—0.06, 0.21) NS

Rescue frequency (#/min/ MT)
9.5 ± 0.41 10.1 ± 0.37 (+6%) —0.61 (—1.67, 0.48) NS 2.5 ± 1.2 3.4 ± 1.4 (+35%) —0.88 (—4.62, 2.66) NS

Attenuation duration (min)
0.12 ±
0.004
0.16 ± 0.006
(+32%)
—0.04 (—0.05, —0.03) <0.0001 0.79 ±
0.07
1.4 ± 0.15 (+79%) —0.63 (—0.96, —0.31) 0.0002

Percent time growing (%) 49 40 (—18%) – 79 49 (—38%) –

Percent time shortening (%)
Percent time attenuated (%)
24 19 (—21%) – 6 6 –
27 41 (+52%) – 15 45 (+200%) –

Dynamicity (μm/min/MT) 8.2 ± 0.27 5.0 ± 0.22 (—38%) 3.15 (2.47, 3.84) <0.0001 1.7 ±
0.18
0.95 ± 0.13 (—44%) 0.77 (0.35, 1.22) 0.0008

Fig. 4. MMAE inhibits the assembly of MAP-rich tubulin in vitro and alters the morphology of MT ends. Assembly mixes containing MAP-rich tubulin (~30 μM), 1 mM GTP, and various concentrations of MMAE were prepared on ice and then transferred to a 30 ◦C spectrophotometer. A) Microtubule assembly was monitored by
light scattering at 350 nm. B) After 60 min, samples were centrifuged to collect microtubules and the polymer mass was determined. Each data point in A and B is the average of 2–4 experiments normalized to the pooled average of the control (no MMAE) values. The dotted line indicates 50% of the control value. C) Samples prepared for transmission electron microscopy in parallel with centrifugation show an MMAE concentration-dependent transition from microtubules to tubulin rings.
Scale bar = 250 nm. Note that the 10 μM data show that the tubulin rings do not scatter or sediment during centrifugation. D) MT end morphologies were categorized
as blunt, splayed, or frayed and scored in the absence (CTR) or presence of varying concentrations of MMAE. The width of each set of bars is proportional to the total number of ends scored in each condition (CTR n = 184; 1 μM n = 243; 2 μM n = 160; 2.5 μM 126; 5 μM n = 139; 10 μM n = 2). The bar at the right indicates the proportions across all experimental conditions. E) Logistic regression shows that the proportion of frayed ends increases as a function of MMAE concentration (blue line; p < 0.0001; log-odds increase 0.33 per μM, 95%CL 0.24–0.43). Each point represents an individual imaging field. Excluding the 10 μM data point had only minor effects (red line) and did not change the interpretation. Data in D and E were pooled from 3 independent experiments. (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this article.)

pelletable material harvested at the 60-min plateau time point (Fig. 4B). Importantly, since the tubulin concentration in these reactions was ~16 μM while the IC50 for MMAE was ~3 μM, these data indicate that the mechanism of MMAE action is sub-stoichiometric relative to tubulin, consistent with at least a subset of MMAE action being mediated at MT ends.
For comparison, we also measured the activity of an MMAE-
Antibody Drug Conjugate (ADC; CNJ2985-vc-MMAE) in this assay (Supplemental Fig. 4). The MMAE ADC contains ~3.5 MMAE drugs per antibody. We assembled MTs as above in the presence of 1 μM or 5 μM MMAE-ADC (equivariant to 3.5 μM and 17.5 μM of free MMAE). The presence of the MMAE-ADC resulted in very high A350 light scattering signals, possibly indicative of nonspecific aggregation, and we were therefore unable to assess the effects of the ADC on MT assembly by this

method. As an alternative approach, we centrifuged the samples at the 60-min time point and then used SDS-PAGE to determine what pro- portion of the pelleted material was tubulin. By this method, we found that 1 μM and 5 μM MMAE-ADC only reduced MT mass by 7% and 18%, respectively, compared to 28% and 77% with 1 μM and 5 μM free drug, respectively. Considering that there are ~3.5 MMAE molecules per antibody, the simplest interpretation is that MMAE-ADC is approxi- mately 14-fold less potent upon MT assembly compared to its free form. The difference in activities could result from greatly reduced accessi- bility of MMAE to its targets (tubulin and MTs) when it is conjugated to an antibody. The observed low activity of ADC on MT assembly might be due to a low level of MMAE released from ADC in the reaction mixture.
⦁ MMAE reduces the length of assembled MTs, alters the morphology of MT ends, and promotes formation of tubulin rings during MT assembly
To further characterize the concentration-dependent effects of MMAE on MAP-rich MT assembly in isolated MTs, we imaged MT as- sembly products at the 60-min time point in the above experiment by TEM (Fig. 4C). Control samples had normal, linear MT structures with regular end morphologies (blunt or slightly splayed). In contrast, the presence of MMAE during MT assembly led to (i) a marked reduction in MT length (with a 50% reduction at ~0.5 μM; Supplemental Fig. 5), (ii) an increase in the proportion of MT ends with an abnormal “frayed” appearance, defined as having a projection at least as long as the MT width (Fig. 4C, D, E; see (Best et al., 2019) for more detail on end morphologies), and (iii) a shift from MTs to tubulin ring structures (Fig. 4C; Supplemental Fig. 6). Indeed, MTs were exceedingly rare at 10 μM drug. The formation of tubulin rings is consistent with earlier reports examining the parent compound, dolastatin 10 (Bai et al., 1995).

⦁ MMAE action in cultured MCF7 cells
Given the above-described insights into MMAE action in an isolated system, we next sought to mechanistically examine MMAE action in a cellular system. We chose MCF7 cells stably expressing EGFP-tubulin, as this is a well-characterized system in which we have an established methodology for assessing the cellular effects of MTAs. We evaluated the action and potency of MMAE as a free drug and as an ADC (CNJ2985-vc- MMAE). Cellular uptake of the ADC is followed by intracellular enzy- matic cleavage of the vc peptide linker, releasing free drug. Importantly,
since the epitope recognized by CNJ2985-vc-MMAE is not expressed on MCF7 cells, any effects on these cells are likely due to non-specific cellular uptake.
In order to ensure that experiments were performed when the intracellular drug concentration had reached steady state, we first
determined the rate of MMAE uptake. Cells were plated and treated with 10 nM 3H-MMAE, and then the cellular drug content was determined as
a function of time. As shown in Fig. 5, MMAE accumulation proceeded relatively slowly, requiring 24 h of exposure to reach a plateau.
⦁ MMAE, as a free drug but not as an ADC, potently inhibits cell proliferation, induces mitotic arrest, and alters spindle morphology in MCF7 cells
±
We first determined the effects of free MMAE upon cell proliferation using a sulforhodamine B (SRB) assay following a 96 h drug incubation. The percent inhibition of proliferation was determined over a range of MMAE concentrations. As shown in Fig. 6, MMAE potently inhibited cell proliferation with an IC50 of 0.9 nM 0.24 nM.
≥
Next, we examined the effects of MMAE upon the mitotic index, defined as the percentage of cells in mitosis at any given time. As shown in Fig. 7A, MMAE markedly and concentration-dependently increased the percentage of cells in mitosis, with an IC50 of 0.5 nM. The maximum mitotic index observed was 55% at 10 nM MMAE (Fig. 7B). Comple- mentary cell cycle analyses demonstrated that free MMAE strongly induced G2/M arrest in a concentration-dependent manner. Specifically, 50% of the cells were blocked in G2/M after 24 h of incubation at 1 nM MMAE (Fig. 7C).
Finally, we examined the effects of MMAE upon the morphology of the interphase MT network and the mitotic spindle MTs. As shown in Fig. 8, MMAE altered MT structure in interphase cells and caused mitotic spindle abnormalities in a concentration-dependent manner, beginning at very low concentrations (0.5 nM). Higher concentrations of MMAE led to complete MT depolymerization and cell death (Supplemental Fig. 7).
±
The MMAE ADC also inhibited cell proliferation, increased the per- centage of cells in mitosis, altered the interphase MT network, and caused mitotic spindle abnormalities similarly to free MMAE but with greatly reduced potency relative to the free drug. The ADC inhibited cell proliferation with an IC50 of 149 31 nM, 165-fold less potently than free MMAE (Fig. 6). With respect to inhibition of mitosis, the ADC also

Fig. 5. Intracellular uptake of free MMAE. MCF7 cells were incubated with 10 nM tritiated MMAE ([3H]-MMAE). At the indicated time points, samples of cells were rinsed with fresh buffer, scintillation fluid added, and then radioactivity measured in a scintillation counter. Data are from 4 to 5 experiments.

Fig. 6. Inhibition of MCF7 cell proliferation by MMAE as a free drug (MMAE) and as an antibody- drug conjugate (CNJ2985). The IC50 for cell proliferation at 96 h was 0.9 ± 0.2 nM for MMAE and 149 ± 31 nM for CNJ2985.

increased the percentage of cells in mitosis, but 3–4 orders of magnitude less potently than the free drug (Fig. 7A). Because the effect of the ADC did not reach an upper plateau at the concentrations tested here, its IC50 could not be determined. Similarly, cell cycle analyses showed greatly reduced potency for the MMAE ADC compared with free MMAE (Fig. 7B), and induced G2/M arrest 2–3 orders of magnitude less potently than the free drug (Fig. 7C). The ADC had a moderate effect on interphase MTs, producing similar spindle abnormalities as free MMAE but only at much higher concentrations (200–1000 nM ADC was com- parable to 0.5 nM free drug). Spindle abnormalities were not observed with ADC at 200 nM (but were obvious with free drug at 0.5 nM). At 1 μM, the MMAE ADC produced similar abnormalities to those of 0.5 nM MMAE (Fig. 8). Collectively, these results demonstrate that MMAE potently affects MCF7 cell proliferation, mitosis, and the integrity of the MT network both during interphase and mitosis. The effects of the MMAE ADC, although much less potent than the free drug, were qual- itatively similar. As MCF7 cells lack the epitope that is recognized by the antibody, this suggests that some small subset of the MMAE ADC was non-specifically taken up into the cells and free MMAE released as a result of intracellular proteolysis. Since free MMAE acts on cells in the nM range and the MMAE ADC is 2–3 orders of magnitude less potent than the free drug in cells, we reason that the intracellular concentration of the MMAE resulting from non-specific uptake of the ADC is on the order of pM concentrations.
⦁ MMAE potently suppresses MT dynamics in cells, similar to its effects on purified tubulin in vitro
We sought to quantitate the effects of MMAE upon the dynamics of cellular MTs. MCF7 cells were grown on coverslips and treated for 24 h with 0.5 nM MMAE. We chose this drug concentration because it was the IC50 for mitotic block, and because after 24 h at this concentration, cells maintained the flat morphology required for effective MT imaging. Following exposure to drug or vehicle, MTs were imaged by time- resolved confocal microscopy. As shown in Table 1, MMAE suppressed both the rate and length of growth and shortening events, greatly increased the duration of individual attenuation events, and increased the percentage of time MTs spent in an attenuated state. It also decreased catastrophe frequency without affecting rescue frequency. The strong suppressive effects of MMAE on MT dynamics were also apparent in overall dynamicity. Box plot presentations of these data are presented in Supplemental Fig. 2.
⦁ Discussion
We have conducted a detailed investigation of MMAE actions on MTs and free tubulin heterodimeric subunits, both as a free drug and as an ADC, employing both reconstituted in vitro systems consisting of het- erodimeric tubulin subunits or MTs as well as cultured human MCF7 cells. In the reconstituted in vitro systems, free MMAE (i) binds to sol- uble tubulin heterodimers with a maximum stoichiometry of approxi- mately 1:1, (ii) binds to pre-assembled MTs abundantly along the MT length and with high affinity at the ends, (iii) potently suppresses MT dynamics, (iv) introduces structural defects both at MT ends and along the MT length that may increase the number of drug binding sites, (v) reduces the kinetics and extent of MT assembly, and (vi) reduces the length of assembled MTs while promoting the formation of tubulin rings. In addition, our cellular studies demonstrate that MMAE potently sup- presses (i) proliferation, (ii) mitosis, (iii) the integrity of the MT network (during both mitosis and interphase), and (iv) MT dynamics. The effects of the antibody-conjugated version of MMAE were qualitatively similar to those of free MMAE, although quantitatively much less potent. A reduction in potency is unsurprising given that these cells do not express the antigen recognized by the conjugated antibody (much like would be expected for non-antigen presenting peripheral neurons).
Previous mechanistic work has shown that MMAE-ADC action upon
neuronal MTs in peripheral nerves is mediated by nonspecific uptake of the entire MMAE-ADC into the nerve cells followed by intracellular release of the anti-tubulin payload (i.e. MMAE), leading to nerve degeneration (Stagg et al., 2016). For this reason, the main focus of this work was to understand how MMAE affects MT biochemistry and dy- namics in purified free tubulin/MTs as well as in cells. Taken together, the simplest interpretation of the data is that MMAE binds extensively to both soluble tubulin and MTs and causes severe dysregulation of the MT network. Specifically, MMAE binding to tubulin heterodimers interferes with MT assembly. MMAE binding to pre-formed MTs along the MT length, at MT ends, and perhaps also at new sites exposed by MMAE- induced structural defects along the MT’s length, results in substantial structural damage to the MTs. Since MMAE-treated cells exhibit deficits in MT-mediated events such as proliferation, mitosis, and the appear- ance of the MT network, we conclude that the structural effects observed in purified tubulin/MTs likely underlie MT dysfunction in cells. While these effects are ideal for attacking highly proliferative tumor cells, MT dysfunction will likely also have negative consequences for peripheral nerves. Even though peripheral nerve cells are not proliferating, they are susceptible to damage by MMAE and other MTAs because of the essential role of the neuronal microtubule network for functions such as

Fig. 7. Free MMAE affects mitotic index, cell cycle stage, and induces G2/M arrest at much lower con- centrations than the MMAE-ADC. A) Mitotic arrest at 24 h in MCF7 cells incubated with MMAE as a free
drug (MMAE) or an MMAE-ADC (CNJ2985). Free MMAE had an IC50 = 0.5 nM, and the maximum mitotic index (defined as the percentage of cells
arrested in any stage of mitosis) was 55% at concen- trations ≥ 10 nM. Mitotic arrest also occurred with the MMAE-ADC but did not reach a plateau at the
concentrations tested, preventing an estimate of the IC50. B) Cell cycle analysis after 24 h incubation with free MMAE or CNJ2985. C) Strong induction of G2/M arrest in MCF7 cells by MMAE in a concentration- dependent manner. After 24 h, ~50% of cells treated with 1 nM free MMAE are blocked in G2/M. CNJ2985 also induced G2/M arrest, but the effect did not reach a plateau at the concentrations tested.

axonal transport along the long axonal projections (Morfini et al., 2009, Almeida-Souza et al., 2011).
Although different MTAs interact with the same target (i.e., MTs) and alter MT dynamics to promote tumor cell death, they vary in the fre- quency and severity of CIPN observed in patients. Paclitaxel and
vinblastine (Carlson and Ocean, 2011; Swain and Arezzo, 2008; Younes et al., 2012) are the most commonly used MTAs and induce relatively high levels of severe CIPN, while more recently developed MTAs such as eribulin (Goel et al., 2009; Gradishar, 2011; Tan et al., 2009; Vahdat et al., 2009) and vinflunine (Swain and Arezzo, 2008) have markedly

Fig. 8. Free MMAE alters the interphase MT network and mitotic spindle at much lower concentrations than the MMAE-ADC. A) Free MMAE alters the
interphase MT network (anti- α-tubulin, green) at concentrations ≥ its mitotic IC50 (0.5 nM), similar to other MT-depolymerizing drugs (DNA stained with
DAPI, blue). B) The MMAE-ADC (CNJ2985) has a
moderate effect on interphase MTs at concentrations
≥ 200 nM. C) Free MMAE induces spindle abnor- malities at its mitotic IC50, similar to other MT depolymerizing drugs. D) Top: 200 nM CNJ2985 does
not induce significant spindle abnormalities. Bottom: Spindle abnormalities induced by 1000 nM CNJ2985 are similar to those produced by free MMAE at ~0.5 nM. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

lower incidences. Since MTA-induced peripheral neuropathy is likely due to the disruption of normal MT-dependent events such as axonal transport, understanding how different MTAs bind to different sites on MTs and/or soluble tubulin and differentially affect MT dynamics (Jordan and Kamath, 2007; Jordan and Wilson, 2004) compared to MMAE can provide mechanistic insights into MMAE-ADC peripheral neuropathy.
⦁ How might MMAE-ADCs induce severe peripheral neuropathy?
There are a number of non-mutually exclusive hypotheses that could account for the relatively high level of MMAE-ADC-induced severe pe- ripheral neuropathy observed clinically. One possible hypothesis focuses upon the location and abundance of drug binding sites on MTs, leading to severe dysregulation of the MT network and disruption of MT- dependent events such as axonal transport. Once the MMAE ADC is taken up non-specifically in the peripheral nerve cells and MMAE is released, as demonstrated in this work, MMAE binds to abundant, lower affinity binding sites along the length of MTs as well as with a small number of higher affinity sites at MT ends. This pattern of binding to an abundant number of sites along the length of the MT has been observed with other MTAs that have a high incidence of severe CIPN including vinca alkaloids, paclitaxel and epothilones (Genualdi et al., 2019). Many of the vinca alkaloids (i.e. vinblastine and vincristine) have a very similar binding pattern to MMAE, whereby they bind a large number of relatively low affinity binding sites along the MT length and to a small number of high affinity sites at MT ends (Gigant et al., 2005; Hans et al., 1978; Mirsalis et al., 1999; Wilson et al., 1975). Paclitaxel and the epothilones also bind to abundant sites along the length of the MT, although these binding sites are distinct from MMAE and vinblastine
binding sites since they are located on the inside of the MT (Amos and
Lowe, 1999; Bollag et al., 1995; Diaz et al., 1998; Nettles et al., 2004; Nogales et al., 1998; Nogales et al., 1995; Prota et al., 2013). These various binding patterns are in contrast to eribulin, which binds to only a small number of high affinity sites at MT ends but lacks the abundant, low affinity sites along the length of the MT (Smith et al., 2010). Eribulin also exhibited markedly lower incidences of severe CIPN in patients (Gradishar, 2011). DM1, an approved payload in ado-trastuzumab emtansine (T-DM1) was evaluated for its in vitro binding as a free drug in the form of S-methyl-DM1 and was found to bind similarly to MMAE (high affinity binding sites at the end and low affinity binding sites along the length (Lopus et al., 2010). Consistent with the above model, peripheral neuropathy has been a frequent adverse event with almost all DM1 and DM4 containing conventional ADCs (Bendell et al., 2013; Younes et al., 2012). T-DM1, however, had a very low incidence that was primarily grade 1 and sensory and did not result in dose delays, reductions or discontinuations, which may be attributed to other prop- erties of the ADC (e.g. less uptake of the ADC and/or release of its payload in the peripheral nerves).
In addition, high affinity for MTs may contribute to the frequency of
severe peripheral neuropathy. Indeed, paclitaxel has a very high affinity for MTs, followed by vincristine followed by vinblastine followed by eribulin (S. Lobert et al., 1996; Nogales et al., 1995; Smith et al., 2010; Wilson et al., 1982). This order of affinities correlates with the frequency of severe peripheral neuropathy.
Yet another possible contributor to the frequency of severe periph- eral neuropathy derives from the fact that exposure of pre-formed MTs to either MMAE or vinblastine at high concentrations induces structural defects in the MT lattice along the length of the MT and protofilament peeling/fraying at MT ends (Fig. 3, (Jordan et al., 1986; Wilson et al., 1982). These lattice defects have been proposed to expose additional, otherwise sequestered drug binding sites (Wilson et al., 1982). Since these structural effects occur along the length of the MT lattice and require relatively high drug concentrations, these effects are likely mediated by drug binding at the abundant low affinity sites along the length of the MT. In contrast, exposure of pre-formed MTs to eribulin
results in moderate “splaying” at plus MT ends but no defects along the length (Smith et al., 2010).
Mechanistically, one could speculate that drug-induced structural alterations along the length of MTs would be especially deleterious to neurons, which require especially stable MTs in their axonal compart- ments in order to conduct axonal transport (Baas et al., 2016). Perhaps these defects in the MT lattice interfere with the ability of kinesin and/or dynein to translocate along MTs, thereby interfering with the essential transport of cargo between neuronal cell bodies and their distant syn- apses, such as has been observed in previous studies examining the ef- fects of vincristine, paclitaxel, and ixabepilone on fast axonal transport (LaPointe et al., 2013; Smith et al., 2016). In support of this hypothesis, eribulin, which binds exclusively at MT ends, exhibited weak effects upon the rates of axonal transport relative to vincristine, paclitaxel, and ixabepilone (LaPointe et al., 2013), all of which bind to high abundance sites along the MT length.
While differential MTA binding to MTs (both in terms of binding sites
and affinities for those sites) may underlie their differential clinical in- cidences of CIPN, differences in MTA binding to free tubulin hetero- dimeric subunits may also play a role. MMAE, vinblastine, and eribulin all bind to free tubulin heterodimers. However, there are differences in the details of each binding event, leading to the formation of different MTA-tubulin structures. For example, at relatively high concentrations, MMAE binding to tubulin heterodimers leads to the formation of olig- omeric rings, while vinblastine binding leads to formation of oligomeric spirals (Wilson et al., 1982). This is consistent with both drugs binding along the length of a protofilament and promoting a curved structure by acting as a wedge between longitudinally stacked tubulin subunits (Gigant et al., 2005; Wang et al., 2016). However, the fact that MMAE generates rings while vinblastine generates spirals demonstrates that the exact binding events are distinct from one another. Indeed, even the closely related vinca alkaloid vinflunine demonstrates smaller spirals (and a lower rate of severe CIPN) than closely related vinblastine. In contrast, although eribulin also binds to tubulin heterodimers, no obvious oligomeric structures form (Alday and Correia, 2009; Dabydeen et al., 2006). Mechanistically, different modes of binding to tubulin heterodimer subunits could easily lead to different structural effects and unique downstream impacts upon MT function. Consistent with this
model, vinflunine has been shown to promote the formation of fewer,
shorter oligomeric tubulin structures compared to vinblastine (Sharon Lobert et al., 1998)}.
A second hypothesis for the differential incidence of CIPN among MTAs posits different affinities for different tubulin isotypes. This seems very plausible considering that each MTA has its own precise binding interactions with MTs/tubulin subunits, as demonstrated at the level of x-ray crystallography (Doodhi et al., 2016; Waight et al., 2016; Wang et al., 2016). These subtle structural differences between the in- teractions of different MTAs with differentially expressed α or β tubulin isotypes would be predicted to affect the extent of MTA-mediated MT disruption and resulting cellular consequences. Consistent with this notion, eribulin is less active on MTs containing the neuronal-specific βIII tubulin isoform than on non-neuronal MTs found in other cell types that lack this isoform (Wilson et al., 2015). An analysis of data generated with ixabepilone, paclitaxel or vincristine in similar experi- ments in bovine brain MTs or cultured cells (MCF7 or NSCLC H460) in the absence of neuronal βIII tubulin did not show increased activity (Genualdi et al., 2019). It is possible that MMAE and other MTAs with relatively high frequencies of severe peripheral neuropathy may exhibit relatively strong interactions with the tubulin isotypes expressed in peripheral neurons, resulting in their accompanying deleterious consequences.
A third hypothesis for increased MMAE ADC-induced peripheral neuropathy is a more complete suppression of MT dynamics (i.e. both growth and shortening of MTs) than for some other MTAs, resulting in a more severe dysregulation of MT-dependent events such as axonal transport. At sub-stoichiometric concentrations in isolated MTs and in

cells, both MMAE and vinblastine suppress MT growing and shortening parameters and greatly increase the time that MTs spend in an attenu- ated state ((Dhamodharan et al., 1995; Jordan et al., 1985; Toso et al., 1993); Table 1). Maytansine and its thio-containing derivative DM1 act similarly (Lopus et al., 2010). More specifically, at these sub- stoichiometric concentrations, MMAE and vinblastine both increase the amount of time that MTs spend in an attenuated state by suppressing both the on rate (addition, growth) and the off rate (loss, shortening) of tubulin subunits. This mechanism of action has been proposed to represent “true kinetic suppression,” as opposed to other MTAs that do not directly reduce both on and off rates (Castle et al., 2017). For example, both eribulin (Jordan et al., 2005; Smith et al., 2010) and vinflunine (Ngan et al., 2000), which are both associated with relatively low incidences of CIPN clinically, affect MT growth but not shortening parameters, perhaps underlying their lower toxicities. Importantly, the potent action of all four of these drugs at sub-stoichiometric concen- trations indicates that their suppressive effects on MT dynamics are most likely due to drug binding to, and action upon, high affinity binding sites located at MT ends.
⦁ Future directions
Overall, this work suggests that MMAE causes severe MT dysregu- lation and provides possible hypotheses for MMAE-mediated inhibition of MT-dependent axonal transport leading to severe cases of peripheral neuropathy in patients. To further study these hypotheses, it would be extremely valuable to study MMAE and other MTAs in neuronal cells. One especially attractive route might be to use models of nociceptor neurons such as human iPS cells differentiated into a nociceptor phenotype (Chambers et al., 2012). These systems would allow an assessment of how MMAE and other MTAs affect critical neuron-specific features such as the maintenance of neurites and intracellular, MT- dependent transport. However, the extended morphology of these neuronal systems and their very high axonal MT density make assays of individual MT dynamics challenging, although the use of super- resolution microscopy techniques have allowed for some advances (Dent, 2017). Additionally, it would be very informative to test MMAE for its effects upon cells harboring or lacking the neural-specific βIII isoform of tubulin. As mentioned earlier, eribulin was found to be less active on βIII-containing MTs than on non βIII-containing MTs, perhaps contributing to eribulin’s lower incidence of severe CIPN (Wilson et al., 2015). One might hypothesize similar activity with MMAE on βIII-con-
taining MTs as non βIII-containing MTs given the high incidences of
severe peripheral neuropathy, but this would need to be assessed directly to establish if this is a good indicator of severe peripheral neuropathy risk. In addition, the data with the MMAE ADC in cultured cells that do not express the antigen demonstrated that MMAE ADC can
be taken up into cells to release free MMAE, and this is likely the main
review and editing.
Herb Miller – conceptualization, data curation, formal analysis, investigation, methodology, writing writing original draft.
Christine Genualdi – conceptualization, data curation, formal anal- ysis, investigation, methodology, writing – original draft.
Stephen Chih –data curation, formal analysis, investigation, meth- odology, writing – original draft.
Ben-Quan Shen – data curation, formal analysis, investigation, methodology, writing original draft, writing – review and editing.
Mary Ann Jordan – conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, supervision, writing original draft, writing – review and editing.
Leslie Wilson – conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, supervision, writing original draft, writing – review and editing.
Stuart C. Feinstein – conceptualization, data curation, formal anal- ysis, funding acquisition, investigation, methodology, supervision, writing original draft, writing – review and editing.
Nicola J. Stagg – conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, supervision, writing original draft, writing – review and editing.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments
The authors gratefully acknowledge technical assistance from Andrew Yang, Amairany Benitez, Andrew Lee and Colleen Sweeney. We also appreciate the expertise and instrumentation of the NRI-MCDB Microscopy Facility. Thank you also to Rebecca Rowntree and Thomas Pillow for providing unlabeled MMAE and MMAE ADC (CNJ2985-vc- MMAE) and Ben-Quan Shen’s lab for providing tritium-labeled MMAE for this work. This work was supported by a grant from Genentech to SCF, LW and MAJ.

Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi. org/10.1016/j.taap.2021.115534.

References
Alday, P.H., Correia, J.J., 2009. Macromolecular interaction of halichondrin B analogues

pathway to deliver the cytotoxic agent to peripheral nerves and
eribulin (E7389) and ER-076349 with tubulin by analytical ultracentrifugation.

contribute to the resulting peripheral neuropathy. Finally, efforts to develop new and improved tubulin/MT targeted drugs (MTAs and/or MTA conjugates) with reduced incidence of severe CIPN may be pro- moted by careful consideration of mechanisms and locations of binding events as well as the regulation of MT dynamics, as discussed here and more extensively in Genualdi et al., 2019.
Author contributions
Rebecca L. Best – conceptualization, data curation, formal analysis, investigation, methodology, writing original draft, writing – review and editing.
Rebecca L. Best – conceptualization, data curation, formal analysis, investigation, methodology, writing original draft, writing – review and editing.
Olga Azarenko – conceptualization, data curation, formal analysis, investigation, methodology, writing writing original draft, writing -
Biochemistry 48 (33), 7927–7938. https://doi.org/10.1021/bi900776u.
Almeida-Souza, L., Timmerman, V., Janssens, S., 2011. Microtubule dynamics in the peripheral nervous system: A matter of balance. Bioarchitecture 1 (6), 267–270. https://doi.org/10.4161/bioa.1.6.19198.
Amos, L.A., Lowe, J., 1999. How Taxol stabilises microtubule structure. Chem. Biol. 6 (3), R65–R69. Retrieved from. https://www.ncbi.nlm.nih.gov/pubmed/10074470.
Argyriou, A.A., Marmiroli, P., Cavaletti, G., Kalofonos, H.P., 2011. Epothilone-induced peripheral neuropathy: a review of current knowledge. J. Pain Symptom Manag. 42 (6), 931–940. https://doi.org/10.1016/j.jpainsymman.2011.02.022.
Argyriou, A.A., Bruna, J., Marmiroli, P., Cavaletti, G., 2012. Chemotherapy-induced peripheral neurotoxicity (CIPN): an update. Crit. Rev. Oncol. Hematol. 82 (1), 51–77. https://doi.org/10.1016/j.critrevonc.2011.04.012.
Azarenko, O., Smiyun, G., Mah, J., Wilson, L., Jordan, M.A., 2014. Antiproliferative mechanism of action of the novel taxane cabazitaxel as compared with the parent compound docetaxel in MCF7 breast cancer cells. Mol. Cancer Ther. 13 (8), 2092–2103. https://doi.org/10.1158/1535-7163.MCT-14-0265.
Baas, P.W., Rao, A.N., Matamoros, A.J., Leo, L., 2016. Stability properties of neuronal microtubules. Cytoskeleton (Hoboken) 73 (9), 442–460. https://doi.org/10.1002/ cm.21286.
Bai, R., Pettit, G.R., Hamel, E., 1990. Dolastatin 10, a powerful cytostatic peptide derived from a marine animal. Inhibition of tubulin polymerization mediated through the vinca alkaloid binding domain. Biochem. Pharmacol. 39 (12), 1941–1949. Retrieved from.

Bai, R., Taylor, G.F., Schmidt, J.M., Williams, M.D., Kepler, J.A., Pettit, G.R., Hamel, E., 1995. Interaction of dolastatin 10 with tubulin: induction of aggregation and binding and dissociation reactions. Mol. Pharmacol. 47 (5), 965–976. Retrieved from. http s://www.ncbi.nlm.nih.gov/pubmed/7746283.
Benbow, S.J., Wozniak, K.M., Kulesh, B., Savage, A., Slusher, B.S., Littlefield, B.A., Feinstein, S.C., 2017. Microtubule-targeting agents eribulin and paclitaxel differentially affect neuronal cell bodies in chemotherapy-induced peripheral neuropathy. Neurotox. Res. 32 (1), 151–162. https://doi.org/10.1007/s12640-017-
9729-6.
Bendell, J., Blumenschein, G., Zinner, R., Hong, D., Jones, S., Infante, J., Hassan, R., 2013. Abstract LB-291: First-in-human phase I dose escalation study of a novel anti- mesothelin antibody drug conjugate (ADC). BAY 94-9343, in patients with advanced solid tumors. Cancer Research 73 (8 Supplement). https://doi.org/10.1158/1538- 7445.Am2013-lb-291. LB-291-LB-291.
Best, R.L., LaPointe, N.E., Liang, J., Ruan, K., Shade, M.F., Wilson, L., Feinstein, S.C., 2019. Tau Isoform-specific Stabilization of Intermediate States During Microtubule Assembly and Disassembly, 294(33), pp. 12265–12280. https://doi.org/10.1074/ jbc.RA119.009124.
Bollag, D.M., McQueney, P.A., Zhu, J., Hensens, O., Koupal, L., Liesch, J., Woods, C.M., 1995. Epothilones, a new class of microtubule-stabilizing agents with a taxol-like mechanism of action. Cancer Res. 55 (11), 2325–2333. Retrieved from. http s://www.ncbi.nlm.nih.gov/pubmed/7757983.
Carlson, K., Ocean, A.J., 2011. Peripheral neuropathy with microtubule-targeting agents: occurrence and management approach. Clin. Breast Cancer 11 (2), 73–81. https:// doi.org/10.1016/j.clbc.2011.03.006.
Castle, B.T., McCubbin, S., Prahl, L.S., Bernens, J.N., Sept, D., Odde, D.J., 2017.
Mechanisms of kinetic stabilization by the drugs paclitaxel and vinblastine. Mol. Biol. Cell 28 (9), 1238–1257. https://doi.org/10.1091/mbc.E16-08-0567.
Chambers, S.M., Qi, Y., Mica, Y., Lee, G., Zhang, X.J., Niu, L., Studer, L., 2012. Combined small-molecule inhibition accelerates developmental timing and converts human pluripotent stem cells into nociceptors. Nat. Biotechnol. 30 (7), 715–720. https:// doi.org/10.1038/nbt.2249.
Dabydeen, D.A., Burnett, J.C., Bai, R., Verdier-Pinard, P., Hickford, S.J., Pettit, G.R., Hamel, E., 2006. Comparison of the activities of the truncated halichondrin B analog NSC 707389 (E7389) with those of the parent compound and a proposed binding site on tubulin. Mol. Pharmacol. 70 (6), 1866–1875. https://doi.org/10.1124/ mol.106.026641.
Dent, E.W., 2017. Of microtubules and memory: implications for microtubule dynamics in dendrites and spines. Mol. Biol. Cell 28 (1), 1–8. https://doi.org/10.1091/mbc. e15-11-0769.
Dhamodharan, R., Jordan, M.A., Thrower, D., Wilson, L., Wadsworth, P., 1995.
Vinblastine suppresses dynamics of individual microtubules in living interphase cells. Mol. Biol. Cell 6 (9), 1215–1229. Retrieved from. https://www.ncbi.nlm.nih. gov/pubmed/8534917.
Diaz, J.F., Valpuesta, J.M., Chacon, P., Diakun, G., Andreu, J.M., 1998. Changes in microtubule protofilament number induced by Taxol binding to an easily accessible site. Internal microtubule dynamics. J. Biol. Chem. 273 (50), 33803–33810.
=
Retrieved from. http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd Retrieve
= = =
&db PubMed&dopt Citation&list_uids 9837970.
Doodhi, H., Prota, A.E., Rodriguez-Garcia, R., Xiao, H., Custar, D.W., Bargsten, K., Steinmetz, M.O., 2016. Termination of protofilament elongation by eribulin induces lattice defects that promote microtubule catastrophes. Curr. Biol. 26 (13), 1713–1721. https://doi.org/10.1016/j.cub.2016.04.053.
Field, J.J., Kanakkanthara, A., Miller, J.H., 2014. Microtubule-targeting agents are clinically successful due to both mitotic and interphase impairment of microtubule function. Bioorg. Med. Chem. 22 (18), 5050–5059. https://doi.org/10.1016/j. bmc.2014.02.035.
Gell, C., Bormuth, V., Brouhard, G.J., Cohen, D.N., Diez, S., Friel, C.T., Howard, J., 2010.
Microtubule dynamics reconstituted in vitro and imaged by single-molecule fluorescence microscopy. Methods Cell Biol. 95, 221–245. https://doi.org/10.1016/ S0091-679X(10)95013-9.
Genualdi, C., Feinstein, S.C., Wilson, L., Jordan, M.A., Stagg, N.J., 2020. Assessing the utility of in vitro microtubule assays for studying mechanisms of peripheral neuropathy with the microtubule inhibitor class of cancer chemotherapy. Chem Biol Interact 315, 108906. https://doi.org/10.1016/j.cbi.2019.108906.
Gigant, B., Wang, C., Ravelli, R.B., Roussi, F., Steinmetz, M.O., Curmi, P.A., Knossow, M., 2005. Structural basis for the regulation of tubulin by vinblastine. Nature 435 (7041), 519–522 doi:nature03566 [pii]. https://doi.org/10.1038/nature03566.
Goel, S., Mita, A.C., Mita, M., Rowinsky, E.K., Chu, Q.S., Wong, N., Takimoto, C.H., 2009. A phase I study of eribulin mesylate (E7389), a mechanistically novel inhibitor of microtubule dynamics, in patients with advanced solid malignancies. Clin. Cancer Res. 15 (12), 4207–4212. https://doi.org/10.1158/1078-0432.CCR-08-2429.
Gradishar, W.J., 2011. The place for eribulin in the treatment of metastatic breast cancer.
Curr. Oncol. Rep. 13 (1), 11–16. https://doi.org/10.1007/s11912-010-0145-9.
Hans, F.O., Dickerson, R.M., Wilson, L., Owellen, R.J., 1978. Differences in the binding properties of vinca alkaloids and colchicine to tubulin by varying protein sources and methodology. Biochem. Pharmacol. 27 (1), 71–76. Retrieved from. htt ps://www.ncbi.nlm.nih.gov/pubmed/563724.
Hyman, A., Drechsel, D., Kellogg, D., Salser, S., Sawin, K., Steffen, P., Mitchison, T., 1991. Preparation of modified tubulins. Methods Enzymol. 196, 478–485. Retrieved from. http://www.ncbi.nlm.nih.gov/pubmed/2034137.
Jordan, M.A., Kamath, K., 2007. How do microtubule-targeted drugs work? An overview.
Curr. Cancer Drug Targets 7 (8), 730–742. https://doi.org/10.2174/
156800907783220417.

Jordan, M.A., Wilson, L., 2004. Microtubules as a target for anticancer drugs. Nat. Rev. Cancer 4 (4), 253–265. Retrieved from. http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?cmd Retrieve&db PubMed&dopt Citation&list_uids 15057285.
= = = =
Jordan, M.A., Himes, R.H., Wilson, L., 1985. Comparison of the effects of vinblastine, vincristine, vindesine, and vinepidine on microtubule dynamics and cell proliferation in vitro. Cancer Res. 45 (6), 2741–2747. Retrieved from. https://www. ncbi.nlm.nih.gov/pubmed/3986806.
Jordan, M.A., Margolis, R.L., Himes, R.H., Wilson, L., 1986. Identification of a distinct class of vinblastine binding sites on microtubules. J. Mol. Biol. 187 (1), 61–73. https://doi.org/10.1016/0022-2836(86)90406-7.
Jordan, M.A., Kamath, K., Manna, T., Okouneva, T., Miller, H.P., Davis, C., Wilson, L., 2005. The primary antimitotic mechanism of action of the synthetic halichondrin E7389 is suppression of microtubule growth. Mol. Cancer Ther. 4 (7), 1086–1095. https://doi.org/10.1158/1535-7163.MCT-04-0345.
LaPointe, N.E., Morfini, G., Brady, S.T., Feinstein, S.C., Wilson, L., Jordan, M.A., 2013. Effects of eribulin, vincristine, paclitaxel and ixabepilone on fast axonal transport and kinesin-1 driven microtubule gliding: implications for chemotherapy-induced peripheral neuropathy. Neurotoxicology 37, 231–239. https://doi.org/10.1016/j. neuro.2013.05.008.
Lobert, S., Vulevic, B., Correia, J.J., 1996. Interaction of vinca alkaloids with tubulin: a comparison of vinblastine, vincristine, and vinorelbine. Biochemistry 35 (21), 6806–6814. https://doi.org/10.1021/bi953037i.
Lobert, S., Ingram, J.W., Hill, B.T., Correia, J.J., 1998. A comparison of thermodynamic parameters for vinorelbine- and vinflunine-induced tubulin self-association by sedimentation velocity. Mol. Pharmacol. 53 (5), 908. Retrieved from. http://molph arm.aspetjournals.org/content/53/5/908.abstract.
Lopus, M., Oroudjev, E., Wilson, L., Wilhelm, S., Widdison, W., Chari, R., Jordan, M.A., 2010. Maytansine and cellular metabolites of antibody-maytansinoid conjugates strongly suppress microtubule dynamics by binding to microtubules. Mol. Cancer Ther. 9 (10), 2689–2699 doi:9/10/2689 [pii]. https://doi.org/10.1158/1535-7163. MCT-10-0644.
Margolin, K., Longmate, J., Synold, T.W., Gandara, D.R., Weber, J., Gonzalez, R., Doroshow, J.H., 2001. Dolastatin-10 in metastatic melanoma: a phase II and pharmokinetic trial of the California Cancer Consortium. Invest. New Drugs 19 (4), 335–340. Retrieved from. https://www.ncbi.nlm.nih.gov/pubmed/11561695.
Masters, J.C., Nickens, D.J., Xuan, D., Shazer, R.L., Amantea, M., 2017. Clinical toxicity of antibody drug conjugates: a meta-analysis of payloads. Investig. New Drugs. https://doi.org/10.1007/s10637-017-0520-6.
Miller, H.P., Wilson, L., 2010. Preparation of microtubule protein and purified tubulin from bovine brain by cycles of assembly and disassembly and phosphocellulose chromatography. In: Wilson, L., Correia, J.J. (Eds.), Methods in Cell Biology, vol. 95. Elsevier Inc, pp. 3–15.
Mirsalis, J.C., Schindler-Horvat, J., Hill, J.R., Tomaszewski, J.E., Donohue, S.J., Tyson, C. A., 1999. Toxicity of dolastatin 10 in mice, rats and dogs and its clinical relevance. Cancer Chemother. Pharmacol. 44 (5), 395–402. https://doi.org/10.1007/ s002800050995.
Morfini, G.A., Burns, M., Binder, L.I., Kanaan, N.M., LaPointe, N., Bosco, D.A., Brady, S. T., 2009. Axonal transport defects in neurodegenerative diseases. J. Neurosci. 29 (41), 12776–12786. https://doi.org/10.1523/JNEUROSCI.3463-09.2009.
Nettles, J.H., Li, H., Cornett, B., Krahn, J.M., Snyder, J.P., Downing, K.H., 2004. The binding mode of epothilone A on alpha,beta-tubulin by electron crystallography. Science 305 (5685), 866–869. https://doi.org/10.1126/science.1099190.
Newman, R.A., Fuentes, A., Covey, J.M., Benvenuto, J.A., 1994. Preclinical pharmacology of the natural marine product dolastatin 10 (NSC 376128). Drug Metab. Dispos. 22 (3), 428–432. Retrieved from. https://www.ncbi.nlm.nih.gov/p ubmed/8070319.
Ngan, V.K., Bellman, K., Panda, D., Hill, B.T., Jordan, M.A., Wilson, L., 2000. Novel Actions of the Antitumor Drugs Vinflunine and Vinorelbine on Microtubules, 60(18), pp. 5045–5051. Retrieved from. http://cancerres.aacrjournals.org/content/canres/ 60/18/5045.full.pdf.
= = =
=
Nogales, E., Wolf, S.G., Khan, I.A., Luduena, R.F., Downing, K.H., 1995. Structure of tubulin at 6.5 A and location of the taxol-binding site. Nature 375 (6530), 424–427. Retrieved from. http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd Retrieve &db PubMed&dopt Citation&list_uids 7760939.
= =
Nogales, E., Wolf, S.G., Downing, K.H., 1998. Structure of the alpha beta tubulin dimer by electron crystallography. Nature 391 (6663), 199–203. Retrieved from. htt p://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd Retrieve&db PubMed&dopt
= =
Citation&list_uids 9428769.
Okouneva, T., Azarenko, O., Wilson, L., Littlefield, B.A., Jordan, M.A., 2008. Inhibition of centromere dynamics by eribulin (E7389) during mitotic metaphase. Mol. Cancer Ther. 7 (7), 2003–2011. https://doi.org/10.1158/1535-7163.MCT-08-0095.
Oroudjev, E., Lopus, M., Wilson, L., Audette, C., Provenzano, C., Erickson, H., Jordan, M. A., 2010. Maytansinoid-antibody conjugates induce mitotic arrest by suppressing microtubule dynamic instability. Mol. Cancer Ther. 9 (10), 2700–2713. https://doi. org/10.1158/1535-7163.MCT-10-0645.
Pettit, G.R., Kamano, Y., Herald, C.L., Tuinman, A.A., Boettner, F.E., Kizu, H., Bontems, R.J., 1987. The isolation and structure of a remarkable marine animal antineoplastic constituent: dolastatin 10. J. Am. Chem. Soc. 109 (22), 6883–6885. https://doi.org/10.1021/ja00256a070.
Pitot, H.C., McElroy Jr., E.A., Reid, J.M., Windebank, A.J., Sloan, J.A., Erlichman, C., Ames, M.M., 1999. Phase I trial of dolastatin-10 (NSC 376128) in patients with advanced solid tumors. Clin. Cancer Res. 5 (3), 525–531. Retrieved from. https
://www.ncbi.nlm.nih.gov/pubmed/10100703.
Poruchynsky, M.S., Komlodi-Pasztor, E., Trostel, S., Wilkerson, J., Regairaz, M., Pommier, Y., Fojo, A.T., 2015. Microtubule-targeting agents augment the toxicity of DNA-damaging agents by disrupting intracellular trafficking of DNA repair proteins.

Proc. Natl. Acad. Sci. U. S. A. 112 (5), 1571–1576. https://doi.org/10.1073/ pnas.1416418112.
Prota, A.E., Bargsten, K., Zurwerra, D., Field, J.J., Diaz, J.F., Altmann, K.H., Steinmetz, M.O., 2013. Molecular mechanism of action of microtubule-stabilizing anticancer agents. Science 339 (6119), 587–590. https://doi.org/10.1126/ science.1230582.
Saber, H., Leighton, J.K., 2015. An FDA oncology analysis of antibody-drug conjugates.
Regul Toxicol Pharmacol 71 (3), 444–452. https://doi.org/10.1016/j. yrtph.2015.01.014.
Smith, J.A., Wilson, L., Azarenko, O., Zhu, X., Lewis, B.M., Littlefield, B.A., Jordan, M.A., 2010. Eribulin binds at microtubule ends to a single site on tubulin to suppress dynamic instability. Biochemistry 49 (6), 1331–1337. https://doi.org/10.1021/ bi901810u.
Smith, J.A., Slusher, B.S., Wozniak, K.M., Farah, M.H., Smiyun, G., Wilson, L., Jordan, M. A., 2016. Structural basis for induction of peripheral neuropathy by microtubule- targeting cancer drugs. Cancer Res. 76 (17), 5115–5123. https://doi.org/10.1158/ 0008-5472.CAN-15-3116.
Stagg, N.J., Shen, B.Q., Brunstein, F., Li, C., Kamath, A.V., Zhong, F., Fine, B.M., 2016.
Peripheral neuropathy with microtubule inhibitor containing antibody drug conjugates: challenges and perspectives in translatability from nonclinical toxicology studies to the clinic. Regul. Toxicol. Pharmacol. 82, 1–13. https://doi.org/10.1016/j. yrtph.2016.10.012.
=
Swain, S.M., Arezzo, J.C., 2008. Neuropathy associated with microtubule inhibitors: diagnosis, incidence, and management. Clin. Adv. Hematol. Oncol. 6 (6), 455–467. Retrieved from. http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd Retrieve&db
= = =
PubMed&dopt Citation&list_uids 18567992.
Tan, A.R., Rubin, E.H., Walton, D.C., Shuster, D.E., Wong, Y.N., Fang, F., Rosen, L.S., 2009. Phase I study of eribulin mesylate administered once every 21 days in patients with advanced solid tumors. Clin. Cancer Res. 15 (12), 4213–4219. https://doi.org/ 10.1158/1078-0432.CCR-09-0360.
Team, R, 2016a. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Retrieved from. https:// www.R-project.org/.
Team, R, 2016b. RStudio:Integrated Development for R. MA, Boston. Retrieved from. http://www.rstudio.com/.
Toso, R.J., Jordan, M.A., Farrell, K.W., Matsumoto, B., Wilson, L., 1993. Kinetic stabilization of microtubule dynamic instability in vitro by vinblastine. Biochemistry

32 (5), 1285–1293. Retrieved from. https://www.ncbi.nlm.nih.gov/pubme d/8448138.
Vahdat, L.T., Pruitt, B., Fabian, C.J., Rivera, R.R., Smith, D.A., Tan-Chiu, E., Blum, J.L., 2009. Phase II study of eribulin mesylate, a halichondrin B analog, in patients with metastatic breast cancer previously treated with an anthracycline and a taxane.
J. Clin. Oncol. 27 (18), 2954–2961. https://doi.org/10.1200/JCO.2008.17.7618. Waight, A.B., Bargsten, K., Doronina, S., Steinmetz, M.O., Sussman, D., Prota, A.E., 2016.
Structural basis of microtubule destabilization by potent auristatin anti-mitotics. PLoS One 11 (8), e0160890. https://doi.org/10.1371/journal.pone.0160890.
Wang, Y., Benz, F.W., Wu, Y., Wang, Q., Chen, Y., Chen, X., Yang, J., 2016. Structural insights into the pharmacophore of vinca domain inhibitors of microtubules. Mol. Pharmacol. 89 (2), 233–242. https://doi.org/10.1124/mol.115.100149.
Wickham, H., 2009. ggplot2:Elegant Graphics for Data Analysis. Springer-Verlag, New Your, NY.
Wilson, L., Creswell, K.M., Chin, D., 1975. The mechanism of action of vinblastine.
Binding of [acetyl-3H]vinblastine to embryonic chick brain tubulin and tubulin from sea urchin sperm tail outer doublet microtubules. Biochemistry 14 (26), 5586–5592. Retrieved from. https://www.ncbi.nlm.nih.gov/pubmed/1203244.
Wilson, L., Jordan, M.A., Morse, A., Margolis, R.L., 1982. Interaction of vinblastine with steady-state microtubules in vitro. J. Mol. Biol. 159 (1), 125–149. Retrieved from. https://www.ncbi.nlm.nih.gov/pubmed/7131559.
Wilson, L., Lopus, M., Miller, H.P., Azarenko, O., Riffle, S., Smith, J.A., Jordan, M.A., 2015. Effects of eribulin on microtubule binding and dynamic instability are strengthened in the absence of the betaIII tubulin isotype. Biochemistry 54 (42), 6482–6489. https://doi.org/10.1021/acs.biochem.5b00745.
Windebank, A.J., Grisold, W., 2008. Chemotherapy-induced neuropathy. J. Peripher. Nerv. Syst. 13 (1), 27–46. https://doi.org/10.1111/j.1529-8027.2008.00156.x.
Yenjerla, M., Lopus, M., Wilson, L., 2010. Analysis of dynamic instability of steady-state microtubules in vitro by video-enhanced differential interference contrast microscopy with an appendix by Emin Oroudjev. Methods Cell Biol. 95, 189–206. https://doi.org/10.1016/S0091-679X(10)95011-5.
Younes, A., Gopal, A.K., Smith, S.E., Ansell, S.M., Rosenblatt, J.D., Savage, K.J., Chen, R., 2012. Results of a pivotal phase II study of brentuximab vedotin for patients with relapsed or refractory Hodgkin’s lymphoma. J. Clin. Oncol. 30 (18), 2183–2189. https://doi.org/10.1200/JCO.2011.38.0410.