View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Toxicology and Applied Pharmacology 273 (2013) 298–313

Contents lists available at ScienceDirect

Toxicology and Applied Pharmacology

journal homepage: www.elsevier.com/locate/ytaap

Preclinical safety profile of trastuzumab emtansine (T-DM1): Mechanism of action of its cytotoxic component retained with improved tolerability

Kirsten Achilles Poon a,⁎, Kelly Flagella a,JosephBeyera, Jay Tibbitts b, Surinder Kaur a,OlaSaada,Joo-HeeYia, Sandhya Girish a,NoelDybdala,1, Theresa Reynolds a,1

a Genentech, Inc., South San Francisco, CA, USA b UCB, Brussels, Belgium

article info abstract

Article history: Trastuzumab emtansine (T-DM1) is the first antibody-drug conjugate (ADC) approved for patients with human Received 13 June 2013 epidermal growth factor receptor 2 (HER2)-positive metastatic breast cancer. The therapeutic premise of ADCs is Revised 30 August 2013 based on the hypothesis that targeted delivery of potent cytotoxic drugs to tumors will provide better tolerability Accepted 3 September 2013 and efficacy compared with non-targeted delivery, where poor tolerability can limit efficacious doses. Here, we Available online 12 September 2013 present results from preclinical studies characterizing the toxicity profile of T-DM1, including limited assessment of unconjugated DM1. T-DM1 binds primate ErbB2 and human HER2 but not the rodent homolog c-neu. There- Keywords: Breast cancer fore, antigen-dependent and non-antigen-dependent toxicity was evaluated in monkeys and rats, respectively, in HER2 both single- and repeat-dose studies; toxicity of DM1 was assessed in rats only. T-DM1 was well tolerated at Ado-trastuzumab emtansine doses up to 40 mg/kg (~4400 μgDM1/m2) and 30 mg/kg (~6000 μgDM1/m2) in rats and monkeys, respective- T-DM1 ly. In contrast, DM1 was only tolerated up to 0.2 mg/kg (1600 μgDM1/m2). This suggests that at least two-fold Toxicology higher doses of the cytotoxic agent are tolerated in T-DM1, supporting the premise of ADCs to improve the ther- Antibody-drug conjugates apeutic index. In addition, T-DM1 and DM1 safety profiles were similar and consistent with the mechanism of action of DM1 (i.e., microtubule disruption). Findings included hepatic, bone marrow/hematologic (primarily platelet), lymphoid organ, and neuronal toxicities, and increased numbers of cells of epithelial and phagocytic or- igin in metaphase arrest. These adverse effects did not worsen with chronic dosing in monkeys and are consistent with those reported in T-DM1-treated patients to date. © 2013 The Authors. Published by Elsevier Inc. Open access under CC BY-NC-ND license.

Introduction domain of the receptor, has made HER2 a suitable target for antibody therapy. The human epidermal growth factor receptor 2 (HER2, also known Trastuzumab (Herceptin®, Genentech, Inc., South San Francisco, CA) as ErbB2) is a transmembrane receptor tyrosine kinase that is part of a is a humanized monoclonal antibody directed against subdomain IV of complex signal transduction network involved in cell differentiation, the extracellular region of HER2 and is indicated for the treatment of proliferation, and survival (Yarden and Sliwkowski, 2001). HER2 is HER2-overexpressing breast cancer and HER2-overexpressing meta- expressed in normal epithelial tissues at relatively low levels in healthy static gastric or gastroesophageal junction adenocarcinoma (Herceptin adults (Press et al., 1990), but it is overexpressed in approximately 20% package insert, 2010; Herceptin Summary of Product Characteristics, of tumors from patients with breast cancer (Dawood et al., 2010; Ross 2010). The mechanisms of action are thought to include one or more et al., 2009). HER2-positive breast tumors are associated with aggres- of the following: interference with signal transduction pathways, sive growth and poor clinical outcomes (Slamon et al., 1987, 1989). impairment of extracellular domain (ECD) cleavage, inhibition of DNA The association between HER2 overexpression and tumor pathogenesis repair, decreased angiogenesis, induction of cell cycle arrest, and in breast cancer, together with the accessibility of the extracellular activation of antibody-dependent cellular cytotoxicity (Hudis, 2007; Sliwkowski et al., 1999; Spector and Blackwell, 2009). Although trastuzumab provides substantial benefits for many patients with HER2-positive breast cancer, a proportion have tumors that either do not respond to trastuzumab or relapse following initial response to treatment (Nahta et al., 2006; Slamon et al., 2001). Following relapse, HER2 overexpression is still present (Spector et al., 2005), suggesting ⁎ Corresponding author at: Development Sciences, 1 DNA Way, South San Francisco, CA that these tumors could still be responsive to treatment with HER2- 94080, USA. Fax: +1 650 225 2797. E-mail address: [email protected] (K.A. Poon). targeted agents. Indeed, two phase III clinical trials have shown in pa- 1 These authors contributed equally to this work. tients with tumors that progressed on trastuzumab-containing therapy

0041-008X © 2013 The Authors. Published by Elsevier Inc. Open access under CC BY-NC-ND license. http://dx.doi.org/10.1016/j.taap.2013.09.003 K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313 299 that the addition of trastuzumab to subsequent therapy resulted in that resulted in equivalent DM1 doses based on body surface area significantly improved clinical outcomes (Blackwell et al., 2010; von (μg DM1/m2). Minckwitz et al., 2009, 2011). An alternative to targeted antibody therapy and/or systemic chemo- Methods therapy is the use of antibody-drug conjugates (ADCs). ADCs are created by chemically linking a cytotoxic agent to a monoclonal antibody that General animal information. All procedures in animals described targets a tumor-enriched or tumor-specific protein. Tumor-specific below were performed in compliance with the Animal Welfare Act, delivery of potent cytotoxic agents has the potential to ameliorate the the Guide for the Care and Use of Laboratory Animals, and the Office systemic toxicity associated with many chemotherapies (Alley et al., of Laboratory Animal Welfare. Protocols were reviewed by the Institu- 2010; Lambert, 2005; Wu and Senter, 2005). tional Animal Care and Use Committees of the relevant facility; either Trastuzumab emtansine (T-DM1, Kadcyla™, Genentech, Inc., South Genentech, Inc. (South San Francisco, CA), or Covance, Inc. (Madison, San Francisco, CA) is an ADC composed of trastuzumab, a nonreducible WI). Sprague–Dawley rats were obtained from Charles River Laborato- thioether linker (4-[N-maleimidomethyl]-cyclohexane-1-carbonyl ries (Portage, MI, or Hollister, CA). Cynomolgus monkeys (Macaca [MCC]) (Lewis Phillips et al., 2008), and the cytotoxic agent DM1 fascicularis) were obtained from Three Springs Scientific, Inc. (Perkasie, (N2′-deacetyl-N2′-(3-mercapto-1-oxopropyl) maytansine) (Blättler PA), or Covance Research Products, Inc. (Alice, TX). All animals were and Chari, 2001; Cassady et al., 2004; Goldmacher et al., 2002) individually housed in stainless-steel cages and were provided with (Fig. 1). DM1 is derived from the highly potent antitumor agent food supplements (that did not require analysis) and various cage- maytansine and inhibits microtubule polymerization (Cabanillas enrichment devices. Monkeys were commingled to provide psycholog- et al., 1978; Chabner et al., 1978; Eagan et al., 1978; Issell and ical enrichment. Animals were assigned to dose groups using a stratified Crooke, 1978; Remillard et al., 1975). T-DM1 comprises trastuzumab randomization scheme based on individual body weight. For scheduled with zero to eight DM1 molecules linked via MCC, primarily to lysine or unscheduled necropsies, rats were euthanized with isoflurane residues, and has an average drug-to-antibody ratio (DAR) of ap- followed by an overdose of a ketamine/xylazine cocktail or sodium proximately 3.5 (Krop et al., 2010). After T-DM1 binds to HER2 on pentobarbital; monkeys were terminally sedated with sodium pento- the cell surface, the T-DM1/HER2 complex is internalized via endo- barbital and euthanized by exsanguination. cytosis and degraded in lysosomes, ultimately leading to the intra- cellular release of lysine-MCC-DM1 (lys-MCC-DM1) (Chari, 2008; Intravenous dose formulations. T-DM1 was formulated in a vehicle Erickson et al., 2006, 2010). In the phase III EMILIA study, patients composed of 10 mM succinate, 100 mg/mL trehalose, and 0.1% polysor- with HER2-positive, metastatic breast cancer previously treated with bate 20, pH 5.0. DM1 was formulated in the T-DM1 vehicle plus 0.5% trastuzumab and a taxane received treatment with either T-DM1 or dimethyl adipimidate and 1 mM ethylenediaminetetraacetic acid. The lapatinib plus capecitabine. Patients who received T-DM1 had significant- average DAR was 3.4 for all studies, with the exception of the repeat- ly improved clinical outcomes compared with patients treated with dose (every 3 weeks [q3w] × eight doses) monkey study where the lapatinib plus capecitabine (Verma et al., 2012). Median progression- DAR was 3.8. free survival was 9.6 months versus 6.4 months in the T-DM1 and con- trol arms, respectively (hazard ratio = 0.65; 95% confidence interval Parameters for the evaluation of toxicity. Toxicity indices consisted of [CI]: 0.55 − 0.77; p b 0.001). Median overall survival also improved daily clinical observations, body weight, food consumption, clinical and was 30.9 months versus 25.1 months in the respective treatment pathology, organ weights, and histopathology. Additionally, physical arms (hazard ratio = 0.68; 95% CI: 0.55 − 0.85: p b 0.001) (Verma examinations (including neurologic assessment) and ophthalmic exam- et al., 2012). inations were conducted in the repeat-dose monkey studies. Necropsies The preclinical and clinical toxicities of trastuzumab and maytansine included examination of the carcass; external body orifices; abdominal, (the parent compound of DM1) have been extensively characterized. In thoracic, and cranial cavities; and organs. Tissues collected at necropsy single- and repeat-dose toxicity studies of trastuzumab in cynomolgus were preserved in 10% neutral-buffered formalin or modified Davidson's monkeys, no adverse effects were observed, although trastuzumab- fixative and were processed for routine histologic examination. Select or- related cardiotoxicity and embryotoxicity had been previously identi- gans were weighed prior to fixation. fied in the clinical setting at low frequencies (Herceptin Summary of Product Characteristics, 2010; Herceptin® package insert, 2010; Rat studies of T-DM1 and DM1. The toxicity of T-DM1 was evaluated Klein and Dybdal, 2003). DM1 and maytansine are closely related in single- and repeat-dose studies in rats (Table 1A). In the first single- chemical entities; DM1 differs from maytansine only in the replace- dose study, rats were dosed intravenously (IV) on day 1 with vehicle ment of a methyl group with a thioethyl group to enable conjugation. alone or 6, 20, or 60 mg/kg T-DM1 (n = 10 animals/sex/group). Six Maytansine-associated preclinical and clinical toxicities included ad- additional animals (n = 3 animals/sex/group) were assigned to each verse effects on the liver, bone marrow, lymphoid organs, gastroin- T-DM1-treated group for toxicokinetic (TK) analyses (described in the testinal tract, and central/peripheral nervous system (Issell and Toxicokinetic analyses section). Body weights and food consumption Crooke, 1978). The established clinical dose of 3.6 mg/kg T-DM1 de- were recorded on days 1, 3, 5, 8, 15, and 22. A necropsy was conducted livers molar equivalent doses of DM1 that are only slightly higher 2 days postdose (day 3) (n = 5 animals/sex/group), and a second (~2.3 mg/m2) than the maytansine MTD but with a significant improve- necropsy was conducted following a 3-week recovery period (day 22). ment in efficacy and safety. Additionally, an unscheduled necropsy was performed for animals de- We sought to characterize the safety of T-DM1 and to understand termined to be moribund on day 5 or 6. In the second single-dose the drivers of toxicity to support its clinical development. The study, rats were injected IV on day 1 with vehicle alone or 46 mg/kg trastuzumab component of T-DM1 binds to nonhuman primate ErbB2 T-DM1 (n = 5 females/group). Body weights were recorded daily, and human HER2 but does not cross-react with the corresponding and a necropsy was conducted 11 days postdose (day 12). In the rodent receptor (c-neu) (data on file at Genentech, Inc., South San repeat-dose study, rats (n = 6 females/group) were administered Francisco CA), limiting the assessment of antigen-dependent effects to T-DM1 weekly for a total of three IV doses that ranged from 4 to 5, non-human primates. As rodents have been shown to be sensitive to 9 to 13, and 18 to 25 mg/kg T-DM1. Body weights were recorded the DM1 parent compound maytansine (Issell and Crooke, 1978), rats daily, and a necropsy was conducted 3 weeks after the third dose were used to evaluate the antigen-independent effects of T-DM1 and (day 36). An unscheduled necropsy was also performed (including compare them with toxicities associated with DM1. Cross-species blood collection for clinical pathology and TK) for animals in the comparisons were possible because of the selection of T-DM1 doses high-dose group determined to be moribund on day 24. The toxicity 300 K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313

DM1 MCC linker Trastuzumab

O O O N N S O Me O Cl MeO O MeO N N O H O

N O H MeO OH

n Where n~3.5 DM1/antibody

Fig. 1. Trastuzumab emtansine is an antibody-drug conjugate composed of trastuzumab (a humanized anti-HER2 IgG1), MCC (a nonreducible thioether linker), and DM1 (a microtubule- inhibitory maytansinoid). The bracketed structure is DM1 plus MCC, which is linked to antibody lysine residues on trastuzumab. The n is, on average, 3.5 per trastuzumab molecule. HER2: human epidermal growth factor 2; IgG1, immunoglobulin G1; MCC: 4-[N-maleimidomethyl]-cyclohexane-1-carbonyl.

of DM1 was evaluated in two single-dose rat studies: a dose range- Monkey studies of T-DM1. The toxicity of T-DM1 was evaluated in finding study and a definitive study. The dose range-finding study single- and repeat-dose studies (q3w × four doses and q3w × eight was conducted at IV doses of 0 mg/kg (vehicle, n = 2), 0.1, 0.2, 0.4, doses) in cynomolgus monkeys (Table 1B). Owing to the anticipated 0.6, 0.8, or 1.0 mg/kg (n = 6 animals/group), followed by a defini- clinical dosing schedule (q3w), we wanted to understand the toxicities tive study at doses of 0 mg/kg (vehicle), 0.07, 0.1, or 0.2 mg/kg following a single exposure before setting doses for a multidose study. (n = 10 animals/sex/group). In the dose range-finding study, body In the single-dose study, monkeys were administered IV doses of weights were measured on day 1 prior to DM1 administration and T-DM1 on day 1 at 0 (vehicle), 3, 10, and 30 mg/kg (n = 6 animals/ daily thereafter. An unscheduled blood collection and necropsy sex/group). Body weights were recorded prior to T-DM1 administration were also performed for animals determined to be moribund on and weekly thereafter. Food consumption was assessed qualitatively on day 3 or 4. The study design for the definitive study of DM1 in rats a daily basis. Necropsies were conducted 2 days postdose (study day 3) was identical to the single-dose study of T-DM1 in rats described and following a 3-week recovery period (study day 22) on the basis previously, with blood collections and necropsy on days 3 and 22. of the intended clinical dosing regimen of q3w administration. TK

Table 1A Summary of study designs and sampling schedules for rat studies of T-DM1 and DM1.

N/group Regimen/recovery Day 1 mean body Dose T-DM1 or DM1 BSA-based equivalent Hematology/serum TK sampling schedule weight range (mg/kg)b dose of DM1 chemistry sampling (days) (μgDM1/m2)c schedule (study day)

T-DM1 10 (F)a Single dose/3 weeks 199–202 g (F) 6 700 (F), 800 (M) 3, 22 1 (10, 24 h postdose), 4, 8, 15, 22 10 (M)a 305–309 g (M) 20 2300 (F), 2600 (M) 60d 6800 (F) d,7800(M)d 5 (F) Single dose/11 days 122 g (F) 46 4400 5, 12 Day 12 6 (F) Weekly × 3 doses/3 weeks 110–117 g (F) 5 500 5, 12, 19, and 24 or 36 Day 24 or 36 11 1000 22 2100 DM1 6 (F) Single dose/11 days 119–121 g (F) 0.1 600 3, 5, 12 No data available 0.2 1100 0.4d 2200d 0.6d 3400d 0.8d 4500d 1.0d 5600d 10 (F)a Single dose/3 weeks 212–214 g (F) 0.07 500 (F), 500 (M) 3, 8, 22 1 (5 min, 1, 6, 24 h postdose), 3, 22 10 (M)a 317–322 g (M) 0.1 700 (F), 800 (M) 0.2 1400 (F), 1600 (M)

a Toxicokinetics sampling from satellite animals (n = 3/sex/group). b mg/kg T-DM1 = mg antibody (trastuzumab) per kg body weight; mg/kg DM1 = mg DM1 per kg body weight. c BSA (body surface area)-based equivalent dose of DM1 (μg/m2) calculated as follows: for T-DM1, [dose level (mg/kg T-DM1) × day 1 group mean body weight (kg) × 0.017 (fraction of DM1 to trastuzumab in T-DM1) × 1000/BSA) (m2)]. For DM1, [dose level (mg/kg DM1) × day 1 group mean body weight (kg) × 1000 / BSA) (m2)]. BSA calculation [(day 1 group mean body weight, [g])0.66 × 9.1 (Meeh's constant) / 10,000]. Fraction of DM1 to trastuzumab in T-DM1 determined using the molecular weights (MW) of DM1 (MW 738) and trastuzumab (MW 145, 167), and the measured drug/antibody ratio (DAR 3.4), and calculated as follows: MW DM1 × DAR) / MW trastuzumab. d Animals found dead or euthanized in a moribund condition. K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313 301

Table 1B Summary of study designs and sampling schedules for monkey studies of T-DM1.

N/group Regimen/recovery Day 1 mean body Dose T-DM1 BSA-based equivalent Hematology/serum chemistry TK/ATA sampling schedule weight range or DM1 dose of DM1 sampling schedule (days) (mg/kg)a (μgDM1/m2)b (study day)

T-DM1 6 (F) Single dose/3 weeks 2.8–2.9 kg (F) 3 600 3, 8, 15, 22 1 (5 min postdose), 3, 4, 8, 15, 22c 6(M) 3.4–3.6 kg (M) 10 2000 30 6000 7(F) Repeat dose 2.8 kg (F) 3 600 3, 22c,24,43c,45,64c,66, 1 (5 min, 6, 10, 24 h postdose), 7(M) (q3w × 4 doses)/3 3.1–3.2 kg (M) 10 2000 85, 94, 106 3, 6, 22c (predose), 43c (predose), or 6 weeks 30 6000 64c (predose), 65, 67, 69, 71, 78, 85c, 94, 106c 6(F) Repeat dose 2.5–2.6 kg (F) 1 200 4, 8, 22c,43c,64c,85c, 1 (5 min, 4, 8, 24 h postdose), 6(M) (q3w × 8 doses)/3 3.0–3.1 kg (M) 3 600 106c,127c,148c, 151, 4, 7, 15, 43d (predose, 5 min postdose), or 6 weeks 10 2000 155, 169, & 190 64d (predose, 5 min, 4, 8, 24 h postdose), 67, 70, 78, 85d (predose, 5 min postdose), 127d (predose, 5 min postdose), 148d (predose, 5 min, 4, 8, 24 h postdose), 151, 154, 162, 169d, 176, 183, 190d

a mg/kg T-DM1 = mg antibody (trastuzumab) per kg body weight. b BSA (body surface area)-based equivalent dose of DM1 (μg/m2) calculated as follows: [dose level (mg/kg T-DM1) × 12 × 0.017 or 0.019 (fraction of DM1 to trastuzumab in T-DM1 for single-dose and 4-dose study, or 8-dose study, respectively) × 1000)]. Fraction of DM1 to trastuzumab in T-DM1 determined using the molecular weights(MW)ofDM1(MW738)and trastuzumab (MW 145, 167), and the measured drug/antibody ratio (DAR 3.4 for single-dose and 4-dose study, or 3.8 for the 8-dose study), and calculatedasfollows:MW DM1 × DAR) / MW trastuzumab. The conversion factor of 12 for cynomolgus monkeys was taken from the Handbook of Toxicology, 2nd Edition, CRC Press, 1995. Equivalents of DM1 are based on a 3 kg animal, and will vary for the body weight of each animal. c Predose. d Blood samples collected predose for ATA (anti-therapeutic antibody) analysis. and anti-T-DM1 antitherapeutic antibody (ATA) analyses were also Pharmacology study of cardiovascular safety of single-dose T-DM1 in performed (assays described in the Toxicokinetic analyses section and cynomolgus monkeys. Female cynomolgus monkeys were admin- Assay to detect antitherapeutic antibody response to T-DM1 section). istered a single IV bolus dose of vehicle alone or 3, 10, or 30 mg/kg In the first repeat-dose study (q3w × four doses), monkeys were T-DM1 (n = 4 animals/group). At least 3 weeks prior to initiation administered IV doses of T-DM1 on days 1, 22, 43, and 64 at 0 (vehicle), of treatment, animals were surgically implanted with a telemetry trans- 3, 10, and 30 mg/kg (n = 7 animals/sex/group). Body weight and food mitter (Data Sciences, International [DSI Dataquest™ Open Art™]and consumption were recorded as described previously for the single-dose Gould/PoNeMah® [P3P (PoNeMah Physiology Platform)] acquisition study. Ophthalmologic examinations were performed prior to treat- system). A large animal ECG and a pressure transmitter (Model #D70 ment and during weeks 3 and 9 using a handheld slit lamp and an indi- PCTP) were implanted into the abdomen. The ECG leads of the transmit- rect ophthalmoscope. Neurologic and physical examinations were ter were arranged in an approximate lead II configuration. One pressure performed prior to treatment and during weeks 3 and 9. Electrocardio- line, which was used to assess blood pressure (BP), was placed in the gram (ECG) rate and heart rate (HR) were recorded prior to treatment abdominal aorta. The second pressure line, which was used to assess in- and during weeks 2, 8, 12, and 15. For ECG examination, routine trathoracic pressure, was placed on the thoracic side of the diaphragm. measurements of HR, and PR, QRS, QT, corrected QT interval (QTc; ECG and BP measurements were recorded as follows: twice in the 2- calculated), and RR intervals were conducted. Qualitative assessments week period prior to T-DM1 administration (at least 7 days apart); of ECG traces for rhythm and abnormalities were also performed. Clini- prior to dosing on day 1 for at least 60 min; continuously through 8 h cal pathology TK (described in the Toxicokinetic analyses section [Dere postdose; for a 10-minute period each hour through 24 h postdose; et al., 2013]) and ATA samples (described in the Assay to detect anti- and on days 3, 4, 5, 8, 15, and 22 for at least 10 min at approximately therapeutic antibody response to T-DM1 section [Carrasco-Triguero the same time of day as dosing. Quantitative evaluation of ECG mea- et al., 2013]) were collected as summarized in Table 1B. Necropsies surements included RR interval (for use in rate correction), QT were conducted 2 days after the fourth dose on day 66 (n = 3 interval, and QTc, using Fridericia's method. BP measurements included animals/sex/group), 3 weeks after the fourth dose on day 85 (n = 2 HR, systolic (SBP), diastolic (DBP), and mean arterial pressure and pulse animals/sex/group), and 6 weeks after the fourth dose on day 106 pressure (systolic–diastolic). Intrathoracic pressure measurements in- (n = 2 animals/sex/group). In the second repeat-dose study (q3w × cluded respiratory rate and a qualitative assessment of respiratory eight doses), monkeys were administered IV doses of T-DM1 on days depth. 1, 22, 43, 64, 85, 106, 127, and 148 at 0 (vehicle), 1, 3, and 10 mg/kg (n = 6 animals/sex/group) to evaluate the chronic effects of T-DM1. In vitro human ether-à-go-go-related gene assay (DM1). The Body weight and food consumption were recorded as described previ- concentration–response relationship of the effect of DM1 on the potas- ously for the single-dose study. Ophthalmologic, physical, and neuro- sium channel current of the human ether-à-go-go-related gene (hERG) logic examinations; body temperature; respiration rates; and pulse was evaluated at ChanTest® (ChanTest Corporation, Cleveland, OH) oximetry measurements were performed prior to T-DM1 administra- (ChanTest® hERG IC50 Concentration Response Assays, 2013)usingstan- tion, once after the first dose, after the eighth dose, and again during dard hERG IC50 (half-maximal inhibition determined by electro- the last week of the recovery period. ECG and HR were recorded twice physiology) methodologies. DM1 concentrations of 2.6, 8.8, and prior to T-DM1 administration: once after the first dose, fourth dose, 29.5 μM were evaluated. The positive control was 60 nM of terfenadine. eighth dose, and during the last week of the recovery period. ECG mea- surements and qualitative assessments were performed as described Tissue cross-reactivity. The immunohistochemical cross-reactivity and for the repeat-dose (q3w × four doses) toxicity study in monkeys. Clin- distribution of reactivity of T-DM1 were assessed on cryosections of a ical pathology TK and ATA samples were collected as summarized in complete panel of human and cynomolgus monkey tissues. T-DM1 bind- Table 1B. Necropsies were conducted 7 days after the final dose (n = ing was evaluated at concentrations of 1.0 μg/mL and 10.0 μg/mL and de- 3 animals/sex/group) on day 155, and after a 6-week recovery period tected immunohistochemically with an anti-DM1/biotinylated antimouse (n = 3 animals/sex/group) on day 190. immunoglobulin G (IgG) secondary/tertiary antibody complex. 302 K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313

Toxicokinetic analyses. Toxicokinetic (TK) analyses were conducted in screening cut point was set to give an untreated positive rate of approx- the single-dose study of T-DM1 in rats and cynomolgus monkeys, in the imately 8%. The relative sensitivity of the assay using anti-T-DM1 definitive single-dose study of DM1 in rats, and in the repeat-dose and affinity-purified antibodies from cynomolgus monkeys was 65 ng/mL. chronic toxicity studies of T-DM1 in cynomolgus monkeys. Parameters, The tolerance of the assay to interference by T-DM1 was demonstrated including but not limited to area under the concentration–time curve by the detection of 382 ng/mL antibodies in the presence of 50 μg/mL of

(AUC), maximum concentration, terminal half-life (t½), clearance (CL), T-DM1. and volume of distribution, were estimated, as applicable. In rats, inten- sive blood sampling was performed using satellite animals that were ad- Statistical analyses. For the single- and repeat-dose DM1 and T-DM1 ministered the same doses as the animals dosed for toxicity evaluations. toxicity studies, continuous clinical pathology values, organ weights, In monkeys, intensive sampling occurred during the first and last dose food consumption values, body weights, and ECG/BP/HR data were cycles (cycles 1, 4, or 8) with a collection of peak and trough samples analyzed by factorial analysis of variance/covariance with repeated for the interim dose cycles. Validated assays (developed by Genentech, measures (when appropriate). The factors used were main effect and Inc.) were used to measure specific analytes in the samples and included sex; when treatment interactions were significant, the sexes were an enzyme-linked immunosorbent assay (ELISA) for T-DM1 conjugate, analyzed separately. Group comparisons were evaluated at the 5.0% an ELISA for total trastuzumab, and a liquid chromatography with tan- two-tailed probability level. TK parameter estimates were summarized dem mass spectrometry (LC–MS/MS) assay for DM1 (see the following as mean ± standard deviation (SD). For the cardiovascular safety phar- sections). macology study, descriptive statistics included means and SD at each time interval. For statistical analysis, the predose means were combined T-DM1 ELISA. The T-DM1 ELISA was designed to measure all T-DM1 into an overall predose value. Repeated measure analyses of variance DARs except DAR 0 in serum. Anti-DM1 monoclonal antibody was were performed for ECG, HR, SBP, DBP, mean BP, pulse pressure, and re- used as the capture reagent, and biotinylated recombinant HER2 ECD spiratory rate, with subsequent Bonferroni-adjusted t-test for pairwise and horseradish peroxidase-conjugated streptavidin was used for comparisons between treatment and control. Statistical evaluation detection. The minimum quantifiable concentrations in rat and cyno- was at the 5.0% two-tailed probability level. molgus monkey serum were 30 ng/mL and 40 ng/mL, respectively. Results Total trastuzumab ELISA. The total trastuzumab ELISA was designed to measure all T-DM1 DARs, including conjugated T-DM1, as well as Effect of T-DM1 and DM1 in rats partially unconjugated and fully unconjugated T-DM1 in serum. Recom- binant HER2 ECD was used as the capture reagent and peroxidase- A single IV dose of T-DM1 was well tolerated in rats at doses up conjugated F(ab′)2 goat antihuman IgG Fc was used for detection. The to 46 mg/kg (~4400 μgDM1/m2), but at 60 mg/kg T-DM1 (~6800– minimum quantifiable concentration in both rat and cynomolgus mon- 7800 μgDM1/m2 — DM1 dose determined based on female and male key serum was 40 ng/mL. day 1 group mean body weights) decreased body weight (Fig. 2A), mor- For both ELISAs, serum samples were quantified against an assay bidity (decreased food consumption, abnormal clinical signs; data not standard curve prepared from T-DM1 with an average DAR of approxi- shown), and/or mortality occurred 4 or 5 days postdose (Fig. 2A; mately 3.5 (Dere et al., 2013). Table 2). In contrast, a single IV dose of DM1 was only tolerated up to 0.2 mg/kg (~1400–1600 μgDM1/m2); doses ≥0.4 mg/kg (~2200 μg Unconjugated DM1. The liquid chromatography with LC–MS/MS assays DM1/m2) were associated with mortality, 2 or 3 days postdose, and were designed to measure DM1 and any disulfide-bound forms of re- body weight decreases (Fig. 2A; Table 2), suggesting that the acute leased DM1 (e.g., dimers, glutathione, cysteine, and albumin adducts) tolerability of DM1 is improved at least two-fold when conjugated in rat and cynomolgus monkey lithium heparin plasma; the assay did to trastuzumab and administered as T-DM1. The repeat dose study of not measure DM1 conjugated to trastuzumab via MCC-DM1. Because T-DM1 in rats showed that weekly doses (weekly for three doses DM1 contains a free sulfhydryl, any DM1 released from T-DM1 is total) ranging from approximately 5 mg/kg to 11 mg/kg (~500 to expected to rapidly dimerize or react with other thiol-containing mole- 1000 μgDM1/m2, respectively) were well tolerated (data not shown), cules in plasma. Therefore, to avoid underquantification of released while animals showed signs of morbidity/mortality approximately DM1, plasma samples were treated with a reducing agent (tris [2- 1 week after the third dose at 22 mg/kg (~2100 μgDM1/m2, respec- carboxyethyl]phosphine) to release disulfide-bound DM1. The free tively). The cumulative total dose and the timing of adverse effects in thiol was then blocked with N-ethylmaleimide (NEM) to prevent any this repeat-dose study were in line with the doses tolerated after a further reactions. The LC–MS/MS assays quantified DM1-NEM. DM1 single dose of T-DM1. was extracted from plasma samples by protein precipitation using The primary toxicities in rats were comparable between T-DM1 acetonitrile and online solid-phase extraction for all studies noted and DM1 and included dose-dependent effects on liver, bone marrow/ except the chronic toxicity study. For the chronic monkey toxicity hematologic systems (mainly platelets), and lymphoid organs. Hemato- study, DM1 was extracted from plasma by protein precipitation using logic and serum chemistry findings were primarily associated with bone acetonitrile. DM1, derivatized with NEM, was analyzed using electrospray marrow suppression and hepatotoxicity (Table 3). Hematologic changes LC–MS/MS. The LC–MS/MS assays in rat and cynomolgus monkey plasma included decreased platelet (Fig. 2B), reticulocyte, and lymphocyte had a lower limit of quantification (LLOQ) of 1.00 nM (0.737 ng/mL). For counts, as well as increased absolute neutrophil counts, which correlat- the chronic toxicity study, LLOQ was 0.50 nM (0.37 ng/mL). ed with microscopic observations of mild bone marrow hypocellularity and minimal to moderate lymphoid depletion or necrosis in the lymph Assay to detect antitherapeutic antibody response to T-DM1. A validat- nodes, thymus, and spleen (Tables 4 and 5), accompanied by increased ed bridging antibody electrochemiluminescence assay (Bioveris®, San spleen weights in some animals (data not shown). Changes in serum Diego, CA) was used to detect ATAs to T-DM1 in cynomolgus monkey chemistry were reflective of adverse effects on the liver, including in- serum samples. The assay was designed to detect all ATA responses creases in serum alanine aminotransferase (ALT) (Fig. A.1), aspartate directed against T-DM1 by using biotinylated and ruthenium-labeled aminotransferase (AST) (Fig. 2C), alkaline phosphatase (ALP), gamma- T-DM1 to form bridging complexes with ATAs in the sample. Data glutamyl transpeptidase, and total bilirubin. Increased liver weights from a panel of serum samples from 72 T-DM1-naive cynomolgus (data not shown) and histopathologic findings were associated with monkeys were used to establish the screening threshold or cut point these elevations in liver enzymes and consisted of minimal to severe he- for the assay. To minimize the potential for false-negative results, the patocellular degeneration and necrosis (most severe at high doses) with K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313 303

A T-DM1 6 mg/kg (700 – 800 µg DM1/m 2)

2 110 T-DM1 20 mg/kg (2300 – 2600 µg DM1/m ) T-DM1 46 mg/kg (4400 µg DM1/m 2) T-DM1 60 mg/kg (6800 – 7800 µg DM1/m 2)† 100 DM1 0.07 mg/kg (500 µg DM1/m 2) DM1 0.1 mg/kg (700 – 800 µg DM1/m 2) T-DM1 0.2 mg/kg (1400 – 1600 µg DM1/m 2) 90 T-DM1 0.4 mg/kg (2200 µg DM1/m 2)† T-DM1 0.6 mg/kg (3400 µg DM1/m 2)†

† T-DM1 0.8 mg/kg (4500 µg DM1/m 2)† 80 † † † T-DM1 1.0 mg/kg (5600 µg DM1/m 2)† †

Body weight (% of control) 70 0 5 10 15 20 25 Time point (day) B 150

100

50 Platelets (% of control) 0

500 2200 3400 4500 5600 700–800 700–800 2300–26006800–7800 1400–1600

T-DM1 Free DM1 Dose (µg DM1/m 2)

C 2000

1500

1000 AST (% of control) 500

0

500 2200 3400 4500 5600 700–800 700–800 2300–26006800–7800 1400–1600

T-DM1 Free DM1 Dose (µg DM1/m 2)

Fig. 2. BW by dose over time (A), day 3 platelet levels (B), and day 3 AST levels (C) from the single-dose intravenous toxicity studies of T-DM1 and DM1 in rats (data for DM1 combined from two studies). Based on the differences in timing of peak body weight decreases and mortality with administration of T-DM1 versus DM1, it is possible that the peak effects on clinical pathology parameters (and histology) may have occurred slightly later (2- to 3-day delay) in T-DM1-treated rats. Importantly, the toxicity profile is very similar between T-DM1 and DM1 despite differences in tolerability. Group mean body weight data and individual platelet and AST data are expressed as a percent of control. A single T-DM1, DM1, or vehicle dose was given on day 1. T-DM1 and DM1 doses are presented as dose equivalents of DM1 per body surface area, μgDM1/m2. AST, aspartate aminotransferase; BW, body weight; T-DM1, trastuzumab emtansine. † indicates group euthanized in moribund condition/animals found dead ( DM1 0.4 mg/kg (2200 μgDM1/m2), DM1 0.6 mg/kg (3400 μgDM1/m2), DM1 0.8 mg/kg (4500 μgDM1/m2), DM1 1.0 mg/kg (5600 μgDM1/m2), and T-DM1 60 mg/kg (6800–7800 μgDM1/m2)groups). hypertrophy and vacuolation of Kupffer cells, prominent extramedullary Renal tubular epithelium was also affected in rats given either hematopoiesis, and increased mitotic figures (cells arrested in meta- T-DM1 or DM1 with minimal to slight degeneration, necrosis, phase) for hepatocytes, biliary epithelium, and Kupffer cells (see and increased mitotic arrest (see Tables 4 and 5). However, Tables 4 and 5). these findings were not accompanied by other indicators of renal injury 304 K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313

Table 2 lymphocytolysis, and cellular necrosis in multiple tissues (hepatocytes Summary of mortality in rats administered a single-dose of T-DM1 or DM1. and Kupffer cells in the liver, renal tubular epithelium and urothelium, BSA-based equivalent dose of DM1, μgDM1/m2 skeletal muscle and cardiac myofiber, thyroid follicular epithelium, (mg/kg T-DM1 or DM1) pancreatic endocrine and exocrine ductular epithelium, and small and T-DM1 DM1 large intestine mucosal epithelium). Although the majority of findings were similar for rats administered – 500 (0.07) – 600 (0.1) T-DM1 or DM1, there were subtle differences in toxicity (e.g., severity or 700 (6) 700 (0.1) nature/type of findings) that are likely related to differences in pharma- 800 (6) 800 (0.1) cokinetics, drug distribution, and/or cellular uptake between the mole- – 1100 (0.2) cules. Testicular degeneration, necrosis, and hemorrhage of corpora – 1400 (0.2) lutea, and necrosis of mammary glands were seen in rats given the – 1600 (0.2) 2 a – μ – 2200 (0.4) day 4 highest doses of T-DM1 (60 mg/kg; ~6800 7800 gDM1/m), but 2300 (20) – they were not observed after any dose of DM1, presumably since rats 2600 (20) – could not tolerate DM1 doses greater than 1400–1600 μgDM1/m2 – a 3400 (0.6) day 3 (see Fig. 2A). The administration of DM1 did not result in the 4400 (46) – – 4500 (0.8)a day 3 same widespread tissue distribution of cells in mitotic arrest (the – 5600 (1.0)a day 3 expected pharmacologic action of maytansinoids as microtubule in- 6800 (60)a day 5 or 6 – hibitors) that was observed for T-DM1 (Tables 4 and 5), and these a 7800 (60) day 5 or 6 – cells were only found in the liver and kidney. At tolerated doses of a Animals found dead or euthanized in a moribund condition. T-DM1 (≤20 mg/kg; 2300–2600 μgDM1/m2)orDM1(≤0.2 mg/kg; 1400–1600 μgDM1/m2), all findings were partially or completely re- (e.g., elevated blood urea nitrogen, creatinine, and electrolyte versed after a 3- or 4-week recovery period, which demonstrates that alterations). the toxic effects of T-DM1 are reversible after a sufficient interval of At DM1 doses ≥0.04 mg/kg (~2200 μgDM1/m2) where clinically time (data not shown). significant toxicity and morbidity occurred, the histologic findings T-DM1 and DM1 exposures were confirmed in the single-dose rat were present in a similar host of tissues (data not shown) but were studies of T-DM1 (Table A.1)andDM1(Table A.2). Because T-DM1 significantly more severe when compared with the non-tolerated dose has the molecular characteristics of both large-molecule biologic agent of T-DM1, 60 mg/kg (~6800–7800 μg DM1/m2). Findings included and small-molecule drug, we considered it important to characterize dose-dependent bone marrow hypoplasia, lymphoid depletion and the TKs by measuring three different analytes: (1) T-DM1 conjugate

Table 3 Notable hematologic and clinical chemistry parameters at day 3 from the single-dose T-DM1 and DM1 studies in Sprague–Dawley rats.

Parameter, mean (SD) Single-dose T-DM1

Male Female

Vehicle alone 6 mg/kg 20 mg/kg 60 mg/kg Vehicle alone 6 mg/kg 20 mg/kg 60 mg/kg (800 μg (2600 μg (7800 μg (700 μg (2300 μg (6800 μg DM1/m2) DM1/m2) DM1/m2) DM1/m2) DM1/m2) DM1/m2)

Hematology, thousands/μL Platelet countsd 1250 (124.4) 1141 (76.1) 711b (314.3) 759b (123.3) 1226 (222.7) 1273 (125.8) 944a (102.6) 769b (144.4) Reticulocyte, absoluted 385.6 (49.16) 310.1b (17.71) 309.8b (40.59) 137.4b (15.49) 298.0 (42.99) 285.8 (53.64) 275.7 (30.18) 129.6b (28.65) Neutrophil, absoluted 1.40 (0.450) 2.15 (1.421) 5.58b (1.903) 6.68b (1.159) 1.26 (0.280) 0.92 (0.349) 3.67a (1.130) 7.43b (2.179) Lymphocyted 11.19 (4.187) 9.43 (2.064) 8.35 (2.324) 4.61b (1.049) 9.97 (1.274) 7.51a (0.618) 8.40 (1.860) 4.38b (1.090) Clinical chemistry, U/L ALTd 37 (10.2) 39 (4.7) 115 (42.7) 424b (165.3) 30 (4.3) 28 (3.7) 62 (9.1) 299b (165.9) ASTd 189 (39.1) 193 (44.0) 353 (81.8) 947b (244.9) 217 (61.2) 199 (61.2) 259 (33.0) 789b (375.6) ALPd 237 (41.8) 226 (53.5) 411a (132.8) 566b (99.1) 161 (43.3) 149 (49.4) 221 (60.2) 342b (68.4) GGTc 0 (0.0) 0 (0.0) 1 (2.2) 1 (1.7) 0 (0.0) 0 (0.0) 0 (0.0) 2 (2.0)a Total bilirubin 0.1 (0) 0.1 (0) 0.1 (0) 0.3a (0.05) 0.1 (0) 0.2 (0.04) 0.2 (0.04) 0.9a (0.68)

Parameter, mean (SD) Single-dose DM1

Male Female

Vehicle alone 0.05 mg/kg 0.1 mg/kg 0.2 mg/kg Vehicle alone 0.05 mg/kg 0.1 mg/kg 0.2 mg/kg

Hematology, thousands/μL Platelet countsd 1175 (69.1) 926b (132.5) 981a (90.2) 884b (193.4) 1257 (163.4) 964a (144.8) 1083 (177.9) 792b (186.7) Reticulocyte, absoluted 335.2 (63.28) 45.4b (8.91) 42.8b (12.61) 31.7b (6.81) 262.7 (42.03) 51.1b (26.90) 48.9b (14.63) 33.7b (4.68) Neutrophil, absoluted 0.76 (0.188) 0.45 (0.096) 1.19 (0.470) 2.02b (0.484) 0.65 (0.122) 0.48 (0.243) 1.09 (0.766) 1.29 (0.598) Lymphocytec 6.30 (1.825) 4.64a (0.899) 4.32a (0.662) 3.84b (0.742) 5.33 (1.405) 6.57 (2.244) 3.91 (0.448) 4.02 (2.052) Clinical chemistry, U/L ALTd 33 (0.8) 35 (2.8) 65b (16.8) 91b (17.0) 31 (2.6) 37 (8.3) 55 (38.8) 72b (13.3) ASTd 139 (21.8) 154 (20.4) 246b (39.5) 331b (47.0) 125 (29.3) 148 (29.5) 170 (47.3) 316b (44.3) ALPd 259 (77.7) 204 (22.9) 231 (22.9) 172b (28.6) 168 (43.4) 150 (33.5) 192 (29.6) 146 (12.6) GGT 1 (0.4) 2 (0.8)a 1 (0.8) 1 (0.4) 2 (0.5) 2 (0.9) 2 (.07) 2 (0.0)

ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; SD, standard deviation; T-DM1, trastuzumab emtansine. a p b 0.05 compared with control. b p b 0.01 compared with control. c p b 0.05 combined treated compared with combined control. d p b 0.01 combined treated compared with combined control. K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313 305

Table 4 Notable microscopic findings following single-dose administration of 6, 20, and 60 mg/kg T-DM1 in Sprague–Dawley rats at necropsy on days 3, 5, or 6.a

Microscopic findings Organ/tissue Necropsy results

Day 3b Day 5 or 6d N (dose, mg/kgc) N (dose, mg/kgc)

Degeneration and/or necrosis Liver (hepatocytes, periportal/midzonal) 10 (20), 10 (60) 10 (60) Spleen (lymphocytes) 4 (60) 3 (60) Thymus (lymphocytes) 1 (20), 6 (60) 4 (60) Kidney (tubule) 5 (20), 6 (60) 10 (60) Duodenum (crypt epithelium) 8 (60) Testis (seminiferous tubules) 5 (60)e 5 (60)e Female mammary 3 (60)f 4 (60)g Male mammary 3 (20), 5 (60)h 3 (60)i Mesenteric lymph node (lymphocytes) – 2 (60) Ovary (corpus luteum) – 2 (60)f Increased mitotic figures/arrested metaphase Adrenal cortex (diffuse) 4 (60) 10 (60) Adrenal medulla (diffuse) 2 (60) 3 (60) Pituitary (pars distalis) 1 (20), 8 (60) 1 (60) Thyroid (follicular epithelium) 5 (60) 1 (60) Heart (subepicardial connective tissue) 3 (60) Tongue (epithelium) 3 (60) 1 (60) Liver (biliary epithelium) 8 (20), 10 (60) 10 (60) Liver (hepatocytes) 3 (6), 10 (20), 10 (60) 10 (60) Liver (Kupffer cells) 7 (6), 10 (20), 10 (60) 10 (60) Kidney (tubule) 2 (6), 8 (20), 10 (60) 10 (60) Eye (choroid) 2 (60) Eye (corneal epithelium) 8 (60) 9 (60) Skin (epidermis) 5 (60) 5 (60)j Injection site (epidermis/epidermis adnexa) 4 (60) 5 (60) Hypertrophy/vacuolation Liver (Kupffer cells) 4 (6), 10 (20), 10 (60) 10 (60) Spleen (reticuloendothelial cells; with increased mitoses) 4 (20), 10 (60) 10 (60) Hypocellular Femur marrow 3 (20), 10 (60) 10 (60) Sternum marrow 10 (60) 10 (60) Lymphoid depletion Spleen 8 (20), 9 (60) 9 (60) Thymus – 7 (60) Mesenteric lymph node – 7 (60) Other Luminal debris Epididymis (cellular) 5 (60)e 5 (60)e Hyperkeratosis Nonglandular stomach – 4 (60) Hemorrhage Testis – 3 (60)e

T-DM1, trastuzumab emtansine. a No gross or microscopic observations at recovery necropsy (day 22). b Data from males and females combined, unless otherwise noted; n = 10 for each dose evaluated. c 6, 20, and 60 mg/kg T-DM1 doses are equivalent to ~700–800, ~2300–2600, and ~6800–7800 μgDM1/m2, respectively. d Data from males and females combined, unless otherwise noted; n = 10 evaluated at T-DM1 60 mg/kg. e Data from males only; n = 5 for each dose evaluated. f Data from females only; n = 5 for each dose evaluated. g Data from females only; n = 4 evaluated at T-DM1 60 mg/kg. h Data from males only; n = 3 evaluated at T-DM1 20 mg/kg; n = 5 evaluated at T-DM1 60 mg/kg. i Data from males only; n = 3 evaluated at T-DM1 60 mg/kg. j Data from five males and four females, each evaluated at T-DM1 at 60 mg/kg.

and (2) total trastuzumab (all T-DM1 DARs, including conjugated Effect of T-DM1 in cynomolgus monkeys T-DM1, as well as partially unconjugated and fully unconjugated T-DM1), both measured in rat serum by ELISA, and (3) free DM1 mea- T-DM1 was well tolerated in cynomolgus monkeys with either sured in plasma by LC–MS/MS. An integrated analysis revealed that the single or repeat (q3w × four doses) IV doses up to 30 mg/kg TK profile of T-DM1 in rats is consistent with other therapeutic antibod- (~6000 μgDM1/m2) or after chronic administration (q3w × eight 2 ies, as characterized by a terminal t1/2 of approximately 3 to 5 days, a doses) up to the highest dose tested, 10 mg/kg (~2000 μgDM1/m). slow CL ranging from 13 to 15 mL/day/kg, and a central compart- The majority of adverse findings in cynomolgus monkeys given ment volume approximating the plasma volume. Total trastuzumab T-DM1 were concordant with T-DM1-related toxicity in rats, including exhibited approximately two-fold slower CL than T-DM1 at all doses hepatic, bone marrow/hematologic (primarily platelet), lymphoid (data not shown). This phenomenon is also observed in monkeys organ toxicities, and increased numbers of cells of epithelial and (Fig. 3)andpatients(Girish et al., 2012) and is hypothesized to be phagocytic origin in metaphase arrest. Hematologic and lymphoid due to deconjugation to lower DAR species, faster clearance of high organ toxicity was composed of decreased platelet counts, decreased DAR species, or a combination of these processes. Free DM1 concentra- red cell mass (i.e., erythrocytes, hemoglobin, and hematocrit) with tions (resulting from degradation and release from T-DM1) were at increased reticulocytes (a rebound response), and lymphoid depletion their peak (8–154 ng/mL) immediately after administration of T-DM1 in the spleen and thymus (Tables 6, 7, Fig. 4A). Findings pertinent to and decreased in a manner similar to that for T-DM1. In contrast the liver included transient elevations of AST, ALT (see Table 6; to the pharmacokinetics of DM1 when conjugated to trastuzumab, Figs. 4B and C), and ALP, and, microscopically, centrilobular vacuolation, unconjugated DM1 exhibited a large apparent volume of distribution hypertrophy and/or hyperplasia of Kupffer cells, increased sinusoidal (N5000 mL/kg) and rapid CL, which is consistent with the pharmacoki- leukocytes, multinucleated hepatocytes, and increased numbers of netic profile expected for a small-molecule compound. Kupffer cells arrested in mitosis (see Table 7). Similar to rats, monkeys 306 K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313

Table 5 tissue binding was detected primarily in epithelial cells of multiple Notable microscopic findings following single-dose administration of 0.07, 0.1, and tissue types (data not shown), consistent with the known expression 0.2 mg/kg DM1 in Sprague–Dawley rats at necropsy on day 3.a of HER2 on normal epithelium (Press et al., 1990), providing support Microscopic finding Organ/tissue Necropsy results that mitotic arrest in those tissues is due to antigen-dependent binding Nb (dose, mg/kg)c of T-DM1. However, cellular mitotic arrest was even more widespread in rats, affecting tissues beyond those identified in the monkey. Degeneration and/or Liver (hepatocytes, periportal/ 2(0.07),9(0.1),10(0.2) necrosis midzonal) This indicates that antigen-independent uptake of T-DM1 can occur Liver (biliary epithelium) 3 (0.1), 9 (0.2) either by Fc-mediated or nonspecific endocytotic mechanisms, or Kidney (tubule) 6 (0.07), 10 (0.1), 9 (0.2) through T-DM1 catabolism and subsequent uptake of DM1-containing Thymus (lymphocytes) 3 (0.1), 9 (0.2) catabolites. Duodenum (lamina propria) 1 (0.1), 3 (0.2) A notable adverse finding observed only in monkeys was microsco- Jejunum (lamina propria) 1 (0.2) Rectum (epithelium) 1 (0.1), 6 (0.2) pic axonal degeneration in the sciatic nerve and dorsal funiculus of the Increased mitotic Liver (biliary epithelium) 10 (0.07), 10 (0.1), 10 (0.2) spinal cord after four doses of 10 or 30 mg/kg (see Table 7, Fig. 5)or figures/arrested Liver (hepatocytes) 9 (0.07), 10 (0.1), 10 (0.2) eight doses of 1, 3, or 10 mg/kg (data not shown) that were not revers- metaphase Liver (Kupffer cells) 6 (0.1), 10 (0.2) ible within the 6-week recovery period. These observations were Kidney (tubule) 10 (0.07), 10 (0.1), 10 (0.2) Hemorrhage Femur marrow 8 (0.07), 10 (0.1), 10 (0.2) supported by tissue cross-reactivity data from cynomolgus monkey Sternum marrow 8 (0.07), 10 (0.1), 10 (0.2) and human frozen tissue sections in which low-intensity membranous Hypertrophy/ Liver (Kupffer cells) 2 (0.07), 3 (0.1), 7 (0.2) staining was detected in glial cells and peripheral nerve spindle cells vacuolation Spleen (reticuloendothelial 8(0.2) (presumptive Schwann cells). However, these microscopic changes cells; with increased mitoses) did not translate to effects noted during neurologic examinations or Hypocellular Femur marrow 6 (0.07), 10 (0.1), 10 (0.2) Sternum marrow 8 (0.07), 10 (0.1), 10 (0.2) other in-life observations. Nonetheless, the high incidence (14 of 14 Other animals) and severity (mild to severe) of this neurologic finding at the Lymphoid depletion Spleen 5 (0.2) dose of 30 mg/kg was the basis for the calculation of the starting dose a Increased mitosis in adrenal cortex of two out of five females given 0.2 mg/kg seen at for the phase I clinical study. recovery necropsy (day 22); no other gross or microscopic observations at recovery. The TK analyses of the repeat-dose study (q3w × four doses) in b Data from males and females combined, unless otherwise noted; n = 10 for each cynomolgus monkeys confirmed exposure to T-DM1 at 3, 10, and dose evaluated. 30 mg/kg (see Fig. 3; Table 8), and dose-proportional TK in the two c 0.7, 0.1, and 0.2 mg/kg DM1 doses are equivalent to ~500, ~700–800, and ~1400–1600 μgDM1/m2, respectively. higher-dose groups (10 and 30 mg/kg). CL of T-DM1 was faster than CL of total trastuzumab. Animals treated with 3 mg/kg demonstrated nonlinear TK with ~50% faster CL of T-DM1 (see Table 8) and total trastuzumab (data not shown) compared with that in animals treated administered T-DM1 had increased numbers of cells arrested in meta- with 10 or 30 mg/kg. This observation is consistent with target- phase mitosis that extended to tissues beyond the liver, indicating mediated disposition. CL of T-DM1 (see Table 8), total trastuzumab that antigen-dependent or -independent uptake of T-DM1 occurred (data not shown), and DM1 (data not shown) was similar between with subsequent DM1-driven disruption of cellular division. Tissue sexes at all dose levels. Little or no accumulation of T-DM1 was observed cross-reactivity results from human and cynomolgus monkey frozen in any dose group (data from female animals treated with T-DM1 tissue sections showed that membranous staining consistent with 30 mg/kg are presented in Fig. 3). Following the administration of T-DM1, DM1 concentrations were shown to be at least 50-fold lower

(on a molar basis) than T-DM1 at all times (e.g., T-DM1 Cmax =5.2μM versus DM1 C =0.09μM) (see Fig. 3). Exposure to DM1 (maximum 10000 max concentration, area under the time–concentration curve for all values [AUC ], and/or partial AUC) was proportional to the T-DM1 dose 1000 ALL administered (see Table 7; data not shown). As was observed in rats, CL of T-DM1 conjugate was faster than CL of total trastuzumab. This 100 phenomenon is also observed in patients (Girish et al., 2012). DM1 exposures were two-fold higher in rats than in monkeys when given 10 T-DM1 containing molar equivalent doses of DM1, which may explain, at least in part, the tolerability differences seen between rats and mon- 1 keys. The TK profile in cynomolgus monkeys after chronic administra- tion of T-DM1 (q3w × eight doses) was comparable to the profile in 0.1 monkeys administered four doses of T-DM1 (data not shown). The Total Tmab chronic toxicology study was the only study in which serum ATAs T-DM1 Concentration (µg/mL) 0.01 were detected, and the finding was limited to four of 36 monkeys DM1 (11%) (n = 3 at 1 mg/kg, n = 1 at 10 mg/kg) (Carrasco-Triguero 0.001 et al., 2013). There was no apparent effect of ATAs on TK. Although animal models are not always predictive of immunogenicity in humans 0.0001 (Swanson and Bussiere, 2012), to date, immunogenicity rates have been 0 21 42 63 84 105 low in T-DM1 clinical trials (Girish et al., 2012). Time (day) Cardiovascular safety Fig. 3. Toxicokinetic analyses during repeat-dose administration of T-DM1 in female cynomolgus monkeys. Mean (±SD) concentrations of total trastuzumab, T-DM1, and When telemeterized monkeys were administered a single IV dose of DM1 following repeat-dose administration (days 1, 22, 43, and 64) of T-DM1 30 mg/kg 0, 3, 10, or 30 mg/kg T-DM1, ECG examinations revealed no drug- (6000 μgDM1/m2). DM1 concentrations were at least 50-fold lower (on a molar basis) related effects on RR or QTc (see Fig. 6A). One individual animal in the than T-DM1 at all times (e.g., T-DM1 C =5.2μMversusDM1C =0.09μM). SD, max max fi standard deviation; T-DM1, trastuzumab emtansine; Tmab, trastuzumab monoclonal 30 mg/kg group had a notably prolonged QTc that was rst noted on antibody. day 4 and persisted through day 22. The relationship of this finding to K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313 307

Table 6 Selected hematologic and clinical chemistry parameters at day 3 from the repeat-dose administration of T-DM1 studies in cynomolgus monkeys.

Parameter, mean (SD) Repeat-dose T-DM1

Male Female

Vehicle 3mg/kg 10 mg/kg 30 mg/kg Vehicle 3 mg/kg 10 mg/kg 30 mg/kg alone (600 μgDM1/m2) (2000 μgDM1/m2) (6000 μgDM1/m2) alone (600 μgDM1/m2) (2000 μgDM1/m2) (6000 μgDM1/m2)

Hematology, thousands/μL Platelet counts 434 395 361 380a 418 460 359 452 (119.1) (96.5) (86.7) (95.0) (128.7) (79.5) (65.1) (151.9) Reticulocyte, absolute 75.4 78.8 114.5 56.2 73.0 71.0 70.8 60.7 (31.34) (14.43) (27.76) (26.92) (34.10) (18.64) (15.75) (18.78) Neutrophil, absolute 4.05 5.62 6.87a 7.47a 3.72 6.69 7.28a 8.53a (1.647) (2.526) (1.687) (2.363) (1.370) (2.631) (1.775) (2.909) Lymphocyte 4.64 4.69 6.38 5.26 6.99 4.54 4.12 4.40a (1.429) (1.657) (2.625) (2.479) (1.776) (1.105) (1.339) (1.352) Clinical chemistry ALT, U/L 47 69 64 82 45 84 96 169a (15.2) (21.0) (22.1) (27.0) (16.7) (17.1) (52.2) (157.5) AST, U/L 43 70 78a 153a 41 72 98 199a (12.7) (15.0) (20.7) (37.5) (6.9) (25.1) (34.3) (101.6) ALP U/L 459 552 605 951 285 246 307 642a (113.5) (124.4) (168.1) (508.5) (95.0) (70.9) (125.8) (232.1) GGT, U/L 64 81 59 69 52 64 48 65 (19.5) (29.6) (9.8) (9.3) (14.5) (13.4) (13.8) (12.2) Total bilirubin, mg/dL 0.3 0.5 0.3 0.3 0.3 0.4 0.3 0.3 (0.14) (0.63) (0.11) (0.08) (0.04) (0.13) (0.07) (0.14)

ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; SD, standard deviation; T-DM1, trastuzumab emtansine. a p ≤ 0.05 compared with control.

treatment with T-DM1 is unclear, but evidence of poor health in this an- dose toxicity studies — data not shown) suggest that the effects on QT imal (decreased body weight, elevated clinical pathology parameters, interval were confounded by factors not related to T-DM1. However including white blood cells, lymphocytes, AST, ALT, and total bilirubin) in this same study, potential T-DM1-related effects on hemodynamic that arose during acclimation (posttelemetry implant) and persisted parameters were identified. These effects were limited to the high- through the end of the study, combined with a lack of similar changes dose group of 30 mg/kg (about seven times the clinical exposure ob- in the other animals given 30 mg/kg T-DM1 (including data from the served in patients following 3.6 mg/kg T-DM1 administration based safety pharmacology study and external ECG data from the repeat- on AUC) (Girish et al., 2012) and consisted of modest increases in

Table 7 Notable microscopic findings following repeat-dose administration of 3, 10, and 30 mg/kg T-DM1 in cynomolgus monkeys at necropsy on days 66, 85, and 106.

Microscopic finding Organ/tissue Day 66 Na Necropsy results (dose, mg/kg)b Day 85 Na Day 106 Na (dose, mg/kg)b (dose mg/kg)b

Axonal degeneration Spinal cord (dorsal funiculus) 3 (10), 6 (30) 3 (10), 4 (30) 2 (10), 4 (30) Sciatic nerve 3 (10), 6 (30) 1 (10), 4 (30) 2 (10), 4 (30) Hypertrophy and/or hyperplasia Sciatic nerve (Schwann cells) 6 (30) 4 (30) 4 (30) Liver (Kupffer cells) 4 (3), 6 (10), 6 (30) 2 (30) Spleen (reticuloendothelial cells) 1 (3), 2 (10), 6 (30) Increased mitotic figures/arrested metaphase Liver (Kupffer cells) 4 (3), 6 (10), 6 (30) 2 (30) 3 (30) Spleen (reticuloendothelial cells) 1 (3), 4 (10), 6 (30) 1 (10), 4 (30) Tongue (basal epithelium) 6 (3), 6 (10), 6 (30) Kidneys (tubular epithelial cells) 6 (30) Mandibular salivary gland (acinar cells) 1 (30) Skin (basal epithelium) 6 (3), 6 (10), 6 (30) IV site, right saphenous (basal epithelium) 6 (3), 6 (10), 6 (30) Lymphoid depletion Spleen (follicular center) 3 (30) Thymus (grades 1 to 4) 1 (0), 3 (3), 5 (10), 6 3 (0), 3 (3), (30) 3 (10), 4 (30) Mesenteric lymph node 1 (3), 2 (30) Other Increased sinusoidal leukocytes Liver 4 (10), 5 (30) 2 (30) Multinucleated hepatocytes Liver 1 (10), 5 (30) 1 (30) Vacuolation, centrilobular Liver 1 (3), 2 (10), 3 (30) Increased cellularity, red pulp Spleen 1 (3), 1 (10), 3 (30) Decreased cytoplasmic granules Mandibular salivary gland 2 (30)

IV, intravenous; T-DM1, trastuzumab emtansine. a Data from males and females combined; n = 6 (day 66), 4 (day 85), and 4 (day 106) for each dose evaluated. b T-DM1 doses of 3, 10, and 30 mg/kg are equivalent to ~600, ~2000, and ~6000 μgDM1/m2, respectively. 308 K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313

A The potential of T-DM1 to induce QT prolongation was investigated 40 0 mg/kg by evaluating the in vitro effects of DM1 on the hERG channel current, a 30 3 mg/kg 10 mg/kg surrogate for IKr, the rapidly activating, delayed rectifier cardiac potas- 20 30 mg/kg sium current (data not shown). Blockage of this channel is associated 10 with QT prolongation and torsades de pointes (Raschi et al., 2009). 0 When DM1 concentrations of 2.6, 8.8, and 29.5 μM (~1.9, ~6.5, and

–10 ~21.8 ng/mL) were tested, the inhibitory concentration at 20% (IC20) –20 and IC50 for the inhibitory effect of DM1 on hERG potassium current –30 could not be determined and was estimated to be N29.5 μM. Based on

Mean platelet counts –40 an IC50 of ≥29.5 μM, average plasma DM1 concentrations detected in (% change from baseline) –50 T-DM1-treated patients (6 ng/mL), and 93% plasma protein binding, a Baseline 322 4324 45 64 66 85 94 106 safety margin of at least 30-fold was calculated, which is considered Time point (day) adequate for even nononcology therapeutics (Redfern et al., 2003). B 300 0 mg/kg Discussion 3 mg/kg 250 10 mg/kg 30 mg/kg T-DM1 was approved by the US Food and Drug Administration in 200 February 2013 for the treatment of patients with HER2-positive meta- static breast cancer who have received prior treatment with trastuzumab 150 and a taxane, either separately or in combination (Kadcyla package 100 insert, 2013). Preclinical characterization of toxicities associated with T-DM1 and DM1 provided early insights into a safety profile that has Mean ALT (units/L) 50 been consistent with the most common adverse events in patients to date (Burris et al., 2011; Kadcyla package insert, 2013; Verma et al., 0 –21 –5 3 2422 43 45 6664 85 94 106 2012). Furthermore, the toxicity studies in rats comparing T-DM1 with Time point (day) DM1 further support the theory that ADCs have the potential to widen the therapeutic window of potent chemotherapeutic agents. C 400 0 mg/kg The mechanism of toxicity of T-DM1 is consistent with the pharma- 350 3 mg/kg cology of DM1 as an inhibitor of tubulin polymerization (Lopus et al., 10 mg/kg 300 30 mg/kg 2010; Oroudjev et al., 2010). The concordance of toxicities observed in rats (antigen-nonbinding species) and cynomolgus monkeys (antigen- 250 binding species) treated with either T-DM1 or DM1 indicates that the 200 toxicities are primarily antigen-independent and consistent with the 150 mechanism of action and pharmacologic activity of DM1. In both AST (units/L) 100 species, the hallmark DM1 signature, cellular mitotic arrest, is seen 50 histologically in the affected target organs. This DM1-driven mechanism of toxicity for T-DM1 is supported by a similar nonclinical toxicity pro- 0 fi –21 –5 322 24 43 456664 85 94 106 le for cantuzumab mertansine (C242-DM1), another ADC containing Time point (day) DM1 as its cytotoxic component (Tolcher et al., 2003). Cantuzumab mertansine was tolerated by monkeys at doses comparable with Fig. 4. Percent change from baseline in platelet counts and mean ALT and AST levels over T-DM1 and resulted in increased liver enzymes, axonal degeneration, time from the repeat-dose monkey study (dosed on days 1, 22, 43, and 64 at 0, 3, 10, and and hematopoietic effects. In contrast, trastuzumab has been extremely 30 mg/kg T-DM1 (~600, ~2000, and ~6000 μgDM1/m2, respectively); male and female data combined). (A) Platelet counts were decreased 2 days postdose during each dose well tolerated in extensive nonhuman primate safety evaluations and cycle for animals administered 30 mg/kg T-DM1, with partial or complete reversibility lacks any of the findings associated with T-DM1 administration. Taken at the end of each dose cycle. p ≤ 0.05 decreases/increases relative to control for males together, these data illustrate that the toxicities observed with T-DM1 on days 3, 22, and 24, and for females on days 45, 64, and 66. (B) Mean (SD) ALT levels are related to DM1, rather than the antibody or chemical linker, and were elevated 2 days postdose during each dose cycle for animals administered 10 or appear to be generally independent of the target antigen. 30 mg/kg T-DM1, with partial or complete reversibility at the end of each dose cycle. p ≤ 0.05 increases relative to control at 30 mg/kg for females only on days 3, 22, 24, 43, T-DM1-related trends in transient cyclic elevations in liver enzymes 45, and 66 and (C) mean AST (SD) levels were elevated 2 days postdose during each and anemia have been consistent between nonclinical species (espe- dose cycle for animals administered 10 or 30 mg/kg T-DM1, with partial or complete cially monkeys) and patients to date. However, the minimal to mild de- ≤ reversibility at the end of each dose cycle.) p 0.05 increases relative to control at crease in platelet counts observed in the nonclinical species manifested 30 mg/kg for males and females on days 3, 24, 43, 45, 64, 66 and 85. Similar cyclic eleva- tions in alkaline phosphatase were also observed at 30 mg/kg T-DM1 (data not shown). as the dose-limiting toxicity in patients in the phase I clinical trial at ALT, alanine aminotransferase; AST, aspartate aminotransferase; SD, standard deviation; 4.8 mg/kg (Krop et al., 2010). The reason for this difference in sensitiv- T-DM1, trastuzumab emtansine. ity between nonclinical species and patients remains unclear, but in vitro investigations have contributed to a better understanding of the mechanism of thrombocytopenia (Mahapatra et al., 2011). Experi- systolic (see Fig. 6B), diastolic (Fig. 6C), mean arterial, and pulse pres- ments of platelet function using platelet-rich plasma or washed plate- sures (data not shown) that were variable in onset, magnitude, and lets have shown that T-DM1 does not have a direct effect on platelet duration in individual monkeys. Peak changes in BP were most consis- function (Mahapatra et al., 2011). However, platelet formation experi- tently observed on day 5 with a trend in reversal by day 22 for three ments using human hematopoietic stem cells (CD133+/CD34+)have of four animals. Elevations persisted through day 22 for one animal. shown that both T-DM1 and an irrelevant binding control ADC (MAb- The toxicologic impact of these hemodynamic changes is unclear. Tissue DM1) can impair megakaryocyte maturation and subsequent platelet changes (e.g., tissue congestion or vascular or glomerular changes) that release. The similarity of effects for both T-DM1 and the MAb-DM1 might reflect sustained hypertensive effects were not identified at nec- ADC supports the hypothesis that the observed thrombocytopenia is ropsy or in the microscopic examination of tissues from the monkeys in an antigen-independent toxicity resulting from exposure to DM1 the repeat-dose studies. (Mahapatra et al., 2011). Although unconjugated DM1 was not included K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313 309

AB

Control, day 66 30 mg/kg × 4, day 66 CD

10 mg/kg × 4, day 66 10 mg/kg × 4 + 3-week recovery

Fig. 5. Axonal degeneration of sciatic nerves following repeat-dose administration of T-DM1 in cynomolgus monkeys. At day 66 (2 days after final dose), compared with (A) control, slight to severe axonal degeneration was observed following (B) repeat-dose treatment with 30 mg/kg T-DM1 (6000 μgDM1/m2) and (C) Minimal lesions were noted in animals treated with 10 mg/kg (2000 μgDM1/m2). (D) Axonal degeneration was not reversible following a 3-week (day 85) or 6-week (day 106) recovery period (day 106, not shown). T-DM1, trastuzumab emtansine. in the in vitro platelet formation assessment, decreased platelet levels Peripheral sensory and motor neuropathy related to nerve axo- were observed in rats administered either T-DM1 or DM1, indicating nal degeneration is a known toxicity of microtubule-inhibiting that this phenomenon can occur whether DM1 is in conjugated or chemotherapeutic agents, including taxanes, vinca alkaloids, and unconjugated form. In another experiment, megakaryocytes were incu- auristatins (Adcetris package insert, 2012; Park et al., 2008). Given bated with Alexa488-conjugated trastuzumab or T-DM1, and surface that maytansines and DM1 share this mechanism of action, it was antic- binding and internalization were confirmed by immunofluorescence ipated that DM1 and possibly T-DM1 would cause some degree of neu- and flow cytometry. Preincubation with anti-CD32, which binds rologic injury. Monkeys given 10 or 30 mg/kg T-DM1 (one- to 10-fold and blocks FcRγIIb, decreased Alexa488-conjugated trastuzumab above the patient exposure range) exhibited varying degrees of axonal and T-DM1 antibody binding and uptake by approximately two- degeneration. Although this histologic finding did not translate to ob- fold, suggesting that FcRγIIb may also partially contribute to the inter- servable neurologic deficits, the high incidence (14 of 14 animals) and nalization process and could contribute to the enhanced sensitivity in severity (mild to severe) of this finding in monkeys administered patients (Mahapatra et al., 2011). 30 mg/kg T-DM1 was the basis for calculation of the starting dose for

Table 8

T-DM1 TK parameters and DM1 plasma Cmax following repeat-dose administration of T-DM1 (q3w × four doses) in cynomolgus monkeys.

TK parameter (mean ± SD) T-DM1 dose

3 mg/kg 10 mg/kg 30 mg/kg

Males Females Males Females Males Females

AUC0–inf (day · μg/mL) 189 ± 26.5 195 ± 19.8 871 ± 107.0 930 ± 125.0 2900 ± 552.0 3150 ± 162.0

Cmax (μg/mL) 75.7 ± 6.51 78.0 ± 3.16 266 ± 31.8 265 ± 20.7 787 ± 76.9 776 ± 39.9 CL (mL/day/kg) 16.1 ± 2.25 15.3 ± 1.5 11.5 ± 1.3 10.7 ± 1.3 10.5 ± 2.3 9.4 ± 0.6

t1/2 effective (days) 2.43 ± 0.23 2.66 ± 0.12 3.76 ± 0.51 4.13 ± 0.29 4.63 ± 0.80 5.16 ± 0.23

Vss (mL/kg) 56.5 ± 10.0 58.4 ± 3.7 61.6 ± 6.6 63.7 ± 9.4 68.1 ± 4.7 70.0 ± 4.7

DM1 Cmax, ng/mL (mean ± SD) T-DM1 dose 3 mg/kg 10 mg/kg 30 mg/kg

Males Females Males Females Males Females

Dose 1 5.95 ± 0.770 5.85 ± 1.13 21.9 ± 1.90 19.6 ± 1.00 74.6 ± 9.27 58.6 ± 2.12 Dose 2 6.98 ± 1.03 6.72 ± 1.23 23.6 ± 10.1 24.2 ± 3.39 81.8 ± 9.82 65.1 ± 10.5 Dose 3 8.04 ± 0.574 7.09 ± 1.01 25.1 ± 3.99 23.6 ± 3.88 86.1 ± 11.1 66.1 ± 4.70 Dose 4 7.02 ± 0.957 7.53 ± 1.46 23.6 ± 2.15 23.4 ± 1.10 76.8 ± 7.55 59.7 ± 4.77

AUC0–inf, area under the concentration versus time curve from time 0 to infinity; CL, clearance; Cmax, maximum observed concentration; q3w, every 3 weeks; SD, standard deviation; t1/2 effective, effective half-life; T-DM1, trastuzumab emtansine; TK, toxicokinetic; Vss, volume of distribution at steady state. 310 K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313 A 0 mg/kg 30 mg/kg

450 I03812 450 I03812 I03813 I03813 I03814 I03814 I03815 I03815 400 400

350 350

300 300 QTc interval (mSec) QTc interval (mSec) QTc 250 250

0 5 10 15 20 25 0 5 10 15 20 25 Time point (day) Time point (day)

B 0 mg/kg 30 mg/kg 160 I03812 160 I03813 I03814 I03815 140 140

120 120 (mm Hg) (mm Hg)

100 100 Aortic systolic pressure Aortic systolic pressure 80 80

0–2 2 4 6 8 10 12 14 16 18 20 22 24 0–22 4 6 8 10 12 14 16 18 20 22 24 Time point (day) Time point (day) Predose Predose –1 hour –1 hour

C 0 mg/kg 30 mg/kg

120 I03812 120 I03813 I03814 I03815

100 100

80 80 (mm Hg) (mm Hg)

60 60 Aortic diastolic pressure Aortic diastolic pressure

–20 2 4 6 8 1012141618202224 0–22 4 6 8 10 12 14 16 18 20 22 24 Time point (day) Time point (day) Predose Predose –1 hour –1 hour

Fig. 6. Individual corrected QT interval (QTc) (A), individual aortic systolic pressure (B), and aortic diastolic pressure (C) for 30 mg/kg T-DM1 group versus vehicle control from days 1 to 22 of the single dose cynomolgus monkey cardiovascular safety pharmacology study. Similar changes were observed at 30 mg/kg (6000 μgDM1/m2) for mean arterial pressure and pulse pressure (data not shown). Plotted time points: (predose (−1 h), 24 h postdose, and days 3, 4, 5, 8, 15, and 22).

the phase I clinical study. Peripheral neuropathy has been reported in rats, and impairment of the central and peripheral nervous system in pa- patients with breast cancer who were administered the established tients) (Issell and Crooke, 1978; Rebert et al., 1984; Sieber et al., 1978) clinical dose of 3.6 mg/kg T-DM1 (Burris et al., 2011; Krop et al., suggest that the neurologic toxicity of T-DM1 is antigen-independent 2012), but, to date, the incidence and severity (mainly ≤grade 2) has and driven by the DM1 component of the molecule. This theory is also been low, and the condition has been well managed with dose reduction supported by findings of neurologic toxicity in cynomolgus monkeys or dose holiday. The mechanisms of T-DM1-induced neurotoxicity treated with non-HER2 targeting cantuzumab mertansine (Tolcher can likely be attributed to both antigen-dependent and -independent et al., 2003). In studies of T-DM1 and DM1 in rats and T-DM1 in mechanisms. The neurologic effects induced by repeat doses of cynomolgus monkeys, however, the findings suggest that there may maytansine (hind limb paralysis in mice, decreased motor activity in be an antigen-dependent contribution to the neurologic toxicity as K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313 311 well. Only monkeys (a HER2-binding species) exhibited axonal de- trastuzumab. The single-dose rat toxicity studies indicate that the generation, and the neurologic effects seen with maytansine were not tolerated dose of DM1 when conjugated to trastuzumab is approxi- reproduced in rats given either T-DM1 or DM1. Based on tissue cross- mately two-fold greater than the dose-equivalent of unconjugated reactivity experiments, low-level HER2-dependent staining was ob- DM1 alone. When mortality occurred in rats given DM1 at served in monkey and human glial cells and peripheral nerve spindle 2200 μgDM1/m2, T-DM1 was tolerated at equivalent DM1 doses up cells (presumptive Schwann cells), indicating that T-DM1 effects on to 4400 μg DM1/m2. A similar comparison can be made clinically be- the nerve may be partially antigen-dependent and may, in part, explain tween T-DM1 and the parent compound of DM1, maytansine, which the presence of this finding in monkeys. The absence of neurologic tox- was evaluated in phase I and phase II clinical trials in the late 1970s. icity in rats may be explained both by the limited exposure to DM1 from Maytansine proved to have too narrow a therapeutic index, which a single dose of T-DM1 or DM1 and by the lack of antigen-dependent resulted in the discontinuation of its development. Although anti- contributions to this pathology. Based on the neurologic effects induced tumor responses were observed at doses of 1.6 and 2.0 mg/m2 in pa- in mice and humans with maytansine, the axonal degeneration observed tients with solid tumor or hematologic cancers, the beneficial effects in monkeys given T-DM1 or cantuzumab mertansine (HER2 antigen- were accompanied by dose-limiting toxicities after a limited number binding and -nonbinding DM1-ADCs, respectively), and low-grade pe- of treatment cycles, including profound weakness, severe nausea and ripheral neuropathy observed in humans administered T-DM1, it is likely vomiting, and neurologic toxicity (Chabner et al., 1978; Chahinian that a combination of antigen-dependent and -independent mecha- et al., 1979). Given the chemical similarity and comparable potency of nisms underlie the neurologic toxicity of T-DM1 observed in these maytansine and DM1, it is likely that administration of DM1 would re- studies. sult in a similar clinical profile. By contrast, 3.6 mg/kg T-DM1 (approx- Although cardiotoxicity was also not identified in trastuzumab pre- imately 2.3 mg DM1/m2) has been well tolerated in patients treated clinical studies (Klein and Dybdal, 2003), clinically, trastuzumab has over multiple cycles (Verma et al., 2012). The findings from the acute been associated with an increased risk of cardiac dysfunction, and this toxicity assessments in rats support the apparent differences in tolera- risk is increased when trastuzumab is given in combination with bility seen clinically between maytansine and T-DM1. anthracyclines (De Keulenaer et al., 2010; Seidman et al., 2002). Given In summary, T-DM1 had a favorable safety profile in preclinical this phenomenon, we hypothesized that the conjugation of a cytotoxic studies. The differences in species and patient sensitivity and/or component to trastuzumab had the potential to enhance this liability. toxicity profile may be due to differences in exposure to T-DM1 The results from the cynomolgus monkey cardiovascular safety study and/or DM1 and the impact of antigen distribution and binding, did not reveal any significant changes to ECG parameters, including but overall the toxicity profile translated well from rat to monkey QTc (see Fig. 6A). Also, there was no evidence of cardiomyocyte damage to human. The toxicities associated with T-DM1 and DM1 are in the ventricle, atrium, or atrial–ventricular valve evaluated histopath- well characterized, predictable, and consistent with the mechanism ologically in the single- or multiple-dose T-DM1 studies in monkeys. In of action and pharmacologic activity of DM1. Finally, the studies of addition, DM1 did not inhibit the hERG channel current at concentra- T-DM1 and DM1 in rats, and the emerging differences in patient tions up to 30-fold higher than the fraction of unbound DM1 in clinically tolerability for T-DM1 versus maytansine, uphold the premise that relevant doses of trastuzumab. These results are consistent with find- ADCs have the potential to improve the therapeutic window for ings from a phase II dedicated QT study in which T-DM1 had minimal ef- cytotoxic agents. fects on QT prolongation at the proposed therapeutic dose (3.6 mg/kg Supplementary data to this article can be found online at http://dx. T-DM1 q3w) (Gupta et al., 2013). T-DM1-related effects on hemody- doi.org/10.1016/j.taap.2013.09.003. namic parameters were identified in the monkey safety pharmacolo- gy study (Figs. 6B and C), but the mechanism of this effect is unclear and was variable in onset, magnitude, and duration in individual Conflict of interest statement monkeys. Chronic administration of T-DM1 was not associated with evidence of hypertension in monkeys, and the incidence of KAP, KF, JB, SK, OS, J-H Yi, SG, ND, and TR are employees of hypertension in patients has been low and primarily grade 1 or 2 to Genentech, Inc., and own stock in F. Hoffmann-La Roche Ltd. date (Kadcyla package insert, 2013). In total, results from the preclinical JT was an employee of Genentech, Inc., at the time the experiments toxicity and safety pharmacology studies showed that T-DM1, like were being conducted. He owns stock in F. Hoffmann-La Roche Ltd. trastuzumab, is not associated with acute changes in cardiac conduction or hemodynamic parameters. However, cardiac dysfunction resulting in reduced left ventricular ejection fraction that is observed in some Acknowledgments trastuzumab patients and results in drug discontinuation was not measured in these studies. The authors would like to thank: Fiona Zhong (Genentech Inc.), Angela DM1 and maytansine are closely related chemical entities of the Hendricks (Genentech Inc.) Geoffrey Ganem (Genentech Inc.), Michelle same maytansinoid class and are considered equipotent based on data McDowell (Genentech Inc.), Trung Nguyen (Genentech Inc.), Kristin Lewis from acute toxicity studies: The LD10 of maytansine is ~0.4 mg/kg, and (previously at Genentech Inc. but now affiliated with Ohio State University DM1 is not tolerated at ≥0.4 mg/kg (2400 μgDM1/m2)(Issell and Veterinary School), Khiem Tran (Genentech Inc.), Kevin Williams (Covance Crooke, 1978; Mugera and Ward, 1977)(seeFig. 2). Maytansine is Laboratories, Madison, WI), Lisa Biegel (Covance Laboratories, Madison, embryotoxic, teratogenic, and clastogenic in mice at single intraperito- WI), Jay Herman (Covance Laboratories, Madison, WI), Mingyi Trimble neal doses of 0.1 mg/kg, 0.2 mg/kg, and 0.25 mg/kg (Issell and Crooke, (Covance Laboratories, Madison, WI),AmyJerde(CovanceLaboratories, 1978; Sieber et al., 1978). Importantly, when rats are administered sim- Madison, WI), Leon Stankowski (Bioreliance), Yong Xu (Covance Labora- ilar doses of DM1 (0.07 mg/kg, 0.1 mg/kg, or 0.2 mg/kg) intravenously, tories, Greenfield, IN), Christina Satterwhite (Charles River Laboratories, their plasma DM1 concentrations (5–28 ng/mL) are comparable with Reno, NV), and Cynthia Wladyka (ChanTest, Cleveland, OH). the range of plasma DM1 concentrations achieved in patients given The study was funded by Genentech, Inc. The authors, all of whom 3.6 mg/kg T-DM1 q3w (Girish et al., 2012). Hence, patients will be are current or former Genentech employees, designed the experiments; exposed to concentrations of DM1 that have been (indirectly) shown collected, analyzed, and interpreted the data; and decided to submit the to lead to fetal harm, irrespective of whether the fetus is exposed to article for publication. The first author wrote the manuscript, and all significant levels of T-DM1 during organogenesis. other authors reviewed and critically evaluated the manuscript. Support Another outcome of the studies reported here was the demonstrated for third-party writing assistance for this manuscript was provided by improvement in tolerability that resulted from conjugation of DM1 to Genentech, Inc. 312 K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313

Appendix A

Table A.1

T-DM1 toxicokinetic parameters and DM1 plasma Cmax measurement following single-dose administration of T-DM1 in Sprague–Dawley rats (total trastuzumab data not shown).

TK parameter (mean ± SDa) T-DM1 dose

6 mg/kg 20 mg/kg 60 mg/kgb

Males Females Males Females Males Females

AUC0–inf (day · μg/mL) 426 ± 28.1 436 ± 18.4 1520 1330 ––

AUC0–3 (day · μg/mL) 206 ± 3.34 198 ± 3.09 683 ± 14.0 598 1820 ± 86.8 1500 ± 98.91 CL (mL/day/kg) 13.8 ± 0.896 13.5 ± 0.581 13.1 15.0 ––

Cmax (μg/mL) 162 ± 11.0 168 ± 3.88 489 ± 27.7 432 1720 ± 81.9 1510 ± 223

t1/2 effective (days) 3.80 ± 0.126 4.08 ± 0.196 4.66 4.57 ––

Vss (mL/kg) 56.5 ± 10.0 58.4 ± 3.7 61.6 ± 6.6 63.7 ± 9.4 68.1 ± 4.7 70.0 ± 4.7

T-DM1 dose

6 mg/kg 20 mg/kg 60 mg/kg

Males Females Males Females Males Females

DM1 Cmax, ng/mL 13.2 ± 2.11 8.27 ± 0.858 43.9 ± 1.38 59.6 ± 0.778 124 ± 12.9 154 ± 13.9

AUC0–inf, area under the concentration versus time curve from time 0 to infinity; AUC0–3, area under the concentration versus time curve from 0 to 3 days; Cmax, maximum observed concentration; CL, clearance; SD, standard deviation; t1/2, effective, effective half-life; T-DM1, trastuzumab emtansine; TK, toxicokinetic; Vss, volume of distribution at steady state. a SD not available for some parameters/doses because n = 2/sex. b Only exposure parameters (AUC0–3 and Cmax) are reported at 60 mg/kg because of animal mortality.

Table A.2 DM1 toxicokinetic parameters following single-dose administration of DM1 in Sprague–Dawley rats.

TK parameter (mean ± SD) DM1 dose

0.07 mg/kg 0.1 mg/kg 0.2 mg/kg

Males Females Males Females Males Females

AUC0–inf (min · ng/mL) 1290 ± 140 1530 ± 528 2940 ± 883 3950 ± 664 6340 ± 906 9160 ± 1090

Cmax (ng/mL) 9.67 ± 2.03 8.71 ± 0.0316 13.8 ± 1.36 13.9 ± 1.13 22.5 ± 2.62 28.2 ± 2.89 CL (mL/min/kg) 52.0 ± 6.19 46.7 ± 16.1 39.2 ± 9.38 27.3 ± 4.70 31.6 ± 4.83 21.4 ± 1.80

t1/2 terminal (min) 214 ± 28.4 304 ± 116 377 ± 126 546 ± 105 548 ± 62.8 673 ± 87.1

Vss (mL/kg) 14,300 ± 1340 17,100 ± 1070 18,600 ± 1300 19,900 ± 1450 23,700 ± 4740 19,800 ± 2550

AUC0–inf, area under the concentration versus time curve from 0 to infinity; CL, clearance; Cmax, maximum observed concentration; SD, standard deviation; t1/2 terminal, terminal half-life; TK, toxicokinetic; Vss, volume of distribution at steady state.

References Chari, R.V., 2008. Targeted cancer therapy: conferring specificity to cytotoxic drugs. Acc. Chem. Res. 41, 98–107. Adcetris [package insert]. Bothell, WA: Seattle Genetics; 2012. Dawood, S., Broglio, K., Buzdar, A.U., Hortobagyi, G.N., Giordano, S.H., 2010. Prognosis of Alley, S.C., Okeley, N.M., Senter, P.D., 2010. Antibody-drug conjugates: targeted drug de- women with metastatic breast cancer by HER2 status and trastuzumab treatment: livery for cancer. Curr. Opin. Chem. Biol. 14, 529–537. an institutional-based review. J. Clin. Oncol. 28, 92–98. Blackwell, K.L., Burstein, H.J., Storniolo, A.M., Rugo, H., Sledge, G., Koehler, M., Ellis, C., De Keulenaer, G.W., Doggen, K., Lemmens, K., 2010. The vulnerability of the heart as a Casey, M., Vukelja, S., Bischoff, J., Baselga, J., O'Shaughnessy, J., 2010. Randomized pluricellular paracrine organ: lessons from unexpected triggers of heart failure in study of lapatinib alone or in combination with trastuzumab in women with targeted ErbB2 anticancer therapy. Circ. Res. 106, 35–46. ErbB2-positive, trastuzumab-refractory metastatic breast cancer. J. Clin. Oncol. 28, Dere, R., Yi, J.H., Lei, C., Saad, O.M., Huang, C., Li, Y., Baudys, J., Kaur, S., 2013. PK assays for 1124–1130. antibody-drug conjugates: case study with ado-trastuzumab emtansine. Bioanalysis Blättler, W.A., Chari, R.V.J., 2001. Drugs to enhance the therapeutic potency of anticancer 5, 1025–1040. antibodies: antibody-drug conjugates as tumor-activated prodrugs. In: Ojima, I., Vite, Eagan, R.T., Ingle, J.N., Rubin, J., Frytak, S., Moertel, C.G., 1978. Early clinical study of an G., Altmann, K. (Eds.), Anticancer Agents: Frontiers in Cancer Chemotherapy. intermittent schedule for maytansine (NSC-153858): brief communication. J. Natl. American Chemical Society, Washington, DC, pp. 317–338. Cancer Inst. 60, 93–96. BurrisIII,H.A.,Rugo,H.S.,Vukelja,S.J.,Vogel,C.L.,Borson,R.A.,Limentani,S.,Tan-Chiu, Erickson, H.K., Park, P.U., Widdison, W.C., Kovtun, Y.V., Garrett, L.M., Hoffman, K., Lutz, R.J., E.,Krop,I.E.,Michaelson,R.A.,Girish,S.,Amler,L.,Zheng,M.,Chu,Y.W.,Klencke, Goldmacher, V.S., Blättler, W.A., 2006. Antibody-maytansinoid conjugates are activated B., O'Shaughnessy, J.A., 2011. Phase II study of the antibody drug conjugate in targeted cancer cells by lysosomal degradation and linker-dependent intracellular trastuzumab-DM1 for the treatment of human epidermal growth factor receptor processing. Cancer Res. 66, 4426–4433. 2 (HER2)-positive breast cancer after prior HER2-directed therapy. J. Clin. Oncol. Erickson, H.K., Widdison, W.C., Mayo, M.F., Whiteman, K., Audette, C., Wilhelm, S.D., 29, 398–405. Singh, R., 2010. Tumor delivery and in vivo processing of disulfide-linked Cabanillas, F., Rodriguez, V., Hall, S.W., Burgess, M.A., Bodey, G.P., Freireich, E.J., 1978. and thioether-linked antibody-maytansinoid conjugates. Bioconjug. Chem. 21, Phase I study of maytansine using a 3-day schedule. Cancer Treat. Rep. 62, 425–428. 84–92. Carrasco-Triguero, M., Yi, J.H., Dere, R., Qiu, Z.J., Lei, C., Li, Y., Mahood, C., Wang, B., Leipold, Girish, S., Gupta, M., Wang, B., Lu, D., Krop, I.E., Vogel, C.L., Burris III, H.A., Yi, J.H., Saad, O., D., Poon, K.A., Kaur, S., 2013. Immunogenicity assays for antibody-drug conjugates: Tong, B., Chu, Y.W., Holden, S., Joshi, A., 2012. Clinical pharmacology of trastuzumab case study with ado-trastuzumab emtansine. Bioanalysis 5, 1007–1023. emtansine (T­DM1): an antibody­drug conjugate in development for the treatment Cassady, J.M., Chan, K.K., Floss, H.G., Leistner, E., 2004. Recent developments in the of HER2­positive cancer. Cancer Chemother. Pharmacol. 69, 1229–1240. maytansinoid antitumor agents. Chem. Pharm. Bull. (Tokyo) 52, 1–26. Goldmacher, V.S., Blättler, W.A., Lambert, J.M., Chari, R.V.J., 2002. Immunotoxins and Chabner, B.A., Levine, A.S., Johnson, B.L., Young, R.C., 1978. Initial clinical trials of antibody-drug conjugates for cancer treatment. In: Muzykantov, V.R., Torchilin, V.P. maytansine, an antitumor plant alkaloid. Cancer Treat. Rep. 62, 429–433. (Eds.), Biomedical Aspects of Drug Targeting. Kluwer Academic Publishers, Boston, Chahinian, A.P., Nogeire, C., Ohnuma, T., Greenberg, M.L., Sivak, M., Jaffrey, I.S., Holland, Massachusetts, pp. 291–310. J.F., 1979. Phase I study of weekly maytansine given by IV bolus or 24-hour infusion. Gupta, M., Wang, B., Carrothers, T., LoRusso, P.M., Chu, Y.-W., Shih, T., Loecke, D., Joshi, A., Cancer Treat. Rep. 63, 1953–1960. Saad, O., Yi, J.-H., Girish, S., 2013. Effects of trastuzumab emtansine (T-DM1) on QT in-

ChanTest, 2013. ChanTest hERG IC50 Concentration Response Assays [Internet]. [cited terval and safety of pertuzumab plus T-DM1 in patients with previously treated 2013 Apr 25]. Available from: http://www.chantest.com/screening-services/safety- human epidermal growth factor receptor 2-positive metastatic breast cancer. Clin. services/pre-clinical-safety/herg-concentration-response-assay. Pharmacol. Drug Dev. 2, 11–24. K.A. Poon et al. / Toxicology and Applied Pharmacology 273 (2013) 298–313 313

Herceptin [package insert]. South San Francisco, CA: Genentech, Inc.; 2010. Accessed Remillard, S., Rebhun, L.I., Howie, G.A., Kupchan, S.M., 1975. Antimitotic activity of the November 7, 2012. potent tumor inhibitor maytansine. Science 189, 1002–1005. Herceptin European Summary of Product Characteristics [Internet]. [cited 2013 Apr 25]. Ross, J.S., Slodkowska, E.A., Symmans, W.F., Pusztai, L., Ravdin, P.M., Hortobagyi, G.N., Available from: http://www.ema.europa.eu/docs/en_GB/document_library/ 2009. The HER-2 receptor and breast cancer: ten years of targeted anti–HER-2 EPAR_-_Product_Information/human/000278/WC500074922.pdf. therapy and personalized medicine. Oncologist 14, 320–368. Hudis, C.A., 2007. Trastuzumab — mechanism of action and use in clinical practice. N. Seidman, A., Hudis, C., Pierri, M.K., Shak, S., Paton, V., Ashby, M., Murphy, M., Stewart, S.J., Engl. J. Med. 357, 39–51. Keefe, D., 2002. Cardiac dysfunction in the trastuzumab clinical trials experience. Issell, B.F., Crooke, S.T., 1978. Maytansine. Cancer Treat. Rev. 5, 199–207. J. Clin. Oncol. 20, 1215–1221. Kadcyla [package insert]. South San Francisco, CA: Genentech, Inc.; 2013. Sieber, S.M., Whang-Peng, J., Botkin, C., Knutsen, T., 1978. Teratogenic and cytogenetic ef- Klein, P.M., Dybdal, N., 2003. Trastuzumab and cardiac dysfunction: update on preclinical fects of some plant-derived antitumor agents (vincristine, colchicine, maytansine, studies. Semin. Oncol. 30 (5 Suppl. 16), 49–53. VP-16-213 and VM-26) in mice. Teratology 18, 31–47. Krop, I.E., Beeram, M., Modi, S., Jones, S.F., Holden, S.N., Yu, W., Girish, S., Tibbitts, J., Yi, J.H., Slamon, D.J., Clark, G.M., Wong, S.G., Levin, W.J., Ullrich, A., McGuire, W.L., 1987. Sliwkowski, M.X., Jacobson, F., Lutzker, S.G., Burris, H.A., 2010. Phase I study of Human breast cancer: correlation of relapse and survival with amplification of trastuzumab-DM1, an HER2 antibody-drug conjugate, given every 3 weeks to pa- the HER-2/neu oncogene. Science 235, 177–182. tients with HER2-positive metastatic breast cancer. J. Clin. Oncol. 28, 2698–2704. Slamon, D.J., Godolphin, W., Jones, L.A., Holt, J.A., Wong, S.G., Keith, D.E., Levin, W.J., Stuart, Krop, I.E., LoRusso, P., Miller, K.D., Modi, S., Yardley, D., Rodriguez, G., Guardino, E., Lu, M., S.G., Udove, J., Ullrich, A., 1989. Studies of the HER-2/neu proto-oncogene in human Zheng, M., Girish, S., Amler, L., Winer, E., Rugo, H.S., 2012. A phase II study of breast and ovarian cancer. Science 244, 707–712. trastuzumab emtansine in patients with human epidermal growth factor receptor Slamon, D.J., Leyland-Jones, B., Shak, S., Fuchs, H., Paton, V., Bajamonde, A., Fleming, T., 2-positive metastatic breast cancer who were previously treated with trastuzumab, Eiermann, W., Wolter, J., Pegram, M., Baselga, J., Norton, L., 2001. Use of chemothera- lapatinib, an anthracycline, a taxane, and capecitabine. J. Clin. Oncol. 30, 3234–3241. py plus a monoclonal antibody against HER2 for metastatic breast cancer that Lambert, J.M., 2005. Drug-conjugated monoclonal antibodies for the treatment of cancer. overexpresses HER2. N. Engl. J. Med. 344, 783–792. Curr. Opin. Pharmacol. 5, 543–549. Sliwkowski, M.X., Lofgren, J.A., Lewis, G.D., Hotaling, T.E., Fendly, B.M., Fox, J.A., 1999. Lewis Phillips, G.D., Li, G., Dugger, D.L., Crocker, L.M., Parsons, K.L., Mai, E., Blättler, W.A., Nonclinical studies addressing the mechanism of action of trastuzumab (Herceptin). Lambert, J.M., Chari, R.V., Lutz, R.J., Wong, W.L., Jacobson, F.S., Koeppen, H., Schwall, Semin. Oncol. 26, 60–70. R.H., Kenkare-Mitra, S.R., Spencer, S.D., Sliwkowski, M.X., 2008. Targeting HER2- Spector, N.L., Blackwell, K.L., 2009. Understanding the mechanisms behind trastuzumab positive breast cancer with trastuzumab-DM1, an antibody-cytotoxic drug conjugate. therapy for human epidermal growth factor receptor 2-positive breast cancer. Cancer Res. 68, 9280–9290. J. Clin. Oncol. 27, 5838–5847. Lopus, M., Oroudjev, E., Wilson, L., Wilhelm, S., Widdison, W., Chari, R., Jordan, M.A., 2010. Spector, N.L., Xia, W., Burris III, H., Hurwitz, H., Dees, E.C., Dowlati, A., O'Neil, B., Maytansine and cellular metabolites of antibody-maytansinoid conjugates strongly Overmoyer, B., Marcom, P.K., Blackwell, K.L., Smith, D.A., Koch, K.M., Stead, A., suppress microtubule dynamics by binding to microtubules. Mol. Cancer Ther. 9, Mangum, S., Ellis, M.J., Liu, L., Man, A.K., Bremer, T.M., Harris, J., Bacus, S., 2005. 2689–2699. Study of the biologic effects of lapatinib, a reversible inhibitor of ErbB1 and ErbB2 Mahapatra, K., Darbonne, W., Bumbaca, D., Shen, B., Du, X., Tibbitts, J., Olsen, S., tyrosine kinases, on tumor growth and survival pathways in patients with advanced Sliwkowski, M., Girish, S., Hartley, D., Dambach, D., Ramakrishnan, V., Uppal, H., malignancies. J. Clin. Oncol. 23, 2502–2512. 2011. T-DM1-induced thrombocytopenia results from impaired platelet production Swanson, S.J., Bussiere, J., 2012. Immunogenicity assessment in non-clinical species. Curr. in a HER2-independent manner [abstract]. Mol. Cancer Ther. 10 (11 Suppl.), A135. Opin. Microbiol. 15, 337–347. Mugera, G.M., Ward, J.M., 1977. Acute toxicity of maytansine in F344 rats. Cancer Treat. Tolcher, A.W., Ochoa, L., Hammond, L.A., Patnaik, A., Edwards, T., Takimoto, C., Smith, L., de Rep. 61, 1333–1338. Bono, J., Schwartz, G., Mays, T., Jonak, Z.L., Johnson, R., DeWitte, M., Martino, H., Nahta, R., Yu, D., Hung, M.C., Hortobagyi, G.N., Esteva, F.J., 2006. Mechanisms of disease: Audette, C., Maes, K., Chari, R.V., Lambert, J.M., Rowinsky, E.K., 2003. Cantuzumab understanding resistance to HER2-targeted therapy in human breast cancer. Nat. mertansine, a maytansinoid immunoconjugate directed to the CanAg antigen: a Clin. Pract. Oncol. 3, 269–280. phase I, pharmacokinetic, and biologic correlative study. J. Clin. Oncol. 21, 211–222. Oroudjev,E.,Lopus,M.,Wilson,L.,Audette,C.,Provenzano,C.,Erickson,H.,Kovtun,Y., Verma, S., Miles, D., Gianni, L., Krop, I.E., Welslau, M., Baselga, J., Pegram, M., Oh, D.Y., Chari, R., Jordan, M.A., 2010. Maytansinoid-antibody conjugates induce mitotic Diéras, V., Guardino, E., Fang, L., Lu, M.W., Olsen, S., Blackwell, K., the EMILIA Study arrest by suppressing microtubule dynamic instability. Mol. Cancer Ther. 9, Group, 2012. Trastuzumab emtansine for HER2-positive advanced breast cancer. N. 2700–2713. Engl. J. Med. 367, 1783–1791. Park, S.B., Krishnan, A.V., Lin, C.S., Goldstein, D., Friedlander, M., Kiernan, M.C., 2008. von Minckwitz, G., du Bois, A., Schmidt, M., Maass, N., Cufer, T., de Jongh, F.E., Maartense, Mechanisms underlying chemotherapy-induced neurotoxicity and the potential for E., Zielinski, C., Kaufmann, M., Bauer, W., Baumann, K.H., Clemens, M.R., Duerr, R., neuroprotective strategies. Curr. Med. Chem. 15, 3081–3094. Uleer, C., Andersson, M., Stein, R.C., Nekljudova, V., Loibl, S., 2009. Trastuzumab be- Press, M.F., Cordon-Cardo, C., Slamon, D.J., 1990. Expression of the HER-2/neu proto- yond progression in human epidermal growth factor receptor 2-positive advanced oncogene in normal human adult and fetal tissues. Oncogene 5, 953–962. breast cancer: a German breast group 26/breast international group 03-05 study. Raschi, E., Ceccarini, L., De Ponti, F., Recanatini, M., 2009. hERG-related drug toxicity and J. Clin. Oncol. 27, 1999–2006. models for predicting hERG liability and QT prolongation. Expert Opin. Drug Metab. von Minckwitz, G., Schwedler, K., Schmidt, M., Barinoff, J., Mundhenke, C., Cufer, T., Toxicol. 5, 1005–1021. Maartense, E., de Jongh, F.E., Baumann, K.H., Bischoff, J., Harbeck, N., Lück, H.J., Rebert, C.S., Pryor, G.T., Frick, M.S., 1984. Effects of vincristine, maytansine, and cis- Maass, N., Zielinski, C., Andersson, M., Stein, R.C., Nekljudova, V., Loibl, S., GBG 26/ platinum on behavioral and electrophysiological indices of neurotoxicity in the rat. BIG 03-05 study group and participating investigators, 2011. Trastuzumab beyond J. Appl. Toxicol. 4, 330–338. progression: overall survival analysis of the GBG 26/BIG 3-05 phase III study in Redfern, W.S., Carlsson, L., Davis, A.S., Lynch, W.G., MacKenzie, I., Palethrope, S., Siegl, P.K., HER2-positive breast cancer. Eur. J. Cancer 47, 2273–2281. Strang, I., Sullivan, A.T., Wallis, R., Camm, A.J., Hammond, T.G., 2003. Relationships Wu, A.M., Senter, P.D., 2005. Arming antibodies: prospects and challenges for between preclinical cardiac electrophysiology, clinical QT interval prolongation and immunoconjugates. Nat. Biotechnol. 23, 1137–1146. torsade de pointes for a broad range of drugs: evidence for a provisional safety Yarden, Y., Sliwkowski, M.X., 2001. Untangling the ErbB signalling network. Nat. Rev. Mol. margin in drug development. Cardiovasc. Res. 58, 32–45. Cell Biol. 2, 127–137.