The Vedotin Antibody–Drug Conjugate Payload Drives Platform-Based Nonclinical Safety and Pharmacokinetic Profiles

Abstract

Nonclinical safety and pharmacokinetic data for monomethyl auristatin E (MMAE) and 14 vedotin antibody–drug conjugates (ADC) were evaluated to determine patterns of toxicity, consistency of pharmacokinetic results, and species differences between rats and monkeys. Most nonclinical toxicities were antigen-independent, common across ADCs, and included hematologic, lymphoid, and reproductive toxicity related to MMAE pharmacology. Hematologic toxicity was the dose-limiting toxicity (DLT) or predominant toxicity for the majority of vedotin ADCs in both species. Tissue expression of the targeted antigen of an ADC rarely correlated with DLT; only two ADCs had antigen-dependent skin DLTs. For two additional ADCs, antigen-dependent delivery of MMAE in the bone marrow may have exacerbated the antigen-independent hematologic DLT. The highest tolerated doses and pharmacokinetics were similar within a given species, with rats tolerating higher doses than monkeys. Studies longer than 1 month in duration detected the same or fewer toxicities than 1-month studies and had no additional findings that affected the human risk assessment. These data support opportunities to streamline ADC toxicity assessments without compromising human starting dose selection or target organ identification.

Introduction

Antibody–drug conjugates (ADC) are designed to reduce systemic toxicity by delivering potent payloads via a tumor target-specific antibody. Vedotin ADCs consist of a mAb bound to monomethyl auristatin E (MMAE) by a protease-cleavable valine–citrulline (vc) linker (1–7). The vc-MMAE drug linker is conjugated to native mAb cysteines, resulting in a heterogeneous average drug-to-antibody ratio of 4 (8). MMAE causes cell death through microtubule disruption in rapidly dividing tissue (4, 5, 9). Vedotin technology developed by Seagen, Inc. is utilized in 4 of the 13 FDA-approved oncology ADCs for both hematologic and solid tumor malignancies.

The clinical pharmacokinetic (PK) and safety profiles for vedotin ADCs across patient populations are generally consistent. A recent analysis of eight vedotin ADCs demonstrated that the PK properties of antibody-conjugated MMAE, total antibody (TAb), and unconjugated MMAE were highly similar at 2.4 mg/kg when given to patients once every 3 weeks (Q3W; ref. 10). Furthermore, an FDA-based review of vedotin ADCs showed that the human maximum tolerated doses (MTD) were consistently between 1.8 and 2.4 mg/kg (Q3W), indicating toxicities are primarily driven by the MMAE payload rather than antibody effector functions (11). Vedotin ADC clinical safety profiles often share overlapping toxicities including bone marrow toxicity or neuropathy, which are generally considered antigen independent (12, 13).

A systematic analysis of the nonclinical toxicity and PK profiles of vedotin ADCs has not yet been reported. In this publication, we present data generated in Good Laboratory Practice (GLP)–compliant studies including 29 repeat-dose toxicology studies with MMAE and 14 vedotin ADCs, along with embryofetal development (EFD) toxicity and genotoxicity studies. These data support efforts to streamline development of novel vedotin ADCs without compromising patient safety. Furthermore, this platform approach sets a foundation of principles that should be considered for ADCs with linker–payloads other than vedotin.

Materials and Methods

Data sources

Data are primarily sourced from reports from GLP-compliant studies sponsored by Seagen/collaborators over an 18-year period (2005–2023). Data from one non-GLP study was included to provide a comparison for weekly dosing of brentuximab vedotin (BV; GLP studies were dosed Q3W). Information for disitamab vedotin (DV; ref. 14), polatuzumab vedotin (PV), and PV_surr (15) was accessed from peer-reviewed publications. For FDA-approved drugs, the summary basis of approvals available at drugs@FDA was also reviewed (16–19).

Definition of studies

Some study reports included multiple test articles and/or dosing regimens. Therefore, a study herein is defined as one set of animals dosed with one specific test article. Using this definition, 29 repeat-dose general toxicity studies (seven rats and 22 monkeys) were included in this analysis, along with additional genotoxicity and EFD studies.

Ethical considerations

All animal experiments were conducted in facilities accredited by the Association for Assessment and Accreditation of Laboratory Animal Care under Institutional Animal Care and Use Committee guidelines and appropriate animal research approvals. Studies were conducted in accordance with the Department of Agriculture Animal Welfare Act and the National Research Council’s Guide for the Care and Use of Laboratory Animals.

Test articles

Test articles in this report included the small molecule MMAE and 14 vedotin ADCs: one nontargeted control and 13 antigen-targeting (therapeutic) ADCs (Table 1). Most ADCs utilized a humanized IgG1 mAb backbone except BV (mouse–human chimeric), and enfortumab vedotin (EV) and tisotumab vedotin (TV; fully human). SGN-PDL1V has an Fc mutation (LALA) that decreases Fcγ receptor (FcγR) binding. All the therapeutic ADCs except PV bind to the corresponding monkey homolog of the human antigen (15, 20). For PV, a cynomolgus monkey binding surrogate ADC (PV_surr; ref. 20) was used to evaluate antigen-dependent toxicities (15). EV and SGN-B7H4V also bind to the rat homologs of the human antigens. h00-1006(4) is a control vedotin ADC with intact Fc-binding capability but with a Fab engineered to eliminate antigen binding. Lack of binding to human antigens via the CDR for h00 was confirmed with a Retrogenix Cell Microarray Technology assay [Charles River Laboratories, Research Resource Identifier (RRID):SCR_003792; ref. 21].

Table 1.

Dose-limiting or predominant toxicities for MMAE and vedotin ADCs.

Test article	Target antigen	Dose-limiting or predominant toxicity
		Antigen independent (hematotoxicity)	Antigen dependent (hematotoxicity or skin toxicity)
MMAE	Small molecule, not applicable	X	—
h00-1006(4)a	Null (nontargeting)	X	—
BV	TNFRSF8 (CD30)	X	Xb
DV	HER-2	X	—
EV	Nectin-4	—	X
LV	SLC39A6 (LIV-1)	X	—
LIFA	NaPi2b	X	—
PV	CD79b, clinical antibody does not bind to monkey antigen	Xc	—
PVsurr	CD79b, nonclinical antibody binds to monkey antigen	X	Xb
SGN-ALPV	ALPP and ALPG (ALPPL2)	X	—
SGN-B6A	ITGB6	X	—
SGN-B7H4V	VTCN1 (B7-H4)	Xc	—
SGN-PDL1V	CD274 (PD-L1)	X	—
SGN-STNV	STn	X	—
TV	F3 (TF)	—	X

Abbreviations: ALPG, alkaline phosphatase, germ cell; ALPP, alkaline phosphatase placental; ALPPL2, alkaline phosphatase, placental like 2; CD, cluster of differentiation; F3, coagulation factor III; ITGB6, integrin subunit beta 6; LV, ladiratuzumab vedotin; NaPi2b, sodium-dependent phosphate transport protein 2B; SLC39A6, solute carrier family 39 member 6; STn, sialyl-Thomsen nouveau; TF, tissue factor; TNFRSF8, tumor necrosis factor receptor superfamily member 8; TV, tisotumab vedotin; VTCN1, V-set domain containing T cell activation inhibitor 1.

^ah00-1006(4) is a vedotin ADC with intact Fc-binding capability engineered to have no target antigen, confirmed with a Retrogenix Cell Microarray Technology assay (Charles River Laboratories) that detected no binding of h00 to human antigens via its CDR (21).

^bTarget expression and subsequent antigen-dependent delivery of MMAE in the hematopoietic system may have added to the antigen-independent effects.

^cLiver toxicity was also dose limiting for rats.

Genotoxicity evaluations (MMAE only)

Two in vitro and two in vivo assays characterized the genotoxic potential of MMAE.

In vitro genotoxicity

An in vitro bacterial reverse mutation assay was performed with MMAE at 75, 200, 600, 1,800, or 5,000 µg per plate using Salmonella typhimurium tester strains TA98, TA100, TA1535, TA1537, and Escherichia coli tester strain WP2 uvrA in the presence and absence of Aroclor-induced rat liver S9 enzymes.

An in vitro assay was conducted to evaluate the ability of MMAE to induce forward mutations at the thymidine kinase locus in the mouse lymphoma L5178Y thymidine kinase^+/− (TK^+/−) cell line. The assay was conducted in the presence and absence of an exogenous metabolic activation system (S9), and mutagenic potential was assayed by colony growth in the presence of 5-trifluorothymidine (resistance to the drug). MMAE was tested at concentrations of 0.0500 to 70 ng/mL with S9 (4-hour treatment) and 0.00100 to 6.00 ng/mL without S9 (24-hour treatment).

Rat bone marrow micronucleus assays

MMAE was evaluated for in vivo clastogenic activity and/or disruption of the mitotic apparatus by detecting micronuclei in polychromatic erythrocytes (PCE) in male (8-week-old) CD (SD) rat bone marrow. For the first assay, rats were administered either vehicle control (i.v. injection), positive control (60 mg/kg cyclophosphamide, oral gavage), or MMAE (0.01, 0.1, or 0.2 mg/kg, i.v. injection). The micronucleus frequency following exposure for 24 or 48 hours was determined by analyzing the number of micronucleated PCEs from at least 2,000 PCEs per animal. The PCE:NCE ratio was determined by scoring the number of PCEs and normochromatic erythrocytes (NCE) observed while scoring at least 500 erythrocytes per animal.

A confirmatory assay with antikinetochore analysis was performed with MMAE (0.1, or 0.2 mg/kg, i.v. injection), vehicle, 60 mg/kg cyclophosphamide, and an additional positive control, carbendazim (1,250 or 1,500 mg/kg, oral gavage, once daily for 2 days prior to bone marrow collection). Bone marrow was harvested at 24 or 48 hours (carbendazim only) post treatment. Bone marrow from tibias was harvested for centromere analysis using immunofluorescence; micronuclei in PCE and NCE were classified as centromere-positive if one or more fluorescent signals were present. The test article was considered aneugenic if it induced predominantly centromere-positive micronuclei (similar to carbendazim) or clastogenic if it induced predominantly centromere-negative micronuclei (similar to cyclophosphamide).

EFD studies

Effects of MMAE, BV, or EV on development were evaluated using time-mated female rats dosed intravenously on gestational day 6 and 13 at doses of 0.2 mg/kg MMAE (molar equivalent to 10 mg/kg/dose BV), 0.3, 1, 3, or 10 mg/kg BV, and 2 or 5 mg/kg EV. Study endpoints included maternal mortality, clinical signs, body weight and weight change, food consumption, hematology and anatomic pathology, and cesarean section evaluation, embryofetal evaluation (viability, growth, and development), and fetal evaluations (external, soft tissue, and skeletal). For BV and MMAE, both maternal and fetal blood samples, as well as amniotic fluid, were collected for toxicokinetic and immunogenicity evaluation. EV was evaluated in a preliminary EFD study consistent with ICH S9 Q&A, S5, and M3 guidance with reduced group sizes, no fetal blood or amniotic fluid collections, and limited maternal blood samples analyzed to verify EV exposure.

General toxicity study procedures and endpoints

Species selection and relevance

All in vivo studies were conducted in purpose-bred, naïve Sprague–Dawley rats or cynomolgus monkeys (Macaca fascicularis). Studies in laboratory animals provide relevant data for extrapolation to humans and are required to support regulatory submissions. Cynomolgus monkeys are closely related phylogenetically and physiologically to humans and were the only pharmacologically relevant species for antibody binding to the intended target for most of the vedotin ADCs. EV and SGN-B7H4V cross-react with the rat orthologs of human nectin-4 and B7-H4, respectively, with comparable affinity and tissue distribution; therefore, the Sprague–Dawley rat was an additional appropriate species for nonclinical toxicity evaluations of these ADCs. Rat studies with other ADCs (BV and PV) were performed prior to the updated ICH S9(Q&A) guidance (2018) and/or were used to assess antigen-independent toxicities only. Cynomolgus monkey studies were performed for all vedotin ADCs. All targeted vedotin ADCs demonstrated specific binding to the monkey homolog of the target antigen, with the exception of PV. A monkey study with PV_surr was conducted to further characterize antigen‐dependent and ‐independent pharmacology, pharmacokinetics, and toxicity to support development of the ADC (22).

Test article administration and endpoints

Test articles were administered by IV bolus injection or infusion via a peripheral vein to Sprague–Dawley rats (RRID: RGD_70508) and/or cynomolgus monkeys using clinically relevant dosing regimens. In general, studies were only performed in the pharmacologically relevant species. Rat studies were conducted for BV and PV even though these ADCs do not bind to the rat homolog; this was due to the novel drug format and unknown regulatory expectations at the time the studies were performed [subsequently clarified in ICH S9 (2009) and S9 Q&A (2018) guidances]. The studies varied in experimental design (Tables 2–4), but endpoints were generally consistent and included standard in-life, clinical pathology, toxicokinetic, safety pharmacology, and histopathology assessments (Supplementary Table S1).

Table 2.

Key toxicities associated with MMAE and h00-1006(4).

Species	Rat	Monkey
Test article	MMAE	MMAE			h00-1006(4)
Dosing schedule	Q1Wx4	Q1Wx4	Q3Wx4	Q2Wx7	Q1Wx4
Recovery duration	4 weeks	6 weeks	5 weeks	6 weeks	4 weeks
Dose level(s) in mg/kg	0.0097, 0.097, 0.194	0.1093/0.0545a	0.058	0.12	3, 4, 5
STD10	NR	NA	NA	NA	NA
HNSTD	NA	NE	NE	NR	5
Main study animals (n/sex/group)	10	3	3	3	3
Recovery study animals (n/sex/group)	5b	2	2	2	2
Antidrug antibodies	NA	NA	NA	NA	—
Early deaths due to test article	—	✓	—	—	—
Hematotoxicity	✓	✓	✓	✓	✓
Skin toxicity	—	✓c	—	—	—
Lymphoid toxicity	✓	✓d	✓	✓	✓
Reproductive toxicity
Male	✓	—	—	—	—
Female	—	—	—	—	✓
Liver and/or biliary toxicity	✓	—	—	—	—
Acute phase protein changes	—	✓	—	—	✓
Lung toxicity	—	—	—	—	—

The vedotin ADC payload MMAE resulted in rat and/or monkey hematologic, lymphoid, reproductive, liver, and skin toxicity and acute phase protein changes. Predominant toxicities of the nontargeted vedotin ADC were similar to those observed in monkeys administered MMAE and included hematopoietic, lymphoid, and female reproductive toxicity and acute phase protein changes.

Abbreviations: NA, not applicable; NE, not established; NR, not reported; STD₁₀, severely toxic dose in 10% of the animals. Dash (—) indicates finding not present. Gray box, not applicable because of species difference in risk assessment.

^aThe first two doses of MMAE on days 1 and 8 were administered at 0.1093 mg/kg; the dose level was reduced to 0.0545 mg/kg beginning with the day 15 dose (∼ 5.4 or 2.7 mg/kg ADC, respectively).

^bVehicle, mid, and high dose only.

^cSkin findings (skin missing and injection site changes) were present in one animal removed from study early on day 19 because wound could not be repaired.

^dLymphoid changes with MMAE included increased spleen weights and spleen mixed cell infiltrates, in contrast to the unchanged or decreased cellularity in lymphoid organs with most other test articles.

Table 4.

Key toxicities associated with targeted vedotin ADCs in monkeys.

Test article:		BV		DV	EV		LV	LIFA	PV	PV Surr	SGN-ALPV	SGN-B6A			SGN-B7H4V	SGN-PDL1V	SGN-STNV		TV
Antibody target		CD30		HER2	Nectin-4		LIV1	NaPi2b	CD79b	CynoCD79b1	ALPP/ALPPL2	Integrin β-6			B7-H4	PDL1	STn		Tissue factor
Dosing schedule		Q3Wx4	Q3Wx9	Q2Wx7	Q1Wx4		Q3Wx5	Q3Wx5	Q3Wx4	Q3Wx4	Q1Wx4	Q1Wx4	Q3Wx2	2Q3Wx5	Q1Wx4	Q1Wx4	Q1Wx4	2Q3Wx5	Q3Wx5
Recovery duration		5 weeks	6 weeks	6 weeks	6 weeks		NA	6 weeks	9 weeks	9 weeks	NA	NA	NA	NA	NA	6 weeks	NA	NA	6 weeks
Dose level(s) in mg/kg		1, 3, 6	1, 3	2.3, 4.6, 9.12	1, 3, 6	3	1, 3, 5	1, 3, 6	1, 3, 5	3, 5	3, 4, 5	3, 4, 5	6	3	3, 4, 5	3, 4, 5	3, 4, 5	3	13, 3, 5
HNSTD		3	3	5	3	3	34	3	3*	NR	5	5	6	3	5	5	4	3	3
Main study animals (n/sex/group)		3–4	3–65	36	3	2–37	4	3	3	3	3	3	3	3	3	3	3	3	38
Recovery study animals (n/sex/group)		2–3	2–49	210	2	211	2	2	0	2	0	0	0	0	1	2	0	0	2
Antidrug antibodies		✓	✓	✓	✓	✓	✓	✓	✓	✓	—	—	—	✓	—	✓	—	✓	✓
Early deaths due to test article		✓	—	✓	✓	—	✓	✓	—	✓	—	—	—	—	—	—	✓	—	✓
Hematotoxicity		✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓
Skin toxicity		—	—	—	✓	✓	—	—	—	—	U	—	—	—	—	—	—	—	✓
Lymphoid toxicity		U	—	✓	✓12	—	—	✓	—	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓12
Reproductive toxicity
Male	—	—	—	—	—	✓	—	—	—	—	—	—	—	—	—	—	—	—	✓
Female	—	—	—	—	—	—	—	—	—	✓	✓	✓	✓	✓	✓	✓	✓	—
Liver/biliary toxicity		—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
Acute phase protein changes		—	—	—	✓	✓	—	—	—	—	✓	✓	✓	—	✓	✓	—	—	—
Lung toxicity		—	—	—	—	—	—	U	—	—	—	—	✓	—	—	—	—	—	—

Antigen-independent toxicities (hematopoietic, lymphoid, reproductive, and embryofetal toxicity) were common across vedotin ADCs at high doses. For antigen-dependent toxicities, BV had more pronounced hematologic toxicity and PV_surr had more lymphoid toxicity. EV and TV had skin toxicity, which was dose-limiting.

✓, Finding was observed; — , finding not observed; U, uncertain relationship to test article; NR, not reported; NA, not applicable.

Antibody target: ¹All the targeted ADCs bind to the corresponding monkey homolog of the human antigens, except for PV, so a cynomolgus monkey–binding surrogate antibody (PV_surr) was used to evaluate target-dependent toxicities for PV.

Dose level(s): ²DV: Dose levels differ from those published due to differences in calculation methods between DV and other ADCs. Thus, DV doses were reported at the level consistent with those used with other vedotin ADCs. ³TV: Based on persistently low dosing solution concentrations on each dosing occasion for the nominal 1 mg/kg dose, the actual dose level was between 0.79 and 0.91 mg/kg.

LV HNSTD: ⁴NOAEL and HNSTD in females was 5 mg/kg; HNSTD in males was 3 mg/kg. *5 mg/kg PV was tolerated, however the HNSTD for PV was determined based on the tolerability of the PV surrogate ADC (22).

Details for studies with different numbers across groups: ⁵n_vehicle = 3/sex/group; n_BV = 6/sex/group. ⁷n_vehicle = 2/sex/group; n_EV = 3/sex/group. ⁹n_vehicle = 2/sex/group; n_BV = 4/sex/group. ¹⁰n = 2/sex/group for vehicle, mid, and high dose only. ¹¹n_vehicle = 0/sex/group; n_EV = 2/sex/group.

Necropsy timepoints: In the majority of studies, necropsy was performed 1 week following last dose. ⁶Necropsy of DV main study animals was performed the day after the last dose. ⁸Necropsy of TV main study animals was performed 12 days after the last dose.

Lymphoid changes: ¹²Lymphoid changes with EV and TV included increased cellularity (EV) or increased mononuclear cell infiltrates (TV) of lymph node/GALT, in contrast to the unchanged or decreased cellularity in lymphoid organs with most other test articles.

Toxicology assessments and definition of antigen dependency

Findings deemed test article-related by the study scientists are the focus of this report. In general, minor test article-related findings that were attributed solely to stress (e.g., adrenal changes), considered secondary to clinical condition, or that are commonly observed nonspecific findings were excluded from these analyses (Supplementary Table S2). Histologic diagnoses of increases in mitotic figures, abnormal mitoses, and/or apoptosis or single-cell necrosis are associated with MMAE but were recorded inconsistently and were therefore not included in the results. Changes secondary to bone marrow effects (e.g., presence of bacterial colonies) were categorized as hematotoxicity and not as organ-specific toxicities. Histologic diagnoses describing equivalent pathologic processes were consolidated (Supplementary Table S3).

Antigen-independent findings were defined as test article-related changes that occurred with MMAE or the nonbinding vedotin ADC, h00-1006(4). Antigen-dependent findings were defined as test article-related changes in tissues expressing the target antigen (Supplementary Table S4) and/or changes that were more pronounced with targeted ADCs than with MMAE and/or h00-1006(4). Findings of uncertain relationship either did not occur with MMAE and/or h00-1006(4) or occurred with targeted ADCs in tissues that did not express the target. Target expression was based on a proprietary mRNA database with 25 cynomolgus monkey tissues (three males and three females), literature reviews, public databases, and IHC experiments.

PK analyses

TAb and unconjugated MMAE components were quantitated in a consistent manner across all vedotin ADC studies and thus were used for the retrospective PK exposure comparisons. As dose levels and schedules varied across studies, the maximum observed concentration (C_max) and area under the concentration–time curve from 0 to 7 days (AUC_{0–7 days}) after the first dose of 3 mg/kg were selected as the most representative parameters for comparison. To assess the impact of immunogenicity on systemic exposure for a subset of vedotin ADCs, the AUC_{0–21 days} for MMAE and ADC after the first and last 3 mg/kg IV doses were used. TAb (including all forms of conjugated and unconjugated antibody) was quantitated by ELISA or a comparable fluorometric immunoassay on the Gyrolab platform (Gyros Protein Technologies, Uppsala, Sweden). ADC (measuring drug-conjugated antibody) was quantitated by ELISA. Unconjugated MMAE was quantitated by LC-MS/MS. Immunogenicity was evaluated when applicable using ELISA or electrochemiluminescence-based bioanalytical methods. PK parameters, including the C_max and AUC_{0–7 days} or AUC_{0–21 days}, were estimated by noncompartmental analysis using Phoenix WinNonlin version 8.2 (Certara USA, Inc., Princeton, NJ).

Data availability

The data included in this study are not publicly available due to the proprietary nature of the data for products currently in clinical trials and partnership requirements but are available upon reasonable request of the corresponding author.

Results

General information

Data are presented from 15 test articles (MMAE and 14 vedotin ADCs) administered to monkeys (Table 1), five of which were also administered to rats (MMAE, BV, EV, PV, and SGN-B7H4V). In general, findings were typical of toxicity studies with other biopharmaceutical agents in that the severity and incidence of findings increased with higher doses.

Genotoxicity of MMAE

MMAE caused centromere-positive micronuclei in vivo, indicating aneugenic (chromosome lagging) micronuclear formation (16), consistent with the microtubule disrupting mechanism of MMAE. No evidence of mutagenicity was observed with MMAE in either a bacterial reverse mutation assay with or without metabolic activation up to 5,000 μg/plate or in a L5178Y TK^+/− mouse lymphoma forward mutation assay. Therefore, MMAE and vedotin ADCs are considered aneugenic.

EFD toxicity of MMAE

Embryofetal toxicity and the teratogenic potential of MMAE were evaluated in rat (16, 17). Intravenous administration of MMAE at 0.2 mg/kg on gestational days 6 and 13 resulted in clinical observations in the dams of red/black vaginal discharge associated with fetal loss, decreased body weights and food consumption, and changes in hematology parameters analogous to findings in the general toxicology studies. Additional findings included significant increases in total resorptions, postimplantation loss, early delivery, and loss of viable fetuses. There were no MMAE-related changes in mean corpora lutea, mean implantation sites, or preimplantation loss. MMAE was measurable in fetal serum and resulted in fetal external malformations but not fetal soft tissue or skeletal malformations.

General toxicity of MMAE

Repeat-dose general toxicity studies with intravenous administration of MMAE were conducted in Sprague–Dawley rats (one study) and cynomolgus monkeys (three studies) as summarized in Table 2.

The toxicity of MMAE in rats was evaluated following once weekly administration (Q1Wx4) at doses up to 0.2 mg/kg. Predominant findings included hematopoietic, lymphoid, reproductive, and hepatic/biliary toxicity, and acute phase protein changes. Hematotoxicity was characterized by decreased red blood cell mass, reticulocytes, and leukocytes, and increased platelets, with decreased cellularity of bone marrow. Lymphoid toxicity was characterized by decreased lymphocytes in the thymus, sometimes with necrosis. Reproductive toxicity included degeneration/atrophy of the testes with secondary decreased sperm in the epididymis. Hepatic/biliary toxicity was characterized by increased clinical chemistry liver parameters [total bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and gamma-glutamyltransferase (GGT)] and increased incidence of hepatic coagulative necrosis at a high dose. Acute-phase protein changes were limited to decreased albumin. Findings reversed within a 4-week recovery period except for the male reproductive tract effects and partial recovery of red blood cell count.

The toxicity of MMAE in monkeys was evaluated with three dosing regimens (every 3 weeks (Q3W)x4, Q1Wx4, and every 2 weeks (Q2W)x7) generally at the molar equivalent of the highest dose level used in the corresponding ADC arm of the study (0.058–0.12 mg/kg). Predominant MMAE-related findings in monkeys were generally similar across studies and to those observed in rats. Differences in monkeys compared with rats included no histologic or clinical chemistry evidence of liver/biliary toxicity (clinical chemistry changes of increased AST, but decreased ALP (alkaline phosphatase) and ALT were not considered evidence of liver/biliary toxicity), additional acute phase protein (decreased albumin to globulin ratio and increased fibrinogen and globulin) and bone marrow (decreased myeloid to erythroid ratio) changes, and changes in spleen (either increased weights and mixed cell infiltrates in red pulp or decreased cellularity) and lymph node (no findings or increased macrophage infiltrate). Following a 5- to 6-week recovery period, MMAE-related findings partially (eosinophils) or completely recovered.

Early death with MMAE

One monkey at the weekly dose regimen of MMAE was euthanized due to a small area of missing skin that could not be surgically repaired; the reason for the missing skin was not given (Supplementary Table S5). There were no histologic changes in the routinely collected section of skin in this animal and skin changes were not observed in the other 29 monkeys administered MMAE.

Toxicity of the nontargeted vedotin ADC, h00-1006(4)

To assess the antigen-independent toxicity of vedotin ADCs, h00-1006(4), a nontargeted vedotin ADC, was administered to monkeys at doses up to 5 mg/kg Q1Wx4 with a 4-week recovery period (Table 2).

In monkeys, the predominant toxicities of h00-1006(4) were similar to MMAE and included hematopoietic, lymphoid, and reproductive toxicity, and acute phase protein changes. Changes in peripheral blood and bone marrow and acute phase proteins were the same as those in monkeys administered MMAE, except platelets were not increased. Lymphoid toxicity was limited to decreased cellularity in thymus, spleen, and lymph node/GALT without other histologic changes. Reproductive toxicity was observed only in females and was limited to decreases in secondary and tertiary follicles of ovaries, with sparing of primordial follicles. No skin changes were observed with h00-1006(4). Most test article-related changes partially (ovary and spleen) or fully recovered during the 4-week recovery period, except for thymus lymphoid cellularity.

Toxicities of targeted vedotin ADCs

There were six rat studies and 18 monkey studies conducted with 13 targeted ADCs. Highest non-severely toxic dose (HNSTD) values ranged from 5 to 10 mg/kg in rats and 3 to 6 mg/kg in monkeys (Fig. 1; Tables 2–4). In monkeys, the no observed adverse effect level was only reported for BV at 1 mg/kg (Q3Wx4) and 3 mg/kg (Q3Wx9), and for TV at 1 mg/kg (Q3Wx5). High doses in both species were limited primarily by hematotoxicity (11 ADCs) or skin toxicity (EV and TV; Table 1).

Figure 1.

Tolerated doses in cynomolgus monkeys dosed with vedotin ADCs. A, Schematic of a vedotin ADC. Vedotin ADCs consist of a mAb bound to MMAE by a protease-cleavable vc linker. B, Schematic demonstrating the two primary classifications of ADC toxicity used in this analysis: antigen dependent (changes in tissues expressing the target antigen) and antigen independent (changes that occurred with MMAE or the nonbinding vedotin ADC). C, Doses of vedotin ADCs evaluated in studies with either a Q1Wx4 or Q3Wx2 (SGN-B6A only) repeat dosing schedule. D, Doses of vedotin ADCs evaluated in ≥3-month duration studies. 2Q3Wx5: SGN-B6A and SGN-STNV. Q2Wx7: DV only. Studies following a Q3W schedule administered 4 (BV, PV, and PV_surr), 5 (LV, LIFA, and TV) or 9 (BV) total doses. LV, ladiratuzumab vedotin; PV_surr, PV surrogate; TV, tisotumab vedotin.

Early deaths with targeted vedotin ADCs

High doses were associated with test article-related early deaths (animals found dead or euthanized moribund) for three rats and 16 monkeys (Supplementary Table S5). The underlying causes of moribundity/mortality were consistent with toxicity in surviving animals and included hematologic and/or skin toxicity, liver toxicity (rats only), or were undetermined (Tables 3–4; Supplementary Table S5).

Table 3.

Key toxicities associated with targeted vedotin ADCs in rats.

Test article	BV	EV		PV	SGN-B7H4V
Antibody target	CD30	Nectin-4		CD79b	B7-H4
Dosing schedule	Q1Wx4	Q1Wx4	Q1Wx13	Q1Wx4	Q1Wx4	2Q3Wx5
Recovery duration	4 weeks	6 weeks	None	6 weeks	4 weeks	None
Dose level(s) in mg/kg	0.5, 5, 10	2, 5, 10	0.5, 2, 5	2, 6, 10	5, 10, 15	5, 10
STD10	10*	10*	ND	10	15	ND
HNSTD/MTD	5	5	5	NR	NR	10
Main study animals (n/sex/group)	10	10	10	10	10	10
Recovery study animals (n/sex/group)	5a	5	0	5	5	0
Antidrug antibodies	—	✓	✓	NR	✓	✓
Early deaths due to test article	—	✓	—	✓	✓	—
Hematotoxicity	✓	✓	✓	✓	✓	✓
Skin toxicity	—	✓	✓	✓	✓	—
Lymphoid toxicity	✓	✓b	—	✓	✓	✓
Reproductive toxicity
Male	✓	✓	✓	✓	✓	✓
Female	—	—	—	—	✓c	✓
Liver and/or biliary toxicity	✓	✓	✓	✓	✓	✓
Acute phase protein changes	—	✓	✓	—	✓	—
Lung toxicity	—	—	—	—	✓	✓

Antigen-independent toxicities (hematopoietic, lymphoid, reproductive, and embryofetal toxicity) were common across vedotin ADCs at high doses. For antigen-dependent toxicities, BV had more pronounced hematologic toxicity, EV and SGN-B7H4V had skin toxicity, and SGN-B7H4V had lung toxicity.

Abbreviations: ND, not determined as all animals tolerated the highest dose; STD₁₀, severely toxic dose in 10% of the animals. Dash (—) indicates finding not present. *Based on the authors’ assessment, 10 mg/kg would have been the STD₁₀ but submission/review documents reported only a determination of HNSTD/MTD.

^aVehicle, mid, and high dose only.

^bLymphoid histologic changes with EV limited to increased spleen weights without histologic correlate, in contrast to the unchanged or decreased cellularity in lymphoid organs with most other test articles.

^cOvarian toxicity with was limited to hemorrhage in corpora lutea.

Antigen-independent toxicities of targeted vedotin ADCs

Toxicities with targeted vedotin ADCs that were consistent with MMAE and/or h00-1006(4) were considered antigen independent (Tables 2–4). These toxicities were similar across vedotin ADCs and included hematotoxicity and/or lymphoid toxicity (13 of 13 targeted ADCs; Supplementary Fig. S1) and reproductive toxicity (10 of 13 targeted ADCs; Supplementary Fig. S2).

Hematotoxicity profiles were generally consistent among MMAE and non-targeted and targeted vedotins; however, in rats, targeted ADCs did not have increased platelets seen with MMAE alone. In monkeys, differences with targeted ADCs compared with h00-1006(4) included increased platelets, variable monocyte counts and/or myeloid to erythroid ratios, and variable decreases in lymphoid cellularity in thymus, spleen, and/or lymph node/GALT. Hematotoxicity was the predominant toxicity for all test articles except EV and TV (Supplementary Table S6).

Consistent with MMAE and h00-1006(4) findings, lymphoid toxicity (generally characterized by decreased lymphocytes in blood and/or lymphoid organs) was observed with all targeted ADCs regardless of study duration or species (Supplementary Table S6). Additional lymphoid findings reflecting immune stimulation included increased spleen weights without a histologic correlate in rats with EV and increased cellularity (EV) and increased mononuclear/macrophage infiltration (EV and TV) in lymph node/GALT of monkeys. Similar lymphoid findings of mixed cell infiltrate in spleen and macrophage infiltrate in lymph node were also observed in monkeys in one study with MMAE.

In rats, male reproductive toxicity was observed in all studies with MMAE or targeted vedotin ADCs (Tables 2–3). Histological findings in rats and monkeys were similar and included degeneration/atrophy of the testes with secondary decreased sperm in epididymis (Supplementary Fig. S2A). In monkeys, male reproductive toxicity was observed only in two studies (LV, TV); these studies were of longer duration and used sexually mature (4–7 years old) monkeys. Female reproductive toxicity in rats, characterized by hemorrhage in corpora lutea, was only observed following administration of SGN-B7H4V (Table 3). Female reproductive toxicity in monkeys, characterized by decreases in secondary and tertiary follicles of ovaries, occurred in 8 of 18 studies with 5 of 13 targeted ADCs (Table 4; Supplementary Fig. S2B); these studies used relatively younger females (1.8–4 years old) and were conducted between 2020 and 2023, compared with those in which ovarian toxicity was not observed (2–7.3 years old, 2006–2020). Administration of the two vedotin ADCs with target expression in the ovary (tissue factor for TV and LIV-1 for LV) did not result in ovarian toxicity in monkeys.

The EFD toxicities and maternal toxicity observed with BV and EV in rats were generally consistent with those caused by MMAE (16, 17). However, the incidence and severity of changes, including reductions in viable fetuses and fetal weights, and skeletal variations were greater with the ADCs compared with MMAE. BV crossed the placenta (ADC, TAb, and MMAE were found in fetal serum); these endpoints were not evaluated for MMAE or EV.

Other antigen-independent toxicities included liver/biliary toxicity, acute phase changes, and skin discoloration. Liver/biliary toxicity was observed in all rat studies (six of six) but in none of the monkey studies with targeted ADCs. Liver/biliary findings in rats that differed from those observed with MMAE included hepatocyte necrosis or single-cell necrosis, hepatocellular hypertrophy, and a lack of coagulative necrosis. Hepatotoxicity contributed to the early deaths of one rat administered SGN-B7H4V and one rat administered PV. Acute phase protein changes were similar to those observed with MMAE or h00-1006(4) and occurred in 3 of 6 rat studies and 7 of 18 monkey studies with targeted ADCs. Acute phase protein changes were only associated with histologic evidence of inflammation in rats with EV (skin), and in monkeys with EV and TV (skin) and SGN-B6A (kidney and lung).

Antigen-dependent toxicity of targeted ADCs

Antigen-dependent toxicities were infrequent and found in a select number of target-expressing tissues limited to bone marrow (BV), spleen (PV_surr), skin (EV, TV, and SGN-B7H4V), eye (EV and TV), and lung (SGN-B7H4V). At the end of the recovery periods, these target-dependent toxicities had resolved.

Three of the ADCs (BV, PV_surr, and SGN-PDL1V) had target antigens that were highly expressed in hematologic and/or lymphoid cells/tissues of monkeys (Supplementary Table S4). Increased toxicity due to presence of target antigen occurred for BV and PV_surr but not SGN-PDL1V. Although the hematotoxicity of BV in monkeys was generally more pronounced than other ADCs (Supplementary Fig. S1), the HNSTD of BV with Q3W dosing (3 mg/kg) was similar to other targeted ADCs (3–6 mg/kg; Fig. 1; Table 4). Antigen-dependent hematotoxicity of BV became evident with more frequent dosing, as weekly dosing at 3 mg/kg was not tolerated due to severe neutropenia and required doses to be withheld (non-GLP study, Supplementary Table S7), whereas the same dose level and regimen was tolerated for seven other vedotin ADCs (Fig. 1; Tables 2 and 4). The concurrent evaluation of PV and PV_surr allowed direct comparison of toxicity which showed that PV_surr, which binds the monkey homolog of the target, caused an absence of germinal centers in the spleen, whereas in the same study PV, which does not bind in monkeys, did not cause histologic changes. The severity of hematologic and lymphoid toxicity associated with SGN-PDL1V was similar to other targeted ADCs, indicating that presence of the target did not cause additional toxicity or that additional toxicity could not be discerned in this analysis.

Antigen-dependent skin toxicity was observed with SGN-B7H4V in the 4-week study in rats and was dose-limiting for EV (rats and monkeys) and TV (monkeys). Skin toxicity was not observed in the 13-week rat study with SGN-B7H4V or with other ADCs with target expression in the skin (DV and LV).

In rats, EV at 10 mg/kg weekly (Q1W) led to death in one animal after four doses; the cause of death was uncertain although this rat had prior clinical signs of severe skin toxicity. Clinical signs of skin toxicity in rats also occurred after 2 weekly doses of EV and SGN-B7H4V and included observations of abrasions/sores (EV) or scab/crust (SGN-B7H4V). Histologic changes included epidermal hyperplasia and erosion/ulceration. In longer term studies, EV [weekly (Q1W) up to 5 mg/kg] and SGN-B7H4V [twice weekly in a 3-week period (2Q3W) up to 10 mg/kg] were tolerated without clinical or histologic observation of skin toxicity.

In monkeys, EV- and TV-related clinical signs of generalized skin toxicity occurred after one dose (3–6 mg/kg EV; 5 mg/kg TV). Changes with EV consisted of dry skin, generalized reddened skin, and abrasions all over the body including around the eyes. Administration of TV was also associated with findings of reddened skin (on eyelids and around the eyes) and dry flaky skin around the eyes. Clinical signs of skin toxicity necessitated unscheduled euthanasia in three animals at the high dose of each test article. Dosing of EV was discontinued at 6 mg/kg due to the early deaths. Dosing of TV was continued but intermittent supportive care (generally systemic antimicrobial therapy) was provided to monkeys at 5 mg/kg; the effect of this intervention on the tolerability of TV is unknown. The primary histologic skin finding in surviving and early death monkeys for both EV and TV was epidermal degeneration/necrosis, with ulceration and hemorrhage.

Antigen-dependent clinical signs of eye toxicity were observed in monkeys with EV (Q1Wx4) and TV (Q3Wx5). Findings were limited to changes in external appearance noted during clinical observations and were not observed by more detailed evaluation such as ophthalmology examinations or histopathology. For EV, eye changes were limited to clinical signs of sunken eyes, potentially secondary to toxicity in periorbital skin (see skin toxicity above). For TV, eye changes were limited to reddened or partially closed eyes and conjunctivitis which improved after treatment with antibiotic eye drops.

ADC-related changes of uncertain relationship to target

Findings of uncertain relationship to target typically occurred only in one or two animals per study and included transient clinical signs of changes in gait and histologic findings in mammary gland, Harderian gland, or respiratory, gastrointestinal, uterus/vagina, and/or renal/urinary tissues. In general, changes of uncertain relationship to target did not correlate to toxicities in humans, were not dose-limiting, and were generally expected to be reversible.

Lung toxicity was observed in rats treated with SGN-B7H4V (Q1Wx4 and 2Q3Wx5) and may have been associated with target expression. Findings included gross observations of tan foci in lung with histologic changes of alveolar macrophage aggregation, hypertrophy of type II pneumocytes, mononuclear cell infiltration of the alveolar interstitium, and perivascular mononuclear cell infiltrates. Lung findings in rats were also observed with PV and consisted of increased alveolar macrophages at all dose levels and type 2 pneumocyte hyperplasia at high doses. Lung findings in monkeys were limited to single instances of inflammation in monkeys administered lifastuzumab vedotin (LIFA) and SGN-B6A. The low incidence in the context of normal variability in monkey tissues obscures the potential relationship of expression to toxicity for these two test articles.

Single instances of degenerative/regenerative changes in the esophagus, trachea, and cecum with SGN-B6A may have been associated with target expression but were considered equivocal due to the low incidence of these findings and lack of observation in the 3-month study.

Consistency in the highest tolerated doses for vedotin ADCs

Rats tolerated repeat dose administration of vedotin ADCs at doses between 5 and 10 mg/kg (Table 3). For monkeys, the highest tolerated doses ranged between 3 and 6 mg/kg (Fig. 1A and B). Dose ranges for GLP studies were based on exploratory studies and/or non-tolerated dose levels of other vedotin ADCs. As a result, most GLP-compliant monkey studies (11 of 19) did not include a dose that caused early deaths (Tables 2 and 4; Supplementary Table S5).

Longer duration studies do not uncover additional toxicities

Results were analyzed to determine the effect of dosing schedule and study duration on detection of toxicity. In rats, there were five 1-month studies with MMAE or vedotin ADCs, and two longer studies with vedotin ADCs (Tables 2 and 3). In monkeys, there were 10 1-month studies and 12 longer studies with MMAE or vedotin ADCs (Tables 2 and 4). In general, complete toxicity profiles were identified in studies with higher doses, more frequent dosing, and of shorter duration. In rats, most target organs were identified in the 1-month studies. Longer rat studies also identified mammary gland, Harderian gland, and vaginal findings. In rats and monkeys, clinical observations and clinical pathology data collected repeatedly over the course of dosing showed no progression of toxicity with dosing longer than 1 month. In monkeys, early deaths occurred in longer studies only if a 1-month study was not conducted in the same species. Progression of male reproductive toxicity in monkeys could not be assessed as it was only evident in 2 of 22 studies, both of which had longer (>3 month) dosing periods and utilized sexually mature males.

Consistency of nonclinical pharmacokinetics for vedotin ADCs

Concentration–time profiles for TAb and unconjugated MMAE were generally comparable after a single dose of nine vedotin ADCs at 3 mg/kg (Fig. 2A and B). Likewise, the mean C_max and AUC_{0–7 days} were similar (<1.7-fold differences) across these ADCs (Supplementary Table S8).

Figure 2.

Vedotin ADCs have similar pharmacokinetics in monkeys regardless of target. A, Concentration–time profiles for TAb after a single dose at 3 mg/kg. B, Concentration–time profiles for MMAE after a single dose at 3 mg/kg. LV, ladiratuzumab vedotin.

Impact of immunogenicity on pharmacokinetics in longer studies

The impact of immunogenicity on exposures during ≥3-month monkey studies was assessed for studies with BV, LV, and TV. These three studies included a group at 3 mg/kg Q3W for five (LV and TV) or nine (BV) doses. Exposures (AUC_{0–21 days}) of ADC and MMAE following the first and last doses and ADA incidence are presented in Supplementary Table S9. For BV and TV, nearly all animals (19 of 20 and 10 of 10, respectively) were ADA positive at the end of the first dosing cycle (day 22), whereas for LV only one of six animals was ADA positive, with onset at day 64. ADC exposures for ADA-positive animals were generally reduced after the last dose compared with the first. However, MMAE exposures for ADA-positive animals were elevated following the last dose, likely related to increased ADC catabolism. ADC and MMAE exposures for ADA-negative animals were similar across dosing cycles.

Discussion

All vedotin ADCs in this analysis had similar PK properties (Fig. 2; Supplementary Table S8), predominant nonclinical toxicities (Table 1), and HNSTD values (Fig. 1). Only 2 of the 13 targeted ADCs (EV and TV) had dose-limiting antigen-dependent toxicity under the dosing schedules reported; antigen-independent hematotoxicity was dose-limiting for all other ADCs. These data support the conclusion that the vc-MMAE linker–payload generally drives vedotin ADC DLTs in animals.

PK and tolerated dose levels

The PK of nine vedotin ADCs were comparable in monkeys, as assessed by TAb and MMAE C_max and AUC_{0–7 days}, after an initial 3 mg/kg IV dose (Fig. 2; Supplementary Table S8). This finding is consistent with published data comparing first-in human clinical exposures for TAb and antibody-conjugated MMAE for eight different vedotin ADCs following a single dose of 2.4 mg/kg (10). Together, these data indicate that monkey and human PK of vedotin ADCs are consistent across targets and can be predicted.

All vedotin ADCs had similar HNSTD values within a given species (5–10 mg/kg in rat; 3–6 mg/kg in monkeys; Fig. 1; Tables 2 and 4) and all early deaths in monkeys occurred at doses ≥5 mg/kg (Supplementary Table S5). This is consistent with a prior FDA analysis that showed vedotin ADCs administered Q3W to monkeys have HNSTDs ranging from 3 to 6 mg/kg (11). Clinical experience with vedotin ADCs mirrors that of nonclinical studies in that the range of MTDs associated with repeat doses in patients is narrow, ranging from 1.8 to 2.4 mg/kg Q3W (11).

The consistency of HNSTDs in monkeys suggests that future 1-month nonclinical studies with vedotin ADCs could be streamlined with a high dose of 4 or 5 mg/kg and only two dose levels given the steep dose-response curve. Longer duration studies could be refined by the inclusion of a single dose level of 3 mg/kg without a recovery arm. This streamlined approach has recently enabled phase III clinical evaluations for two vedotin ADCs. A single nonclinical study supported both molecules; each ADC was dosed at 3 mg/kg and a common control group was leveraged. These study design modifications prevent unnecessary stress on animals associated with excessive toxicity at higher dose levels and decrease the number of animals required—consistent with 3R initiatives and ICH guidelines—without compromising risk assessment.

Antigen-independent toxicity is generally similar across vedotin ADCs

Most antigen-independent toxicities (hematopoietic, lymphoid, reproductive, and embryofetal toxicity) were common across vedotin ADCs at high doses and expected based on the mechanism of action of MMAE. Although data in this analysis were limited, the toxicity profile was also consistent when the antibody backbone lacked Fc binding (SGN-PDL1V). Other tissues that are highly vascularized (e.g., lung and kidney) or with rapidly dividing cells (gastrointestinal tract) were generally not affected. The reason for this lack of sensitivity is unknown.

Unlike hematologic and lymphoid toxicity, antigen-independent ovarian and testicular toxicity were inconsistently observed with vedotin ADCs in monkeys. Ovarian toxicity was consistently observed in more recent studies (i.e., 2020 and later) using young monkeys dosed weekly, and was not observed in prior studies that mostly used older monkeys and less frequent dosing. Possible reasons for the lack of ovarian findings in the earlier studies include a decreased sensitivity to toxicity in older animals, or differences in sexual maturity, age, and social status that may make evaluation of ovarian follicular development challenging. Testicular toxicity was consistently observed in rats but only observed in monkeys in longer duration studies in which the majority (if not all) of the males were sexually mature. Ovarian and testicular toxicity are consistent with the mechanism of MMAE on rapidly dividing cells.

Results from EFD studies conducted with MMAE, BV, and EV conclusively demonstrate that vedotin ADCs cause antigen-independent embryofetal toxicity. This is consistent with the mechanism of action of MMAE and indicates that there is no need to perform EFD studies with future vedotin ADCs.

Skin toxicity was unique among antigen-independent findings in terms of its low incidence (1 of 30 total monkeys administered MMAE) and anatomic distribution (limited to one small area). Skin toxicity was not observed in the monkey study with h00-1006(4) or with ADCs that did not have target expression in skin. Skin toxicity was observed as an antigen-dependent toxicity with three of five ADCs that had target expression in the skin (EV, TV, and SGN-B7H4; not DV and LV).

Although peripheral neuropathy is a known MMAE-related toxicity in the clinic (4–7, 13), it has not been observed in nonclinical toxicology studies. Numerous approaches have been taken to develop a sensitive preclinical model for MMAE-driven peripheral neuropathy (23), including enhanced neurohistologic techniques, electrophysiology, and extended duration studies in presensitized animals, but thus far none have proven to be predictive and repeatable (Seagen unpublished data). Pulmonary toxicity has been reported with vedotin ADCs in the clinic regardless of target expression in the lung (4, 5, 7, 12). Although pulmonary toxicity was not routinely observed in nonclinical studies with vedotin ADCs, the turnover rate of pulmonary epithelial cells, particularly when increased by injury, may render them sensitive to the antimitotic effects of MMAE.

Antigen-dependent toxicity was infrequent

Antigen-dependent toxicity was observed with only 5 of the 13 targeted ADCs (BV, EV, PV_surr, SGN-B7H4V, and TV). With BV and PV_surr, antigen-dependent toxicity seems to be an exacerbation of antigen-independent toxicity—BV had more pronounced hematologic toxicity and PV_surr had more lymphoid (spleen) toxicity. Skin toxicity observed with EV, TV, and SGN-B7H4V was not observed with other vedotin ADCs.

Antigen-dependent toxicities were observed in a small number of tissues expressing vedotin ADC targets, and the tissues with highest target expression (Supplementary Table S4) did not always predict nonclinical toxicity. This was most clearly demonstrated with EV and TV. Nectin-4 (EV) and tissue factor (TV) are highly expressed in skin, and both ADCs resulted in dose-limiting skin toxicity. However, other tissues with similar Nectin-4 and tissue factor expression levels did not show toxicity. Even tissues with approximately 100x expression that are well perfused and have rapidly dividing cells [e.g., kidney expresses NaPi2b (LIFA) and colon expresses B6A (SGN-B6A)] showed no consistent toxicity, suggesting that expression did not predict toxicity.

Antigen-dependent ocular toxicities (in particular, corneal keratopathies or conjunctival changes) are observed clinically for vedotin ADCs when the target is expressed in ocular tissues (EV and TV; refs. 4, 7). Corneal toxicity has not been observed in monkeys for vedotin ADCs, but increased mitotic figures were observed in the corneas of rats administered EV. Non-GLP studies in rats with h00-1006(4) also show increased corneal mitotic figures at high doses. The relevance of these changes to human corneal toxicity is uncertain because a uniform dataset with consistent counting of corneal mitotic figures across multiple vedotin ADCs is not available.

Increased study duration did not improve human risk assessment for vedotin ADCs

This cross-program analysis indicates that 3-month studies of vedotin ADCs predict the same or fewer toxicities that are relevant to human risk assessment than 1-month studies (Tables 2–4). These results are consistent with Chien and colleagues (24), who analyzed data from first-in-human (FIH)-enabling and chronic toxicity studies for 142 mAbs and found that 15 of 111 (4 of 32 for oncology) 6-month chronic exposure studies uncovered new toxicities that impacted clinical development.

The lack of additional toxicities identified in the longer duration vedotin ADC studies could be due to lower dose levels tested, less-frequent dosing schedules, or the development of ADA and subsequent decreases in ADC exposure. In the context of high ADA incidence, MMAE concentrations were sometimes elevated but were not associated with additional toxicity. It is possible that toxicity may be driven by ADC rather than MMAE exposure or that the MMAE concentrations are below the threshold for toxicity.

Implications for future ADC toxicology programs

The consistency of nonclinical toxicities and PK, and narrow range for HNSTDs indicate that animal use can be reduced by utilizing fewer dose groups in future nonclinical studies with vedotin ADCs without compromising patient safety. Therefore, platform-based approaches leveraging existing nonclinical and clinical data should be employed to set clinical starting doses and predict antigen-independent toxicities for vedotin ADCs. This approach may also be leveraged for other well-characterized ADC payloads, including noncytotoxic molecules. For new payloads/platforms, dose range-finding studies with a non-targeted ADC could characterize antigen-independent toxicities and tolerability limits. Alternatively, rat studies with a targeted, but non–cross-reactive ADC could also establish platform toxicity. Fewer studies in a toxicology program means that ADC development can be expedited, thereby speeding the delivery of novel therapies to patients with significant unmet medical needs.

Supplementary Material

Supplementary Tables 1-9

Supplementary Tables 1-9

Supplementary Figure 1

Supplementary Figure 1. Hematopoietic and lymphoid tissue toxicity in cynomolgus monkeys administered a vedotin ADC or MMAE.

Supplementary Figure 2

Supplementary Figure 2. Reproductive toxicity with vedotin ADCs in cynomolgus monkeys.

Authors’ Disclosures

H.D. Neff-LaFord, S.A. Carratt, C. Carosino, N. Everds, K.A. Cardinal, S. Duniho, M.M. Schutten, C. Frantz, C. Zuch de Zafra, and E.B. Harstad report being an employee of and having equity ownership in Seagen, Inc. (acquired by Pfizer in December 2023) at the time of this work. N. Everds also reports being a consultant for Seagen after her retirement.

Authors’ Contributions

H.D. Neff-LaFord: Conceptualization, formal analysis, supervision, investigation, visualization, writing–original draft, project administration, writing–review and editing. S.A. Carratt: Conceptualization, data curation, formal analysis, investigation, visualization, writing–original draft, writing–review and editing. C. Carosino: Conceptualization, formal analysis, supervision, investigation, writing–original draft, project administration, writing–review and editing. N. Everds: Conceptualization, data curation, formal analysis, investigation, visualization, writing–original draft, writing–review and editing. K.A. Cardinal: Conceptualization, data curation, formal analysis, investigation, visualization, writing–original draft, writing–review and editing. S. Duniho: Conceptualization, formal analysis, visualization, writing–review and editing. M.M. Schutten: Conceptualization, supervision, project administration, writing–review and editing. C. Frantz: Conceptualization, writing–review and editing. C. Zuch de Zafra: Conceptualization, writing–review and editing. E.B. Harstad: Conceptualization, supervision, project administration, writing–review and editing.