Precemtabart tocentecan, an anti-CEACAM5 antibody-drug conjugate, in metastatic colorectal cancer: a phase 1 trial

Abstract

CEACAM5, a cell surface protein, is overexpressed in colorectal cancer (CRC). Precemtabart tocentecan (Precem-TcT, previously M9140) is an anti-CEACAM5 antibody-drug conjugate with the topoisomerase 1 inhibitor exatecan as payload. Precem-TcT demonstrated strong antitumor activity and potent bystander activity in preclinical models. Its toxicity profile in cynomolgus monkeys was consistent with that of exatecan. In the dose-escalation stage of the phase 1 trial of Precem-TcT (PROCEADE-CRC-01), 40 heavily pretreated patients with irinotecan-refractory metastatic CRC received Precem-TcT every 3 weeks across seven dose levels (DLs, 0.6-3.2 mg kg^-1). Primary endpoints were dose-limiting toxicities (DLTs), adverse events and preliminary clinical activity to establish the recommended dose(s) for expansion (RDEs). Secondary endpoints included pharmacokinetic parameters, objective response and median progression-free survival (mPFS). At the planned, end-of-dose-escalation analysis with extended follow-up (cutoff: 1 August 2024), seven patients had experienced DLTs, primarily hematologic events at 3.0 mg kg^-1 and 3.2 mg kg^-1. A treatment-related death, also deemed disease related, was reported in a patient with multiple comorbidities and grade 3 obesity. The maximum tolerated dose was determined to be 2.8 mg kg^-1 every 3 weeks. Total and conjugated antibody pharmacokinetic profiles largely overlapped, indicating stability of the linker-payload (β-glucuronide-exatecan) in circulation. After a median treatment of 19.1 weeks (range: 1.7-48.3), three of 40 patients (7.5%) had confirmed partial responses (15.0% (6/40) unconfirmed), all at DLs ≥2.4 mg kg^-1. mPFS was 5.9 months (95% confidence interval: 4.6-7.2); at DLs ≥2.4 mg kg^-1 (n = 34), mPFS was 6.7 months (95% confidence interval: 4.6-8.8). Four patients (10.0%) remained on treatment at cutoff. These early clinical data corroborate preclinical findings, showing predictable safety and encouraging antitumor activity of Precem-TcT at DLs ≥2.4 mg kg^-1, with no interstitial lung disease or ocular toxicity. The dose-optimization part at the RDEs of 2.4 mg kg^-1 and 2.8 mg kg^-1 (both every 3 weeks) in PROCEADE-CRC-01 is ongoing. ClinicalTrials.gov identifier: NCT05464030 .

Fig. 1. Preclinical data overview.

a, Selective killing of CEACAM5-positive cancer cell lines compared to CEACAM5-negative cell line. Data are mean ± s.d. (n = 3) from a representative replicate of three independent experiments. b, Bystander effect of Precem-TcT and tusa.rav.a. in in vitro co-culture experiments (1 nM ADC). Data are mean from a representative replicate of two independent experiments. c, In vivo efficacy of Precem-TcT in comparison to tusa.rav.a. in two CRC PDX models: REPF210 and COPF230; data are mean ± s.e.m. from n = 6 mice per treatment group and n = 5 mice per treatment group, respectively. d, Potency shift assay using Precem-TcT in comparison to tusa.rav.a. or different payloads with and without Pgp inhibition using verapamil/zosuquidar (LS513 cell line). Left panel: data are mean ± s.e.m. (n = 2) from a representative replicate of two independent experiments; right panel: each data point represents an independent experiment and is the mean of two technical replicates; horizontal black line represents the mean of the two data points. e, Potency shift assay using payloads with and without BCRP inhibition in OUMS23 cell line (left) and SNU5 cell line (right). Each data point represents an independent experiment and is the mean of two technical replicates. Horizontal black line represents the mean of the two data points. f, Antitumor efficacy of Precem-TcT compared to that of irinotecan in the human CRC PDX model CXF4102, derived from a patient with CRC previously treated with irinotecan-based SoC (FOLFIRI/Avastin). Data are mean ± s.e.m. from n = 6 mice per treatment group. w/o, without. Source data

Fig. 2. Patient disposition.

CONSORT diagram for the PROCEADE-CRC-01 dose escalation study.

Fig. 3. Efficacy overview.

a, Best overall response (confirmed). The waterfall plot only includes patients who have a post-baseline assessment (n = 36). Most of the patients had high CEACAM5 expression, as indicated by the dark blue boxes in the ‘Percentage of CEACAM5-positive tumors’ category. Blank boxes indicate that CEACAM5 expression data were not available. b, Progression-free survival. c, Treatment duration and response over time. CI, confidence interval; CR, complete response; N, no; NE, not evaluable; PD, progressive disease; PR, partial response; SD, stable disease; U, unknown; Y, yes.

Extended Data Fig. 1. Bystander effect of Precem-TcT and tusa.rav.a.

In vitro co-culture experiments (1 nM ADC) with lower number of CEACAM5-positive cells and a control of CEACAM5-negative cells only. Data are mean (n = 2) from a representative replicate of 2 independent experiments.

Extended Data Fig. 2. Study design of PROCEADE-CRC-01.

3L, third line; 3 L + , third- or later-line; ADA, anti-drug antibody; AE, adverse event; CRC, colorectal cancer; DC, disease control; DL, dose level; DLT, dose-limiting toxicity; DoR, duration of response; ECG, electrocardiogram; ECOG PS, Eastern Cooperative Oncology Group performance status; Est. estimated; FiH, first-in-human; ICI, immune checkpoint inhibitor; IV, intravenous; la/m, locally advanced/metastatic; MSI-H, microsatellite instability high; MTD, maximum tolerable dose; OR, objective response; OS, overall survival; PFS, progression-free survival; PK, pharmacokinetics; Q2W, every 2 weeks; Q3W, every 3 weeks; RDE, recommended dose for expansion; RECIST v1.1, Response Evaluation Criteria In Solid Tumors version 1.1; TTR, time to response; US, United States. ^aAssessed by investigator per RECIST v1.1. 1. Kopetz S, et al. J Clin Oncol. 2024;42(16_Suppl):3000.

Extended Data Fig. 3. Clinical pharmacokinetics of Precem-TcT.

Panel a. Solid line represents the predicted value, dotted lines show 95% CI. Panel b. Error bars show 95% CIs (assessed as the 2.5^th and 97.5^th quantiles by nominal time points). AUC, area under curve; CI, confidence interval.