CEA-CD3 bispecific antibody cibisatamab with or without atezolizumab in patients with CEA-positive solid tumours: results of two multi-institutional Phase 1 trials

Abstract

Cibisatamab is a bispecific antibody-based construct targeting carcinoembryonic antigen (CEA) on tumour cells and CD3 epsilon chain as a T-cell engager. Here we evaluated cibisatamab for advanced CEA-positive solid tumours in two open-label Phase 1 dose-escalation and -expansion studies: as a single agent with or without obinutuzumab in S1 (NCT02324257) and with atezolizumab in S2 (NCT02650713). Primary endpoints were safety, dose finding, and pharmacokinetics in S1; safety and dose finding in S2. Secondary endpoints were anti-tumour activity (including overall response rate, ORR) and pharmacodynamics in S1; anti-tumour activity, pharmacodynamics and pharmacokinetics in S2. S1 and S2 enrolled a total of 149 and 228 patients, respectively. Grade ≥3 cibisatamab-related adverse events occurred in 36% of S1 and 49% of S2 patients. The ORR was 4% in S1 and 7% in S2. In S2, patients with microsatellite stable colorectal carcinoma (MSS-CRC) given flat doses of cibisatamab and atezolizumab demonstrated an ORR of 14%. In S1 and S2, 40% and 52% of patients, respectively, developed persistent anti-drug antibodies (ADAs). ADA appearance could be mitigated by obinutuzumab-pretreatment, with 8% of patients having persistent ADAs. Overall, cibisatamab warrants further exploration in immunotherapy combination strategies for MSS-CRC.

Fig. 1. Treatment-related adverse events.

a Most frequent treatment-related adverse events (≥ 20% in either study) in S1 and S2. b Median time to onset (middle line; with means as circles, quartiles as boxes, and ranges as bars) of most common treatment-related adverse events (≥20% in either study) in S1 and S2. CIBI Cibisatamab, IRR Infusion-related reaction. Source data can be requested from the authors for academic research purposes.

Fig. 2. Patient with bilateral lung lesions at baseline and lung lesion inflammation and localised perilesional oedema on Cycle 1 Day 3.

The patient received 60 mg cibisatamab on Cycle 1 Day 1 and experienced transient hypoxia and transient dyspnoea 48 h after infusion. CT Computed tomography.

Fig. 3. Combined incidence and severity of IRR- and CRS-related adverse events and associated symptoms.

Data for (a) ADA-negative and (b) ADA-positive patients are shown by infusion for patients receiving 40–600 mg of cibisatamab (flat-dose and step-up dosing cohorts) in S1 and S2. ADA, Anti-drug antibody; CRS, Cytokine-release syndrome; IRR, Infusion-related reaction; M Grade information not provided by site. Source data can be requested from the authors for academic research purposes.

Fig. 4. Patterns of response.

a Spider plot of change in sum of target lesions from baseline according to Response Evaluation Criteria in Solid Tumors version 1.1 in patients with MSS-CRC receiving cibisatamab flat dose 100 mg once weekly (top, n = 20), and 100 mg once every 3 weeks (bottom, n = 19) plus atezolizumab in S2. b Computerised tomography scan images from a RECIST-confirmed partial responder with MSS-CRC taken at baseline and week 16 after receiving cibisatamab 160 mg QW and atezolizumab 1200 mg Q3W. CEACAM5, Carcinoembryonic antigen-related cell adhesion molecule 5; NA Not applicable, NE Not evaluable, PD Progressive disease, PR Partial response, SD Stable disease. Source data can be requested from the authors for academic research purposes.

Fig. 5. CEACAM5 levels and intratumoural changes in CD8 + T cells and PDCD1 gene expression.

a CEACAM5 level vs best overall response according to Response Evaluation Criteria in Solid Tumors version 1.1 in the 100–160 mg cibisatamab cohort (MSS-CRC population) including all patients who received at least one dose of any study treatment (medians shown as middle lines, quartiles as boxes, and ranges as bars). b–f Refer to changes between baseline and on-treatment samples from patients with MSS-CRC in S2 who received doses ≥ 100 mg cibisatamab, by best overall response. b Dual chromogenic IHC for CD8 (yellow) and Ki67 (purple nuclear signal, outlining most tumour cells) for a patient with partial response (upper panel) and a patient with stable disease (lower panel). Scale bar represents 88.5 μm (top left), 77.3 μm (top right), 84.4 μm (bottom left) and 73.0 μm (bottom right). c Changes in CD8 + T cells between baseline and on-treatment tissue were quantified with a KI57/CD8 duplex assay. d Single chromogenic IHC for PD-1 at baseline (left) and on-treatment (right). Scale bar represents 76.7 μm (left) and 77.4 μm (right). e Changes in PDCD1 gene expression between baseline and on-treatment tissue were quantified by RNAseq. f Cibisatamab induces tissue pharmacodynamic changes. IHC for CD8 (yellow) and Ki67 (purple, outlining most tumour cells). The immune-excluded immunophenotype is shown on the left (baseline sample) and the immune-inflamed immunophenotype is shown on the right (on-treatment sample). Scale bar represents 65.4 μm (top left), 60.2 μm (top right), 64.4 μm (middle left), 52.8 μm (middle right), 64.5 μm (bottom left), and 61.0 μm (bottom right). In panels a, c, e, blue colour refers to PR, orange to SD, and red to PD. The N or n refers to individual patients. In subplots (c, e) dots for the same patient are connected by a line. The exploratory test in subplot f is two-sided. Atezo Atezolizumab; BL Baseline tissue; BOR Best overall response; CEACAM5 Carcinoembryonic antigen-related cell adhesion molecule 5; cibi Cibisatamab OT On-treatment tissue; PD Progressive disease; PR Partial response; RPKM Reads per kilobase of transcript per million reads mapped; SD Stable disease.

Fig. 6. Pharmacodynamic changes in gene expression signatures between baseline and on-treatment tissue.

These findings demonstrate increased T-effector gene signatures, with no changes in antigen presentation genes, and support the theorised synthetic immunity mechanism of action of cibisatamab targeting tumour-associated CEA, without the requirement for antigen-specific T cells and acquired immunity for activity. Medians are shown as middle lines, quartiles as boxes, and ranges as bars. N refers to individual patients with the same patients shown for antigen-processing and T-effector. Blue colour refers to PR, orange to SD, and red to PD. Antigen-processing gene set: TAPBP, TAP1, TAP2, PSMB8, PSMB9. T-effector gene set: CD8A, GZMA, GZMB, IFNG, EOMES, PRF1, CXCL9, CXCL10, TBX21. BL Baseline tissue; OT On-treatment tissue; PD Progressive disease; PR Partial response; SD Stable disease.