The clinical efficacy of first-generation carcinoembryonic antigen (CEACAM5)-specific CAR T cells is limited by poor persistence and transient pre-conditioning-dependent respiratory toxicity

Abstract

The primary aim of this clinical trial was to determine the feasibility of delivering first-generation CAR T cell therapy to patients with advanced, CEACAM5⁺ malignancy. Secondary aims were to assess clinical efficacy, immune effector function and optimal dose of CAR T cells. Three cohorts of patients received increasing doses of CEACAM5⁺-specific CAR T cells after fludarabine pre-conditioning plus systemic IL2 support post T cell infusion. Patients in cohort 4 received increased intensity pre-conditioning (cyclophosphamide and fludarabine), systemic IL2 support and CAR T cells. No objective clinical responses were observed. CAR T cell engraftment in patients within cohort 4 was significantly higher. However, engraftment was short-lived with a rapid decline of systemic CAR T cells within 14 days. Patients in cohort 4 had transient, acute respiratory toxicity which, in combination with lack of prolonged CAR T cell persistence, resulted in the premature closure of the trial. Elevated levels of systemic IFNγ and IL-6 implied that the CEACAM5-specific T cells had undergone immune activation in vivo but only in patients receiving high-intensity pre-conditioning. Expression of CEACAM5 on lung epithelium may have resulted in this transient toxicity. Raised levels of serum cytokines including IL-6 in these patients implicate cytokine release as one of several potential factors exacerbating the observed respiratory toxicity. Whilst improved CAR designs and T cell production methods could improve the systemic persistence and activity, methods to control CAR T 'on-target, off-tissue' toxicity are required to enable a clinical impact of this approach in solid malignancies.

Keywords: CEA; Chimeric antigen receptor; Persistence; T cells; Toxicity.

Fig. 1

Increased intensity of patient pre-conditioning results in prolonged lymphodepletion and increased frequency of MFEζ CAR T cell engraftment. Lymphocyte counts (×10⁹ cells/L) of patients in a cohort 1, b cohort 2, c cohort 3 and d cohort 4. The timing of fludarabine pre-conditioning (black box) and cyclophosphamide (grey box) are shown prior to MFEζ T cell infusion on day 0. MFEζ CAR T cell frequency determined by qPCR in blood samples estimated by comparison to β2-microglobulin to determine relative CAR T cell frequency in e cohort 1, f cohort 2, g cohort 3 and h cohort 4

Fig. 2

Consistent decrease in systemic CEACAM5 levels and transient elevations in serum cytokines are seen in cohort 4 post-MFEζ CAR T cell infusion. Patients in each cohort are colour-coded to identify cohorts: cohort 1 in blue, cohort 2 in green, cohort 3 in red and cohort 4 in orange. Serum CEACAM5 levels were determined in patient blood samples prior and post-MFEζ CAR T cell infusion for a cohorts 1–3 and b cohort 4. The same colour-coding is applied to serum cytokine analysis of c IFNγ where analysis was performed at baseline (b) and weekly until day 28. For all other cytokines, d MCP-1, e IL-6, f IL-8, g IL-10 and h IL-1β where baseline, day 7, day 14 and post-treatment (P; day 21 or 28) are shown

Fig. 3

Molecular analysis of fine needle biopsy samples to determine the presence of MFEζ CAR T cells within tumour tissue. MFEζ transgene in tumour biopsies (hatched bars) and peripheral blood (black bars) taken at the same time as the biopsy at week 10 for patent 36004; week 6 for patient 36010; week 14 for patient 36014; week 6 for patient 36015; and week 4 for patient 36017

Fig. 4

Evidence of pulmonary infiltrates consistent with local cytokine release syndrome in patient 36,013. CT images demonstrating non-specific pulmonary infiltrates 10 days after MFEζ T cell infusion (a) and resolution by day 42 (b)