TCR-engineered T cells targeting E7 for patients with metastatic HPV-associated epithelial cancers

Abstract

Genetically engineered T cell therapy can induce remarkable tumor responses in hematologic malignancies. However, it is not known if this type of therapy can be applied effectively to epithelial cancers, which account for 80-90% of human malignancies. We have conducted a first-in-human, phase 1 clinical trial of T cells engineered with a T cell receptor targeting HPV-16 E7 for the treatment of metastatic human papilloma virus-associated epithelial cancers (NCT02858310). The primary endpoint was maximum tolerated dose. Cell dose was not limited by toxicity with a maximum dose of 1 × 1011 engineered T cells administered. Tumor responses following treatment were evaluated using RECIST (Response Evaluation Criteria in Solid Tumors) guidelines. Robust tumor regression was observed with objective clinical responses in 6 of 12 patients, including 4 of 8 patients with anti-PD-1 refractory disease. Responses included extensive regression of bulky tumors and complete regression of most tumors in some patients. Genomic studies, which included intra-patient tumors with dichotomous treatment responses, revealed resistance mechanisms from defects in critical components of the antigen presentation and interferon response pathways. These findings demonstrate that engineered T cells can mediate regression of common carcinomas, and they reveal immune editing as a constraint on the curative potential of cellular therapy and possibly other immunotherapies in advanced epithelial cancer.

Figures

Extended Data Fig. 1 |

CONSORT flow diagram.

Extended Data Fig. 2 |. Durable, complete regression of multiple index and non-index tumors.

Radiographic imaging studies from Patient 5 and Patient 12. a, For Patient 5, computed tomography (CT) scans are shown. b, For Patient 12, the top row are CT scans and the middle and bottom rows are positron emission tomography-CT scans. The timepoint for the scans is indicated above each column. Yellow arrows indicate tumors. c, The number of discrete tumors present at baseline and serial timepoints after treatment are shown.

Extended Data Fig. 3 |. Peripheral blood engraftment of E7 TCR-T cells in Patient 12 at late time points.

The frequency of E7 TCR-T cells in the peripheral blood of Patient 12 at late time points after treatment is graphed.

Extended Data Fig. 4 |. Peripheral blood engraftment of E7 TCR-T cells.

Flow cytometry was used to determine engraftment of E7 TCR-T cells in the peripheral blood of patients at the first response assessment timepoint (6 weeks). a, The frequency of E7 TCR-T cells in peripheral blood correlated with dose (P = 0.0006, Kruskal-Wallis one-way ANOVA) but not with response (P = 0.347, two-sided unpaired t-test). b, The concentration of E7 TCR-T cells in peripheral blood correlated with dose (P = 0.0226, one-way ANOVA) but not with response (P = 0.4796, two-sided unpaired t-test).

Extended Data Fig. 5 |. Serum antibodies against E7 TCR-T cells.

The dotted line indicates the background detection level. Hamster anti-mouse TCR antibody was used as a positive control (Pos). Healthy donor serum was used as the negative control (Neg).

Extended Data Fig. 6 |. Infusion product phenotypic characterization.

The expression of cell surface antigens by E7 TCR-T cells was determined by flow cytometry. a, Stacked bar graph of the frequency of single-positive CD4 and CD8 T cells. b, Pie chart that depict memory T cell subset composition. Tn (naïve, CD45RA+CCR7+), Tcm (central memory, CD45RA−CCR7+), Tem (effector memory, CD45RA−CCR7−), Temra (effector memory RA+, CD45RA+CCR7−). c, Bar graph of the frequency of expression of the chemokine receptors in the symbol legend. d, The frequency of each marker or subset in the E7 TCR-T cells administered to patients who responded (R) or did not respond (NR) to treatment. P-values are two-sided unpaired t-tests.

Extended Data Fig. 7 |. Phenotype of engrafted peripheral blood E7 TCR-T cells.

a, Flow cytometric analysis of memory T cell subsets, chemokine receptors, CD4/CD8 T cell subsets, and Ki-67 expression by engrafted E7 TCR-T cells is shown. The T cell subset or phenotypic marker is indicated above each graph. Gating is on CD3+, mTCRB+, live, lymphocytes. CD4 and CD8 frequencies are for single positive cells. Tn (naïve, CD45RA+CCR7+), Tcm (central memory, CD45RA−CCR7+), Tem (effector memory, CD45RA−CCR7−), Temra (effector memory RA+, CD45RA+CCR7−). b, Phenotype at first response assessment (6 weeks) of engrafted E7 TCR-T cells in patients who responded (R) or did not respond (NR) to treatment. The T cell subset or phenotypic marker is indicated above each graph. Red symbols represent patients with tumor responses. P-values are from two-sided unpaired t-tests.

Extended Data Fig. 8 |. Infusion product functional characterization.

a, Percent cell killing at 8 hours as measured by impedance-based cytolysis assay. The effector to target ratio (E:T) is indicated on the x-axis. The target cell is indicated in the symbol legend. Error bars represent the standard deviation of 2–5 technical replicates. CaSki is HLA-A*02:01+ HPV-16+. 624 is HLA-A*02:01+ HPV-16−. b, The frequency of transduced cells that expressed Ki-67 as determined by flow cytometry. c, Cytokine production following coculture of the infusion product from the patient indicated in the graph title with 293-A2 cells pulsed with the target peptide indicated in the symbol legend. Error bars represent the standard deviation of 2–3 independent experiments (with 2 technical replicates in each experiment); all replicates are shown. d, ELISPOT assay measurement of cytokine-secreting cells in the infusion product. The cytokine measured is indicated by the graph title. The target cells and symbol legend are the same as in panel C. 2 independent experiments are shown. e-h, Functional characteristics of the infusion products administered to patients who responded (R) or did not respond (NR) to treatment. P-values are two-sided unpaired t-tests.

Extended Data Fig. 9 |. Serum cytokines and chemokines following E7 TCR-T cell infusion.

Serum concentrations of IL-1α, IL-1β, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12/23 p40, IL-12 p70, IL-13, IL-15, IL-16, IL-17A, IL-21, IL-22, IL-23, IL-27, IL-31, IFNγ, TNFα, TNFβ, GMCSF, VEGF-A, Eotaxin, Eotaxin-3, CXCL10, MCP-1, MCP-4, MIP-1α, MIP-1β, TARC, and MIP-3α were determined. Chemokines and cytokines for which the peak values were significantly different between responding and non-responding patients are shown. a, Graph of serum IL-17 and VEGF-A concentrations in patients indicated in symbol legend at the time points indicated on the x-axis. Red color indicates patients with tumor responses. b, Peak IL-17 and VEGF-A serum concentrations in patients who responded (R) or who did not respond (NR) to treatment. The symbol legend is shown in panel A. Statistical significance was determined by a two-sided unpaired t-test.

Extended Data Fig. 10 |. Infiltration of tumor biopsy specimens with E7 TCR-T cells.

Serial samples from endoscopic biopsy of Patient 12’s tumor were examined. Immunohistochemistry was performed to detect CD3 protein, and RNAscope was performed to detect the E7 TCR α-chain (mTRAC) and HPV-16 E7 transcripts. A sequential labeling technique was employed to simultaneously examine protein and RNA expression on a single slide (each sample was stained once and is shown). The time point for each biopsy is indicated at the top of each row. The protein or transcript labeled is indicated to the left of each column. The color legend indicates labeling for the overlays on the bottom two rows. The yellow arrows point to E7 TCR-T cells as detected by CD3 protein membrane labeling and cytoplasmic mTRAC transcript labeling.

Fig. 1 |. E7 TCR-T cells demonstrate robust clinical activity in epithelial cancers.

a, Waterfall plot of the best clinical response for each evaluable patient. b, Spider plot of the change in the sum of the diameters of each patient’s index lesions over time. Patients with objective clinical responses are indicated in red. c, Positron emission tomography (PET) scans from patient 1, who had metastatic vulvar cancer with extensive pulmonary, retroperitoneal, pelvic and thigh lesions. d, Computed tomography (CT) scans from patient 5, who had metastatic anal cancer with numerous pulmonary, pleural and kidney lesions. e, CT scans (top row) and PET-CT scans (middle row and bottom row) from patient 12, who had metastatic cervical cancer with chest wall, retroperitoneal and rectal lesions. Yellow arrows point to tumors.

Fig. 2 |. Engineered T cells displayed in vivo persistence and function, and tumors displayed genetic defects in crucial immune-related genes.

a, Persistence of E7 TCR-T cells based on flow cytometric analysis of peripheral blood. White, gray and black symbols indicate dose levels 1, 2 and 3, respectively. b, The frequency of E7 TCR-T cells and the reactivity of peripheral blood T cells against E7 before treatment and at the first clinical response assessment (that is, at the 6-week time point). The left y axis and bars indicate interferon-γ (IFN-γ) production as measured by FluoroSpot assay. Data are presented as mean ± standard deviation of three technical replicates. The right y axis and circle symbols indicate the frequency of E7 TCR-T cells per peripheral blood T cell. Persistence of E7 TCR-T cells correlated with peripheral blood T cell reactivity against E7 (Spearman r = 0.6545, P = 0.0336). c–e, Flow cytometric analysis of the E7 TCR-T cells expression of inhibitory receptors. c, Expression of inhibitory receptors by E7 TCR-T cells in infusion products (infusion) compared to E7 TCR-T cells in peripheral blood at the 6-week time point (6 weeks). The frequency of transduced E7 TCR-T cells that express the inhibitory receptor indicated in the y-axis label is shown. P values represent two-sided paired t-tests (P = 0.78 for PD-1, P < 0.0001 for LAG-3 and P = 0.0018 for TIM-3). d, Comparison of inhibitory receptor expression by E7 TCR-T cells in patient who responded (R) or did not respond (NR) to treatment. The frequency of transduced E7 TCR-T cells that express the inhibitory receptor indicated in the y-axis label is shown. The top row shows infusion products. The bottom row shows peripheral blood at the 6-week time point. P values represent two-sided unpaired t-tests. e, Expression of inhibitory receptors at the 6-week time point by peripheral blood T cells expressing (E7 TCR+) or not expressing (E7 TCR−) the E7 TCR. P values represent two-sided paired t-tests. f, Heatmap indicating genetic defects in molecular pathways related to antigen processing (top section) and IFN response (bottom section). Each column represents a distinct tumor. The patient, site of disease, biopsy timing and response to treatment for that specific tumor are indicated above the columns. Only damaging mutations are shown. Copy number variation (CNV) analysis was performed with whole exome sequencing (WES) data. CNV loss was defined as copy number less than tumor ploidy. Decreased expression was defined as a twofold or greater decrease by RNA-sequencing (RNA-seq) analysis. LN, lymph node. g, Diagram illustrating immune-related genetic defects in antigen processing and IFN response in the tumor biopsy from patient 3. h, Illustration depicting identified damaging mutations to the target HLA complex for each patient. Patient 5 had a missense mutation and patient 4 had a nonsense mutation in HLA-A*02:01. Patient 12 had CNV loss of B2M combined with a start-loss point mutation in the remaining copy of B2M.