Purpose Adoptive transfer of genetically modified T cells is being explored as a treatment for patients with metastatic cancer. Most current strategies use genes that encode major histocompatibility complex (MHC) class I-restricted T-cell receptors (TCRs) or chimeric antigen receptors to genetically modify CD8+ T cells or bulk T cells for treatment. Here, we evaluated the safety and efficacy of an adoptive CD4+ T-cell therapy using an MHC class II-restricted, HLA-DPB1*0401-restricted TCR that recognized the cancer germline antigen, MAGE-A3 (melanoma-associated antigen-A3). Patients and Methods Patients received a lymphodepleting preparative regimen, followed by adoptive transfer of purified CD4+ T cells, retrovirally transduced with MAGE-A3 TCR plus systemic high-dose IL-2. A cell dose escalation was conducted, starting at 107 total cells and escalating at half-log increments to approximately 1011 cells. Nine patients were treated at the highest dose level (0.78 to 1.23 × 1011 cells). Results Seventeen patients were treated. During the cell dose-escalation phase, an objective complete response was observed in a patient with metastatic cervical cancer who received 2.7 × 109 cells (ongoing at ≥ 29 months). Among nine patients who were treated at the highest dose level, objective partial responses were observed in a patient with esophageal cancer (duration, 4 months), a patient with urothelial cancer (ongoing at ≥ 19 months), and a patient with osteosarcoma (duration, 4 months). Most patients experienced transient fevers and the expected hematologic toxicities from lymphodepletion pretreatment. Two patients experienced transient grade 3 and 4 transaminase elevations. There were no treatment-related deaths. Conclusion These results demonstrate the safety and efficacy of administering autologous CD4+ T cells that are genetically engineered to express an MHC class II-restricted antitumor TCR that targets MAGE-A3. This clinical trial extends the reach of TCR gene therapy for patients with metastatic cancer.