This study tests the safety and preliminary anti-tumor activity of an investigational dual-target chimeric antigen receptor natural killer (CAR-NK) cell therapy in adults with advanced breast cancer. After a tumor antigen assessment (HER2/ERBB2, MUC1, ROR1, and in some TNBC cases mesothelin), each participant will receive the most suitable dual-target CAR-NK product for their tumor profile, following short-course lymphodepleting chemotherapy.

Natural killer (NK) cells can recognize and kill abnormal cells as part of the innate immune system. CAR engineering can enhance NK-cell recognition of tumor-associated antigens and may improve anti-tumor activity in solid tumors. This is a two-part, first-in-program study. Part A uses dose escalation to identify a safe and feasible dose of a dual-target CAR-NK cell product. Part B evaluates preliminary efficacy in biomarker-defined expansion cohorts for HER2-positive breast cancer and triple-negative breast cancer (TNBC).

Target selection is biomarker-guided. A fresh or archival tumor sample is tested by immunohistochemistry (IHC) for HER2/ERBB2, MUC1, and ROR1 expression. Participants are assigned to one of three dual-target CAR-NK constructs based on a predefined algorithm prioritizing highest and most homogeneous target expression. In TNBC, mesothelin testing may be performed to support an exploratory sub-cohort.

All participants receive lymphodepleting chemotherapy (fludarabine + cyclophosphamide) before CAR-NK infusion. The CAR-NK product is an allogeneic, cryopreserved NK-cell therapy engineered to express a dual-specific CAR and an inducible safety switch; the product is administered intravenously.

Inclusion Criteria:

• Histologically confirmed breast carcinoma that is locally advanced, unresectable, or metastatic.

• Disease subtype: HER2-positive breast cancer or triple-negative breast cancer (TNBC).

• Progression after, intolerance to, or ineligibility for standard therapies appropriate for the disease subtype and line of therapy.

• At least one measurable lesion per RECIST v1.1.

• Tumor antigen assessment available (fresh or archival): expression of at least one candidate target antigen (HER2/ERBB2, MUC1, or ROR1). For TNBC, mesothelin assessment may be performed for exploratory analyses.

• ECOG performance status 0-1.

• Adequate organ function (example thresholds): ANC ≥ 1.0 x 10^9/L; platelets ≥ 75 x 10^9/L; hemoglobin

  • 8 g/dL; AST/ALT ≤ 3x ULN (≤ 5x with liver metastases); total bilirubin ≤ 1.5x ULN; creatinine clearance
  • 50 mL/min.

• Left ventricular ejection fraction (LVEF) ≥ 45% and no uncontrolled cardiac arrhythmia.

• Negative pregnancy test for participants of childbearing potential; agreement to use effective contraception during study treatment and for 6 months after last CAR-NK infusion.

• Ability to understand and willingness to sign informed consent.

Exclusion Criteria:

• Active, untreated central nervous system (CNS) metastases or leptomeningeal disease. Patients with treated CNS metastases may be eligible if clinically stable for ≥ 4 weeks and off high-dose steroids.
• Prior gene-modified cellular therapy (e.g., CAR-T or CAR-NK) within 6 months or unresolved grade ≥ 2 toxicity from prior cellular therapy.
• Clinically significant active autoimmune disease requiring systemic immunosuppression (physiologic steroid replacement permitted).
• Uncontrolled infection, including uncontrolled HBV, HCV, or HIV infection (controlled infections may be eligible per investigator).
• History of severe hypersensitivity to fludarabine or cyclophosphamide.
• Pregnant or breastfeeding.
• Concurrent participation in another interventional study that could confound safety or efficacy assessments.
• Any condition that, in the investigator's judgment, would make the participant unsuitable for the study (e.g., uncontrolled comorbidity, inability to comply with protocol procedures).

Open-label design; no blinding of participants, investigators, or outcome assessors.