br CAR NK cells per

CAR-NK Temozolomide per injection. The mice were followed with
serial weekly imaging to assess the tumor burden. (A)
Effects of NK cells electroporated with NKG2Dp mRNA
on tumor burden over time in mice with HCT116-Luc
xenografts. Tumor burden over time by BLI is shown.
Each mouse is represented by one line. (B) Kaplan-Meier
analysis of survival in the HCT116-Luc tumor model.
Statistical analysis of survival between groups was per-
formed using the log-rank test. Medium survival days are
shown below.
B
including CAR-NK cell release criteria and the cell viability of the CAR-NK cells after electroporation are summarized in Table 1 and Figures S4 and S5. Autologous NK cells were prepared with 100 mL of blood from the first patient (1001), and allogeneic NK cells were prepared with 200 mL of blood from haploidentical family donors for the second and third patients (1002 and 1003). NK cell expansion was performed with K562 feeder cells expressing mbIL-15, 4-1BBL, and mbIL-21. Flow cytometry analysis was carried out before and after NK cell expansion and before and following transduction with NKG2Dp CAR construct. For the three patients, the mean percentage of CD56+ CD3 cells increased from initial 11.7% to 35.1% to 96.9% (±2.83%) after expansion and CD3+ T cell depletion. Electroporation was used to transfect NKG2Dp CAR mRNA into NK cells. Successful transfection was demonstrated by increase of NKG2D median fluo-rescent intensity (MFI) (Table 1) and in vitro cytotoxicity assays that confirmed the robust lysis of NKG2D ligand-positive HCT116 human colorectal carcinoma cells by CAR-NK cells (Table 1). Cells in the final products were above 90% viable before infusion. The treat- 
ment scheme for the three patients is detailed in Table 1 and illustrated in Figure S6.
Adverse Events Experienced from CAR-NK Cell Therapy
We evaluated adverse events of any grade attrib-utable to any cause in the three patients according to Common Terminology Criteria for Adverse Events (CTCAE) v4.0 (Table 2). No CAR-NK dose reduction was required and, therefore, all patients received scheduled doses. There were no dose-limiting toxicities. No serious adverse effects (Rgrade 3 adverse events) were noted in any of the three patients. Only grade 1 cytokine release syndrome was observed, and none of the patients experienced neurologic symptoms. The most common treatment-related adverse events reported included fever (Figures S7–S9), fatigue, and anorexia. Patient 1003 developed
fever after the second, fourth, fifth, and sixth infusions that peaked at 40.0 C (Figure S9), associated with increased plasma interleukin (IL)-6 levels peaked at 224 pg/mL and C-reactive protein (CRP) levels peaked at 117.6 mg/L (Figure S10). The fever resolved after ice-pack application. GVHD was not observed in the two patients treated with haploidentical NK cells.
Adoptive Transfer of CAR-NK Cells into the Patient’s Peritoneal
Cavity Shows a Clinical Benefit in Controlling Malignant Ascites
Two patients, 1001 and 1002, with substantial disease burdens and malignant ascites (Table 3) received multiple rounds of intraperitoneal NK cell infusions in a dose-escalation manner (Table 1; Figure S6). Patient 1001 received two infusions of CAR-NK cells (2 107 and 1 108) prepared with autologous NK cells, while patient 1002 was treated with four infusions of CAR NK cells (1 108, 3 108,
5 108, and 7 108) prepared with allogeneic haploidentical NK cells. Direct local injection into the peritoneal cavity allows the quick access of CAR-NK cells to tumor sites, facilitating the interaction between
Molecular Therapy
Table 1. NKG2Dp CAR-NK Cells Used in the Clinical Trial
NK Cell
Specification
Cell Expansion
Autologous or haploidentical
autologous
haploidentical
haploidentical
Blood vol. collected (mL)
initial
before T cell depletion
after T cell depletion
before infusion
NK cell expansion fold
CAR-NK Cell Release Criteria
Mycoplasmaa
negative
negative
negative
negative
Sterility (bacterial)a
negative
negative
negative
negative
NKG2D MFI after transfection
Cancer cell killing activity
Treatment Scheme
Treatment frequency
once a week
once a week
twice a week
Total infusion/injection times
Dose level (E8 cells per infusion)
MFI, median fluorescent intensity.
aResults were not available at the time of infusion. bHCT116 human colorectal carcinoma cell line.