Chimeric Antigen Receptor-Glypican-3 T-Cell Therapy for Advanced Hepatocellular Carcinoma: Results of Phase 1 Trials

Purpose: Our preclinical studies demonstrated the potential of chimeric antigen receptor (CAR)-glypican-3 (GPC3) T-cell therapy for hepatocellular carcinoma (HCC). We report herein the first published results of CAR-GPC3 T-cell therapy for HCC. Materials and Methods: In two prospective phase 1 studies, adult patients with advanced GPC3 + HCC (Child-Pugh A) received autologous CAR-GPC3 T-cell therapy following cyclophosphamide- and fludarabine-induced lymphodepletion. The primary objective was to as-sess the treatment’s safety. Adverse events were graded using the Common Terminology Criteria for Adverse Events (version 4.03). Tumor responses were evaluated using the Response Evaluation Criteria in Solid Tumors (version 1.1). Results: A total of 13 patients received a median of 19.9 × 10 8 CAR-GPC3 T cells by a data cutoff date of July 24, 2019. We observed pyrexia, decreased lymphocyte count, and cytokine release syndrome (CRS) in 13, 12 and 9 patients, respectively. CRS (grade 1/2) was reversi-ble in 8 patients. One patient experienced grade 5 CRS. No patients had grade 3/4 neurotoxicity. The overall survival rates at 3 years, 1 year, and 6 months were 10.5%, 42.0%, and 50.3%, respectively, according to the Kaplan-Meier method. We confirmed two partial responses. One patient with sustained stable disease was alive after 44.2 months. CAR T-cell expansion tended to be positively associated with tumor response. Conclusions: This report demonstrated the initial safety profile of CAR-GPC3 T-cell therapy. We observed early signs of antitumor activity of CAR-GPC3 T cells in patients with advanced HCC. rapidly exhibited elevation of IL-2, IL-6, IFN-γ and IL-10 levels. CRS was diagnosed on day 1 and treated with the tocilizumab and high-dose corticosteroid. The peak CAR-GPC3 DNA copy number was 10,713 copies/μg genomic DNA on day 7. The patient’s condition rapidly deterio-rated and the patient died on day 19.


Introduction
Hepatocellular carcinoma (HCC) is the most common histological subtype of liver cancer, which is the sixth most common cancer and the fourth leading cause of cancer-related death worldwide (1). Half of all liver cancer cases and deaths are estimated to occur in China (2). In the United States, the incidence of liver cancer has more than tripled since 1980 (3).
Considerable challenges exist in the clinical management of HCC. Only 15-20% of HCC cases are diagnosed at an early stage that may be suitable for curative treatment, such as surgical resection, liver transplantation, local ablation with percutaneous ethanol injection, microwave ablation, and radiofrequency ablation (4,5,6,7). Meanwhile, the majority of HCC patients have underlying chronic liver disease, and resection in this population is fraught with the potential for complications. Patients with intermediate-stage disease may benefit from local therapies such as transarterial chemoembolization. However, relapse is a frequent and expected event after systemic and local therapies (8). Also, many patients are diagnosed with unresectable advanced-stage HCC, with macroscopic vascular invasion and extrahepatic spread. Sorafenib, a targeted kinase inhibitor, was the first systemic therapeutic approved by the U.S. Food and Drug Administration for HCC based on improvement in overall survival (OS) duration by about 3 months (9,10). In addition, regorafenib and two programmed cell death protein 1 (PD-1) inhibitors, pembrolizumab and nivolumab, were recently approved for HCC treatment in patients who experienced progression after taking sorafenib (11,12,13). Most recently, lenvatinib, cabozantinib, and ramucirumab were approved for HCC patients who had progression after sorafenib treatment (14,15,16,17). The approval of anti-programmed cell death protein 1 inhibitors, pembrolizumab, and nivolumab, demonstrates that HCC is an immunosensitive tumor. Furthermore, tremelimumab, an anti-cytotoxic T-lymphocyte-associated protein 4 (CTL-4) antibody, produced response rates 6 of 17-26% in HCC patients (18,19). A pilot randomized trial of perioperative immunotherapy with nivolumab and ipilimumab (anti-CTLA-4) for resectable HCC is ongoing, with preliminary results demonstrating good antitumor activity (20). Despite progress in available therapies, effective systemic treatment options for HCC are still limited. Thus, its 5-year survival rate is a dismal 18%; this rate decreases to 11% and 5% if cancer has spread into surrounding tissues and distant parts of the body, respectively (21).
In recent years, chimeric antigen receptor (CAR) T-cell therapy has achieved significant efficacy in the treatment of hematologic malignancies (22,23,24,25,26). Although a breakthrough in disease control using CAR T-cell therapy for solid tumors has yet to be achieved, researchers have observed encouraging antitumor activity in early-phase clinical trials of therapies targeting interleukin (IL)-13R2 (27), mesothelin (28), and claudin18.2 (29,30), suggesting the feasibility of CAR T-cell therapy for solid tumors.
Glypican-3 (GPC3) is a member of the heparan sulfate proteoglycan family and attaches to cell surfaces via a glycosylphosphatidylinositol anchor. Recent studies demonstrated that GPC3 may be a prognostic marker for HCC, with greater GPC3 expression in tumor cells associated with worse prognosis (31). GPC3 is a molecule that is not fully understood in the role of proliferation and suppression of cell growth in normal tissues and abnormal or cancerous tissues.
Of note, our previous studies demonstrated that GPC3 was highly expressed in HCC and squamous non-small cell lung cancer but showed no expression in the kidney and gastric glands (32,33). These findings suggest that GPC3 is a rational immunotherapeutic target for HCC. Indeed, anti-GPC3 monoclonal antibodies had good safety profiles in previous studies (34,35), although significant clinical benefit has yet to be established in phase 2 clinical trials. Recently, we demonstrated that GPC3-targeted CAR T cells could eliminate GPC3 + HCC cells in vitro and 7 eradicate GPC3 + HCC tumor xenografts in mice (36,37,38). Therefore, in the present prospective phase 1 studies, we explored the safety and potential efficacy of CAR-GPC3 T-cell therapy in adult Chinese patients with advanced GPC3 + HCC. We hypothesized that CAR-GPC3 T cells approach was feasible and could be safely tolerated among patients with advanced HCC.

Clinical trial design and patients
Here we report two sequential phase I studies: Patients who met all screening criteria underwent leukapheresis to obtain peripheral blood mononuclear cells (PBMCs) for the generation of autologous CAR-GPC3 T cells. As described previously (36), CAR-GPC3 T cells (product code Y035) consisted of a humanized anti-GPC3 single-chain variable fragment, CD8 hinge domain, CD8 transmembrane domain, CD28 intracellular domain, and CD3 intracellular signaling domain that were cloned into a lentiviral backbone (Fig. 1A). After the CAR-GPC3 cells were manufactured, the patients were admitted to the hospital. Two to six days before CAR-GPC3 T-cell infusion, patients underwent lymphodepletion (Fig. 1B). In study 1, the lymphodepletion regimen used either (1) cyclophosphamide (Cy) 500-1000 mg/m 2 /day for 1-2 days, or (2) the combination of Cy (500 mg/m 2 /day for 1-3 days) and fludarabine (Flu, 25-30 mg/m 2 /day for 2-4 days). In study 2, all patients received the combination regimen of Cy (500 mg/m 2 /day for 1-2 days) and Flu (20-25 mg/m 2 /day for 3-4 days). The lymphodepletion regimen could be adjusted by the treating physician according to patient's disease condition (Supplementary Table S1). One subject (patient P5) was determined to be unfit for lymphodepletion due to poor clinical condition, with a heavy tumor burden and extremely elevated -fetoprotein (AFP) level of 60,500 ng/mL. After lymphodepletion, patients received a cycle of Y035 CAR-GPC3 T-cell therapy under observation and were not discharged until their absolute neutrophil counts recovered to at least 500 cells/mm 3 . In study 1, GPC3-CAR T cells were administrated from a starting split dose   Table 1).
After T-cell infusion, each patient's CAR T-cell DNA copy number and cytokine levels were monitored. The incidence and severity of adverse events (AEs) were graded using the Common Terminology Criteria for Adverse Events (version 4.03). Special attention was given to CAR T-cell-related infusion reactions, cytokine release syndrome (CRS), and preconditioningrelated infectious diseases. Early diagnosis and management of CRS were based on Lee's criteria (40). Tumor response was assessed using RECIST (version 1.1). Computed tomography (CT) or magnetic resonance imaging (MRI) was performed during follow-up. Patients were evaluated every 4-8 weeks for the first 6 months and then about every 3 months per the standard of care.

Quantitative real-time polymerase chain reaction analysis of CAR-GPC3 DNA copy numbers
Real-time fluorescent quantitative polymerase chain reaction (qPCR) was applied to determine the CAR-GPC3 DNA copy numbers in peripheral blood as previously described (41).

Immunological assays to measure blood concentrations of cytokines and tumor biomarkers
Both plasma and serum samples were collected from patients before and at different time points after CAR T-cell infusion. Specific circulating cytokines testing was performed on the same type of samples. The plasma concentrations of IL-6, IL-12p70, interferon (IFN)-γ, Regulated upon Activation, Normal T cell Expressed, and Secreted (RANTES), macrophage inflammatory protein (MIP)-1β and monocyte chemoattractant protein (MCP)-1 were measured using Cytometric Bead Array (CBA) according to the manufacturer's instructions (BD Biosciences; catalog numbers 558276, 558283, 560111, 558324, 558288, and 558287, respectively). The serum concentration of C-reactive protein (CRP) was measured using a high-sensitivity CRP assay with a BN II System (Siemens Healthineers). The serum AFP concentrations were measured using a chemiluminescent immunoassay according to the manufacturer's instructions (ARCHITECT i2000 chemiluminescence immunoassay analyzer, automated; Abbott Diagnostics).

Statistical analysis
Patients in study 1 and study 2 were analyzed together. Descriptive statistics consisted of medians with ranges and means with standard deviations for continuous variables and counts and percentages for categorical variables. AE terms were coded using the Medical Dictionary for Regulatory Activities (version 21.1). Analyses of the association of DNA copy numbers or cytokine levels with CRS and time points of tumor response were performed using SAS software (version 9.4; SAS Institute). Progression-free survival (PFS) and OS were analyzed using the Kaplan-Meier method. SAS software (version 9.4; SAS Institute) was used to calculate pharmacokinetic parameters using trapezoidal rule.

Patient characteristics
A total of 13 HCC patients received Y035 CAR-GPC3 T-cell therapy from December 2015 to August 2018. P1-P8 were enrolled in study 1. P9-P13 were enrolled in study 2. The patients (11 male and 2 female) had a median age of 51 years (range, 34-70 years) ( Table 1). All patient tumors were positive for GPC3 according to immunohistochemical staining, including 10 patients with a staining intensity score of 3+. All patients had Child-Pugh class A disease and three patients had cirrhosis. Ten patients had extrahepatic disease, nine had longer than a 1-year HCC history, and ten had Barcelona Clinic Liver Cancer stage C/D disease. Patients had received prior surgical resection, local therapy, or systemic therapy (Supplementary Table S2

Characteristics of Y035 CAR-GPC3 T cells
We successfully generated Y035 CAR-GPC3 T cells for all 13

Adverse effects
Y035 CAR-GPC3 T-cell therapy was generally tolerable in patients who had low tumor burdens and were in good clinical condition, even with doses greater than 20.0 × 10 8 CARpositive cells. All but one patient experienced an expected transient grade 3/4 decrease in lymphocyte count resulting from chemotherapy-induced lymphodepletion. We observed CRS in 9 patients ( Table 2, Supplementary Table S4). No patients experienced grade 3/4 neurotoxicity and no patients experienced CAR T-cell-related infusion reactions.

Persistence of Y035 CAR-GPC3 T cells in peripheral blood
The CAR-GPC3 vector copy number was closely monitored in the first 2 weeks after the initial infusion, weekly in the first month after the last infusion, and monthly thereafter. The median CAR-GPC3 DNA copy number in the peripheral blood of all patients increased rapidly, reaching a peak of 360.4 copies/g genomic DNA (range, 28.0-23,358.0 copies/g genomic DNA) after a median period of 10.5 days (mean, 13.8 days) and lasting for a median duration  (Fig. 3A). The areas under the curve (AUCs) for the CAR-GPC3 DNA copy numbers in two patients who had objective responses tended to be higher than those in the other 10 patients (Fig. 3B), with median AUCs of 98,016.0 and 5,423.1, respectively (P = 0.0606). In the two responders, we observed CAR T-cell DNA copy number peaks about 2 weeks after Y035 CAR-GPC3 T-cell infusion and before the first partial response (PR) (Fig. 3C).
The longest CAR T-cell DNA copy number duration was about 140 days.

Cytokine profile and cytokine release syndrome
We monitored patient serum levels of 12 cytokines during the studies. The median fold increase in IFN- and IL-6 level over baseline was 68.2 and 34.6, respectively. The median peak value of IFN- and IL-6 level post infusion was 3.98 and 172.82 pg/ml, respectively ( Fig. 3D-E).
The levels of CRP, IL-6, RANTES, and MCP-1 in patient P3 peaked at the same time as the patient's body temperature increase, which was consistent with clinical observations of CRS in patients who received Y035 Tcell infusion. In addition, we observed IL-6, IL-10, IL-15, and IFN- peaks before the CRS event in patient P13 and significant decreases in these levels after the event (Fig. 3F).
CRS is a major clinical concern for patients receiving CAR T-cell therapy (22, 25,26). Eight subjects experienced low-grade CRS (grade 1 or 2), which was self-limiting and reversible. All patients experienced fever: eleven at grade 1 or 2 and two at grade 3. We did not see a clinically meaningful difference in the incidence and severity of CRS when comparing pa-

Antitumor activity of CAR-GPC3 T-cell therapy
We evaluated all 13 patients for antitumor activity. Two patients were alive at the time of data analysis. The survival probabilities at 3 years, 1 year, and 6 months were 10.5%, 42.0%, and 50.3%, respectively, according to the Kaplan-Meier method, with a median OS duration of 278 days (39.7 weeks) (95% confidence interval, 48-615 days) (Fig. 4A-B).
Two patients (P3 and P13) had confirmed PRs (Fig. 4C), and one (P1) had a sustainable stable disease (SD). The target lesions in the two patients with PRs exhibited significant shrinkage. Their OS durations were 615 and 385 days, respectively, and their PFS durations were 111 and 99 days, respectively. Patient P13 was alive at the time of data analysis with an AFP level of 1.97 ng/mL ( Supplementary Fig. S1B). In addition, patient P1, despite having less than a 30% reduction in the size of target lesions, exhibited a significant clinical benefit of T-cell therapy, with a remarkable reduction in serum AFP level and continued long-term survival. The patient remained alive with a serum AFP level in the normal range at the time of data analysis ( Table 1, Supplementary Fig. S1B). Of note, this patient had vessel invasion (inferior vena cava tumor thrombus and right atrium tumor thrombus) and metastatic lymph nodes at baseline, which generally indicate a poor prognosis.
We monitored serum AFP levels throughout the studies. Patients P1, P3 and P13, who had clinically meaningful antitumor activity of PD or sustainable SD, had high-percentage reductions in serum AFP levels after Y035 CAR-GPC3 T-cell infusion (Supplementary Fig. S1A). We observed two objective responses to the Y035 CAR-GPC3 T-cell therapy. Additionally, one patient with stable disease also experienced long-term survival (44.2 months). Although a higher GPC3 expression in tumor cells is associated with a worse prognosis for HCC, we did not find a significant correlation between the staining intensity and clinical efficacy in the studies. This correlation will be intensively investigated in the ongoing studies. The antitumor activity of CAR-GPC3 T cells we reported here was more promising than the efficacy reported in the phase 1 study of monoclonal antibody (mAb) GC33. In the GC33 mAb study, the median OS duration The serum AFP level is an important biomarker of tumor response and disease relapse in HCC patients (42,43). Normalization of serum AFP levels in patients with high baseline AFP levels is a useful indicator of the success of surgical resection. Furthermore, decreased serum AFP level is associated with improved OS after treatment with sorafenib (44). Indeed, in the present study, the three patients with clinically meaningful benefits of the CAR T-cell treatment had greater reductions in serum AFP levels than did the other patients. This finding is consistent with the favorable antitumor activity of the treatment observed in the imaging of these patients. This report confirms that new CAR T-cell approaches with GPC3 should be considered for HCC. For example, in our previous study of immunocompetent and immunodeficient HCC mouse models (36), we demonstrated the combined antitumor effects of sorafenib and CAR-GPC3 T cells, suggesting that this combination therapy can be applied clinically to HCC. We are investigating whether a different lymphodepletion regimen in the combination of CAR-T cells and a tyrosine kinase inhibitor can be an alternative approach to replace the standard lymphodepletion regimen for patients with HCC, who commonly develop compromised hepatic function. Further improvements of the CAR-GPC3 with next-generation technology are also promising and may not require lymphodepletion pretreatment. For instance, our recent study on July 8, 2020. © 2020 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. demonstrated that IL-12 armored GPC3-redirected CAR T cells could greatly improve the antitumor activities in mouse model even with large tumor burdens (41). We also showed that introducing an IL-4/21 inverted cytokine receptor into the CAR-GPC3 construct improved the CAR T-cell potency in vitro and in vivo (37). In another recent study, our group disrupted PD-1 gene expression in CAR-GPC3 T cells using the CRISPR/Cas9 gene-editing system. This disruption enhanced the in vivo activity of CAR T cells against HCC and improved the persistence and infiltration of CAR T cells in mice bearing HCC (45). These new CAR-GPC3 T-cell therapeutic approaches are promising and should be explored in future HCC clinical trials.

Discussion
In summary, we report herein the initial safety profile of Y035 CAR-GPC3 T-cell therapy in patients with advanced HCC. Early signs of antitumor activities were observed. Future studies are warranted to further confirm the safety and efficacy of CAR-GPC3 T cells.    was diagnosed on day 10 and resolved after tocilizumab and corticosteroid administration. The peak CAR-GPC3 DNA copy number was 23,358 copies/μg genomic DNA on day 15. P12 rapidly exhibited elevation of IL-2, IL-6, IFN-γ and IL-10 levels. CRS was diagnosed on day 1 and treated with the tocilizumab and high-dose corticosteroid. The peak CAR-GPC3 DNA copy number was 10,713 copies/μg genomic DNA on day 7. The patient's condition rapidly deteriorated and the patient died on day 19.