Chimeric Antigen Receptor T-cell Therapies for Lymphoma

Abstract

New therapies are needed for patients with Hodgkin and Non-Hodgkin lymphomas that are resistant to standard therapies. Lack of response to standard cytotoxic chemotherapy and relapse after autologous stem cell transplant are especially poor prognostic indicators. Chimeric antigen receptor (CAR) T cells are emerging as a novel treatment for these patients. In this review, we discuss the efficacy, toxicity, limitations, and future directions of CAR T-cell therapy for lymphoma. Clinical trial data of anti-CD19 CAR T cells have demonstrated potent activity against multiple B-cell lymphoma subtypes, including diffuse large B-cell lymphoma, follicular lymphoma, mantle cell lymphoma, and marginal zone lymphoma. Importantly, anti-CD19 CAR T cells have impressive activity against chemotherapy-refractory lymphoma. Some patients with chemotherapy-refractory diffuse large B-cell lymphoma have obtained durable complete remissions lasting 2 years or more after anti-CD19 CAR T-cell infusions. CAR T cells are associated with significant toxicities, including cytokine-release syndrome and neurologic toxicity. New target antigens for CAR T-cell therapies of lymphoma, including CD20 and CD22 for B-cell lymphomas and CD30 for Hodgkin and T-cell lymphomas, are being studied. CAR T cells are being tested in industry-sponsored multicenter clinical trials, and CAR T cells will likely become a standard lymphoma therapy soon. Future directions include improved toxicity management, CAR design changes aimed at improving the efficacy to toxicity ratio of CAR T cells, and combination therapies of CAR T cells plus other novel agents.

New therapies are needed for relapsed and refractory lymphoma

Non-Hodgkin lymphoma (NHL) is a heterogeneous malignancy that comprises approximately 4% of all cancers in the United States¹. While the development of the anti-CD20 monoclonal antibody rituximab has improved treatment outcomes for multiple B-cell NHL subtypes^2–4, prognoses for patients with chemotherapy-refractory or multiply-relapsed B-cell NHL remain poor^5–8. The standard approach for DLBCL patients with residual or relapsed lymphoma after first-line therapy is salvage chemotherapy followed by auto-HSCT^{9, 10}. Diffuse large B-cell lymphoma (DLBCL) is curable with front-line anthracycline-based regimens, but a substantial number of patients have relapsed or primary-refractory lymphoma^{6, 11}. Primary-refractory DLBCL, which is defined as a response of less than partial remission to a first-line regimen, has especially poor outcomes; only 23–29% of patients with DLBCL that is refractory to first-line chemotherapy respond to second-line treatment^{5, 12}, with a median progression-free survival (PFS) of 3 months¹². Patients with DLBCL refractory to second-line therapy have a median overall survival (OS) of only 4 months⁵, with an estimated response to third-line chemotherapy of only 14%¹³. The prognosis of patients with DLBCL that has relapsed after autologous stem cell transplantation (auto-HSCT) is also grim, with median OS of 10 months^{7, 14}. While the majority of patients with follicular lymphoma (FL) respond to first-line chemo-immunotherapy^15–17, these conventional therapies are not curative. A subgroup of approximately 20% of FL patients with progressive lymphoma within 24 months of initial treatment have poor long-term outcomes, with a median OS of less than 5 years⁸. Mantle cell lymphoma (MCL) has traditionally been managed with aggressive chemotherapy regimens followed by auto-HSCT in younger patients, an approach with significant toxicity and without curative potential^{18, 19}. The introduction of novel targeted therapies for MCL, such as bortezomib²⁰, lenalidomide²¹, and ibrutinib²², yield responses that are usually not durable beyond 2 years²³.

Standard first-line therapy for peripheral T-cell lymphoma (PTCL) consists of anthracycline-based regimens^{24, 25}. Unfortunately, the response to therapy with T-cell lymphomas is generally not as robust as it is in B-cell NHL, with a lower overall survival in T-NHL compared with B-NHL^26–28. Only a minority of cases of relapsed or refractory PTCL respond to second-line therapy with targeted agents^29–31.

Hodgkin lymphoma (HL) is often cured with first-line anthracycline-based regimens with or without radiation therapy, but approximately half of patients with relapsed or primary-refractory HL do not obtain durable remissions with the standard treatment of salvage chemotherapy followed by auto-HSCT^{32, 33}. In recent years, novel therapies have provided alternatives for these patients. Brentuximab monotherapy has shown an overall response rate (ORR) of 75%³⁴, but the 5-year PFS has been reported at 22%³⁵. While reported ORR are high with the checkpoint inhibitors nivolumab^{36, 37} and pembrolizumab³⁸, complete remissions (CRs) are the minority of the responses, and median duration of response to nivolumab is just 13.1 months³⁹.

While allogeneic hematopoietic stem cell transplantation (allo-HSCT) offers the potential for cure for multiple lymphoma subtypes, transplant-related mortality remains high, and long-term sequelae, including chronic graft versus host disease (cGVHD), have the potential for substantial negative impact on quality of life^40–43.

Chimeric Antigen Receptors

Chimeric antigen receptors (CARs) are fusion proteins that include both antigen recognition moieties and T-cell activation domains (Figure 1)^44–47. The antigen-recognition domain is usually a single-chain variable fragment (scFv) derived from a monoclonal antibody^{44–46, 48–51}. T-cell signaling domains are of two types, costimulatory domains such as CD28 and 4–1BB and T-cell activation domains such as CD3ζ^{48, 52–55}. The gene encoding a CAR is transferred to T cells by using a gene therapy vector, which is usually a replication-incompetent gamma-retrovirus^56–61, or a lentivirus^62–64, although transposon systems have also been used⁶⁵. Most CAR T-cell infusions are preceded by a conditioning chemotherapy regimen, because extensive evidence from murine studies indicates that recipient lymphocyte depletion enhances the activity of adoptively-transferred T cells^66–69. The first preclinical studies of CAR T cells as cancer treatment were conducted in the early 1990s⁷⁰, and the first preclinical studies of anti-CD19 CARs were published in 2003^{71, 72}. In this review, we will discuss the clinical results of CAR T-cell therapy for lymphoma and discuss therapeutic challenges and ongoing improvements of CAR T-cell therapy. CAR T cells have been shown to have impressive anti-lymphoma activity, but also substantial toxicity.

Figure 1: Chimeric antigen receptor structure.

1a. A diagram is shown of a second-generation CAR, including a single chain variable region (scFv), hinge and transmembrane regions, a costimulatory domain, and a T-cell activation domain. 1b. First generation CARs incorporate a T-cell activation domain but lack a costimulatory domain. Second generation CARs include both a T-cell activation domain and a single costimulatory domain. In this example the costimulatory domain is either CD28 or 4–1BB. Third generation CARs incorporate two costimulatory domains, as well as a T-cell activation domain.

Early Clinical Results with CAR T-cell Therapy in Indolent B-cell Lymphomas

The first published report of biological activity of anti-CD19 CAR T cells in humans was published in 2010⁷³. This report described a patient with multiply-relapsed FL who was treated at the U.S. National Cancer Institute (NCI)⁷³. This patient received a conditioning chemotherapy regimen of cyclophosphamide and fludarabine followed by an infusion of autologous T cells transduced with an anti-CD19 CAR containing a CD28 costimulatory domain; this CAR was designated FMC63–28Z⁴⁸. The CAR T-cell infusion was followed by a course of high-dose interleukin-2 (IL-2)⁷³. The patient achieved a partial remission (PR) lasting 7 months. The lymphoma then progressed, and the patient was re-treated in an identical manner with chemotherapy, anti-CD19 CAR T cells and IL-2. This patient’s lymphoma continues to be progression-free 7 years after this second treatment, and the patient is free of any toxicities⁷⁴. He has received no further lymphoma therapy since his second CAR T-cell infusion in March, 2010. The patient also experienced prolonged B-cell aplasia due to the CAR T-cell-mediated eradication of all cells carrying the CD19 antigen⁷³. This same group reported successful treatment of more patients with indolent lymphomas⁷⁴. Patients were treated with cyclophosphamide and fludarabine, followed by infusion of anti-CD19 CAR T-cells and high-dose IL-2⁷⁴. This report confirmed that CAR T cells can eradicate lymphoma masses, as well as malignancy in the bone marrow⁷⁴. This report also described a correlation between serum interferon gamma (IFNγ) and tumor necrosis factor-α and overall toxicity experienced by patients⁷⁴. This syndrome of toxicity after CAR T-cell infusions that was associated with elevated serum cytokine levels was named cytokine-release syndrome (CRS)^{57, 58, 75}.

Other early studies of anti-CD19 CAR T cells for lymphoma were also conducted. Two patients with follicular lymphoma treated on a clinical trial of T cells transduced with an anti-CD19 CAR incorporating a CD3ζ T-cell activation domain, without a costimulatory domain, did not achieve anti-lymphoma responses⁷⁶. In another trial, 5 patients with relapsed or refractory B-cell NHL simultaneously received 2 different CAR T cell products. These patients received T cells transduced with a CAR containing an anti-CD19 sFv and a CD3ζ T-cell activation domain along with T cells transduced with a CAR that was identical except for addition of a CD28 costimulatory domain⁵⁹. While the best malignancy response was stable disease, the CAR T cells incorporating a costimulatory domain exhibited superior in vivo expansion and persistence⁵⁹.

CAR T-cell therapy for diffuse large B-cell lymphoma and indolent lymphomas

A trial that evaluated the FMC63–28Z CAR⁴⁸ at the NCI was the first to demonstrate effectiveness of anti-CD19 CAR T cells against DLBCL⁵⁶. Patients received lymphocyte-depleting cyclophosphamide and fludarabine chemotherapy followed by a single infusion of CAR T cells. The study included 11 patients with NHL: 4 with DLBCL not otherwise specified (NOS), 4 with PMBL, one with splenic marginal zone lymphoma, one with Richter’s transformation of chronic lymphocytic leukemia (CLL), and one with indolent NHL not otherwise specified. Infiltration of malignant lymph nodes by CAR T cells was demonstrated⁵⁶. Notably, remissions were achieved in patients with chemotherapy refractory DLBCL NOS or PMBL, populations of patients with poor prognoses and few standard treatment options^{5, 12, 77, 78}. High levels of blood CAR T-cell persistence were shown to not be necessary for sustained remissions. While B-cell aplasia was common, recovery of normal polyclonal B cells was demonstrated in absence of lymphoma recurrence⁵⁶.

Thirty-two patients with NHL received infusions of T cells transduced with an anti-CD19 CAR containing a 4–1BB costimulatory domain⁷⁹. The CAR T cells were produced in a defined 1:1 ratio of CD4⁺/CD8⁺ cells with the intent of administering a more homogenous and possibly more efficacious cell product. Responses were achieved in all lymphoma subtypes including DLBCL, transformed follicular lymphoma (TFL), FL, and MCL, with an ORR of 63% and a CR rate of 33%⁷⁹. Twenty patients receiving fludarabine as part of their conditioning chemotherapy achieved an ORR of 72% and a CR rate of 50%, compared to an ORR of 50% and CR rate of 8% among 12 patients not receiving fludarabine⁷⁹. Patients receiving fludarabine-containing conditioning chemotherapy regimens experienced greater CAR T-cell expansion and higher serum levels of IL-15 and IL-7 than patients receiving chemotherapy regimens without fludarabine prior to the cell infusion⁷⁹. Higher CAR T-cell doses posed added risk of CRS and neurotoxicity⁷⁹. Two deaths, due to pontine hemorrhage and gastrointenstinal hemorrhage, occurred at the 2 × 10⁷ CAR⁺ T cells/kg dose, leading the investigators to administer lower cell doses in all subsequent patients⁷⁹. Poor CAR T-cell proliferation in patients who did not receive fludarabine was found in some patients to be associated with development of a cytotoxic T-cell immune response directed against the CAR transgene⁷⁹. Similar immune responses against the CARs have been reported from other clinical trials of NHL⁷⁶ and ALL^{64, 80}.

Recent results from the NCI demonstrated that a low-dose cyclophosphamide plus fludarabine conditioning regimen administered prior to infusion of anti-CD19 CAR T cells was sufficient to deplete lymphocytes and increase recipient serum cytokines, including IL-15 and IL-7⁶¹. Twenty-two patients with various NHL subtypes received low-dose cyclophosphamide and fludarabine followed by an infusion of 1, 2 or 6 × 10⁶ CAR⁺ T cells/kg. The ORR was 73% and the CR rate was 55% for the total study group⁶¹. Nineteen patients had one of the various types of DLBCL, and the ORR rate among this group was 68%, with 47% CRs⁶¹. The low-dose chemotherapy used in this study would not be expected to cause significant antimalignancy activity against chemotherapy-refractory DLBCL. Anti-CD19 CAR T cells produced durable remissions in this study, with ongoing CR durations of 24 to 7 months at the time of publication⁶¹. Of the 12 CRs obtained in this study, 11 are currently ongoing⁶¹. Twelve-month PFS for all participants was 63%. This study showed that CRs were possible in patients with c-myc translocations and in so-called “double-hit” or c-myc and blc-2 translocated DLBCL⁶¹. High peak IL-15 levels were associated with high peak blood CAR⁺ cell levels⁶¹. Both peak blood CAR⁺ cell levels and high levels of serum IL-15 were associated with patients obtaining lymphoma responses of PR or CR⁶¹. Toxicities related to CRS and to neurologic effects were manageable with supportive care and intensive care in some cases. Transfusion-dependent thrombocytopenia occurred in only 2 of 22 patients, and most patients did not experience a neutrophil count less than 500/μL⁶¹. The most prominent toxicities were neurologic with 55% of patients experiencing Grade 3 or 4 neurologic toxicities that resolved in all cases⁶¹. Recovery of normal polyclonal B cells in patients in CR with absence of malignancy recurrence was demonstrated⁶¹.

Recent data from the University of Pennsylvania confirmed that anti-CD19 CAR T cells with a 4–1BB costimulatory domain (CTL019) have significant antimalignancy efficacy in various NHL subtypes, including DLBCL, FL, and MCL, with an ORR of 68% and PFS of 62% at 11.7 months⁸¹.

Recent results have demonstrated the safety and feasibility of administration of anti-CD19 CAR T cells following auto-HSCT for patients with DLBCL, TFL and MCL⁸². These results included data derived from two trials of anti-CD19 CAR central memory-enriched T cells⁸². Published clinical trial results of CARs targeting CD19 as a treatment for lymphoma are summarized in Table 1. Anti-CD19 CAR T-cell clinical trials that are currently recruiting lymphoma patients are listed in Table 2.

Table 1:

Selected published clinical trials of autologous anti-CD19 CAR T-cell therapy for patients with NHL

Publication	Lymphoma subtypes included	Costimulatory domain	Conditioning chemotherapy	Cell dose	ORR and CR rates in NHL patients*
Jensen et al., ASBMT 201076	FL	None	Flu 25 mg/m2 daily for 5 days starting after the first of multiple CAR T-cell infusions	4–5 escalating doses of 1 × 108 – 2 × 109 cells/m2	ORR: 0/2 (0%)
Kochenderfer et al., Blood 201073	FL	CD28	60 mg/kg Cy daily for 2 days followed by Flu 25 mg/m2 daily for 5 days	1 × 108 CAR+ T cells one day, followed by 3 × 108 CAR+ T cells the following day	ORR: 1/1 (100%) CRR: 0/1 (0%)
Savoldo et al., J. Clin Invest 201159	DLBCL TFL PCNSL, systemic relapse	One product without a costimulatory domain and one product with a CD28 costimulatory domain	None	1–2 infusions of two simultaneous cell products in escalating doses of 2 × 107 – 2 × 108 cells/m2	ORR: 0/5 (0%)
Kochenderfer et al., Blood 201274	FL SMZL	CD28	60 mg/kg Cy daily for 2 days followed by Flu 25 mg/m2 daily for 5 days	0.3 – 3.0 × 107 CAR+ T cells/kg	ORR: 3/3 (100%)** CRR: 0/3 (0%)
Kochenderfer et al., JCO 201556	DLBCL PMBL SMZL indolent NHL NOS DLBCL transformed from CLL	CD28	Cy total dose either 120 or 60 mg/kg followed by Flu 25 mg/m2 daily for 5 days	1 – 5 × 106 CAR+ T cells/kg	ORR: 8/9 (89%)*** CRR: 5/9 (56%)
Wang et al., Blood 201682	DLBCL MCL	One clinical trial (NHL1) of a CAR not incorporating a costimulatory domain and one clinical trial (NHL2) of a CAR incorporating a CD28 costimulatory domain	Auto-HSCT conditioninĝ	First generation CAR trial (NHL1): 25, 50, or 100 × 106 total T cells per infusion Second generation CAR trial (NHL2): 50 or 200 × 106 CAR+ T cells per infusion	NHL1 ORR:^^ 7/8 (88%) NHL1 CRR: 5/8 (63%) NHL2 ORR:^^^ 8/8 (100%) NHL2 CRR: 8/8 (100%)
Turtle et al., Sci Transl Med 201679	Aggressive B-cell NHL TFL PMBL MCL FL	4–1BB	One of 4 regimens: Cy 2–4 g/m2 on day 1 Cy 2–4 g/m2 on day 1 and etoposide 100–200 mg/m2 per day on days 1–3 Cy 60 mg/kg on day 1 and Flu 25 mg/m2 daily on days 2–4. Cy 60 mg/kg on day 1 and Flu 25 mg/m2 daily on days 2–6.	2 × 105, 2 × 106, or 2 × 107 CAR+ T cells/kg	ORR: 19/30 (63%) CRR: 10/30 (33%)
Locke et al., Mol Ther 201783	DLBCL	CD28	Cy 500 mg/m2 IV daily for 3 days and Flu 30 mg/m2 daily for 3 days	1 – 2 × 106 CAR+ T cells/kg	ORR: 5/7 (71%) CRR: 4/7 (57%)
Kochenderfer et al. JCO 201761	DLBCL TFL FL PMBL MCL	CD28	Cy 300 mg/m2 or Cy 500 mg/m2 IV daily for 3 days and Flu 30 mg/m2 daily for 3 days	1, 2 or 6 × 10 6 CAR+ T cells/kg	ORR: 16/22 (73%) CRR: 12/22 (55%)

Auto-HSCT: autologous stem cell transplant. CLL: chronic lymphocytic leukemia. Cy: cyclophosphamide. CRR: completion remission rate. DLBCL: diffuse large B-cell lymphoma. Flu: fludarabine. FL: follicular lymphoma. MCL: mantle cell lymphoma. NHL: Non-Hodgkin lymphoma. NOS: Not otherwise specified. ORR: overall response rate. PMBL: primary mediastinal B-cell lymphoma. PCNSL: primary central nervous system lymphoma. SMZL: splenic marginal zone lymphoma. TFL: transformed follicular lymphoma.

^*Response rates take into account only patients with NHL treated on study. Patients with other B-cell malignancies treated in each trial are not included in overall response rate.

^**One patient’s response was not evaluable due to death on study.

^***Two patients’ response not evaluable, one due to death on study and one due to noncompliance with follow up.

^{^}All patients received bis-chloroethylnitrosourea, etoposide, cytarabine, and melphalan high-dose chemotherapy as conditioning and received CAR T-cell infusions following Auto-HSCT.

^{^^}Prior to Auto-HSCT, 3 patients were in CR and 2 patients were in PR.

^{^^^}Prior to Auto-HSCT, 6 patients were in CR and 2 patients were in PR.

Table 2:

Open Clinical Trials of anti-CD19 CAR T cells for Lymphoma

Trial identifier	Trial title	Lymphoma subtype	Center; location	Comments
NCT02529813	CD19+ CAR T Cells for Lymphoid Malignancies	CD19+ B-cell NHL	University of Texas MD Anderson Cancer Center; Houston, Texas, United States
NCT02030834	Phase IIa Study of Redirected Autologous T Cells Engineered to Contain Anti-CD19 Attached to TCRz and 4–1BB Signaling Domains in Patients With Chemotherapy Relapsed or Refractory CD19+ Lymphomas	DLBCL FL TFL MCL	University of Pennsylvania; Philadelphia, Pennsylvania, USA
NCT01865617	Laboratory Treated T Cells in Treating Patients With Relapsed or Refractory Chronic Lymphocytic Leukemia, Non-Hodgkin Lymphoma, or Acute Lymphoblastic Leukemia	CD19+ B-cell NHL	University of Washington Cancer Consortium; Seattle, Washington, USA
NCT02842138	Autologous CD19 CAR T Cells in Relapsed or Refractory B-cell Lymphoma	CD19+ B-cell NHL	Peking University Cancer Hospital; Beijing, China
NCT02976857	A Phase 1 Study Evaluating Safety and Efficacy of C-CAR011 Treatment in DLBCL Subjects (C-CAR011)	DLBCL	People’s Hospital of Jiangsu Province; Nanjing City, Jiangsu, China
NCT02081937	CART-19 Immunotherapy in Mantle Cell Lymphoma	MCL	Chinese PLA General Hospital; Beijing, China
NCT02728882	Study Evaluating the Efficacy and Safety With CAR-T for Recurrent or Refractory Diffuse Large B Cell Lymphoma (EECBL)	DLBCL	The First Affiliated Hospital of Anhui Medical University; Hefei, Anhui, China
NCT02659943	T Cells Expressing a Fully-human Anti-CD19 Chimeric Antigen Receptor for Treating B-cell Malignancies	CD19+ B-cell NHL	National Cancer Institute; Bethesda, Maryland, USA
NCT01087294	Administration of anti-CD19-chimeric-antigen-receptor-transduced T cells from the original transplant donor to patients with recurrent or persistent B-cell malignancies after allogeneic stem cell transplantation	All B-cell NHL	National Cancer Institute; Bethesda, Maryland, USA	Allogeneic hematopoietic stem cell transplant donors are the T-cell source.
NCT01430390	In Vitro Expanded Allogeneic Epstein-Barr Virus Specific Cytotoxic T-Lymphocytes (EBV-CTLs) Genetically Targeted to the CD19 Antigen in B-cell Malignancies	CD19+ B-cell NHL	Memorial Sloan Kettering Cancer Center; New York, New York, USA	Donor-derived allogeneic CAR T cells
NCT02652910	Memory-enriched CAR-T Cells Immunotherapy for B Cell Lymphoma (MeCAR)	CD19+ B-cell NHL	Xinqiao Hospital of Chongqing; Chongqing, China
NCT01853631	Activated T-Cells Expressing 2nd or 3rd Generation CD19-Specific CAR, Advanced B-Cell NHL, ALL, and CLL (SAGAN) (SAGAN)	CD19+ B-cell NHL	Houston Methodist Hospital and Texas Children’s Hospital; Houston, Texas, USA
NCT02247609	Evaluation of 4th Generation Safety-designed CAR T Cells Targeting High-risk and Refractory B Cell Lymphomas (4SCAR19273)	CD19+ B-cell NHL	Peking University Cancer Hospital; Beijing, China
NCT02624258	Pilot Study of Non-Viral, RNA-Redirected Autologous T Cells in Patients With Refractory or Relapsed Hodgkin Lymphoma	HL	Children’s Hospital of Philadelphia
NCT02431988	Evaluation of CAR19 T-cells as an Optimal Bridge to Allogeneic Transplantation (COBALT)	CD19+ DLBCL	University College London Hospital; London, United Kingdom
NCT02050347	Activated T Lymphocytes Expressing CARs, Relapsed CD19+ Malignancies Post-Allo HSCT(CARPASCIO)	CD19+ NHL	Houston Methodist Hospital and Texas Children’s Hospital; Houston, Texas, USA
NCT02650999	Phase I/II Study of Pembrolizumab in Patients Failing to Respond to or Relapsing After Anti-CD19 Chimeric Antigen Receptor Modified T Cell Therapy for Relapsed or Refractory CD19+ Lymphomas	FL DLBCL MCL	University of Pennsylvania; Philadelphia, Pennsylvania, USA	Eligible patients must have received CTL019/CTL119 CAR T cells at the University of Pennsylvania.
NCT02706405	JCAR014 and Durvalumab in Treating Patients With Relapsed or Refractory B-cell Non-Hodgkin Lymphoma	DLBCL PMBL tDLBCL	University of Washington Cancer Consortium; Seattle, Washington, USA
NCT02348216	A Phase 1–2 Multi-Center Study Evaluating KTE-C19 in Subjects With Refractory Aggressive Non-Hodgkin Lymphoma (ZUMA-1) (ZUMA-1)	DLBCL PMBL TFL	Multi-center study, international; sponsored by Kite Pharma, Inc.
NCT02601313	A Phase 2 Multicenter Study Evaluating Subjects With Relapsed/Refractory Mantle Cell Lymphoma (ZUMA-2)	MCL	Multi-center study, all sites in USA; sponsored by Kite Pharma, Inc.
NCT02926833	A Study Evaluating KTE-C19 in Combination With Atezolizumab in Subjects With Refractory Diffuse Large B-Cell Lymphoma (DLBCL) (ZUMA-6)	DLBCL	Multi-center study, all sites in USA; sponsored by Kite Pharma, Inc.
NCT02445248	Study of Efficacy and Safety of CTL019 in Adult DLBCL Patients (JULIET)	DLBCL	Multi-center study, international; sponsored by Novartis Pharmaceuticals
NCT02631044	Study Evaluating the Safety and Pharmacokinetics of JCAR017 in B-cell Non-Hodgkin Lymphoma (NHL)	DLBCL PMBL tDLBCL FL grade 3B MCL	Multi-center study, all sites in USA; sponsored by Juno Therapeutics, Inc.

DLBCL: diffuse large B-cell lymphoma. FL: follicular lymphoma. MCL: mantle cell lymphoma. NHL: Non-Hodgkin lymphoma. tDLBCL: Diffuse large B-cell lymphoma transformed from indolent histology. TFL: transformed follicular lymphoma.

Information accessed from www.clinicaltrials.gov accessed on March 23, 2017.

Multicenter Industry Sponsored Clinical Trials Show That CAR T-cell Therapies Can be Applied to Large Numbers of Patients

FMC63–28Z-expressing CAR T cells^{48, 56, 61} have been designated axicabtagene ciloleucel for commercial development. A phase I multicenter trial of axicabtagene ciloleucel demonstrated feasibility of cell manufacturing and high overall responses rates (71%) in a small group of 7 patients with DLBCL either refractory to last chemotherapy or relapsed within 12 months of auto-HSCT⁸³. CRS and neurotoxicities were substantial, but could be managed across centers⁸³. Activity was shown in both the GCB and non-GCB subtypes of DLBCL⁸³. A phase II multicenter study of 101 patients with refractory aggressive NHL receiving axicabtagene ciloleucel reported an ORR of 82% and CR rate of 54%, demonstrating that CAR T-cell therapy can successfully be delivered to large numbers of patients^{84, 85}. At a median follow up of 8.7 months, the ongoing response rate was 44% with a CR rate of 39%⁸⁵. Early results of a multicenter trial of an anti-CD19 CAR T-cell product with a 4–1BB costimulatory domain, JCAR017, showed an ORR of 80% in 28 NHL patients with 60% attaining a CR, with no severe CRS and manageable neurologic toxicity⁸⁶. CTL019, CAR T cells expressing an anti-CD19 CAR with a 4–1BB costimulatory moiety, is now being evaluated in an international multicenter study for patients with DLBCL. Interim results including 51 patients with three or more months of follow-up demonstrate an ORR of 59% and CR rate of 43%, with median duration of response not reached⁸⁷.

Anti-CD19 CAR T cell therapies for lymphoma progressing after allogeneic stem cell transplantation

Patients with progressive B-cell malignancies after allo-HSCT have very limited treatment options^{88, 89}. Relapsed B-cell malignancies after allo-HSCT is often treated with infusions of unmanipulated allogeneic lymphocytes from the original transplant donor (DLI)^{90, 91}. DLIs have inconsistent efficacy and often cause graft-versus-host disease (GVHD), an immunologic attack against normal recipient tissues⁹². Transducing allogeneic T cells to express anti-CD19 CARs might have an advantage over standard DLIs by specifically targeting malignant cells. Administration of allogeneic anti-CD19 CAR T cells from an allo-HSCT recipient’s stem cell donor has been demonstrated to be feasible and safe in two clinical trials^93–95.

In one trial, eight patients with acute lymphoblastic leukemia (ALL) or CLL who had prior allo-HSCT received virus-specific cytotoxic T cells transduced with an anti-CD19 CAR incorporating a CD28 costimulatory domain⁹⁴. Two of 6 evaluable patients exhibited objective anti-malignancy activity⁹⁴. Toxicities were mild, and, interestingly, the allogeneic anti-CD19 CAR T cells did not induce GVHD⁹⁴.

Another trial of allogeneic anti-CD19 CAR T cells evaluated allogeneic T cells expressing the previously described FMC63–28Z CAR^{48, 93, 95}. Twenty patients with various B-cell malignancies, including ALL, CLL, DLBCL, and MCL were treated with donor-derived anti-CD19 CAR T cells⁹⁵. Unlike most CAR T-cell clinical trials, conditioning chemotherapy was not administered. Responses were observed in all groups⁹⁵. Importantly, no patient developed acute GVHD, and only one patient had progression of previously-existing chronic GVHD⁹⁵. Because this study did not include chemotherapy, it clearly demonstrated the anti-malignancy activity of anti-CD19 CAR T cells alone^{93, 95}. The pronounced lack of GVHD seen in this trial compared with standard DLIs is compelling and not fully explained. Persistence of CAR T cells was shown to be limited to 1 to 2 months, which is the most likely explanation for the lack of GVHD⁹⁵. Very rapid induction of remissions occurred on this trial; we hypothesize that remissions were induced and CAR T cells were eliminated before GVHD could occur⁹⁵. Subsequently, murine studies supported this hypothesis by showing CD28-costimulated CAR T cells undergo increased activation followed by exhaustion and deletion when stimulated with CD19; deletion of the T cells protected against GVHD⁹⁶. Other investigators are using gene-editing techniques to produce CAR T cells that lack T-cell receptor genes⁹⁷. The goal of these studies is to use CAR T cells from unrealated cell donors that are not HLA-matched with the recipient⁹⁷. In another clinical trial of anti-CD19 CAR T cells in the transplant setting, a transposon system was used to genetically-modify T cells for use as adjuvant therapy after autologous or allogeneic hematopoietic stem cell transplantation⁶⁵.

Targeting lymphoma antigens other than CD19

There are 2 main reason to target antigens other than CD19 for treatment of lymphoma. First, many types of lymphoma including Hodgkin lymphoma, T-cell lymphomas, and even some cases of B-cell lymphoma, do not express CD19⁹⁸. Second, treatment of CD19⁺ lymphomas with T cells might fail if CD19 expression is lost from some or all of the lymphoma cells after anti-CD19 CAR T-cell therapy. For these reasons, investigators are developing CARs targeting antigens other than CD19^{60, 76, 99–102}.

The success of the CD20-targeting monoclonal antibody rituximab makes CD20 an attractive target for CAR T-cell therapy in B-cell NHL. Loss of CD20 expression following rituximab therapy^{103, 104} may represent a possible challenge in anti-CD20 CAR T cell therapy. Two patients with DLBCL received T cells transduced with a first-generation anti-CD20 CAR immediately following auto-HSCT, making the response to the CAR T-cell infusion difficult to evalaute⁷⁶. One patient remained in an ongoing CR for 9 years, not uncommon from transplant alone, at time of publication and the other patient achieved a CR lasting 19 months⁷⁶. A clinical trial was conducted using T cells transduced through plasmid electroporation to express an anti-CD20 CAR with both CD28 and 4–1BB costimulatory domains. Patients with MCL and FL received conditioning chemotherapy with cyclophosphamide and 3 infusions of anti-CD20 CAR T cells followed by 14 days of subcutaneous IL-2⁹⁹. Two patients with no evaluable lymphoma prior to cell infusion remained in CR, and one patient had an objective PR. CAR T cells persisted for 9 to 12 months⁹⁹.

A third study was conducted with T cells transduced with a lentiviral vector to express an anti-CD20 CAR with a 4–1BB costimulatory domain¹⁰⁰. This trial included 7 patients with heavily pretreated DLBCL, who were given variable chemotherapy conditioning regimens prior to receiving escalating doses of CAR T cells over 3–4 days¹⁰⁰. Five out of 6 evaluable patients had an objective response¹⁰⁰. One patient who entered the study in partial remission achieved an ongoing CR lasting 14 months at time of publication¹⁰⁰. This patient received no conditioning chemotherapy prior to cell infusion. This patient had detectable anti-CD20 CAR T cells in the blood more than 6 months after cell infusion. CRS toxicities were observed in 4 patients, with notably delayed occurrence 3–8 weeks after cell infusion. Other notably adverse events included delayed tumor lysis syndrome and 2 episodes of gastrointestinal (GI) hemorrhage, one resulting in death, in patients with GI tract involvement of lymphoma¹⁰⁰. Results of a phase II study of the same anti-CD20 CAR in DLBCL and indolent NHL showed an ORR of 82% and a CR rate of 55%¹⁰¹. Responses occurred in one patient with FL and one patient with MCL¹⁰¹. Clinical results are early, but anti-CD20 CAR T cells have significant therapeutic potential.

Because the kappa (κ) light chain antigen is expressed on κ-restricted B-cell lymphomas, but is not expressed on all normal B cells, κ is an attractive target for CAR T-cell therapy as complete B-cell aplasia should be avoided. In a phase I trial, 7 patients with various NHL subtypes that expressed κ-light chain, including DLBCL, TFL, MCL and lymphoplasmacytic lymphoma (LPL), were treated with a κ light chain-targeted CAR⁶⁰. Cyclophosphamide conditioning was administered prior to CAR T-cell infusions when patients were not already lymphopenic⁶⁰. One patient achieved a PR and 2 patients achieved CRs. One CR was ongoing 32 months after CAR T-cell infusion. No adverse events were attributable to CRS⁶⁰.

The B-cell marker CD22 presents another target for B-cell malignancies. Anti-CD22 CAR T cells have undergone preclinical development and are currently being tested in early clinical trials for acute lymphoid leukemia and NHL^{105, 106}. Clinical trials including patients with CD22-expressing NHL are ongoing (NCT02721407, NCT02315612, NCT02794961, NCT02935153, NCT02903810).

Targeting CD30 in Hodgkin Lymphoma and Peripheral T-cell lymphoma

CD30 is a transmembrane receptor and a member of the TNF receptor superfamily¹⁰⁷. The malignant Reed-Sternberg cells of HL strongly express CD30¹⁰⁸. The efficacy of targeting CD30 to treat HL^{34, 109–112} and PTCL^113–115 has been demonstrated in multiple clinical trials of the antibody-drug conjugate brentuximab vedotin.¹¹⁶ A minority of HL patients receiving brentuximab monotherapy are progression-free at 5-year follow-up³⁵, suggesting that CD30-targeting treatments could be improved to increase the durability of remissions.

First-generation anti-CD30 CAR T cells were developed in the 1990s. Preclinical models demonstrated lysis of CD30-expressing cell lines in vitro^{117, 118}. Notably, the presence of soluble CD30 did not attenuate cytolysis, suggesting that CD30 that is shed into the blood from HL cells would not inhibit the efficacy of anti-CD30 CAR T cells in vivo¹¹⁷. Epstein Barr virus-specific cytotoxic T cells transduced with an anti-CD30 CAR have been shown to have activity against CD30⁺ tumor cell lines in vitro as well as in an in vivo xenograft mouse model^{119, 120}.

A phase I clinical trial of anti-CD30 CAR T cells with a CD28 costimulatory domain for patients with HL and ALCL demonstrated responses of 1 CR and 1 PR in a group of 9 patients, 7 with HL and 2 with ALCL^{121, 122}. No patients received conditioning chemotherapy, so the responses were clearly attributable to the anti-CD30 CAR T cells. There was no evidence of CRS¹²². Another phase I trial tested an anti-CD30 CAR with a 4–1BB costimulatory domain¹⁰². Eighteen patients (17 with HL and one with cutaneous ALCL) were treated on this trial; patients received a variety of chemotherapy conditioning regimens prior to cell infusion¹⁰². Seven patients had a response of PR, with a median PFS of 6 months¹⁰². CRS toxicities similar to those seen in clinical trials of anti-CD19 CAR T cells occurred. CAR T-cell persistence was less than 8 weeks in this study, but tumor biopsies showed efficient trafficking of T cells to lymphoma sites¹⁰². Published clinical trial results of CARs targeting lymphoma antigens other than CD19 are summarized in Table 3, and currently open and recruiting clinical trials of CAR T cells targeting antigens other than CD19 are listed in Table 4.

Table 3:

Selected published clinical trials of CAR T cells targeting antigens other than CD19 for patients with lymphoma

	CAR target antigen	Lymphoma subtypes included	Costimulatory domain	Conditioning chemotherapy	Cell Dose	ORR and CR rate
Jensen et al., ASBMT 201076	CD20	DLBCL	None	Patients underwent Auto-HSCT 28–37 days prior to CAR T-cell infusion	3 escalating doses of 1 × 108 – 1 × 109 CAR+ T cells/m2	ORR: 2/2 (100%)* CRR: 2/2 (100%)*
Till et al., Blood 201299	CD20	MCL FL	CD28 and 4–1BB	Cy 1000 mg/m2 on Day −2.	3 cell infusions 2–5 days apart at incremental doses of 1 × 108, 1 × 109, and 3.3 × 109 cells/m2	ORR: 3/3 (100%)** CRR: 2/3 (67%)**
Wang et al., Clin Immunol 2014100	CD20	DLBCL	4–1BB	One of 4 regimens: COED COD CHODE ESHAP	3–5 escalating daily doses of cells, total dose 0.36 – 2.35 × 107 CAR+ T cells/kg	ORR: 5/6 (83%)*** CRR: 1/6 (17%)***
Zhang et al., Signal Transduction and Targeted Therapy 2016101	CD20	DLBCL FL MCL PCMZL	4–1BB	One of 6 regimens: ° CHOP MACH FC EOCH CHOD CHODE or No conditioning	3–5 escalating daily doses of cells, total dose 0.41 – 1.46 × 107 CAR+ T cells/kg	ORR: 9/11 (82%) CRR: 6/11 (55%)
Ramos et al., J Clin Invest 201660	Kappa (κ) light chain	DLBCL TFL LPL	CD28	Cy 12.5 mg/kg 4 days prior to cell infusion or no conditioning	1.7 × 107 – 1.9 × 108 CAR+ T cells/m2	ORR: 3/7 (43%) CRR: 2/7 (29%)
Wang et al., Clin Cancer Res 2017102	CD30	HL C-ALCL	4–1BB	One of 7 regimens: FC PC GEMC EAMC GMC MAMC GMMC or No conditioning	1.1 – 2.1 × 107 CAR+ T cells/kg	ORR: 7/18 (39%) CRR: 0/18 (0%)

Auto-HSCT: Autologous stem cell transplant. C-ALCL: cutaneous anaplastic large cell lymphoma. CLL: chronic lymphocytic leukemia. CHODE: cyclophosphamide, doxorubicin, vincristine, dexamethasone, etoposide. COD: cyclophosphamide, vincristine, dexamethasone. COED: cyclophosphamide, vincristine, etoposide, dexamethasone. Cy: cyclophosphamide. CRR: completion remission rate. DLBCL: diffuse large B-cell lymphoma. EAMC: etoposide, cytarabine, mustargen, cyclophosphamide. EBV: Epstein-Barr Virus. FC: fludarabine and cyclophosphamide. ESHAP: methylprednisolone, etoposide, carboplatin, high-dose cytosine arabinoside. Flu: fludarabine. FL: follicular lymphoma. GEMC: gemcitabine, epirubicin, mustargen, cyclophosphamide. GMC: gemcitabine, mustargen, cyclophosphamide. GMMC: gemcitabine, mitoxantrone, mustargen, cyclophosphamide. MAMC: mitoxantrone, cytarabine, mustargen, cyclophosphamide. LPL: lymphoplasmacytic lymphoma. ORR: overall response rate. NHL: Non-Hodgkin lymphoma. PC: nab-paclitaxel and cyclophosphamide. PCMZL: primary cutaneous marginal zone lymphoma. TFL: transformed follicular lymphoma.

^*Response represents combined effect of Auto-HSCT followed by CAR T-cell infusions. Of two treated patients, one patient remained in CR for 9 years at time of publication and the other patient achieved a CR lasting 19 months.

^**Two patients with no evaluable disease prior to cell infusion remained in CR; one patient had an objective partial remission.

^***One patient not evaluable due to death on study.

^°Specifics of regimens not fully defined.

Table 4:

Open Clinical Trials of CAR T cells for lymphoma targeting antigens other than CD19

Trial identifier	Trial title	CAR Antigen	Lymphoma subtype	Center; location	Comments
NCT02965157	Pilot Study of Anti-CD20-CAR-engineered T Cells in Patients With Chemotherapy Resistant or Refractory CD20+ Lymphoma	CD20	CD20+ B-cell lymphoma	Chinese Academy of Medical Sciences Tumor Hospital; Beijing, China
NCT02721407	Anti-CD22 CAR-T Therapy for CD19-refractory or Resistant Lymphoma Patients (MendCART)	CD22	CD19+/CD22+ B-cell NHL	Xinqiao Hospital; Chongqing, China	Eligible patients must have received anti-CD19:TCRz:CD28 CAR-T cells at Xinqiao Hospital
NCT02259556	CD30-directed Chimeric Antigen Receptor T (CART30) Therapy in Relapsed and Refractory CD30 Positive Lymphomas (CART30)	CD30	CD30+ HL and NHL	Chinese PLA General Hospital; Beijing, China
NCT03049449	T Cells Expressing a Fully-Human Anti-CD30 Chimeric Antigen Receptor for Treating CD30-Expressing Lymphomas	CD30	CD30+ HL and NHL	National Cancer Institute; Bethesda, Maryland, USA
NCT02274584	CAR T Cells Targeting CD30 Positive Lymphomas (4SCAR30273)	CD30	CD30+ lymphoma	University of Florida; Gainesville, Florida, United States, Peking University Cancer Hospital; Beijing, China
NCT02690545	Study of CD30 CAR for Relapsed/Refractory CD30+ HL and CD30+ NHL	CD30	CD30+ HL and NHL	Lineberger Comprehensive Cancer Center at University of North Carolina - Chapel Hill; Chapel Hill, North Carolina, United States
NCT02958410	A Clinical Research of CD30-Targeted CAR-T in Lymphocyte Malignancies	CD30	CD30+ Lymphoid malignancy	Southwest Hospital of Third Military Medical University; Chongqing, China
NCT02663297	Administration of T Lymphocytes for Prevention of Relapse of Lymphomas	CD30	HL, CD30+ NHL, or CD30+ LPD	Lineberger Comprehensive Cancer Center at University of North Carolina - Chapel Hill; Chapel Hill, North Carolina, United States	Anti-CD30 CAR T cells are administered following autologous stem cell transplant
NCT02706392	Genetically Modified T-Cell Therapy in Treating Patients With Advanced ROR1+ Malignancies	ROR1	ROR1+ MCL	University of Washington Cancer Consortium; Seattle, Washington, USA

HL: Hodgkin lymphoma. LPD: lymphoproliferative disorder. MCL: Mantle cell lymphoma. NHL: Non-Hodgkin lymphoma.

Information accessed from www.clinicaltrials.gov accessed on March 23, 2017.

Summary of CAR T-cell Toxicities

Toxicity is currently the biggest barrier to more effective CAR T-cell therapies for lymphoma. Toxicities associated with CAR T-cell therapies mostly resolve within 2 weeks of CAR T-cell infusion, but these short-lived toxicities can be severe, and in rare cases fatal.^{56, 79, 123} Moreover, toxicities hamper efforts to increase anti-malignancy efficacy of CAR T cells because approaches such as administering higher doses of CAR T cells or enhancing the activity of CAR T cells might worsen toxicity. Early CAR studies reported a syndrome with prominent signs of fever, tachycardia, and hypotension, and other toxicities after CAR T-cell infusions; this syndrome was designated cytokine-release syndrome (CRS)^{57, 74, 75, 124}. CRS also includes many other abnormalities including but not limited to decreased cardiac ejection fraction, fatigue, delirium, dyspnea, renal failure, elevated liver enzymes, nausea, diarrhea, and elevated creatine phosphokinase indicative of muscle damage (Table 5). CRS was associated with increased levels of many serum cytokines^{57, 123–126} Coagulopathy and hemorrhage have been described in studies of CAR T cells in ALL^{63, 80}. Similarly, patients with NHL receiving CAR T cells directed against varying antigen targets have experienced hemorrhagic events, including fatal gastrointestinal hemorrhange^{79, 100} and fatal intracranial hemorrhage^{79, 83}. Prolonged B-cell aplasia often occurs in patients with NLH in the months following CAR T-cell infusions^{56, 73, 74, 79, 83}. Recovery of normal polyclonal B cells in patients with NHL can occur without relapse of lymphoma^{56, 61, 74, 95}.

Table 5:

Grade 3–4 Toxicities occurring in selected published clinical trials of CAR T-cell therapy for lymphomas^*

Clinical Trial	Target Antigen	Costimulatory Domain	Hemodynamic changes	Neurologic Toxicity	Cardiac Toxicity	Renal and Hepatic Toxicity	Cytopenias	Infections	Hemorrhage and Coagulopathic or Prothrombotic Events	Other
Jensen et al., ASBMT 201076	CD19	None, first generation CAR	None	None	None	None	None	None	None	None
Jensen et al., ASBMT 201076	CD20	None, first generation CAR	None	None	None	None	None	None	None	None
Kochenderfer et al., Blood 201073	CD19	CD28	None	None	None	None	Cytopenias, not graded	None	None	None
Savoldo et al., J. Clin Invest 201159	CD19	One product without a costimulatory domain and one product with a CD28 costimulatory domain	None	None	None	None	None	None	None	None
Kochenderfer et al., Blood 201274	CD19	CD28	Fatigue Hypotension Dyspnea Hypoxia	Obtundation Cerebral infarction	Nonbacterial thrombotic endocarditis	Acute renal failure Electrolyte abnormalities Elevated liver enzymes Hyperbilirubinemia	None	Influenza Herpes zoster E. coli bacteremia	None	Fatigue Anorexia Hypoalbuminemia Diarrhea Headache Pleural effusion Capillary leak syndrome
Till et al., Blood 201299	CD20	CD28 and 4–1BB	Hypoxia	None	None	None	Thrombocytopenia Neutropenia	Cellulitis	None	None
Kochenderfer et al., Blood 201393°	CD19	CD28	Hypotension	None	None	None	None	None	None	Headache
Wang et al., Clin Immunol 2014100	CD20	4–1BB	Grade 3 CRS Dyspnea Hypoxia	None	None	Tumor lysis syndrome	None	None	Gastrointestinal hemorrhage	Serous cavity effusion
Kochenderfer et al., JCO 201556	CD19	CD28	Fever Tachycardia Hypotension Hypoxia Dyspnea Pneumonitis	Confusion Encephalopathy Aphasia Facial nerve palsy Obtundation Myoclonus Neuropathy Gait Disturbance	None	Acute renal failure Elevated creatinine	None	Urinary tract infection Bacteremia Influenza	None	Headache Nausea Malaise Fatigue Vascular leak syndrome Upper extremity thrombosis
Zhang et al., Signal Transduction and Targeted Therapy 2016101	CD20	4–1BB	None	None	None	Hypokalemia	None	Herpes zoster	None	None
Brudno et al., 201695°	CD19	CD28	Fever Sinus tachycardia Sinus bradycardia Hypotension	None	None	Bilirubin increased AST increased ALT increased Acidosis Hypercalcemia Hypophosphatemia	Anemia Thrombocytopenia Neutropenia	None	None	Headache Pain
Wang et al., Blood 201682	CD19	One clinical trial (NHL1) of a CAR not incorporating a costimulatory domain and one clinical trial (NHL2) of a CAR incorporating a CD28 costimulatory domain	None	None	None	None	Cytopenias as expected following auto-HSCT	None	None	None
Ramos et al., J Clin Invest 201660	Kappa (κ) light chain	CD28	None	None	None	Hypokalemia, not graded Hyperkalemia, not graded Elevated AST	Anemia, not graded Leukopenia, not graded	None	None	Fatigue, not graded
Turtle et al., Sci Transl Med 201679	CD19	4–1BB	Hypotension Grade 3–5 CRS reported**	Grade 3–5 neurotoxicity reported	Cardiac failure Cardiac arrhythmia	None	Cytopenias present, not graded	None	Gastrointestinal hemorrhage Intracranial hemorrhage	None
Wang et al., Clin Cancer Res 2017102	CD30	4–1BB	None	None	Left ventricular systolic dysfunction	AST and ALT increased γ-GGT increased	None	None	None	None
Locke et al., Mol Ther 201783	CD19	CD28	Fever Hypotension Hypoxia	Encephalopathy Tremor Somnolence Agitation Delirium Restlessness	Cardiac failure	Acute kidney injury Metabolic acidosis Hypocalcemia Hyponatremia Hypophosphatemia AST increased	Anemia Thrombocytopenia Neutropenia Febrile neutropenia	Urinary tract infection Oral herpes Pseudomonal sepsis	Intracranial hemorrhage	Ascites Fatigue
Kochenderfer et al. JCO 201761	CD19	CD28	Fevers Hypotension Hypoxia Dyspnea	Encephalopathy Confusion Cognitive disturbance Agitation Dysphasia Ataxia Motor neuropathy Generalized muscle weakness Tremor Somnolence Psychosis seizure	Cardiac left ventricular dysfunction*** Supraventricular tachycardia	Hyponatremia AST and ALT increased Bilirubin increased	Anemia Thrombocytopenia neutropenia Febrile neutropenia	Catheter-related infection UTI	Partial thromboplastin time increased	MDS, likely due to prior cytotoxic therapy Vision loss related to fludarabine toxicity Syncope Dysphagia Urinary incontinence

ALT: alanine aminotransferase. AST: aspartate aminotransferase. Auto-HSCT: Autologous stem cell transplant. CR: complete remission. MDS: myelodysplastic syndrome. UTI: urinary tract infection.

^*Does not include lymphopenia or cytopenias other than anemia, thrombocytopenia, or neutropenia.. Only toxicities occurring in patients with NHL are included.

^**CRS grading per Lee et al. “Current concepts in the diagnosis and management of cytokine release syndrome.” Blood 2014. 124: 188–195.

^***Systolic dysfunction may be due to prior doxorubicin.

^°Both studies reporting on the same clinical trial.

Many clinical and immunological associations with CRS have been reported. The severity of CRS toxicity has been associated with peak serum CAR T-cell levels⁷⁹ and serum cytokine levels, including IFN-γ^{56, 74, 79}, interleukin-6 (IL-6)^{56, 79}, and TNF-α⁷⁴, as well as peak serum C-reactive protein (CRP)⁷⁹. In clinical trials of CAR T cells for ALL, CRS severity has been associated with burden of disease^{63, 80}. Not surprisingly, higher doses of conditioning chemotherapy worsen the severity of neutropenia and thrombocytopenia for patients with NHL receiving CAR T cells⁶¹. Clinical results suggest that higher doses of CAR T cells have been related to increased rates of severe CRS^{56, 79}.

Neurologic toxicities including but not limited to confusion, tremors, ataxia, aphasia, obtundation, and seizures have been prominent in anti-CD19 CAR T-cell trials for lymphoma^{56, 61, 74, 79, 83}, with reported frequencies of grade 3–5 neurologic toxicity of 25–57%^{61, 79, 83}. Fatal cerebral edema occurred in a patient with DLBCL participating in a multicenter clinical trial of axicabtagene ciloleucel¹²⁷. The exact etiology of neurologic toxicity remains unknown. Neurologic toxicity may or may not occur at the same time as CRS^{83, 128}, suggesting differing pathophysiology. CD19 has not been shown to be expressed in the CNS^{56, 129}. Investigators have reported some associations with presence or absence of neurological toxicity. Notably, very low rates of neurologic toxicity have been reported in a clinical trial of anti-CD19 CAR T cells given with no prior conditioning chemotherapy^{93, 95}. Severe neurologic toxicity has been associated with higher CAR T-cell doses⁷⁹. Higher peak levels of serum IL-2⁷⁹, IL-6⁷⁹, IL-10⁶¹, IL-15⁷⁹, IL-18⁷⁹, interferon-γ⁷⁹, ferrtin⁷⁹, CRP⁷⁹, TIM-3⁷⁹, and granzyme B⁶¹, as well as higher peak blood CAR T-cell levels^{61, 79}.

Generally, CRS is managed with supportive care and immunosuppressive agents such as the IL-6 receptor antagonist tocilizumab and corticosteroids^{123, 125}. Tocilizumab has been used successfully to attenuate CRS-related toxicities in lymphoma patients^{79, 83}, though CRS refractory to tocilizumab can also occur^{56, 123}. Neurologic toxicities are sometimes preferentially managed with corticosteroids alone, due to a hypothetical risk of increased CNS IL-6 following saturation of peripheral IL-6 receptors¹²⁵.

Suicide genes are being investigated as possible avenues of inducing abrupt cessation of T-cell mediated toxicities; inducible caspase-9 suicide switches are the most advanced suicide gene approach^{130, 131}. A disadvantage of this approach is that complete elimination of all CAR T cells is likely to attenuate malignancy response, so this approach should probably be reserved for toxicities that cannot be controlled by other approaches.

Promising General Areas of Improvement for CAR T Cells

CAR T-cell therapies are in an early stage of development, and while past results are impressive, there are several areas with great potential for improvement. Gaining a better understanding of the importance of CAR T-cell persistence is an important area for future studies. The goal of CAR T-cell therapy is to provide patients with durable remissions of their malignancies. There are two general hypotheses to explain how CAR T cells could cause a lasting remission. First, a powerful T-cell response could completely eliminate all malignant cells early after infusion. In this case, long-term persistence of CAR T cells would not be needed. Second, CAR T cells are not capable of eliminating all malignant cells, but they are capable of eliminating some malignant cells. In this case persistence of CAR T cells might be important if the long-term persisting T cells could continuously detect and kill low levels of malignant cells to prevent a clinically-evident progression of malignancy. Acceptance of the second hypothesis requires that CAR T cells cannot eliminate all malignancy early after infusion, but can eliminate malignant cells long after CAR T-cell infusion at a time when CAR T-cell levels are much lower than early after infusion.^{61, 79} Persistence of T cells at tumor sites might be much more important than persistence of T cells in the blood, and activation-induced cell death (AICD) might be an important factor limiting persistence of CAR T cells^132–134. CAR T-cell persistence in humans is widely variable⁶¹. Persistence of anti-CD19 CAR T cells of several months to years in the peripheral blood has been reported^{74, 79}. Persistence of anti-CD19 CAR T cells for patients with NHL ranges from a mean of 20 days⁸² and less than 100 days^{56, 61} to greater than 300 days⁷⁹, with an appearance of greater persistence of CAR T cells incorporating a 4–1BB costimulatory domain versus a CD28 costimulatory domain, which is consistent with preclinical models^{53, 135, 136}. Comparison of two early phase lymphoma clinical trials of anti-CD19 CAR T cells, one with a CAR containing 4–1BB and one with a CAR containing CD28, reveals that efficacy is likely similar between costimulatory domains, with overall response rates of 72%⁷⁹ versus 73%⁶¹ and CR rates of 50%⁷⁹ versus 55%⁶¹ in patients receiving conditioning chemotherapy including fludarabine. While these trials had many factors in common, it is impossible to draw any firm conclusions by comparing these or other trials because the trials differ in many variables aside from the costimulatory domain used. One critical factor that is often overlooked is the importance of the hinge and transmembrane portions of CARs, so comparing CARs that have different costimulatory molecules, but also differ in other aspects is not a valid way to compare costimulatory domains¹³³.

Immunologic rejection of CAR T cells is another potential problem with CAR T cells. Most CARs in current clinical use contain antibody variable regions from murine antibodies. These sequences are potentially immunogenic. In addition, the junctions between different CAR domains are artificial sequences that could be immunogenic. Decreasing the immunogenicity of CARs might be particularly important if multiple doses of CAR T cells are administered because immunologic memory could be generated after a first dose of CAR T cells. T-cell responses against the murine components of the single chain variable fragment (sFv) of the CAR have been reported in multiple clinical trials of anti-CD19 CARs for both ALL and NHL^{64, 76, 79, 80}.

Loss of target antigen expression, termed “antigen escape” is a well described mechanism of malignancy progression in clinical trials of anti-CD19 CAR T cells for ALL¹³⁷, and one case of CD19 negative progression in a patient with lymphoma following two infusions of anti-CD19 CAR T cells has been reported¹²⁸.

Improving CAR T-cell Therapy

There are many avenues for improving CAR T-cell therapies (Figure 2). Improved CAR designs have potential to improve many of the current limitations of CAR T-cells. We have shown that holding all variables constant except changing the hinge and transmembrane regions can have an important impact on CAR T-cell cytokine production with higher levels of inflammatory cytokines produced by T cells expressing anti-CD19 CARs with hinge and transmembrane domains from CD28 compared with T cells expressing anti-CD19 CARs with CD8α hinge and transmembrane domains¹³³. AICD is a potential factor limiting CAR T-cell persistence, and AICD is increased in T cells expressing anti-CD19 CARs with hinge and transmembrane domains from CD28 compared to T cells expressing anti-CD19 CARs with CD8α hinge and transmembrane domains¹³³. CARs with humanized or fully human variable regions might have decreased immunogenicity, which may decrease CAR T-cell destruction by the host immune system and promote CAR T-cell expansion and persistence. Use of less immunogenic CARs may allow for greater efficacy of multiple cell infusions, and CARs with human variable regions are under development^{128, 138}. Clinical studies are now investigating humanized and fully human CAR T cells for patients with ALL and NHL (NCT02374333, NCT02659943, NCT03049449). Bispecific CARs simultaneously targeting 2 antigens offer one possible solution to loss of antigens targeted by CARs^{139, 140}.

Figure 2: Improvements to CAR T-cell therapy.

CAR T-cell therapy is being improved at multiple steps in the overall CAR T-cell therapy process. Areas of improvement include modifications in all aspects of CAR design, incorporation of different genes to enhance CAR T-cell activity or control toxicity into CAR gene therapy vectors, use of gene-editing technologies such as CRISPR/Cas9, changes to cell culture methods, optimization of conditioning regimens, and the addition of post-infusion pharmacologic agents.

Gene editing holds great promise for improving cell therapies¹⁴¹. For example, clustered regularly interspaced short palindromic repeats/crisper-associated system9 (CRISPR/Cas9) genome editing has allowed targeting of CAR gene integration to the T-cell receptor-α constant (TRAC) locus¹⁴². This modification decreased T-cell exhaustion; this resulted in superior anti-malignancy efficacy in a mouse leukemia model¹⁴².

Administration of CAR T-cell products composed of a defined 1:1 ratio of CD4⁺/CD8⁺ CAR T cells has been investigated in multiple clinical trials in ALL⁶⁴ and NHL⁷⁹. This strategy has not yet been directly compared with the use unselected cell products containing variable portions of CD4⁺ and CD8⁺ T cells in a clinical trial. Use of less differentiated T-cell subsets such as stem memory T cells is under investigation¹⁴³.

Fludarabine has been used in many successful CAR clinical trials^{56, 64, 79, 144}. There is evidence that improved lymphocyte depletion with a fludarabine-containing conditioning chemotherapy regimen may facilitate greater T-cell expansion and persistence, especially when multiple doses of T cells are administered^{64, 79}. Patients with NHL receiving fludarabine-containing chemotherapy prior to a second cell infusion had robust anti-CD19 CAR T-cell expansion and did not develop cytotoxic T-cell responses to the sFv of the CAR. Patients receiving fludarabine-containing chemotherapy have also been reported to have improved PFS compared with patients who received cyclophosphamide alone or cyclophosphamide and etoposide⁷⁹.

CAR T cell activity may be increased by the administration of pharmaceutical agents. The Bruton’s tyrosine kinase inhibitor ibrutinib has been shown to boost CAR T-cell anti-malignancy efficacy in MCL in xenograft mouse models¹⁴⁵. Immunologic checkpoint inhibitors have clinical efficacy in HL^{36, 38}, with some activity in NHL¹⁴⁶. Following cell infusion, anti-CD19 CAR T cells acquire a more differentiated and exhausted phenotype, with increased expression of PD-1^{56, 95}. Interruption of the PD-1/PD-L1 pathway, through PD-1 antibody checkpoint blockade, can restore CAR T-cell effector function in in vitro models¹⁴⁷. A response to a checkpoint inhibitor in a patient with PMBL and progressive disease following CAR T-cell infusion has been reported¹⁴⁸. A clinical trial is underway in which patients with refractory or relapsed DLBCL, FL, or MCL following anti-CD19 CAR T-cell infusion are administered the checkpoint inhibitor pembrolizumab in an attempt to re-activate exhausted CAR T cells (NCT02650999). Administration of IL-15 after CAR T-cell infusions might be of benefit given the association of high serum IL-15 levels with remissions of lymphoma in a recent study⁶¹.

Another aspect of clinical application improvement could be to move CAR T-cell therapy earlier in the line-up of lymphoma therapies. The majority of CAR T-cell trials for lymphoma include only patients who have had multiple lines of therapy, chemotherapy-refractory lymphoma, or relapse after auto-HSCT. CAR T cells have clear antimalignancy activity in these heavily pretreated patients^{56, 61, 79, 83}. Use of CAR T cells earlier in the course of lymphoma might increase the chance of successful treatment if for no other reason than patients are more likely to be healthy enough to undergo CAR T-cell therapy early in their courses of treatment. CAR T-cell therapy earlier in a patient’s treatment course may avoid some of the negative consequences of prolonged exposure to cytotoxic agents, such as development of myelodysplastic syndrome and other secondary malignancies, which are significant risks, especially following auto-HSCT^{149, 150}. Another potential advantage to moving CAR T-cell therapies earlier in the course of lymphoma treatment is that patients who have never been exposed to chemotherapy might have higher blood T-cell counts and perhaps T cells that are more fit for CAR T-cell therapy. Both of these factors might facilitate CAR T-cell production.

The feasibility of administering anti-CD19 CAR T cells immediately following auto-HSCT to patients with DLBCL has been explored^{76, 82}. The standard approach for fit patients with DLBCL and residual or relapsed lymphoma after first-line therapy is salvage chemotherapy followed by auto-HSCT^{9, 10}. Anti-CD19 CAR T cells may eventually supplant, or be used in tandem with, auto-HSCT as best second-line therapy for DLBCL.

Conclusions

While CAR T-cell therapy for lymphoma is in its early stages, this form of adoptive immunotherapy has already been shown to have the ability to cause durable complete remissions of chemotherapy-refractory lymphoma in many patients.. Anti-CD19 CAR T cells are likely to become the new standard of care for patients with chemotherapy-refractory DLBCL given the very limited options available for these patients currently^{7, 12, 14}. Anti-CD19 CAR T cells may also eventually supplant, or be used in tandem with, auto-HSCT as best second-line therapy for DLBCL. Anti-CD19 CAR T cells will almost certainly soon become widely-used to treat FCL and MCL. Further optimization of CAR design, improved toxicity management, and development of combination therapy will allow CAR T cells to be used safely and effectively for more patients.

Key Points.

• New treatments are needed for patients with chemotherapy-refractory or multiply relapsed lymphoma.

• Chimeric antigen receptor (CAR) T cells targeting CD19 have demonstrated efficacy in multiple B-cell lymphoma subtypes, with activity seen in cases of chemotherapy-refractory lymphoma. Durable remissions are possible.

• Multicenter clinical trials have demonstrated that centralized cell processing is feasible, with early response rates similar to those reported in single center studies.

• CARs targeting new antigens, such as CD20, CD22, CD30 and kappa light chains, are in development and will extend CAR T-cell therapy to patients with Hodgkin lymphoma, T-cell lymphoma, and CD19-negative B-cell lymphoma.

• Cytokine release syndrome and neurologic toxicity are problems associated with CAR T-cell therapies for lymphoma, and reducing toxicity is a major avenue for improving CAR T-cell therapies.

• CAR T-cell therapy is likely to become safer and more effective, and CAR T-cell therapies are likely to become standard treatment options for relapsed and primary chemotherapy-refractory lymphoma in the near future.

Biographies

Jennifer N. Brudno, M.D. received her medical doctorate from the Emory University School of Medicine. She completed internal medicine residency training at the Johns Hopkins University Hospital. She trained in medical oncology and hematology at the National Institutes of Health. She is currently a staff clinician in the Experimental Transplantation and Immunology Branch of the National Cancer Institute. Dr. Brudno’s research focuses on clinical development of chimeric antigen receptor (CAR) T-cell therapies for hematologic malignancies.

James Kochenderfer, M.D. is a Tenure-track Investigator in the Experimental Transplantation and Immunology Branch (ETIB) of the National Cancer Institute (NCI). Dr. Kochenderfer received his M.D. from West Virginia University and then completed internal medicine training at Vanderbilt University. Next he completed a medical oncology fellowship at the M.D. Anderson Cancer Center and a hematology fellowship at Baylor College of Medicine. He then undertook a period of postdoctoral research training in T-cell immunotherapy in the laboratories of Dr. Ronald Gress and Dr. Steven Rosenberg at the National Cancer Institute. For the past 10 years, Dr. Kochenderfer has focused on a full range of chimeric antigen receptor (CAR) research from designing and constructing new CARs to preclinical testing of new CARs to conducting clinical trials. His work emphasizes developing CAR T-cell therapies for lymphoma and multiple myeloma.