Advances in cellular and humoral immunotherapy - Implications for the treatment of poor risk childhood, adolescent, and young adult B-cell non-Hodgkin lymphoma

Summary

Patients with relapsed, refractory or advanced stage B non-Hodgkin lymphoma (NHL) continue to have a dismal prognosis. This review summarises current and novel cellular and immunotherapy for these high-risk populations, including haematopoietic stem cell transplant, bispecific antibodies, viral-derived cytotoxic T cells, chimeric antigen receptor (CAR) T cells, and natural killer (NK) cell therapy, as discussed at the 6^th International Symposium on Childhood, Adolescent and Young Adult Non-Hodgkin Lymphoma on September 26^th-29th 2018 in Rotterdam, the Netherlands, and explores the future of NK/CAR NK therapies.

Introduction

Relapsed, refractory and advanced stage disease remain challenges in childhood, adolescent, and young adult (CAYA) B-cell non-Hodgkin lymphoma (B-NHL). The prognosis for B-NHL is excellent with multiagent chemotherapy, as delineated by the French-American-British (FAB)/lymphomes malins B (LMB) 96 results, in which the 5-year event-free survival (EFS) for those patients with limited disease, intermediate risk and advanced stage was 97 ± 0.5%, 89 ± 1.2% and 79 ± 2.7%, respectively (Cairo, et al 2007, Cairo, et al 2012, Gerrard, et al 2013). There was, however, a small cadre of 104 CAYA amongst the 1,111 studied who had primary refractory disease or disease relapse after first complete remission (CR). The probability of 1- and 2-year overall survival (OS) in this group was a dismal 31.5% and 23.3%, respectively (Cairo, et al 2018). Those patients with lactate dehydrogenase ≥2 times upper limit of normal at diagnosis, patients with relapsed or refractory disease within 6 months of diagnosis and/or those patients with relapsed or refractory disease with bone marrow involvement had a significantly decreased OS (Cairo, et al 2018). Cellular and humoral immunotherapy for these high-risk populations include haematopoietic stem cell transplantation (HSCT), bispecific antibodies, viral-derived cytotoxic T cells, chimeric antigen receptor (CAR) T cells and natural killer (NK) cell therapy.

Stem cell transplantation for childhood NHL

Stem cell therapies, comprised of autologous bone marrow transplantation, allogeneic bone marrow transplantation or tandem autologous/allogeneic stem cell transplantation, are utilised with varying levels of success in treating this difficult-to-treat group, as delineated in Table I.

Table I.

Stem cell transplantation in childhood NHL

Author	Centre/group	N	Age (years)	NHL histology	Donor source	Conditioning regimen	DFS/EFS  (%)
Loiseau et al (1991)	Institut Gustave Roussy	24	NA	16 B-NHL 8 T-NHL	Autologous	33
Won et al (2006)	Korea	33	1.7–16	6 B-NHL 13 LL 14 LCL	Autologous	BEAM BEAC CBV TBI regimen	59
Philip et al (1988)	SFOP	15	NA	B-NHL	14 Autologous 1 Allogeneic	BEAM/BEAC, other	27
Ladenstein et al (1997)	EBMT	89	2.8–16.2	B-NHL	Autologous	BACT 31 BEAM 23 BU/CY 9 Other 26	44 (sensitive relapse)
Kobrinsky et al (2001)	CCG	50	<21	N/A	Autologous	N/A	50
Levine et al (2003)	CIBMTR	128	2–67	LL	Autologous	N/A	39
		76	5–53		Allogeneic	N/A	36
Fanin et al (1999)	EBMT	64	3.2–53	ALCL	Autologous	N/A	47
Gross et al (2010)	CIBMTR	90 92		52 DLBCL 53 LL 41 BL 36 ALCL	90 Autologous 92 Allogeneic	NA	DLBCL 52 (auto) 50 (allo) BL 27 (auto) 31 (allo) ALCL 35 (auto) 46 (allo) LL 4 (auto) 40 (allo)
Harris et al (2011)	COG	10	4.2–19.9	NA	Autologous	CBV	70
Woessmann et al (2006)	BFM	20	1–15.8	ALCL	Allogeneic	TBI/CY/VP-16	75
Bureo et al (1995)	Spain	46	1–17	21 LL 19 B-NHL 6 LCL	14 Allogeneic 32 Autologous	BEAM CY/TBI CY/TBI/ARA-C	58
Giulino-Roth et al (2013)	MSKCC (US)	21 15	3.5–20.9	12 LL 5 BL 4 DLBCL 13 ALCL 1 PTCL 1 u-NHL	21 Allogeneic 15 Autologous	CY/TBI CY/THIO/TBI TBI/other Non-TBI based	53
Jourdain et al (2015)	SFOP	8 33	NA	NA	8 Allogeneic 33 Autologous	BEAM, BAM, BU/MEL, TBI/CY, other	49 (auto) 38 (allo)
Satwani et al (2015)	Multicentre US trial	10	7–33	3 ALCL 2 DLBCL 3 BL 1 LL 1 PTCL	Tandem autologous allogeneic	BEC (auto) BU/FLU ± rATG (allo)	70
Burkhardt et al (2018)	International trial	153 248	0.3–18	152 BL 101 T-LL Remaining 148: DLBCL + PMBL + B-LL + other	153 Autologous 248 Allogeneic	NA	55 (auto) 47 (allo)
Gardenswartz et al (2018)	Multicentre US trial	13	7–33	8 BL 3 DLBCL 2 PMBL	Tandem autologous allogeneic	BEC (auto) BU/FLU ± rATG (allo 12) BEAM (allo 1)	91

ALCL: anaplastic large cell lymphoma; allo: allogeneic; ARA-C: cytarabine; auto: autologous; B-NHL: B-cell non-Hodgkin lymphoma; BACT: bischlorethyl-nitrosurea, cytarabine, cyclophosphamide, 6-thioguanine; BAM: busulfan, cytarabine, melphalan; BEAC: carmustine, etoposide, cytarabine, cyclophosphamide; BEAM: carmustine, etoposide, cytarabine, melphalan; BEC: carmustine, etoposide, cyclophosphamide; BFM: Berlin–Frankfurt–Münster; BL: Burkitt lymphoma; BU: busulfan; CBV: cyclophosphamide, carmustine, etoposide; CCG: Children’s Cancer Group; CIBMTR: Center for International Bone Marrow Transplant Registry; COG: Children’s Oncology Group; CY: cyclophosphamide; DFS: disease-free survival; DLBCL: diffuse large B-cell lymphoma; EBMT: European Society for Blood and Marrow Transplantation; EFS: event-free survival; FLU: fludarabine; IBMTR: International Bone Marrow Transplant Registry; LCL: large cell lymphoma; LL: lymphoblastic lymphoma; MEL: melphalan; MSKCC: Memorial Sloan Kettering Cancer Center; N/A: not applicable; NA: not available; NHL: non-Hodgkin lymphoma; PMBL: primary mediastinal B-cell lymphoma; PTCL: peripheral T-cell lymphoma; rATG: rabbit antithymocyte globulin; SFOP: Société Française di’Oncologie Pédiatrique; T-NHL: T-cell non-Hodgkin lymphoma; TBI: total body irradiation; THIO: thiotepa; u-NHL: undifferentiated NHL; VP-16: etoposide.

Adapted from Bradley, M. B., & Cairo, M. S. (2008). Stem cell transplantation for pediatric lymphoma: Past, present and future. Bone Marrow Transplantation, 41(2), 149–158. doi:1705948 [pii]

Autologous transplantation

Myeloablative conditioning (MAC) therapy followed by autologous HSCT is traditionally considered the standard of care in relapsed/refractory NHL (Bradley and Cairo 2008, Cairo, et al 2013). In the Children’s Oncology Group (COG) prospective study designed to determine the safety and efficacy of cyclophosphamide, carmustine and etoposide (CBV) conditioning and autologous peripheral blood HSCT in children with relapsed or refractory Hodgkin lymphoma (HL) and NHL, the 3-year EFS from study entry for NHL patients was only 30% (Harris, et al 2011). At the 6^th International Symposium on CAYA NHL, Burkhardt et al (2018) presented a large retrospective study analysing the role of transplant in relapsed/refractory NHL in patients diagnosed after the year 2000 who were less than 18 years of age, in 24 countries. Survival for the 241 patients who did not undergo HSCT in Burkhardt’s study was a dismal 9 ± 2%. OS was 55 ± 5% for the 153 patients treated with autologous HSCT. The 5-year cumulative incidences of transplant-related mortality (TRM) and death from disease were 7 ± 2% and 31 ± 4% in this group (Burkhardt, et al 2018).

Allogeneic transplantation

Allogeneic stem cell transplantation in relapsed/refractory NHL capitalizes on the potential graft-versus lymphoma (GvL) effect. Jones et al (1991) were the first to establish a GvL effect and Woessmann et al (2006) demonstrated this effect in paediatric anaplastic large cell lymphoma (ALCL). In a small retrospective analysis from the Center for International Bone Marrow Transplant Registry, Gross et al (2010) showed a superior EFS in patients with lymphoblastic lymphoma receiving allogeneic vs. autologous HSCT. This superior EFS, however, was not demonstrable in the other NHL subtypes (Gross, et al 2010). In the recently reported international study (Burkhardt, et al 2018), OS was 48 ± 3% for the 248 patients treated with allogeneic HSCT. The 5-year cumulative incidences of TRM and death from disease were 16 ± 2% and 34 ± 3%, respectively.

Tandem autologous/allogeneic transplantation

Although, in theory, a GvL effect in allogeneic transplant should yield superior EFS and OS across histological subtypes, this has not been actualised, largely due to TRM in the setting of MAC. Carella et al (2000) pioneered the myeloablative autograft reduced intensity conditioning (RIC) allograft approach in adult relapsed/refractory lymphoma patients in an attempt to glean the benefits of both modalities of cell therapy while minimizing the risks. In their cohort of 15 patients (10 HL and five NHL) they demonstrated a complete remission in 11 patients, nine of whom had only achieved a partial remission (PR) post-autologous HSCT (Carella, et al 2000). Chen et al (2015) reported the largest prospective series of tandem autologous HSCT followed by allogeneic HSCT in high-risk lymphoma. Twenty-nine of 42 enrolled patients (69%) proceeded to a RIC allogeneic HSCT. The 2-year progression-free survival (PFS) and OS for patients who underwent tandem HSCT were impressive, at 72% and 89%, respectively (Chen, et al 2015). Satwani et al (2015) were the first to perform a prospective study utilising MAC autologous HSCT with subsequent RIC allogeneic HSCT in CAYA patients with relapsed/refractory lymphoma. They reported an overall 10-year EFS of 50.0% in an intent-to-treat analysis of all enrolled NHL patients versus a 70% EFS in those patients who received a tandem MAC autologous HSCT and RIC allogeneic HSCT (Satwani, et al 2015). At the symposium, Cairo’s group reported a 91% EFS in a cohort of 13 CAYA patients with relapsed/refractory B-NHL (five of whom were part of Satwani’s cohort) who underwent MAC autologous HSCT with subsequent radioimmunotherapy with Zevalin® followed by RIC allogeneic HSCT with a median follow-up time of 48 months. There has been no TRM and no reports of myelodysplastic syndrome or secondary leukaemia in this cohort to date (Gardenswartz, et al 2018).

There is an important role for HSCT in CAYA patients with relapsed/refractory B-NHL. It is doubtful that large scale studies comparing autologous vs. allogeneic HSCT will be performed due to a very small number of relapsed patients. It is important to evaluate HSCT in the context of NHL histological subtype. While tandem autologous/allogeneic HSCT produced excellent EFS (Satwani et al 2018; Gardenswartz, et al 2018), the numbers are small and not adequate to perform EFS by histological subtype. A large multicentre trial of tandem MAC autologous HSCT and RIC allogeneic HSCT could confirm Cairo’s group promising results. Given transplant-related toxicities and relapses following HSCT, other means of cellular immunotherapy are promising as alternatives to HSCT or as combination therapies with HSCT.

Cellular immunotherapy

Expansion of cellular therapies beyond autologous and allogeneic HSCT provides curative options for patients with relapsed or refractory NHL, particularly B cell disease. NHL derived from B-lymphocytes express B-lymphocyte differentiation antigens on the surface of the malignant cells. CD20 is a well-established target for immunotherapy in mature B cell disease, as is CD30 for ALCL (Dunleavy, et al 2013, Younes, et al 2010). CD19 is expressed in B cell lymphomas, including diffuse large B-cell lymphoma (DLBCL), Burkitt lymphoma (BL), primary mediastinal B-cell lymphoma (PMBCL), follicular lymphoma (FL) and other indolent NHL (Wang, et al 2012). CD19 is also expressed on normal maturing B-lymphocytes but no other cells (Otero, et al 2003). This makes it an excellent target for cellular therapies, including bispecific antibodies and chimeric antigen receptor engineered T cells (CAR T). T cell therapy can also be directed against Epstein–Barr virus (EBV)-driven lymphomas, including NK/T cell lymphoma and post-transplant lymphoproliferative disease (PTLD). Cellular immunotherapy also encompasses using NK or CAR/NK cell therapies to target malignant cells (Fig. 1).

Figure 1. Strategies for use of cell therapies for CAYA patients with relapsed/refractory B-NHL.

These cell therapies include autologous/ allogeneic stem cell transplantation, EBV-activated CTL, CMV/anti-CD19 CAR bispecific T cells, CAR T and CAR NK cells, the combination of T cells with bispecific T-cell engager Blinatumomab, the combination of NK/CAR NK cells with an IL-15 superagonist ALT-803, the combination of NK cells with a novel targeted fusion protein 2B8T2M, and the combination of NK cells with a novel type II antibody, obinutuzumab.

ADCC: antibody-dependent cellular cytotoxicity; B-NHL: B-cell non-Hodgkin lymphoma; CAR: chimeric antigen receptor; CAYA: children, adolescents and young adults; CMV: cytomegalovirus; CTL: cytotoxic T lymphocytes; EBV: Epstein–Barr virus; IFN-γ: γ-interferon; IL-15: interleukin 15; NK: natural killer.

Bispecific antibodies

Bispecific antibodies redirect T cells to epitopes on malignant cells, including CD19 (Kontermann and Brinkmann 2015). Blinatumomab is a 55-kDa bispecific T-cell engager (BiTE) mouse monoclonal antibody that binds to CD19 and CD3 (Loffler, et al 2000). Blinatumomab showed impressive clinical results for CD19-positive B-cell precursor acute lymphoblastic leukaemia (B-ALL), resulting in its US Food and Drug Administration (FDA) approval for use in the treatment of adult and paediatric patients with B-ALL (Velasquez, et al 2018). The anti-tumour effects of blinatumomab were further confirmed in B-NHL patients. In a phase 1/2 clinical study of 76 patients with DLBCL, mantle cell lymphoma and FL, blinatumomab at the maximum tolerated dose of 60 μg/m²/day resulted in an overall response rate (ORR) of 69% across NHL subtypes and 55% for DLBCL with a median response duration of 404 days (Goebeler, et al 2016). A multicentre phase II study of blinatumomab for relapsed/refractory DLBCL found that stepwise dose escalation was necessary to avoid neurological side effects and the ORR was 43% (Viardot, et al 2016). New bispecific antibody constructs are in clinical trials (Bacac, et al 2018). Numerous bispecific antibodies that direct T or NK cells to tumour antigens are undergoing preclinical and clinical evaluation (Felices, et al 2016, Velasquez, et al 2018). The major barriers to bispecific antibody therapy are neurotoxicity and the requirement for continuous infusion. There are no reported clinical studies specifically in CAYA patients with NHL. The Cairo group performed a preclinical study to evaluate the anti-tumour effects of blinatumomab against rituximab-sensitive/resistant BL and PMBCL and found that blinatumomab significantly enhances T cell-mediated in vitro cytotoxicity and T cell cytokine secretion/activation against BL/PMBCL. These preliminary studies suggest that blinatumomab is a novel agent to treat CAYA patients with relapsed/refectory CD19⁺ BL/PMBCL (Awasthi, et al 2018a).

Epstein–Barr virus cytotoxic T cells (EBV-CTL)

EBV is highly immunogenic and produces latent infections modulated by host EBV-specific T cells. In conditions of severely suppressed immunity, genetic immunodeficiency or post-solid organ/HSCT transplantation, EBV infection can cause lymphoproliferative disease (type III latency). In an immunologically normal host, EBV-driven malignant cells express the EBV antigen latent membrane protein (LMP) 1, LMP2, and EBNA1 (type II latency).

EBV cytotoxic T lymphocyte (CTL) infusions restore immune function post-HSCT with minimal side effects (Leen and Heslop 2008). Patient-specific EBV CTL treat and prevent PTLD (Bollard, et al 2012). Patients with active PTLD responded (11/13) and of 100 patients who received preventative EBV CTL infusions, none developed PTLD (Heslop, et al 2010). EBV CTL infusions have also been studied in tumours that exhibit type II latency, including DLBCL, NK/T lymphomas and HL. EBV CTL augment T cell response against the type II latency proteins expressed on EBV-positive tumours and generates memory cells, resulting in persistence (Bollard, et al 2014). In a series of EBV-positive malignancies, seven patients with DLBCL received EBV CTL and three became long-term survivors, three died of complications of chemotherapy/post-relapse transplant and a single patient did not respond. Eleven patients with NK/T lymphoma received EBV CTL. All NK/T patients receiving cells in remission achieved a durable CR, the longest being 4 years from infusion (Bollard, et al 2014). Cho et al (2015) reported a series of patients receiving EBV CTL for NK/T lymphomas. Ten patients in remission after varied treatment for NK/T lymphomas received EBV CTL without acute toxicity. The 4-year OS was 100% and PFS was 90% with a follow-up of 55.5 months. The toxicity of EBV CTL is minimal (Bollard, et al 2014). EBV CTL used post-HSCT are not associated with induction of graft-versus-host disease, (GvHD (Bollard, et al 2012).

Two of the major barriers to EBV CTL are production time and tumour evasion. EBV CTL are generated using the patient’s autologous peripheral blood collected by apheresis. Third-party EBV-specific CTL, bypassing the delay of manufacturing autologous CTL, are now available (Vickers, et al 2014). The COG is currently conducting an open trial for the use of third-party EBV CTL in paediatric patients with PTLD who are not in a CR after three doses of rituximab [ANHL1522 (NSC#782666)]. Targeting programmed death mechanisms and modulation of the tumour microenvironment represent approaches to inhibiting immune evasion by the EBV-positive tumours. Bollard et al (2002) developed modified T cells that express a dominant-negative transforming growth factor (TGF)-beta type II receptor (DNRII) that blocks the effect of endogenous TGF-beta. This approach is published in a phase 1 study of patients with HL as proof of principle that the approach is safe with no acute toxicities and can produce responses and persistence (Bollard, et al 2018).

Summary of EBV CTL in CAYA patients with NHL

In summary, the advantage for using EBV CTL for EBV-positive paediatric B-NHL is that the infusions are safe and very well tolerated. There is experience in using them for post-transplant prevention and treatment of PTLD in paediatric and adolescent patients. There is no chemotherapy required prior to administration and the therapy does not deplete normal B cell populations. Immunity is rapidly restored after cell infusion. Barriers to EBV CTL are the delay in production for patient-specific EBV CTL and the development of tumour evasion, both amenable to further study.

Chimeric antigen receptor T cells (CAR T)

The ability to direct a T cell against a specific antigen, crossing major histocompatibility complex barriers, using synthetic chimeric antigen receptors, created a paradigm shift in the field of targeted cellular therapy for malignancies (Salter, et al 2018). CAR Ts link a monoclonal antibody targeting a tumour-specific antigen and T cell signalling pathways for T-cell activation and/or co-stimulation. Four generations of CAR T are now developed and in preclinical and clinical evaluation (Barth, et al 2016) (Fig. 2). The initial generation of CAR T constructs linked a single-chain variable fragment (scFv) from a monoclonal antibody to the intracellular portion of the T-cell receptor CD3-ζ. Second generation CAR T use T-cell co-stimulatory receptor signalling domains CD28 or 4–1BB and the T-cell activation domain CD3-ζ (Brudno and Kochenderfer 2018). The primary clinical differences between co-stimulation with CD28 and 4–1BB constructs are CD28 co-stimulation creates a higher expansion peak and 4–1BB co-stimulation produces a higher degree of persistence (van der Stegen, et al 2015). Third-generation CAR T cells are currently in clinical trials for adults with B-cell lymphoma. A recently published study of a third-generation construct uses a CH2-CH3 hinge domain to link CD28, 4–1BB and CD3-ζ (Enblad, et al 2018).

Figure 2. Chimeric antigen receptors.

The first-generation CAR Ts only have CD3 zeta signalling domain; the second-generation CARs include one CD28 or 4–1BB co-stimulatory components combined with CD3 zeta signalling domain; the third-generation CARs include two co-stimulatory domains; the fourth-generation CAR Ts are designed with new elements, including a controllable suicide gene, such as caspase 9, or loaded with IL-12 secretion.

CAR, chimeric antigen receptor; IL-12, interleukin-12.

Reproduced with permission from Barth MJ, Chu Y., Hanley PJ, Cairo MS (2016), Immunotherapeutic approaches for the treatment of childhood, adolescent and young adult non-Hodgkin lymphoma. Br J Haematol.173(4):597–616. doi: 10.1111/bjh.14078.

Multiple single institution trials using CAR T targeting CD19 have established efficacy, acute toxicity profile, persistence characteristics and long-term effects in patients with B-NHL. Kochenderfer et al (2012) reported the first use of CAR T cell therapy (CD28/CD3-ζ construct) in a patient with multiply relapsed FL and subsequently reported results in eight patients with B-cell disease. Other early trials of CAR T therapy in NHL using constructs involving CD28 co-stimulation produced mixed response results, but did establish the increased persistence of the CAR T cells with CD28 co-stimulation (Jensen, et al 2010, Savoldo, et al 2011).

Further studies at the National Cancer Institute showed CR in two of four patients with refractory DLBCL and in two of four patients with PMBCL receiving CD28/C3-ζ CD19 CAR T. Notably, investigators demonstrated infiltration of the CAR T cells into a lymph node mass in a patient with chronic lymphocytic leukaemia (Kochenderfer, et al 2015). Low dose conditioning with fludarabine/cyclophosphamide and CAR T was given to 22 adult patients with B-NHL and showed an ORR of 73% and CR rate of 55%. For patients with DLBCL, the ORR was 68% and CR rate was 47% (n=19). CR persisted 9–24 months after infusion. The 12-month PFS for all B-NHL patients was 63% (Kochenderfer, et al 2017).

Turtle et al (2016) studied CAR T cell infusions with CD19 CAR T with defined CD4+: CD8+ ratio in 32 patients with B-NHL of all subtypes. This study demonstrated that CR rate improved (50%) when fludarabine/cyclophosphamide preceded the infusion compared to cyclophosphamide alone (8%). Patients receiving fludarabine prior to CAR T infusion had higher expansion of CAR T cells (Turtle, et al 2016).

Another study reported results using 4–1BB/CD3-ζ CAR T therapy in 28 adults with B-cell lymphomas showing an ORR of 64%, CR of 43% for DLBCL and CR of 71% for FL. The therapy produced sustained remissions in 86% of patients with DLBCL and 89% of patients with FL. Of note, all subtypes of DLBCL responded - germinal centre, non-germinal centre, double hit and transformed FL (Schuster, et al 2017). This single arm phase II study led to the FDA approval for tisagenlecleucel for adults with relapsed/refractory DLBCL.

A phase 1 trial of KTE-C19 CD28/CD-3ζ CAR T (ZUMA 1) established the feasibility of central production and safety profile in a multicentre trial. In patients with DLBCL, the ORR was 71% (5/7) and CR was 57% (4/7) one month after infusion. Three phase 1 patients with CR at 12 months persisted to 24 months (Neelapu, et al 2017). Consistent with prior CAR T cell trials, cytokine release syndrome (CRS) and neurotoxicity are the major toxicities (Locke, et al 2017). Neelapu et al (2017) reported the final phase II results with 101 patients receiving CAR T therapy. The ORR was 82% and the CR 54%. The OS at 18 months was 52%. When the results of the phase I patients were added to the phase II, the ORR was 82% and CR rate was 58%. The study documented late responses in patients with partial responses and stable disease as late as 15 months after infusions. At 15.4 months following infusion of CAR T cells, 42% of patients continued to have a response, with 40% maintaining a CR (Neelapu, et al 2017). The FDA approved Axicabtagene Ciloleucel in October 2017 for adults with refractory or relapsed DLBCL. The use of tocilizumab or steroids for CRS did not affect the ORR. As Schuster et al (2017) also reported, the ZUMA1 trial did not find a difference in response based on cell of origin of DLBCL (germinal centre vs. non-germinal centre) (Neelapu, et al 2017).

Other CD19 CAR T constructs are in clinical trials. Preliminary results for JCAR017 have been reported in abstract form (Abramson, et al 2018). A third-generation CD19 CAR T (CD28/4–1BB/CD-3 ζ with CH2-CH3 hinge) produced results similar to those of previous CAR T trials, with a complete response rate in 36% of B cell lymphoma patients (Enblad, et al 2018).

Anti-CD19 CAR T is standard of care for chemotherapy refractory DLBCL in adults. For use post-HSCT, Wang showed the infusion of CAR T cells post-autologous HSCT is safe and feasible (Wang, et al 2016). Using CAR T for relapse following allogeneic HSCT is also safe and efficacious. No GvHD was detected post-infusion. Persistence of CAR T cells was short-lived (1–2 months) (Kochenderfer, et al 2013). This finding of a GvL effect without GvHD was also noted by the Memorial Sloan Kettering group and opens the way to future combinations of allogeneic HSCT and targeted CAR T post-transplant therapy (Ghosh, et al 2017).

Cytomegalovirus (CMV) infection is also very immunogenic. In a mouse model, Wang shows that bispecific T cells recognizing CMV-specific antigens and CD19 CAR T cells can be stimulated to expand and increase killing of CD19 tumour cells after vaccination with a CMV peptide (pp65) or by repeated exposure to CD19 CAR T cells (Wang, et al 2015). A new vaccine targeting multiple antigens (modified vaccinia Ankara) is in phase II trials for patients following allogeneic HSCT (NCT03354728). Future trials are planned using this bispecific CAR T in humans with CD19+ malignancy. In pre-clinical work by Nakumura, presented at this meeting, CMV specific CAR T cells showed life-long persistence and potent cytotoxicity that could be augmented by vaccination (Nakamura, 2018).

There are several other targets for CAR T cells in CAYA NHL. CD20 is possible target of CAR T cells. The use of CD20 CAR T was originally reported in FL and marginal zone lymphoma (Till, et al 2008). Wang et al (2014) reported on seven patients with DLBCL treated with a CD20 4–1BB co-stimulatory domain CAR T. Five out of six evaluable patients had a response. This study demonstrated two unique results – CRS was delayed 3–8 weeks after infusion and local involvement of tumour resulted in gastrointestinal haemorrhage and significant pulmonary inflammation (Wang, et al 2014). A subsequent study of eleven patients with relapsed disease showed an ORR of 81.8% with 55% CR. Spleen and testicular disease did not respond to the CAR T therapy (Zhang, et al 2016). Kappa light chain antigen is another potential target in CAR T cells. The advantage of using kappa light chain is that it avoids B-cell depletion (Ramos, et al 2016). CD22 CAR T therapy is being used on trial for patients with B-ALL with good CR rate and similar toxicity profile to CD19 CAR T treatment (Fry, et al 2018). There are multiple clinical trials open using CD22 CAR T for adults with B-cell lymphomas (NCT02721407, NCT02315612, NCT02935153).

The acute toxicity of CAR T cell infusions is significant and well established, with a high incidence of CRS and neurotoxicity (Maude, et al 2018). There is also a risk of infection in the short- and long-term with B-cell depletion. Hill et al (2018) reported that 23% of patients develop a grade 3+ infection within 28 days of CAR T infusion. Most are bacterial, but there is a 5% reporting of invasive fungal disease and 4% of infections are life threatening or fatal. Risk of infection was higher in patients with acute leukaemia, with higher CAR T cell dose, and in patients that developed CRS (Hill, et al 2018).

Barriers to CAR T therapy are the acute CRS and the neurotoxicity. Investigators continue to rapidly improve management of the acute complications. They are working to understand the mechanisms of the neurotoxicity (Taraseviciute, et al 2018). Genetic engineering can decrease the inflammatory responses produced or can place a suicide gene into the construct for turning off the activation in cases of severe CRS or neurotoxicity (Zhou and Brenner 2016). The depletion of normal B-cell populations requires regular infusion of gamma globulin to prevent infections. Several studies note a late recovery of B-cell function without relapse of disease (Bhoj, et al 2016). Genetic engineering can also spare normal B cell development. Better understanding of acute infectious risk can inform supportive care measures. Patients must be able to tolerate peripheral blood apheresis for most products and stay in disease control during the time for production. Notably, there is now a third-generation CAR T product manufactured from 30 ml of peripheral blood with a manufacturing success rate of 94% (Enblad, et al 2018). As with EBV CTL, third-party CAR T cells are in development, thus avoiding the waiting time for manufacturing. The ability of tumour to evade CAR T remains a challenge. For patients receiving CD19 CAR T, relapses are often CD19-negative. There are many innovative approaches to minimize evasion including the use of checkpoint inhibitors or small molecule inhibitors (Cherkassky, et al 2016, Chong, et al 2017). There are multi-targeted CAR T in development (i.e. CD19 + CD22) (Cho, et al 2018, Fry, et al 2018). Switch CAR T cells can be activated with a small molecule (Rodgers, et al 2016). Persistence is also an ongoing issue that can be addressed with better understanding of the constructs and the immunology involved.

Despite extensive experience with CAR T in paediatric ALL, there is very limited use of CAR T therapy for paediatric or adolescent patients with NHL. Rivers et al (2018) presented eight paediatric patients receiving CAR T therapy for DLBCL, PMBCL, BL and grey zone lymphoma. Five of eight had a CR or PR with three of eight achieving a CR. The infusions were well tolerated, with a low incidence of CRS (grade 2 maximum) and neurotoxicity. In some patients, CR occurred after day 21 but early responses were not always durable despite persistence of CAR T cells. CD19 negative recurrence was also a challenge (Rivers, et al 2018).

Summary of CAR T for NHL

In summary, the advantage of CAR T for paediatric NHL is that the infusions require minimal additional chemotherapy prior to administration. There is a robust experience in using them for ALL in paediatric and adolescent patients. There are many available targets and the technology of manufacturing will continue to improve at a rapid rate. They currently are associated with a significant toxicity profile that is rapidly being addressed as the field advances (Table II) (Ghobadi 2018).

Table II.

Summary of ZUMA-21, JULIET, and TRANSCEND trials.

	ZUMA-1 Yescarta	JULIET Kymriah	TRANSCEND JCAR-017
Sponsor	Kite/Gilead	Novartis	Juno/Celgene
Source	Phase 1/2 N Engl J Med 2017;377:2531–44 (NCT02348216)	Phase2 ASH 2017 # 577: Blood 2017;130:577 (NCT02445248)	Phase 1 ASH 2017 # 581: Blood 2017;130:581 (NCT02631044)
Population	• 76% DLBCL NOS; 16% TFL; 8% PMBCL • 79% refractory • 21% relapsed post auto-HCT • Median age: 58 years (23–76 years)	• DLBCL NOS or TFL (≥ 2 lines of therapy) • 53% relapsed/refractory • 47% relapsed post auto-HCT • Median age: 56 years (range, 22–76)	(CORE; N = 49) • DLBCL NOS; TFL • 66% chemorefractory • 46% relapsed post auto-HCT • Median age: 61 years (range, 26–82)
Enrollment	• Phase 1: 6 enrolled • Phase 2: 111 enrolled; 101 infused and evaluable • No bridging chemotherapy	• 147 enrolled; 99 infused, 81 evaluable • 90% Bridging chemotherapy	• CORE data set: 49 • Bridging chemotherapy: allowed
CAR	• Second generation, CD28 • Retroviral vector	• Second generation, 41BB • Lentiviral vector	• Second generation, 41BB, defined ratio of CD4+ and CD8+ cells • Lentiviral vector
Dose	2.0 × 106 CAR T cells/kg >100 kg: 2.0 × 108 fixed dose	Median, 3.1 × 108transduced cells (range, 0.1–6.0 × 108transduced cells)	DL 1 5.0 × 107 CAR T cells DL 2 1.0 × 108 CAR T cells
Lymphodepleting chemotherapy	Flu 30 mg/m2 and Cy 500 mg/m2 on days −5, −4, and −3	• Flu 25 mg/m2 and Cy 250 mg/m2/day for 3 days (73%) • Benda 90 mg/m2/day for 2 days (19%)	Flu 30 mg/m2 and Cy 300 mg/m2 for 3 days
Efficacy	mITT = 108 • Median follow-up 15.4 m • ORR: 82%; CR: 58% • Ongoing response: 42%, 40% CR • Median DOR: 11.1 m	Median follow up: 5.6 m • ORR: 53.1%; CR: 39.5% • 6-mo rate: ORR: 37%; CR: 30% • Median DOR and OS NR	(CORE; N = 49) • ORR: 84%; CR: 61% • 6-m rate: 57%; 52% CR • Median OS: NR; 6-m OS: 88% • Median DOR: 9.2 m (NR for CR)
Safety	• Gr ≥ 3 CRS: 13% • Gr ≥ 3 NT: 28% • Gr 5 AE: 3%	• Gr ≥ 3 CRS: 23% • Gr ≥ 3 NT: 12% • Gr 5 AE: 0%	• Gr ≥ 3 CRS: 1% • Gr ≥ 3 NT: 14% • Gr 5 AE: 2% (full DLBCL cohort)

AE: adverse event; auto-HCT: autologous stem cell transplant; Benda: bendamustine; CAR: chimeric antigen receptor; CR: complete response; CRS: cytokine release syndrome; Cy: cyclophosphamide; DL: dose level; DLBCL NOS: diffuse large B cell lymphoma not otherwise specified; DOR: duration of response; Flu: fludarabine; Gr: grade; m: month; mITT: modified intention to treat; NR: not reached; NT: neurotoxicity; ORR: overall response rate; OS: overall survival; PMBL: primary mediastinal B-cell lymphoma; TFL: transformed follicular lymphoma.

Reproduced with permission from: Ghobadi A. (2018), Chimeric antigen receptor T cell therapy for non-Hodgkin lymphoma. Current Research in Translational Medicine. 66(2):43–49. Copyright © 2018, published by] Elsevier Masson SAS. All rights reserved.

NK and CAR NK based cellular therapy

NK cells are bone marrow-derived cytotoxic lymphocytes that play a major role in the rejection of tumours and cells infected by viruses, without specific immunization (Vivier, et al 2011). Various activating and inhibitory receptors on the NK cell surface are engaged to regulate NK cell functions and to discriminate target cells from healthy ‘self’ cells (Vivier, et al 2011). Unlike T cells, NK cells kill tumour cells in a major histocompatibility complex-independent manner without the need for prior sensitization (Daher and Rezvani 2018). NK cell-based immunotherapy is associated with a significant NK-versus-leukaemia effect in the absence of the risk of GvHD and a significant decrease in leukaemia relapse when donor/recipient NK KIR-HLA class I mismatch occurs (Ruggeri, et al 2002). NK cells are easily isolated, expanded ex vivo and can be made available as an off-the-shelf allogeneic product for immediate clinical use in adoptive or autologous cell therapies. Concerns about long-term side effects, “off-tumour” effects and the need for a suicide gene to turn off are minimized in NK therapy because of the shorter lifespan than clonally-expanded T cells (Ames and Murphy 2014). Cancer stem cells are recognized as a population of cancer originated cells that are resistant to drugs (Dean, et al 2005). Exciting results have shown that NK cells selectively kill cancer stem cells derived from a variety of human tumours, such as melanoma, glioblastoma and others (Castriconi, et al 2009, Pietra, et al 2009). NK cells are attractive as a therapeutic approach for cancer.

Barriers to NK therapy are the small numbers of active NK cells in unmodified peripheral blood, short lifespan, poor persistence and lack of specific tumour targeting (Barth, et al 2016). Expansion is addressed by multiple groups (Barth, et al 2016). To increase the targeting specificity of expanded NK cells, Chu et al (2015) investigated functional activities of peripheral blood (PB) NK (PBNK) cells modified by mRNA nucleofection with anti-CD20 CAR against CD20+ B-NHL in vitro and in xenografted non-obese diabetic severe combined immunodeficiency gamma (NSG) mice. This non-viral mRNA electroporation approach has great therapeutic potential to transiently express targeted proteins with limited toxicity and long-term low risk of malignant transformation. The authors found that anti-CD20 CAR-expanded PBNK (exPBNK) cells (CAR+ exPBNK) had significantly enhanced in vitro cytotoxicity and extended human BL xenografted NSG survival compared to mock transfected exPBNK cells against BL cells (Chu, et al 2015). Consistent with previous reports that NK cells do not persist after adoptive transfer and they were detectable in the circulation for only 1–2 weeks without cytokine support, the administered expanded human NK cells survived around 2 weeks in NSG mice (Chu, et al 2015). The short lifespan/persistence of adoptively transferred NK cells has limited the therapeutic efficacy.

Enhancing NK/CAR NK persistence and functions with interleukin-15 (IL-15) superagonist ALT-803

IL-15 is a critical factor that supports the development, proliferation, survival and trafficking of several lymphocyte subsets, including NK, CD8⁺ and γδ T cells (Ma, et al 2006). Unlike IL2, which is required to maintain forkhead box P3 (FOXP3)-expressing CD4⁺CD25⁺ Treg cells, IL-15 has little effect on Tregs (Steel, et al 2012). Furthermore, IL-15 promotes the long-term maintenance of CD8⁺CD44hi memory T cells (Waldmann 2015). These advantages attribute IL-15 as an attractive therapeutic agent and it was ranked first as having the greatest potential for use in cancer immunotherapy by the US National Cancer Institute in 2008 (Steel, et al 2012). Unfortunately, the recombinant human IL-15 protein has some limitations, such as a rapid renal clearance and a short plasma half-life, which diminishes its anti-tumour effects (Han, et al 2011). To overcome these limitations, a new IL-15 superagonist (IL-15 SA-[ALT-803]) has been developed by Altor Bioscience (Miramar, FL, USA) with a longer half-life and increased potency (Han, et al 2011). ALT-803 consists of an IL-15 superagonist mutein (IL-15N72D) and a dimeric IL-15 receptor alpha (IL-15Rα)/Fc fusion protein. The pharmacokinetics and biological activity of ALT-803 was significantly improved, such that the superagonist complex has at least 25 times the activity of the native cytokine in-vivo (Zhu, et al 2009) and a significantly longer serum half-life in vivo than wild-type IL-15 (25 h vs. < 40 min). In addition, a single intravenous injection of ALT-803 was capable of inducing mouse CD8⁺ T cell and NK cell proliferation at a dose at which an equal molar dose of IL-15 has no effects, and significantly prolonged survival of C57BL/6 mice bearing syngeneic murine 5T33P myeloma (Han, et al 2011) (Xu, et al 2013). The results of preclinical studies support clinical development of ALT-803 as a novel immunotherapeutic agent. ALT-803 is currently being used in clinical trials to treat patients with myeloma, melanoma and relapsed haematological malignancies. In a recent first-in-human phase 1 trial to evaluate if ALT-803 augments the anti-leukaemia/anti-lymphoma immunity in patients who relapsed >60 days after allogeneic HSCT, the investigators found that ALT-803 was a safe, well-tolerated agent that significantly increased NK and CD8⁺ T cell numbers and function in patients (NCT01885897) (Romee, et al 2018). This study warrants further investigation of the anti-tumour immunity of ALT-803 alone or in combination with other immunotherapies. At the 6^th International NHL symposium, Chu et al (2018a) reported an anti-tumour effect of anti-CD20 CAR NK cells stimulated by ALT-803 against BL using in-vitro cell culture and in-vivo human BL xenografted immune-deficient NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ (NSG) mouse model. They found that ALT-803 significantly promoted exPBNK proliferation, persistence and in-vitro and in-vivo cytotoxicity against paediatric BL cells (Chu, et al 2018a). Further work includes evaluating the CAR NK and ALT-803 combination against other CAYA NHL diseases, generating clinical-grade CAR modified ex-vivo expanded NK cells and evaluating the anti-tumour effect of clinical grade CAR NK with ALT-803 in phase I clinical trial.

Redirecting NK cell-based therapy with a novel antibody-ALT-803 fusion 2B8T2M

Previous studies showed that ALT-803 may act as a versatile protein scaffold to create novel targeted immunotherapeutic agents. Using ALT-803 as a scaffold, 2B8T2M was generated by fusing four single-chains of the tumour-targeting monoclonal antibody, rituximab to the N terminal of IL-15N72D and IL-15RαSuFc (Liu, et al 2016) (Fig. 3). 2B8T2M exhibits tri-specific binding activity through its recognition of CD20 on B lymphoma cells, IL2/15Rβγc on immune cells, and Fc receptor on NK cells and macrophages (Alter, et al 2018, Liu, et al 2016). 2B8T2M also activated NK cells to enhance antibody-dependent cellular cytotoxicity (ADCC), and induced apoptosis of B-lymphoma cells in-vitro and in Daudi xenografted NSG mice (Liu, et al 2016). A biodistribution study of 2B8T2M demonstrated that this fusion protein has the ability to home to lymphoid tissues and retained there for at least 70 h (Liu, et al 2016). 2B8T2M was shown to deplete B cells in cynomolgus monkeys and, importantly, no significant adverse events were observed (Liu, et al 2016). These results demonstrate that fusing a single chain antibody domain to ALT-803 is an effective strategy to potentiate tumour targeting immunotherapeutic molecule without compromising its immunostimulatory capabilities (Alter, et al 2018). Additionally, the ALT-803 scaffold may also provide a vehicle for preferential delivery to lymphoid organs to allow proper and prolonged immune activation (Alter, et al 2018, Liu, et al 2016). Rituximab has been widely used in frontline treatment of B-NHL, however, some patients treated with rituximab, relapse, which limits patient treatment options (Goldman, et al 2013). Dr. Cairo’s group performed in vitro and in vivo preclinical studies to investigate if 2B8T2M significantly enhanced the cytotoxicity of exPBNK against rituximab-sensitive and -resistant BL cells: in this meeting, they reported that 2B8T2M significantly enhanced exPBNK cytotoxicity against rituximab-sensitive and -resistant BL cells compared to controls (Chu, et al 2018b). This study and others suggest that 2B8T2M may be more beneficial compared to a combination of individual molecules for treatment of cancer and that 2B8T2M is an attractive novel agent for CD20⁺ relapsed/refractory CAYA NHL.

Figure 3. ALT-803 as a scaffold for creation of novel fusion molecules.

A schematic diagram showing construction of novel fusion molecules using ALT-803 as a scaffold and a therapeutic antibody as an example. (A) ALT-803 structure:IL-15N72D:IL-15RαSuFc complex consisting of IL-15N72D associated with the dimeric IL-15RαSuFc fusion protein. (B) Structure of a therapeutic antibody showing the variable region consisting of the light and heavy chains. (C) Variable gene segment of the antibody light chain linked to the 5’ end of the variable gene segment of the antibody heavy chain via a Gly4Ser3 linker to create a single chain antibody. (D) A single chain antibody construct genetically fused to the 5’ end of the IL-15N72D mutein and a single chain antibody construct genetically fused to the 5’ end of the IL-15RαSuFc construct. (E) A soluble novel ALT-803-based fusion molecule is produced. The fusion protein is comprised of four single chain antibody domains: two are fused to the two IL-15N72D muteins and two are fused to the two IL-15RαSuFc fusion constructs.

IL-15, interleukin-15.

Reproduced from: Alter S., Rhode P.R., Jeng E.K., Hing C. Wong. H.C. (2018), Targeted IL-15-based Protein Fusion Complexes as Cancer Immunotherapy Approaches. Journal of Immunological Sciences, 2(1): 15–18.

Enhancing antibody-dependent NK cell-mediated cytotoxicity

Obinutuzumab is a humanized, type II anti-CD20 monoclonal antibody glycoengineered to enhance Fc receptor affinity (Herter, et al 2013). It has lower complement-dependent cytotoxicity than rituximab but greater ADCC and phagocytosis and greater direct B-cell killing effects (Herter, et al 2013). Obinutuzumab was first approved by the FDA in 2013 and by the European Medicines Agency in 2014 as a first-line treatment of chronic lymphocytic leukaemia and has been approved for the treatment of patients with relapsed/refractory FL (Cartron and Watier 2017). Obinutuzumab safety and efficacy in B-NHL were noted in the GAUSS study demonstrating a higher ORR compared to rituximab (44.6% vs 26.7%) at the end of induction (week 8) in 149 relapsed/refractory indolent NHL patients with FL, though no increase in PFS was observed (Barth, et al 2016, Sehn, et al 2015). In aggressive B-NHL, obinutuzumab was investigated in relapsed/refractory DLBCL and mantle cell lymphoma at two dose levels (1600/800 and 400/400) with a higher ORR of 32% at the end of treatment in 19 patients in the higher dose group including responses in 4/12 (33%) rituximab refractory patients (Barth, et al 2016, Morschhauser, et al 2013). In this meeting, Barth et al (2018) reported the early results of a recently opened clinical trial by the CAYA NHL Translational Research and Treatment (CAN TREAT) consortium to investigate obinutuzumab in combination with ifosfamide, carboplatin and etoposide chemotherapy in relapsed/refractory CAYA B-NHL (NCT02393157). In this phase 2 study, the addition of obinutuzumab to standard salvage chemotherapy was well tolerated with an ORR of 100% in four patients following a median of 2 cycles (range 1–3) and with no treatment-associated adverse events observed (Barth, et al 2018). Considering the enhanced ADCC of obinutuzumab compared to rituximab in the setting of rituximab-sensitive and -resistant BL (Awasthi, et al 2015), Dr. Cairo’s group performed a preclinical study and reported in this meeting that obinutuzumab combined with expanded NK cells significantly enhances the cell death and NK-mediated ADCC against PMBCL in-vitro and significantly enhanced the survival and decreased tumour burden in PMBCL NSG xenografts compared to rituximab combined with expanded NK cells (Awasthi, et al 2018b). These preliminary studies support the early phase clinical trials of obinutuzumab combined with expanded NK cells for CAYA patients.

Summary

The future is bright for both cellular and humoral therapy in CAYA with relapsed/refractory poor risk mature B-cell lymphomas. A number of therapeutic approaches include next generation monoclonal antibodies, antibody drug conjugates, BiTE, CAR T cells, CAR NK cells and combinations of the above. The challenge will be to determine which strategies to pursue with so few patients available for novel clinical trials and how to leverage international collaboration in the next generation of clinical trials.