Anti-CD20 treatment for B-cell malignancies: current status and future directions

ABSTRACT

Introduction

The introduction of anti-CD20 monoclonal antibody therapy with rituximab in the 1990s greatly improved outcomes for patients with B-cell malignancies. Disease resistance or relapse after successful initial therapy and declining efficacy of subsequent rounds of treatment were the basis for the development of alternative anti-CD20-based antibody therapies.

Areas covered

The novel anti-CD20 antibodies of atumumab, ublituximab, and obinutuzumab were developed to be differentiated via structural and mechanistic features over rituximab. We provide an overview of preclinical and clinical data, and demonstrate ways in which the pharmacodynamic properties of these novel agents translate into clinical benefit for patients.

Expert opinion

Of the novel anti-CD20 antibodies, only obinutuzumab has shown consistently improved efficacy over rituximab in randomized pivotal trials in indolent non-Hodgkin lymphoma and chronic lymphocytic leukemia. The Phase 3 GALLIUM trial demonstrated significant improvements in progression-free survival with obinutuzumab-based immunochemotherapy over rituximab-based immunochemotherapy. Novel combinations of obinutuzumab, including with chemotherapy-free options are being explored, such as with the newly approved combinations of obinutuzumab with venetoclax, ibrutinib, or acalabrutinib. The biggest unmet need remains in the treatment of diffuse large B-cell lymphoma; emerging options in this field include the use of CAR-T cells and T-cell bispecific antibodies.

KEYWORDS:

• Chronic lymphocytic leukemia (CLL)
• diffuse large B-cell lymphoma (DLBCL)
• follicular lymphoma (FL)
• non-Hodgkin lymphoma (NHL)
• obinutuzumab
• ofatumumab
• rituximab
• ublituximab

1. Introduction

CD20 is a 33- to 37-kDa non-glycosylated phosphoprotein expressed on the surface of mature undifferentiated B-cells [1]. Expression starts at the pre–B-cell stage, and persists until terminal differentiation into plasma cells [2]. This pattern, together with consistent and high levels of expression of CD20 on malignant B-cells [3,4], makes CD20 a therapeutic target. The exact biologic function of CD20 is unknown, but it has been suggested to be involved in B-cell receptor activation and proliferation, and Ca²⁺ transport [2,5]. Its structure in complex with rituximab has been recently determined by cryo-electron microscopy (EM) [6].

The outlook for patients with CD20-positive B-cell malignancies improved in the 1990s with the introduction of targeted therapy with the recombinant chimeric murine/human type I anti-CD20 monoclonal antibody (mAb) rituximab (RITUXAN®/MABTHERA®) [7,8]. Rituximab engages Fc receptors on natural killer (NK) cells and macrophages, facilitates complement-dependent cytotoxicity (CDC) and antibody-dependent cell-mediated cytotoxicity/phagocytosis (ADCC/ADCP), and exerts direct antiproliferative and pro-apoptotic effects [9,10].

Rituximab also enhances the in vitro activity of cytotoxic agents in lymphoma cells [11,12]. Its use has resulted in significantly improved outcomes in patients with B-cell malignancies [13–1617], and rituximab-based therapies are among the most commonly used treatments in diseases including follicular lymphoma (FL), diffuse large B-cell lymphoma (DLBCL), and chronic lymphocytic leukemia (CLL) [18–21].

Despite these advances, unmet needs remain in management of CLL and non-Hodgkin lymphoma (NHL). A significant proportion of patients relapse on or become refractory to treatment [22,23]. Resistance to rituximab is not fully understood, but suggested mechanisms include resistance to CDC [24–26], ADCC [27–31], or apoptosis [32,33]; and CD20 loss [34–3536], ‘shaving’ (loss of CD20 following saturation of elimination mechanisms on malignant B-cells that subsequently remain in circulation [37]) [38–40] or modulation via internalization [34,41–44]. In addition, intravenous administration of rituximab is associated with a number of challenges, e.g. lengthy administration times, and this has prompted research into alternative dosing methods [45]. Therefore, there is a need for treatments with (i) improved efficacy over rituximab to prolong response, (ii) the ability to overcome rituximab resistance in refractory patients, and (iii) improved treatment delivery methods. Several new anti-CD20 mAbs are currently under clinical investigation [46,47].

Obinutuzumab (GA101; GAZYVA®/GAZYVARO^TM) is a humanized glycoengineered anti-CD20 antibody that is Food and Drug Administration (FDA)-approved for first-line CLL treatment in combination with chlorambucil, for first-line FL treatment in combination with chemotherapy, and for relapsed or refractory (R/R) FL in combination with bendamustine [48]. It has also been recently approved for use in CLL in combination with venetoclax, ibrutinib, or acalabrutinib [49 –52]. In Europe, obinutuzumab is approved in combination with chlorambucil, venetoclax, or ibrutinib for previously untreated CLL, in combination with chemotherapy for previously untreated FL or with bendamustine in refractory FL [53–55].

Other CD20 mAbs include ofatumumab (Arzerra®, HuMax-CD20) and ublituximab. Ofatumumab is FDA-approved in combination with chlorambucil for first-line CLL treatment where fludarabine is deemed inappropriate, with fludarabine and cyclophosphamide for patients with relapsed CLL, for extended treatment for patients with complete (CR) or partial (PR) response after at least two lines of therapy for recurrent or progressive CLL, and for CLL refractory to fludarabine and alemtuzumab [56]. Unfortunately, no pivotal study data are available for ofatumumab from randomized clinical studies demonstrating superior efficacy versus rituximab. Recently, ofatumumab has been withdrawn from the market in the EU for commercial reasons [57]. Ublituximab (TG-1101) is a chimeric, glycoengineered antibody currently in Phase 3 clinical trials for patients with hematologic malignancies [58].

This review focuses on these three new mAbs, as they are either approved or have reached pivotal trials, and compares and contrasts them with rituximab. We discuss whether differences in preclinical activity translate into clinically meaningful benefit or advantages in patients with CLL, or indolent or aggressive (i/a) NHL. A recently introduced subcutaneous dosing option for rituximab is also discussed.

2. Preclinical experience

Novel anti-CD20 mAbs have structural and pharmacodynamic characteristics that distinguish them from rituximab. In this section, we examine these characteristics and their mechanistic consequences in preparation for later discussion of their clinical consequences.

Anti-CD20 mAbs are broadly divided into types I and II (Table 1 and Figure 1): type I mAbs are noted typically for their CDC and ADCC effects, whereas type II mAbs are associated with enhanced direct cell death effects, reduced CDC, and retained ADCC/ADCP [9,46,47,59]. Ublituximab and obinutuzumab also differ from rituximab in terms of modification of their Fc domains to enhance ADCC/ADCP by Fc-glycoengineering.

Figure 1. CD20 binding of type I and type II anti-CD20 monoclonal antibodies. Adapted from Goede et al. [62] and Klein et al. [47], with permission

Table 1. Characteristics of anti-CD20 monoclonal antibodies (mAbs) [46,47,59,61,90,101]

mAb	Status	Characteristics	ADCC	CDC	Direct effects	Epitope	Activity relative to rituximab*
Rituximab	Approved for NHL, CLL	Type I** Mouse/human chimeric IgG1	++	++	+	Class I Large extracellular loop Core epitope: ^168EPANPS^173 region Secondary epitope from cryo-EM: Ile76 of ECL1 and Pro160 of ECL2	NA
Obinutuzumab	Approved for CLL, FL	Type II** Humanized IgG1	++++	–	++++	Class II Large extracellular loop Core epitope: ^170ANPSEKNSP^178 region	↑ cell death mediated by non-apoptotic pathway involving ↑ ROS production + membrane permeabilization Fc modification ↑ affinity for Fcγ receptors: ↑ ADCC/ADCP Compensates for inhibitory KIR/HLA interactions ↑ overall B-cell depletion in whole blood
Ofatumumab	Approved for CLL	Type I** Human IgG1	++	++++	+	Class I Large extracellular loop Core epitope: ^146FLKMESLNFIRAHT^159 region Small extracellular loop: ^72IPAGIYAPI^80 region	↑ cell membrane proximity and slower dissociation ↑ CDC, maintained in presence of ↓ CD20 expression Does not induce apoptosis Similar ADCC
Ublituximab	In Phase 3 for NHL; Phase 2 CLL	Type I** Chimeric IgG1	++++	++	+	Class I Large extracellular loop Core epitope: ^158HT^159 and ^168EPAN^171 regions	Fc modification ↑ affinity for Fcγ receptors: ↑ ADCC ↑↑ NK cell-mediated ADCC regardless of surface CD20 expression ↑ NK cell-mediated degranulation in presence of CLL cells

+ to ++++ indicates level of cytotoxic activity, from low to very high; – indicates lack of cytotoxic activity; ↑ indicates increase; ↓ indicates decrease.

*See main text for full discussion and references.

**Type I characteristics include CD20 localization to lipid rafts, weak homotypic aggregation, high FcγRIIb-mediated CD20 modulation, and caspase-dependent or independent apoptosis induction; Type II characteristics include no CD20 lipid raft localization, strong homotypic aggregation, low FcγRIIb-mediated CD20 modulation, and caspase-independent apoptosis induction.

ADCC, antibody-dependent cell-mediated cytotoxicity; ADCP, antibody-dependent cell-mediated phagocytosis; CDC, complement-dependent cytotoxicity; CLL, chronic lymphocytic leukemia; cryo-EM, cryo-electron microscopy; FL, follicular lymphoma; HLA, class I human leukocyte antigen; Ig, immunoglobulin; KIR, killer cell immunoglobulin-like receptors; NA, not applicable; NHL, non-Hodgkin lymphoma; ROS, reactive oxygen species.

2.1. Obinutuzumab: type II antibody

2.1.1. Molecular structure

Obinutuzumab was derived by humanization of the parental B-Ly1 mouse antibody and subsequent glycoengineering of the Fc carbohydrate to yield a mAb with a modified Fc region. In addition, obinutuzumab bears a modified elbow-hinge amino acid sequence compared with the parental B-Ly1 antibody [60,61]. It is currently the only type II anti-CD20 approved and under ongoing development. Essentially, the absence of a fucose sugar residue from IgG oligosaccharides in the Fc region increases activity by enhancing binding affinity to FcγRIIIa receptors on immune effector cells [60,62]. Crystal structure analysis has shown that obinutuzumab binds CD20 with an orientation that differs from that of rituximab, despite recognizing an overlapping epitope (Figure 2) [47,61]. The elbow angle of obinutuzumab is 30 degrees greater, resulting in differences in spatial arrangements in CD20 pairs bound to a single molecule of either antibody [47,61]. Protein tomography and confocal microscopy showed different CD20 complex association, and differences in membrane compartmentalization. These conformational differences are thought to underlie the different cellular responses elicited by type I and type II mAbs (see also Table 1 and Figure 1).

Figure 2. Structure and topology of CD20 and epitopes recognized by rituximab, obinutuzumab (GA101), ofatumumab, and ublituximab. Epitopes recognized by the different antibodies are highlighted as follows: ofatumumab (red), rituximab (yellow), obinutuzumab (purple), and the core epitope of ublituximab (boxed). Adapted from Klein et al. [47], with permission

2.1.2. Mechanism of action and preclinical models

Obinutuzumab is associated with strongly increased direct cell death induction compared with rituximab [63–6468] (Table 1). This type II characteristic differentiates obinutuzumab from the type I antibodies rituximab, ofatumumab, and ublituximab, which work primarily by inducing CDC through clustering of CD20 in lipid rafts (Figure 1), and by ADCC (see also Tobinai et al. [9]).

The increased direct cell death (Table 1) seen with obinutuzumab has been demonstrated in NHL and B-lymphoma cell lines [60,64]. It is mediated by a non-apoptotic pathway characterized by actin reorganization and increased homotypic adhesion, lysosomal permeabilization and cathepsin release, and production of reactive oxygen species (ROS) by nicotinamide adenine dinucleotide phosphate oxidase independently of mitochondria [60,64,66]. These type II properties are believed to accrue from elbow-hinge region amino acid substitution and associated conformational changes [47]. Type I mAbs cause only minimal ROS generation and programmed cell death.

Involvement of a non-apoptotic pathway was shown by mitochondrial membrane permeabilization (MMP) and ROS generation in wild-type and BCL-2–overexpressing Raji cells [64]. Cellular mitochondrial oxygen metabolism is a major source of ROS. During apoptosis, MMP and loss of mitochondrial membrane depolarization result in excessive ROS production [69]. BCL-2 is a major inhibitor of MMP and consequently of ROS production during apoptosis [70], but overexpression of BCL-2 in Raji cells did not inhibit obinutuzumab-induced MMP loss and ROS generation when compared with wild-type Raji cells [64].

Glycoengineering of the Fc region of obinutuzumab increases its affinity for activating Fcγ receptor FcγRIII [71], and this is responsible for the superior ADCC and ADCP (including trogocytosis) effects of obinutuzumab over rituximab [60,71–74] (Table 1). Through its Fc modification, obinutuzumab compensates for inhibitory interactions (that affect rituximab) between killer cell immunoglobulin-like receptors and class I human leukocyte antigen, which leads to recruitment of additional NK cells, the main effectors of ADCC [75]. Superior ADCP of glycoengineered mAbs has been linked to improved Kupffer cell-mediated phagocytosis of B-cells in vivo [76].

In contrast, CDC makes no meaningful contribution to the action of obinutuzumab [71] (Table 1). Cryo-EM studies have recently shown that two rituximab-binding Fab domains bind to one CD20 dimer, with an extensive homotypic Fab–Fab interface [6]. This binding geometry results in the assembly of a hexameric rituximab-CD20 complex that is highly compatible with the hexameric CDC-inducing complement factor. Type I mAbs form lipid rafts with FcγRIIb receptors, followed by internalization of CD20. Conversely, type II mAbs are believed to bind a single CD20 dimer and are not associated with lipid raft localization and internalization [47]. This inter- vs intra-dimer binding of Type II vs Type I CD20 antibodies was recently confirmed by cryo-EM structure in a study comparing obinutuzumab, rituximab, and ofatumumab [77]. As CDC activation correlates with the ability of an antibody to translocate CD20 onto lipid rafts [78,79], obinutuzumab shows minimal CDC activity relative to rituximab [71].

Significantly increased and/or faster B-cell depletion relative to rituximab (Table 1) was shown with obinutuzumab in non-human primate models, in whole blood from healthy human donors, and CLL patients, irrespective of prognostic indicators [60,80–82]. Interestingly, superior B-cell depletion was also demonstrated in samples from patients with rheumatoid arthritis or systemic lupus erythematosus [83].

The mechanistic features of obinutuzumab translated into superiority over rituximab in a number of human lymphoma xenograft mouse models, including:

• dose-dependent efficacy with complete tumor regression in a staged aggressive SU-DHL4 model of DLBCL [60]. There were no complete tumor remissions in animals treated with rituximab, even with saturating doses. Second-line obinutuzumab treatment was effective in rituximab-refractory tumors;

• superiority over rituximab in an aggressive disseminated model of MCL using Z138 cells [60];

• superior direct cell death induction, B-cell depletion, and anti-tumor efficacy over ofatumumab in SU-DHL4 and RL xenografts [71].

The important contribution of direct cell death induction to B-cell depletion and in vivo anti-tumor efficacy in xenograft models was confirmed using an Fc-mutated version of obinutuzumab that does not exhibit ADCC, ADCP, or CDC [84].

Obinutuzumab increased inhibition of tumor growth relative to rituximab when used in combination with cytotoxic agents (bendamustine, or fludarabine, chlorambucil or cyclophosphamide with vincristine) in murine Z138 MCL and WSU-DLCL2 DLBCL tumors [85]. Preclinical experiments combining obinutuzumab with targeted agents showed efficacy in the absence of chemotherapy with venetoclax [86], idasanutlin [86,87], Bruton’s tyrosine kinase (BTK) inhibitors [88], and PI3K-delta inhibitors [84,89].

2.2. Ofatumumab: type I antibody

2.2.1. Molecular structure

Ofatumumab is a type I anti-CD20 IgG1 mAb that specifically recognizes a novel epitope that includes the small (including amino acid residues 74 to 80) and large extracellular loops (including residues 159, 163 and 166) of CD20 (Table 1 and Figure 2). It is fully glycosylated at heavy chain amino acid Asn302 [90].

Crystallography data show that ofatumumab has six complementarity-determining region loops that form a deep pocket in the region that binds with CD20 [91]. This hydrophobic pocket interacts with hydrophobic residues on the small and large extracellular loops of CD20, permitting closer binding to the cell membrane which results in enhanced CDC [92,93]. Ofatumumab may also dissociate more slowly from the cell membrane than rituximab [93] (Table 1).

2.2.2. Mechanism of action and preclinical models

As a Type I mAb, ofatumumab induces CD20 clustering onto lipid rafts [93]. In contrast to rituximab, however, ofatumumab does not induce significant B-cell apoptosis [93,94] (Table 1). The recognition of the small extracellular loop that enables very close antibody binding to the cell membrane leads to more potent deposition of complement product and more efficient CDC by ofatumumab [92,93] at lower densities of cell surface CD20 than by rituximab [90,93] (Table 1).

The activity of ofatumumab has been observed in a wide variety of B-lymphoma cell lines (including systems with rituximab-resistant cells):

• enhanced complement activation with ofatumumab over rituximab at reduced concentrations of C1q in Raji, Daudi, and Z138 B-lymphoma cells [95];

• greater CDC with ofatumumab than with rituximab in rituximab-sensitive Raji, Ramos, U2932, and U-698-M lymphoma cells [96];

• activity in rituximab-resistant Raji 4RH and RL 4RH cells [96].

Ofatumumab has been shown to be more effective in vitro than rituximab at killing CLL cells, which are considered relatively resistant to rituximab because of their low CD20 expression [93]. This activity has been shown via studies of NK cell activation and complement inhibitors (e.g. CD55, CD59) to be primarily complement-dependent [97,98].

Greater CDC with ofatumumab than with rituximab has been reported in cytarabine-sensitive and resistant mantle cell lymphoma (MCL) models, with enhanced in vitro CDC over rituximab even where low CD20 levels and/or high complement-inhibitory protein levels were present [99]. Furthermore, prolonged survival with ofatumumab versus rituximab or no treatment, and greater loss of activity with rituximab versus ofatumumab in the presence of CD20 downregulation following long-term mAb exposure, have been shown in a human lymphoma-bearing mouse model [96].

2.3. Ublituximab (TG-1101; EMAB-6): type I antibody

2.3.1. Molecular structure

Ublituximab is a type I chimeric glycoengineered IgG1 mAb targeting a unique epitope on CD20 (Figure 2) [100], and characterized by low Fc region fucose content that improves FcγRIII binding [101–104]. FcγRIIIa is the activating Fc receptor found on NK cells, macrophages, and monocytes [105] (Table 1).

2.3.2. Mechanism of action and preclinical models

Ublituximab induces apoptosis and CDC similarly to rituximab. Similarly to obinutuzumab, it increases NK cell-mediated ADCC 100-fold relative to rituximab in patient–donor CLL cells, regardless of their CD20 expression levels [106]. Preclinical studies of cells from the peripheral blood of CLL patients and of cell lines showed greater in vitro ADCC relative to rituximab against CLL cells, with increased FcγRIIIa-mediated interleukin-2 production by FcγRIIIa+ Jurkat cells, and enhanced NK cell-mediated degranulation, in the presence of autologous CLL cells [103] (Table 1).

Significantly enhanced ADCC with ublituximab over rituximab has been observed in NHL xenografts and cell lines, including cells resistant to rituximab and cytarabine [107]. This has translated into improved anti-tumor efficacy versus rituximab in MCL and FL models [108], and in primary cerebral and intraocular lymphomas [109,110]. Ublituximab was injected intracerebrally or intravitreously 4–7 days after treatment of BALB/c mice with A20.IIA B-lymphoma cells transfected with green fluorescent protein and CD20 [109]. Tumor cell reduction correlated with an increased proportion of CD8 + T-cells, and was observed only in cells expressing CD20.

In a study of NK cells relative to circulating B-cell clones in 47 patients with Waldenström’s macroglobulinemia, greater lytic efficacy of NK cells against B-cell clones was noted in the presence of ublituximab than with rituximab [111].

3. Clinical experience

The following discussion focuses on studies that show how the structural and mechanistic features of recently introduced anti-CD20 mAbs may help to overcome unmet needs in the clinical management of CLL and NHL. Key time-to-event endpoints (i.e. survival) and safety from primary and/or approval-related analyses are summarized in Tables 2–4. Additional details such as tumor response rates, other time-to-event endpoints and hazard ratios (HRs) where available, and results from additional follow-up, are shown in Table S1.

Table 2. Survival and adverse events in key clinical studies of novel anti-CD20 monoclonal antibodies (mAbs) in chronic lymphocytic leukemia (CLL)

Study	No. and type of patients	Treatments*	Survival (median unless stated otherwise)	Significant differences between groups	Safety/tolerability
Obinutuzumab as first-line therapy
CLL11: primary analysis (May 2013) [114] NCT01010061 Phase 3	118	CLB 0.5 mg/kg d1 + d15 × 6 q28d	PFS 11.1 mo		AE grade ≥3 50% IRR grade ≥3 0% Neutropenia grade ≥3 16% Thrombocytopenia grade ≥3 4% Infection grade ≥3 14%SAE 38%Death due to AE 9%
	333	G 1000 mg d1–2, d8, + d15, then d1 c2–6 +  CLB 0.5 mg/kg d1 + d15 q28d	PFS 26.7 mo	p < 0.001 vs. RIT-CLB	AE grade ≥3 70% IRR grade ≥3 20% Neutropenia grade ≥3 33% Thrombocytopenia grade ≥3 10% Infection grade ≥3 12%SAE 39%Death due to AE 4%
	330	RIT 375 mg/m^2 d1 c1, then 500 mg/m^2 d1 c2–6 +  CLB 0.5 mg/kg d1 + d15 q28d	PFS 15.2 mo		AE grade ≥3 55% IRR grade ≥3 4% Neutropenia grade ≥3 28% Thrombocytopenia grade ≥3 3% Infection grade ≥3 14%SAE 32%Death due to AE 6%
CLL11: final analysis (October 2017) [112]	333	G 1000 mg d1–2, d8, + d15, then d1 c2–6 + CLB 0.5 mg/kg d1 + d15 q28d	PFS 28.9 mo	p < 0.0001 vs. RIT-CLB
			OS NR	p = 0.0245 vs. RIT-CLB
	330	RIT 375 mg/m^2 d1 c1, then 500 mg/m^2 d1 c2–6 +  CLB 0.5 mg/kg d1 + d15 q28d	PFS 15.7 mo
			OS 73.1 mo
Ofatumumab in relapsed or refractory disease
Hx-CD20-406 [118] NCT00349349 Phase 2	117 prior RIT (98 RIT-refractory)	OFA 300 mg d1, then 2000 mg wkly × 7, then 2000 mg monthly (12 doses total)	PFS 5.3 (95% CI 4.0–5.7) mo		IRR (related) grade ≥3 5%Neutropenia grade ≥3 16%Infection grade ≥3 31%
			PFS (RIT-refractory) 5.5 (95% CI 4.1–6.3) mo
			OS (all) 15.5 (95% CI 10.4–18.6) mo
	89 RIT-naïve		PFS 5.6 (95% CI 4.8–7.1) mo		IRRs (related) grade ≥3 9%Neutropenia grade ≥3 13%Infection grade ≥3 25%
			OS 20.2 (13.7–22.8) mo
Ublituximab in relapsed or refractory disease
GENUINE [120] NCT02301156 Phase 3	117 high-risk	IBR 420 mg daily	PFS 35.9 (17–NE) mo		Neutropenia grade ≥3 12%
		IBR 420 mg daily + UTX 900 mg d1, 8, 15 c1, then d1 c2–6, then q3c	PFS NR (NE–NE)		IRR grade ≥3 3%Neutropenia grade ≥3 19%

*Two G doses separated by forward slash = day 1/day 8 dose in cycle 1, with the second (day 8) dose given on day 1 of subsequent cycles.

AE, adverse event; c, cycle(s); CLB, chlorambucil; CI, confidence interval; d, day(s); G, obinutuzumab; IBR, ibrutinib; IRR, infusion-related reaction; mo, month(s); NE, not estimable; NR, not reached; OFA, ofatumumab; OS, overall survival; PFS, progression-free survival; q, every; RIT, rituximab; SAE, serious adverse event; UTX, ublituximab; wkly, once weekly.

Table 3. Progression-free and overall survival, and adverse events in key clinical studies of novel anti-CD20 monoclonal antibodies (mAbs) in indolent non-Hodgkin lymphoma (iNHL)

Study	No. and type of patients	Treatments*	Survival (median unless stated otherwise)	Significant differences between groups	Safety/tolerability
Obinutuzumab as first-line therapy
GALLIUM [121] NCT01332968 Phase 3	601 FL	CHOP, CVP, or BEN + G 1000 mg d1 + d8 + d15, then d1 c2–6 or 8 (CTX-dependent), then q2mo if response for ≤2 y	3-y PFS 80.0% (95% CI 75.9–83.6%)	p = 0.001 vs. RIT	AE grade ≥3 74.6% IRR grade ≥3 6.7% Neutropenia grade ≥3 43.9% Thrombocytopenia grade ≥3 6.1% Infection grade ≥3 19.8% SAE 46.1%Death due to AE 4.0%
			3-y OS 94.0% (95% CI 91.6–95.7%)	NS vs. RIT
	601 FL	CHOP, CVP, or BEN + RIT 375 mg/m^2 d1 c1–6 or 8 (CTX-dependent), then q2mo if response for ≤2 y	3-y PFS 73.3% (95% CI 68.8–77.2%)		AE grade ≥3 67.8% IRR grade ≥3 3.7% Neutropenia grade ≥3 37.9% Thrombocytopenia grade ≥3 2.7% Infection grade ≥3 15.6% SAE 39.9%Death due to AE 3.4%
			3-y OS 92.1% (95% CI 89.5–94.1%)
Obinutuzumab in relapsed or refractory disease
GAUSS [127] NCT00576758 Phase 2	88 relapsed (74 FL)	G 1000 mg wkly × 4, then maintenance if response for ≤2 y	PFS 17.6 mo (FL)	NS vs. RIT	IRR grade ≥3 11%Neutropenia grade ≥3 3%SAE 15%18 deaths (10 PD)
	87 relapsed (75 FL)	RIT 375 mg/m^2 wkly × 4, then maintenance if response for ≤2 y	PFS 25.4 mo (FL)		IRR grade ≥3 5%Neutropenia grade ≥3 7%SAEs 15%11 deaths (5 PD)
GADOLIN: primary analysis (September 2014) [128] NCT01059630 Phase 3	202 RIT-refractory (166 FL)	BEN 120 mg/m^2 d1 + d2 c1–6	PFS 14.9 mo		AE grade ≥3 62% IRR grade ≥3 6% Neutropenia grade ≥3 26% Thrombocytopenia grade ≥3 16% Anemia grade ≥3 10% SAE 33%Death due to AE 6%
			OS NR
	194 RIT-refractory (155 FL)	BEN 90 mg/m^2 d1 + d2 c1–6 +  G 1000 mg d1, d8, d15 c1, then d1 c2–6, then q2mo × 2y or until response	PFS NR	p = 0.0001 vs. BEN	AE grade ≥3 68% IRR grade ≥3 11% Neutropenia grade ≥3 33% Thrombocytopenia grade ≥3 11% Anemia grade ≥3 8% SAE 38%Death due to AE 6%
			OS NR	NS vs. BEN
GADOLIN: updated analysis (April 2016) [129] NCT01059630	209 RIT-refractory (171 FL)	BEN 120 mg/m^2 d1 +  d2 c1–6	PFS 14.1 mo		AE grade ≥3 65.5% IRR grade ≥3 3.4% Neutropenia grade ≥3 27.1% Thrombocytopenia grade ≥3 10.8% Infection grade ≥3 19.2% SAE 36.9%Death due to AE 6.4%
			OS NR; events 34.9%
	204 RIT-refractory (164 FL)	BEN 90 mg/m^2 d1 + d2 c1–6 +  G 1000 mg d1, d8, d15 c1, then d1 c2–6, then q2mo × 2y or until response	PFS 25.8 mo	p < 0.001 vs. BEN	AE grade ≥3 72.5% IRR grade ≥3 9.3% Neutropenia grade ≥3 34.8% Thrombocytopenia grade ≥3 15.8% Infection grade ≥3 22.5%SAE 43.6%Death due to AE 7.8%
			OS NR; events 25.5%	p = 0.0269 vs. BEN
Ofatumumab in relapsed or refractory disease
Maloney et al. (HOMER) [132] NCT01200589 Phase 3	205 (179 FL) relapsed after RIT	OFA 1000 mg wkly × 4, then q2mo × 4	PFS (IR) 16.2 mo (95% CI 15.7–21.5)	Study halted for futility	Any AE 93% IRR grade ≥3 4% Neutropenia any grade 3% Pyrexia any-grade 4%SAE NA
			PFS (investigator) 16.1 mo (95% CI 15.5–18.5)
			OS 45.9 mo (95% CI 39.2–NA)
	204 (167 FL) relapsed after RIT	RIT 375 mg/m^2 wkly × 4, then q2mo × 4	PFS (IR) 21.2 mo (95% CI 16.3–25.2)		Any AE 82% IRR grade ≥3 0% Neutropenia 7% Pyrexia any-grade 10% SAE NA
			PFS (investigator) 17.9 mo (95% CI 16.1–21.6)
			OS 47.1 mo (95% CI 45.5–NA)
Czuczman et al. [133] NCT00394836 Phase 3	30 RIT-refractory FL	OFA 300 mg d1, then 500 mg wkly from d8 (8 doses total)	PFS 5.8 mo (overall population) OS NR		IRR grade ≥3 3%Neutropenia grade ≥3 15%Thrombocytopenia grade ≥3 3%Infection grade ≥3 3%Death due to AE 4%
	86 RIT-refractory FL	OFA 300 mg d1, then 1000 mg wkly from d8 (8 doses total)

*Two G doses separated by forward slash = day 1/day 8 dose in cycle 1, with the second (day 8) dose given on day 1 of subsequent cycles.

AE, adverse event; BEN, bendamustine; c, cycle(s); CHOP, cyclophosphamide, doxorubicin, vincristine, and prednisone; CI, confidence interval; CTX, chemotherapy; CVP, cyclophosphamide, vincristine, and prednisone; d, day(s); FL, follicular lymphoma; G, obinutuzumab; IR, independent review; IRR, infusion-related reaction; mo, month(s); NA, not available; NR, not reached; NS, non-significant; OFA, ofatumumab; OS, overall survival; PD, disease progression; PFS, progression-free survival; q, every; RIT, rituximab; SAE, serious adverse event; wkly, once weekly; y, year(s).

Table 4. Survival and adverse events in key clinical studies of novel anti-CD20 monoclonal antibodies (mAbs) in aggressive non-Hodgkin lymphoma (aNHL)

Study	No. and type of patients	Treatments*	Survival (median unless stated otherwise)	Significant differences between groups	Safety/tolerability
Obinutuzumab as first-line therapy
GOYA: primary analysis (April 2016) [134] NCT01287741 Phase 3	706 DLBCL	CHOP + G 1000 mg d1 + d8 + d15, then d1 c2–8	3-y PFS 69.6%	NS vs. RIT	AE grade ≥3 73.7% IRR all grades 36.1% Neutropenia grade ≥3 46.2% Infection grade ≥3 19.2%SAE 42.6%Death due to AE 5.8%
			3-y OS 81.2% (95% CI 77.9–84.1%)	NS vs. RIT
			3-y EFS 69.9% (95% CI 66.2–73.2%)	NS vs. RIT
	712 DLBCL	CHOP + RIT 375 mg/m^2 d1 c1–8	3-y PFS 66.9%		AE grade ≥3 64.7% IRR all grades 23.5% Neutropenia grade ≥3 38.1% Infection grade ≥3 15.5%SAE 37.6%Death due to AE 4.3%
			3-y OS 81.4% (95% CI 78.1–84.3%)
			3-y EFS 66.5% (95% CI 62.7–70.1)
GOYA: final analysis (January 2018) [135]	704 DLBCL	CHOP + G 1000 mg d1 + d8 + d15, then d1 c2–8	5-y PFS 63.8% (95% CI 59.3–68.0)	NS vs. RIT	AE grade ≥3 75.1% Neutropenia grade ≥3 47.9%SAE 44.4%Death due to AE 6.1%
			5-y OS 77.0% (95% CI 73.3–80.3)	NS vs. RIT
			5-y EFS 60.6% (95% CI 56.3–64.6)	NS vs. RIT
	710 DLBCL	CHOP + RIT 375 mg/m^2 d1 c1–8	5-y PFS 62.6% (95% CI 58.1–66.8)		AE grade ≥3 65.8% Neutropenia grade ≥3 39.5%SAE 38.4%Death due to AE 4.4%
			5-y OS 77.7% (95% CI 74.1–80.9)
			5-y EFS 58.9% (95% CI 54.5–63.1)
Ofatumumab in relapsed or refractory disease
van Imhoff et al. (ORCHARRD) [139] NCT01014208 Phase 3	222 DLBCL	OFA 1000 mg d1 + d8, then d22 & d43 + DHAP	2-y PFS 24%	p = 0.33 vs. RIT	Neutropenia grade 4 22%Thrombocytopenia grade 4 35%SAE 52% across study; similar between groups
			2-y EFS 16%	p = 0.35 vs. RIT
			OS 13.9 mo	p = 0.38 vs. RIT
	223 DLBCL	RIT 375 mg/m^2 d1 + d8, then d22 & d43 + DHAP	2-y PFS 26%		Neutropenia grade 4 26%Thrombocytopenia grade 4 35%
			2-y EFS 18%
			OS 13.2 mo

*Two G doses separated by forward slash = day 1/day 8 dose in cycle 1, with the second (day 8) dose given on day 1 of subsequent cycles.

AE, adverse event; c, cycle(s); CHOP, cyclophosphamide, doxorubicin, vincristine, and prednisone; CI, confidence interval; d, day(s); DHAP, cisplatin, cytarabine, and dexamethasone; DLBCL, diffuse large B-cell lymphoma; EFS, event-free survival; G, obinutuzumab; IRR, infusion-related reaction; mo, month(s); NS, non-significant; OFA, ofatumumab; OS, overall survival; PFS, progression-free survival; q, every; RIT, rituximab; SAE, serious adverse event; wkly, once weekly; y, year.

3.1. CLL

3.1.1. Obinutuzumab as first-line therapy

The Phase 3 CLL11 trial (NCT01010061) randomized 781 patients with previously untreated CLL and comorbidities [112–114] (Tables 2 and S1) to chlorambucil plus obinutuzumab, chlorambucil plus rituximab, or chlorambucil alone. Significant increases in progression-free survival (PFS) were seen in the obinutuzumab–chlorambucil group relative to the other two groups at the primary analysis (data cutoff May 2013) [114] and in an updated analysis carried out 11 months later [113]. No significant difference between groups in overall survival (OS) was seen in these analyses. However, after approximately 2 years of additional follow-up and accrual of greater numbers of events, obinutuzumab–chlorambucil was associated with significantly increased OS relative to rituximab–chlorambucil [112] (Tables 2 and S1). Tumor response was also significantly greater with obinutuzumab–chlorambucil than in either of the other two groups (Table S1).

Rates of minimal residual disease (MRD) negativity in bone marrow and peripheral blood were significantly higher after obinutuzumab–chlorambucil treatment than after rituximab–chlorambucil (bone marrow 19.5% vs. 2.6%; peripheral blood 37.7% vs. 3.3%) [114]. The authors noted the likely association between MRD status and improved outcomes in patients treated with obinutuzumab–chlorambucil.

Safety data revealed no notable differences between treatment arms [114] (Table 2). Increased rates of neutropenia with obinutuzumab did not translate into higher infection rates. Grade 3/4 neutropenia was reported most commonly with obinutuzumab–chlorambucil (33%), and least frequently with chlorambucil alone (16%). Rates of grade 3–5 infection ranged from 11% to 14%, with no notable differences between treatment groups. Grade 3–4 infusion-related reactions (IRRs) were more frequent with obinutuzumab–chlorambucil than with rituximab–chlorambucil (20% vs. 4%). These severe reactions were seen during the first infusion of obinutuzumab only, and no deaths were associated with them. All-grade tumor lysis syndrome occurred in 14 patients treated with obinutuzumab–chlorambucil and no patients treated with rituximab–chlorambucil in CLL11; all cases resolved. Newly diagnosed neoplasms were reported in 4–7% of patients across treatment groups. Deaths due to adverse events (AEs) were less frequent in the obinutuzumab–chlorambucil group (4%) than with rituximab–chlorambucil (6%) or chlorambucil alone (9%) [114].

3.1.2. Ofatumumab in relapsed or refractory disease

Clinical activity of ofatumumab in patients with relapsed or rituximab-refractory CLL became evident in early-phase trials [115–117].

Ofatumumab was active in fludarabine-refractory CLL regardless of prior rituximab exposure in a retrospective analysis [118] of Phase 2 data from the Hx-CD20-406 study (NCT00349349), which evaluated the efficacy and safety of ofatumumab in patients with fludarabine- and alemtuzumab-refractory CLL, or with fludarabine-refractory CLL with bulky lymphadenopathy [119]. Overall, ofatumumab monotherapy was effective in refractory CLL in Hx-CD20-406, including patients with previous rituximab exposure (Tables 2 and S1). There were more grade 1–2 ofatumumab-related IRRs in patients previously treated with rituximab than in rituximab-naïve patients, although worst-grade IRRs and grade ≥3 hematologic toxicities were similar between subgroups. Proportions of patients experiencing grade ≥3 infections did not differ between subgroups, and prior rituximab exposure did not appear to be linked to toxicity of ofatumumab [118].

3.1.3. Ublituximab in relapsed or refractory disease

The Phase 3 GENUINE study (NCT02301156) in 117 treated patients with high-risk (17p or 11q deletion and/or TP53 mutation) R/R CLL showed that adding ublituximab to ibrutinib resulted in significantly higher overall response (OR; ublituximab + ibrutinib, 90% vs ibrutinib, 69%, p < 0.01) and PFS (ublituximab + ibrutinib, median not reached vs ibrutinib, median 35.9 months, p = 0.016) [120].

3.2. Indolent NHL

3.2.1. Obinutuzumab as first-line therapy

The efficacy and safety of obinutuzumab with chemotherapy (cyclophosphamide, doxorubicin, vincristine, and prednisone [CHOP], bendamustine, or cyclophosphamide, vincristine, and prednisone [CVP]) plus obinutuzumab maintenance in previously untreated iNHL patients was compared head-to-head with the corresponding rituximab-based regimen in the Phase 3 GALLIUM study (NCT01332968) [121] (Tables 3 and S1). After a median follow-up of 34.5 months, a planned interim analysis showed significantly lower risk of progression, relapse, or death with obinutuzumab than with rituximab in 1202 FL patients (HR for PFS 0.66; p = 0.001) [121]. Estimated percentages of patients alive after 3 years in GALLIUM were similar (94.0% vs. 92.1%) for obinutuzumab and rituximab, respectively (Table S1). All other time-to-event endpoints were consistent with those of the primary endpoint of investigator-assessed PFS (Tables 3 and S1). Obinutuzumab-based therapy was associated with a higher MRD negativity rate [122].

High-grade AEs and serious AEs were seen more frequently with obinutuzumab-based therapy in GALLIUM, but rates of AEs resulting in death were similar in the two groups (Table 3). The most common AEs were IRRs; these were observed in 351/595 patients (59.0%) receiving obinutuzumab and in 292/597 patients (48.9%) treated with rituximab. However, the large majority of IRRs were grade <3 (Table 3). Nausea (47% in both groups, all grades) and neutropenia (obinutuzumab–chemotherapy 48%, rituximab–chemotherapy, 44%; all grades) were both commonly reported [121]. Health-related quality of life (a secondary endpoint in GALLIUM) as assessed by the Functional Assessment of Cancer Treatment-Lymphoma (FACT) questionnaire showed similar improvements in both treatment arms, with no indication that treatment-related AEs interfered with these improvements [123].

Updated analysis of GALLIUM, with a median follow-up of almost 5 years, reinforced the clinically meaningful improvements in outcomes seen with obinutuzumab–chemotherapy [124]. OS remained immature, i.e. there were too few events for any difference to be detected in this population with indolent disease (and therefore relatively good life expectancy), in whom OS data may also be confounded by the effect of subsequent lines of treatment. In addition, safety data were consistent with the primary analysis.

In an additional analysis, early disease progression (within 24 months) was found to be less frequent with obinutuzumab-based therapy (57/601 patients) than with rituximab-based therapy (98/601), with a 46.0% average risk reduction (95% CI 25.0–61.1%; cumulative incidence rate 10.1% vs. 17.4%) [125]. In patients with progression events before 24 months who were still alive at the 24-month landmark, the HR for 24-month OS after the landmark versus those who did not have progression events before 24 months was 12.2 (95% CI 5.6–26.5).

Of patients who were evaluable for MRD status in GALLIUM, MRD-negativity was observed at end of induction (EOI) in 92.6% and 85.2% of obinutuzumab and rituximab patients, respectively (p = 0.0034) [122]. Most MRD-negative patients remained negative during maintenance, with no difference between treatment arms in MRD relapse rates.

Death rates were increased in patients receiving bendamustine in GALLIUM, in both the obinutuzumab and rituximab arms. However, it should be noted that chemotherapy allocation was not randomized in this study and patients receiving bendamustine were more likely to be elderly or frail [126].

3.2.2. Obinutuzumab in relapsed or refractory disease

The Phase 2 GAUSS study (NCT00576758) in 175 patients with relapsed iNHL indicated no significant PFS difference between obinutuzumab and rituximab monotherapy (Tables 3 and S1), although there was a higher tumor response rate by blinded independent review panel assessment in 149 patients with FL [127] (Table S1). AEs (all grades) were seen with similar incidences in each arm. IRRs (74% vs. 51% overall) and cough (24% vs. 9%) were reported more frequently with obinutuzumab than with rituximab. Serious AEs affected 15% of patients in both arms (Table 3).

The benefit of adding obinutuzumab to chemotherapy in rituximab-refractory iNHL was shown in the Phase 3 GADOLIN study (NCT01059630) in 396 patients randomized to bendamustine monotherapy or bendamustine–obinutuzumab followed by 2 years of obinutuzumab maintenance (1000 mg every 2 months) [128] (Tables 3 and S1). GADOLIN involved mostly patients with FL (81%); 365 patients (92%) were refractory to their last treatment regardless of whether it had contained rituximab, and 311 (79%) were refractory to both rituximab and alkylating agents. Patients had received a median of two prior treatments. Prior treatment with bendamustine (within 2 years of the start of cycle 1), and prior treatment with obinutuzumab were not permitted. After a median follow-up of 21.9 months in the obinutuzumab–bendamustine group and 20.3 months in the bendamustine group, PFS was significantly lengthened in the obinutuzumab arm (HR 0.55; p = 0.0001) [128] (Tables 3 and S1). Benefit of obinutuzumab treatment extended to OS (HR 0.67; p = 0.0269) in an updated analysis [129] (Tables 3 and S1), and to health-related quality-of-life measures [130]. The PFS findings were supported by other endpoints, including EFS and time-to-next-treatment (Table S1).

The most frequent grade ≥3 AEs were neutropenia, thrombocytopenia, anemia, and IRRs (Table 3). GADOLIN demonstrated that obinutuzumab is clinically useful in patients with iNHL who are no longer responding to rituximab.

In an additional analysis in MRD-evaluable patients, 41/52 (79%) of patients receiving obinutuzumab–bendamustine were MRD-negative versus 17/36 (47%) with bendamustine alone (p = 0.0029) at mid-induction [131]; at EOI, 86% and 55%, respectively, were MRD-negative (p = 0.0002). MRD-negativity at EOI was associated with improved PFS (HR 0.33, 95% CI 0.19–0.56; p < 0.0001) and OS (HR 0.39, 95% CI 0.19–0.78; p = 0.008).

3.2.3. Ofatumumab in relapsed or refractory disease

Ofatumumab did not show any clinical advantage over rituximab in the Phase 3 HOMER study [132] (NCT01200589), which evaluated efficacy and safety of ofatumumab versus rituximab in iNHL patients (98% FL) who had relapsed following treatment with a rituximab-containing regimen (Tables 3 and S1). All patients were still sensitive to rituximab. This study was halted for futility at a planned interim analysis.

Ofatumumab showed good tolerability and modest activity as monotherapy in a Phase 3 study (NCT00394836) of heavily pretreated rituximab-refractory population of 116 FL patients who received eight weekly doses of ofatumumab (300 mg at dose 1 followed by 500 mg or 1000 mg for doses 2 to 8) [133] (Table 3). Twenty-seven patients were refractory to rituximab monotherapy, 45 to rituximab maintenance, and 44 to rituximab–chemotherapy. Because of slow recruitment and lack of additional toxicity with the 1000 mg dosage, the 500 mg dosage was discontinued by protocol amendment. The study population had received a median four courses of prior therapy, 47% had high-risk disease, and 65% were chemotherapy-refractory. An OR rate of 22% was reported for the 27 patients refractory to rituximab monotherapy (Table S1). Most common AEs were infections, rash, urticaria, fatigue, pruritus, nausea, cough, neutropenia, and pyrexia (Table 3). Three patients had grade 3 IRRs, none of which were classed as serious AEs [133].

3.3. Aggressive NHL

3.3.1. Obinutuzumab as first-line therapy

In the Phase 3 GOYA study (NCT01287741), which compared obinutuzumab–CHOP with rituximab–CHOP in patients with previously untreated advanced DLBCL, investigator-assessed PFS was similar between treatments after a median 29 months’ observation [134] (Tables 4 and S1). Secondary endpoints including independently reviewed PFS, other time-to-event parameters including event-free (EFS) and OS, and tumor response rates, were similar between arms. The most common AEs were neutropenia, IRRs, nausea, and constipation; all except IRRs (Table 2) were seen at similar rates in both groups. The subsequently released final results of GOYA, with median follow-up of 47.7 months, showed no change from the primary analysis, with no significant difference between treatment arms for 5-year OS [135].

The results of GOYA were unexpected in light of preclinical and early clinical data. The authors suggested that differences in clinical and biologic profiles between indolent and more aggressive disease may have played a part. Preplanned subgroup analysis of outcome by cell-of-origin showed a trend toward greater benefit with obinutuzumab in the germinal center B-cell disease subtype (PFS HR 0.72 [95% CI 0.51–1.03]), which is known to be more similar to FL than other DLBCL subtypes [136,137] and to have a more favorable prognosis and different immune microenvironment to activated B-cell and unclassified subtypes.

In the Phase 3 GAINED trial (NCT01659099), obinutuzumab or rituximab was combined with four cycles of intensified (2-weekly) induction chemotherapy with CHOP or doxorubicin, cyclophosphamide, vindesine, bleomycin, and prednisone (ACVBP) [138]. After interim analysis with median follow-up of 25.2 months and enrollment of 670 patients with previously untreated DLBCL, there was no significant difference between antibodies in the primary endpoint of 2-year EFS (positron emission tomography positivity after two or four induction cycles, progression or relapse, modification of planned treatment, or death) or any of the secondary endpoints, including PFS, and the trial was halted. There were 250 and 205 serious AEs, and 10 and one fatal AEs, in the obinutuzumab and rituximab arms, respectively.

3.3.2. Ofatumumab in relapsed or refractory disease

Ofatumumab was compared with rituximab plus cisplatin, cytarabine, and dexamethasone (DHAP) as salvage therapy in the ORCHARRD study (NCT01014208) in patients with aNHL [139]. Patients responding after two treatment cycles received high-dose chemotherapy and autologous stem cell transplantation after their third cycle. ORCHARRD assessed 445 patients with R/R DLBCL or other aNHL, but failed to find any difference in efficacy or safety between ofatumumab and rituximab when either was given with DHAP in a salvage chemotherapy setting (Tables 4 and S1).

4. Research directions

Because of general experience and superior outcomes with obinutuzumab in CLL, more recent clinical trials have compared different obinutuzumab-based regimens as a new standard of care, rather than comparing obinutuzumab- and rituximab-based treatments. Ongoing Phase 3 research focuses on combinations of anti-CD20 agents with other targeted therapies as part of an effort to provide effective chemotherapy-free options. These include the BCL-2-specific small-molecule BH3-mimetic agent venetoclax [140] and proapoptotic BTK inhibitors [141,142]. The feasibility of chemotherapy-free treatment was shown in the MURANO study (NCT02005471) in 194 patients with R/R CLL [143]. After a median 36 months’ follow-up, HRs for PFS and OS were 0.16 (95% CI 0.12–0.23) and 0.50 (95% CI 0.30–0.85) for treatment with rituximab plus venetoclax and bendamustine plus rituximab, respectively. Patients who received rituximab plus venetoclax also achieved higher rates of undetectable MRD.

In the CLL14 study (NCT02242942) in 432 patients with previously untreated CLL and comorbidities, venetoclax-obinutuzumab was associated with longer PFS than chlorambucil–obinutuzumab [144], leading to its approval in both the USA and Europe [50,55]. In addition, PFS was significantly longer with a BTK inhibitor plus obinutuzumab than with chlorambucil–obinutuzumab in two trials in treatment-naïve CLL: the iLLUMINATE study (NCT02264574) of ibrutinib in 229 patients [145], which led to approval of ibrutinib plus obinutuzumab by the FDA and European Medicines Agency (EMA) [49,146], and the ELEVATE-TN trial (NCT02475681) of acalabrutinib (approved with or without obinutuzumab) in 535 patients [51,52].

In patients with NHL, a number of studies have shown promise of chemotherapy-free treatment based on anti-CD20 therapy with the immune modulator lenalidomide [147]. The RELEVANCE Phase 3 trial demonstrated good outcomes in 1030 previously untreated patients with advanced FL treated with rituximab–lenalidomide and rituximab–chemotherapy, although it did not demonstrate superiority of rituximab–lenalidomide [148]. The AUGMENT Phase 3 study in 358 patients with R/R iNHL showed improved efficacy with rituximab–lenalidomide over rituximab–placebo (HR for PFS 0.46, 95% CI 0.34–0.62; p < 0.001) [149]. Studies of obinutuzumab–lenalidomide are also now being conducted. The Phase 2 GALEN trial evaluated lenalidomide–obinutuzumab induction followed by 1 year of lenalidomide–obinutuzumab maintenance and 1 year of further obinutuzumab maintenance in 89 patients with R/R FL [150]. The OR rate at EOI was 79% (95% CI 69–87), with 38% CR (95% CI 28–50). A Phase 1–2 study of obinutuzumab–lenalidomide in 66 patients with relapsed iNHL demonstrated an OR rate of 98% (CR 72%). The estimated 24-month PFS after 17 months’ follow-up was 73% (95% CI 57–83) [151]. In addition, a study of obinutuzumab-lenalidomide in 90 patients with previously untreated advanced FL demonstrated a 2-year PFS of 96% (95% CI 92–100) and OR rate of 98% (85 CR, 1 PR) [152]. Toxicity of the combination of obinutuzumab and lenalidomide has been manageable.

The chemotherapy-free combination of obinutuzumab and the new generation immunomodulatory imide drug CC-122 (avadomide) showed promising Phase 1 clinical activity, with encouraging response rates and median PFS, in 38 patients with R/R FL, DLBCL, or marginal zone lymphoma [153].

5. Improving delivery

Rituximab given conventionally by infusion may cause severe infusion reactions [154,155]. This necessitates dose titration during the initial infusion, and administration over at least 3.5 to 4 hours in the first cycle and 90 minutes thereafter [154,155]. To help overcome the challenges associated with these prolonged infusion times, a subcutaneous formulation of rituximab and recombinant human hyaluronidase has been developed [45]. This formulation has similar efficacy and safety characteristics to its intravenous counterpart in both NHL and CLL [45,156–161], but requires administration over 5–7 minutes without the need to maintain venous access, and only 15 minutes of monitoring [162,163]. It is approved in the USA for patients with previously untreated or R/R FL, previously untreated DLBCL, or previously untreated or treated CLL [163], and in Europe for previously untreated or responsive FL, DLBCL, and previously untreated or R/R CLL [162]. Note that the first dose of rituximab is always given by IV infusion, with close monitoring for safety and tolerability, in all patients.

Use of subcutaneous rituximab has potential quality-of-life and economic benefits that accrue from improved comfort and convenience for patients and time savings for clinics. The prospective, randomized, open-label, crossover ‘PrefMab’ study (NCT01724021) showed a preference for subcutaneous over intravenous rituximab in approximately 80% of patients with previously untreated DLBCL or FL [164]. Median administration times per cycle were 6 minutes versus 170−240 minutes for subcutaneous and intravenous dosing, respectively. Similar patient preference findings have been reported in other studies [165,166]. In the Phase 1b SAWYER study (NCT01292603), which investigated the pharmacokinetics and safety of subcutaneous rituximab in CLL patients, over 52/56 (92.9%) patients and 53/56 (94.6%) nurses preferred subcutaneous over intravenous administration [160]. Reductions in healthcare professional time for preparation and administration, and reduced treatment room times for patients, are also possible when subcutaneous dosing is used [167,168]. In a multinational non-interventional time and motion study, active healthcare professional time was reduced by a mean 11.3 minutes (–32%; p < 0.0001) when rituximab was administered subcutaneously rather than intravenously [168]; mean patient chair time was also significantly shorter with subcutaneous rituximab (67.3 vs. 262.1 minutes; p < 0.0001) [168].

6. Conclusion

The introduction of rituximab brought important benefits to patients with B-cell malignancies. Unmet needs remain, however, in patients receiving anti-CD20 therapy. Relapse and treatment resistance continue to affect many patients, and the problem of decreasing response duration and quality with each successive course of treatment persists. Furthermore, infusion of the mAb carries the potential for IRRs, and lengthy infusion times. Nevertheless, improved understanding of relapse and resistance mechanisms has been accompanied by the development of novel anti-CD20 mAbs with enhanced pharmacodynamic properties relative to rituximab.

These mechanistic improvements have been shown in some trials in patients with CLL or iNHL to translate into improved outcomes, mainly in terms of PFS and depth of response. Data are most abundant for obinutuzumab, with increased OS observed after extended follow up of a head-to-head comparison with rituximab in CLL patients (CLL11), and in rituximab-refractory FL patients (GADOLIN). Ofatumumab and ublituximab have shown evidence of clinical benefit that is likely to accrue from alternative structural and mechanistic characteristics in patients with CLL who have relapsed on, or who are refractory to, rituximab, although the picture for these two agents is less clear in patients with iNHL. Additional benefit has not been shown with these newer agents in aNHL. The reasons for this are unclear, but may relate to the biological characteristics of aggressive lymphomas such as DLBCL. Further research focusing on B-cell subtypes may shed more light on this subject.

7. Expert opinion

The development of the type I mAb rituximab revolutionized the treatment of B-cell malignancies. Following its approval by the FDA in 1997 [169] and EMA in 1998 [155], rituximab has become a standard component of care across FL, DLBCL, and CLL, and has been included on the WHO list of essential medicines since 2015 [170]. Several large clinical trials have demonstrated the benefits of combining rituximab with chemotherapy as part of induction and salvage regimens in NHL and CLL [15,171–180]. In addition, 2-year rituximab maintenance was shown in the PRIMA study to provide PFS benefit over 9 years of follow-up [181]. Although there was no OS advantage, over 50% of patients remained free from progression and required no new anti-lymphoma treatment.

Several additional CD20 antibodies have since been developed; of those, only obinutuzumab has shown consistently improved efficacy over rituximab in randomized pivotal trials in iNHL and CLL, whereas it failed to demonstrate significance over rituximab in patients with DLBCL. The Phase 3 GALLIUM study of obinutuzumab–chemotherapy versus rituximab–chemotherapy demonstrated significant improvements in PFS with obinutuzumab over rituximab [121], a finding that was maintained with longer follow-up [124] and across various high-risk subgroups [121]. CLL11 demonstrated superiority of obinutuzumab–chlorambucil over rituximab–chlorambucil or chlorambucil alone in CLL, leading to its approval in that indication [114]. Obinutuzumab plus chemotherapy is now one of the preferred options in management of FL and CLL.

MCL responds poorly to rituximab in terms of survival (see review by Sun and Zhang [182]). Obinutuzumab monotherapy achieved early steady-state concentrations and clinical activity with an acceptable safety profile in the randomized Phase 2 GAUGUIN trial in patients with R/R DLBCL (n = 25) or MCL (n = 15) [183]. Ofatumumab monotherapy resulted in a PR in one of 12 patients, and stable disease in six, in a Phase 2 study in R/R MCL [184].

Rituximab and obinutuzumab are administered at different dosages (rituximab at 375 mg/m² and obinutuzumab at a fixed dose of 1000 mg), and the cumulative dose of obinutuzumab was greater than the cumulative dose of rituximab in GALLIUM [121], which prompts the question as to whether differences in dosing or exposure might affect efficacy outcomes. However, these two mAbs are distinct drugs with differing mechanistic features, half-lives and clearance, and equivalent dosing would therefore not be expected. Moreover, fixed dosing of obinutuzumab is based on safety, efficacy, and pharmacokinetic data from Phase 1–2 trials [127,183,185–187], and the surface area-dependent dosing of rituximab [127] is based on observations of clinical efficacy.

Pharmacokinetic and pharmacodynamic modeling studies suggest that improved outcomes with obinutuzumab are not dose effects. Semi-mechanistic pharmacokinetic/pharmacodynamic modeling based on temporal changes in drug and lymphocyte counts in 632 patients who participated in CLL11 has indicated that obinutuzumab exhibits superior efficacy in CLL, regardless of modeled alternative dosing schedules [188]. In addition, population pharmacokinetic modeling of data from GALLIUM [121] showed that variability in obinutuzumab exposure did not influence PFS in patients receiving obinutuzumab–bendamustine [189]. PFS improved with increasing obinutuzumab exposure in the obinutuzumab–CHOP/CVP arm, but in these patients increased obinutuzumab clearance may have resulted from poorer prognosis or resistance to chemotherapy rather than being the cause of inferior clinical outcome (as demonstrated for rituximab [190]).

Of note, Phase 2 dose-escalation data in small numbers of patients suggested improved outcomes with higher dosages of rituximab [191], but this effect has not been reproduced or confirmed in larger trials [192–194]. Given that the recommended obinutuzumab dose and schedule ensures target saturation, increased efficacy observed with obinutuzumab is considered to be primarily due to mode of action and not the higher absolute dose given.

Typical infusions of rituximab or obinutuzumab take 3.5–4 hours; lengthy administration times are burdensome for patients. A rituximab short duration infusion (SDI) has been explored, and the FDA has approved the use of a rapid infusion over 90 minutes in patients without grade 3/4 IRRs in cycle 1 [169]. The ongoing GAZELLE study (NCT03817853) is currently examining an SDI for obinutuzumab in patients with untreated advanced FL, with first results expected in Q1 2021.

Recent and ongoing research studies are investigating novel combinations of rituximab and obinutuzumab, including chemotherapy-free options. The Phase 3 CLL14 trial demonstrated significantly improved PFS in previously untreated CLL patients with comorbidities treated with obinutuzumab–venetoclax compared with obinutuzumab–chlorambucil (HR 0.35, 95% CI 0.23–0.53; p < 0.001) [144]. Furthermore, combinations with the BTK inhibitors ibrutinib and acalabrutinib showed improved outcomes in pivotal trials [145]. As discussed earlier, following on from the promising findings with rituximab–lenalidomide in RELEVANCE [148] and AUGMENT [149], high response rates and encouraging PFS results are now being shown in studies of obinutuzumab–lenalidomide in patients with R/R iNHL [150–152].

The biggest unmet need in NHL treatment remains in DLBCL, but it is unlikely that CD20 antibodies in combination with standard chemotherapy will be further investigated in this indication. Emerging therapies in ongoing DLBCL trials involve molecules that target the immune system, including anti-CD19 CAR-T cells [195], T-cell bispecific antibodies (TCBs, e.g. blinatumomab [196], the CD20-TDB [T-cell-dependent bispecific antibody] mosunetuzumab [197,198] and the obinutuzumab-based CD20-TCB glofitamab [199]), and a mAb for CD19, tafasitamab (MOR208), which has shown promising activity as a single agent in patients with R/R DLBCL, MCL, FL, or other iNHL [200]. Tafasitamab is now being investigated further in DLBCL as a combination with lenalidomide in the single-arm Phase 2 L-MIND study in 81 patients [201,202], or with bendamustine in comparison with rituximab–bendamustine in the Phase 2/3 B-MIND study with an enrollment target of 330 [203].

CAR-T cells approved for use in R/R DLBCL include axicabtagene ciloleucel [204,205] and tisagenlecleucel [206,207], while a further agent, lisocabtagene maraleucel, is in clinical trials [208]. Such therapies may also move to the frontline setting by, for example, combining CD20-TCBs with anti-CD20 antibodies. Preliminary data from a Phase 1–2 study investigating glofitamab with obinutuzumab in R/R aNHL or FL demonstrated promising clinical activity with no new safety signals [209]. Similarly, the CD20/CD3 bispecific antibody mosunetuzumab has demonstrated durable efficacy and favorable tolerability in heavily pretreated R/R NHL patients in a recent report from a Phase 1 trial [197].

Article highlights

• The treatment of B-cell malignancies was greatly advanced in the 1990s by the introduction of therapy with the first anti-CD20 monoclonal antibody (mAb) rituximab. Its activity against non-Hodgkin lymphoma (NHL) and chronic lymphocytic leukemia (CLL), and enhancement of activity of cytotoxic chemotherapy in lymphoma, resulted in significantly improved outcomes in patients with B-cell malignancies, and its adoption as standard of care in follicular lymphoma (FL), diffuse large B-cell lymphoma (DLBCL), and CLL.

• Despite these successes, many patients eventually relapse. Subsequent rounds of treatment are associated with decreasing quality and duration of response, and patients may become refractory to treatment. This has led researchers to look for novel anti-CD20 mAbs with structural and mechanistic features that distinguish them from rituximab.

• We describe epitope binding and mechanistic features of the novel anti-CD20 mAbs obinutuzumab, ofatumumab, and ublituximab. We then demonstrate how these features translate into preclinical activity, and from there into patient outcomes in clinical trials, when compared with rituximab.

• Obinutuzumab is a humanized and glycoengineered type II anti-CD20 mAb, which has demonstrated improved outcomes over rituximab-based therapy in FL and CLL. Ofatumumab is a fully human type I anti-CD20 mAb with an unmodified Fc domain; ublituximab is a chimeric type I anti-CD20 mAb with an Fc domain modified by glycoengineering. Whereas Type I mAbs typically have high complement-dependent cytotoxicity (CDC) and antibody-dependent cell-mediated cytotoxicity (ADCC), type II mAbs are associated with enhanced direct cell death effects, with reduced CDC but retained ADCC and antibody-dependent cell-mediated phagocytosis (ADCP).

• Data from the Phase 3 CLL11 and GALLIUM trials have demonstrated improved clinical outcomes for patients with CLL and indolent non-Hodgkin lymphoma, respectively, with obinutuzumab-based immunochemotherapy compared with rituximab-based immunochemotherapy, including progression-free survival (CLL11 and GALLIUM), overall survival (CLL11), time to next treatment (CLL11 and GALLIUM), and minimal residual disease (CLL11 and GALLIUM).

This box summarizes key points contained in the article.

Declaration of interest

C Klein, C Jamois and T Nielsen are employees and hold stocks in F. Hoffmann-La Roche Ltd. C Klein holds patents with F. Hoffmann-La Roche Ltd. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial relationships or otherwise to disclose.

Acknowledgments

Medical writing support was received from Scott Malkin and Lynda McEvoy, Gardiner-Caldwell Communications Ltd, funded by F. Hoffmann-La Roche Ltd.

Supplementary material

Supplemental data for this article can be accessed here.

Additional information

Funding

This work was funded by F. Hoffmann-La Roche Ltd.