Obinutuzumab (GA101) for the Treatment of Chronic Lymphocytic Leukemia and Other B-Cell Non-Hodgkin's Lymphomas: A Glycoengineered Type II CD20 Antibody Open Access

The monoclonal type I antibody rituximab specifically targets the CD20 antigen expressed on the surface of mature and pre-B-cells. Rituximab in combination with chemotherapy prolonged survival times in many B-cell lymphomas, e.g. diffuse large B-cell lymphoma (DLBCL) [1, 2], chronic lymphocytic leukemia (CLL) [3], follicular lymphoma (FL), and mantle cell lymphoma (MCL) [4]. Most patients, however, eventually relapse and may become resistant to therapy [5]. Therefore, the goal of the development of new CD20 antibodies is to obtain improved properties such as a more efficient B-cell depletion to achieve increased clinical activity [6, 7, 8]. In this review, we elucidate the mode of action and present the clinical development of obinutuzumab (GA101, GAZYVA™/GAZYVARO™, F. Hoffmann-La Roche, Basel, Switzerland). Following approval in the US in 2013, obinutuzumab in combination with chlorambucil has also been licensed in Europe since July 2014 for the treatment of adult patients with previously untreated CLL and with comorbidities making them unsuitable for full-dose fludarabine-based therapy.

Obinutuzumab is a glycoengineered, humanized, monoclonal type II CD20 antibody of the IgG1 isotype [9] that was developed in an attempt to enhance both induction of immune effector cell-mediated killing of tumor cells and direct cell death compared with rituximab. To increase antibody-dependent cellular cytotoxicity (ADCC) and phagocytosis (ADCP), binding affinity to the FcγRIII receptor on immune effector cells was enhanced via a posttranslational glycoengineering process: the enzymatic equipment of the production cells enables the generation of an antibody glycovariant lacking fucosylation of the carbohydrate attached to the Fc region [9] (fig. 1A). Indeed, in vitro FcγRIII receptor affinity was shown to be considerably higher for obinutuzumab than for rituximab [9].

The enhanced binding of obinutuzumab to FcγRIII results in an increase in ADCC: the in vitro ADCC activity of obinutuzumab was 35-100 times higher than that of rituximab [9] or ofatumumab [10], and in contrast to rituximab, ADCC of obinutuzumab was neither blocked by physiological concentrations of unspecific IgG [9] nor by complement [11]. Notably, obinutuzumab was shown to abrogate inhibitory signals by inhibitory killer cell Ig-like receptor (KIR)/human leukocyte antigen (HLA) interactions [12]. In addition, combination studies with ibrutinib and idelalisib demonstrated that these kinase inhibitors had only minimal inhibitory impact on the immune effector function of obinutuzumab [13, 14].

Obinutuzumab recruits phagocytic cells such as monocytes, neutrophils, and dendritic cells via Fc-FcγR interactions. Therefore, the glycoengineered structure of obinutuzumab not only augments ADCC, but also increases the phagocytosis and cytotoxic activity effected by monocytes and macrophages [15] through FcγRIIIa as well as by neutrophils through FcγRIIIb [16].

Antibodies against CD20 can be grouped into 2 major classes referred to as type I and type II CD20 antibodies [8] (table 1). In in vitro assays, homotypic aggregation associated with direct cell death is a characteristic feature of type II antibodies. Rituximab and ofatumumab belong to the group of type I antibodies, while obinutuzumab is a type II antibody derived by humanization of the murine B-Ly1 antibody, which induces homotypic cell aggregation to a certain extent [9]. Obinutuzumab was identified in in vitro assays as an antibody variant that effectively induced direct cell death of B-cells [9]. Compared with the murine sequence, leucine 11 was substituted by valine in obinutuzumab, resulting in an elbow angle between the antigen-binding arms (Fab) that is nearly 30° wider than in type I antibodies [17]. In addition, obinutuzumab binds CD20 in a different orientation than type I antibodies, i.e. rotated by 90° around its middle axis and also tilted 70° towards the CD20 epitope [17]. Another typical feature of type II antibodies is that only half as many antibodies bind per B-cell compared with type I antibodies. Presumably, the different binding topology of type II antibodies causes the 2 Fab arms to bind within a single CD20 tetramer, while type I antibodies are assumed to bind different CD20 tetramers with each Fab arm [8]. Upon binding of rituximab, the CD20 antibody complex can be internalized and degraded, resulting in reduced effector cell recruitment and antibody half-life [18]. Type I CD20antibody-mediated CD20 internalization appears to be dependent on binding to the inhibitory FcγRIIb receptor expressed on B-cells in a cis fashion [19]. In contrast, type II CD20 antibodies show only minimal CD20 internalization [20]. Figure 1B depicts a model in which the special binding properties of type II antibodies prevent interaction with FcγRIIb, thus also preventing the accumulation of CD20 in lipid rafts and downregulation of CD20 surface expression, as compared to type I antibodies.

As a consequence of homotypic cell aggregation, obinutuzumab triggers direct non-apoptotic cell death [21, 22] which is associated with actin rearrangement, lysosomal cathepsin release, and the generation of reactive oxygen species via nicotinamide adenine dinucleotide phosphate (NADPH) oxidase. Because of the absence of characteristic hallmarks of apoptosis such as caspase dependency or BCL2 expression, this type of cell death could bypass mechanisms of apoptotic resistance [23]. Furthermore, direct cell death is induced independently of Fc-FcγR interaction. Therefore, obinutuzumab could also be an improved therapy option for patients with impaired Fc function, e.g. patients with low-affinity FcγRIIIa variants [24], as well as patients with effector cell saturation, exhaustion, or depletion.

Since obinutuzumab does not accumulate CD20 molecules in lipid rafts, no Fc clustering in lipid rafts occurs, which leads to decreased activation of complement-dependent cytotoxicity (CDC) via C1q. Accordingly, obinutuzumab demonstrated a > 10-1,000-fold lower CDC activity than rituximab and ofatumumab [10]. CDC is thus not considered a relevant clinical mechanism of obinutuzumab activity.

In in vitro and in vivo studies, obinutuzumab was shown to be superior to rituximab regarding both direct as well as effector cell-mediated cytotoxicity: obinutuzumab demonstrated a more efficient B-cell depletion in whole blood in vitro/ex vivo [9, 10], and superior antitumor activity in xenograft models, even in rituximab-refractory tumors [9, 10, 25, 26]. In particular, clinically relevant doses of obinutuzumab induced complete remission of SU-DHL4 DLBCL tumors, while an identical dose of rituximab merely inhibited tumor progression [9, 10]. In primates, obinutuzumab achieved B-cell depletion superior to rituximab in lymphoid tissue, including lymph nodes and spleen [9]. Combination studies showed that obinutuzumab demonstrates enhanced activity in combination with chemotherapies such as chlorambucil, fludarabine, and bendamustine, resulting in superior antitumor efficacy compared with the respective combination with rituximab [26]. Furthermore, antitumor efficacy can be enhanced by combining obinutuzumab with the Bcl-2-selective inhibitor GDC-199 [27] or the MDM2-selective inhibitor RG7388 [28].

In the phase I part of the GAUGUIN study (BO20999) [29, 30], 21 patients with relapsed/refractory indolent NHL received escalating doses of obinutuzumab monotherapy over 8 21-day cycles [29]. The phase I part of the GAUSS trial (BO21003) was conducted with 22 patients with relapsed NHL, including 5 patients with CLL [31]. Patients were given escalating doses of obinutuzumab once weekly over 4 weeks. These 2 phase I trials reported overall response rates (ORR) of 32-43%. No dose-limiting toxicity was observed, and infusion-related reactions (IRR) were the most common adverse events (AE) [29, 31]. Similarly, neither an obinutuzumab-associated dose-limiting toxicity nor any unexpected AE were seen in the phase Ib GAUDI study (BO21000), in which patients with relapsed/refractory FL were treated with obinutuzumab in combination with either CHOP (cyclophosphamide, doxorubicine, vincristine, prednisone) or FC (fludarabine, cyclophosphamide) [32]; also no unexpected AE were reported, including in previously untreated patients with FL, who were given obinutuzumab plus either CHOP or bendamustine [33].

The phase II stage of the GAUGUIN study included 40 patients with relapsed/refractory indolent NHL [34] and 40 patients with relapsed/refractory DLBCL or MCL [35]. Among the 2 dose groups that were tested, particularly the higher dose group (1,600/800 mg obinutuzumab) showed promising efficacy results and an acceptable safety profile. The randomized controlled phase II GAUSS trial compared obinutuzumab or rituximab monotherapies followed by maintenance therapies in 175 patients with relapsed indolent NHL, including 149 patients with FL [36]. Based on investigator assessment, higher ORR were achieved with obinutuzumab than with rituximab (43.2 vs. 35.6% in the overall population, 43.2 vs. 38.7% in patients with FL). No difference was found regarding the secondary end point progression-free survival (PFS); however, the trial was not powered to detect a difference in PFS. Obinutuzumab was well tolerated, although IRR occurred more frequently in the obinutuzumab arm.

Dose-finding studies such as GAUGUIN demonstrated that upon administration of a 1,600/800 mg dose of obinutuzumab, plasma concentrations increased more rapidly than with lower doses, leading to an early steady state indicative of CD20 target saturation [35]. Based on both clinical data and pharmacokinetic simulations, a fixed dose of 1,000 mg obinutuzumab on days 1, 8, and 15 of the first 21-day cycle and on day 1 of subsequent cycles was selected as the dose to achieve adequate exposure levels in a similarly rapid manner with less interindividual variability in phase III trials [37].

In the phase II GATHER study (GAO4915g), 80 untreated patients with advanced DLBCL were treated with 6 cycles of a CHOP regimen, plus 8 cycles of obinutuzumab dosed as described above [38].

Details on treatment regimens and results of the clinical studies described in this section are listed in table 2.

For the 13 patients with relapsed/refractory CLL in the phase I part of the GAUGUIN study, an ORR of 62% was reported at the end of treatment [30]. For the 20 patients in the phase II part of GAUGUIN, a lower ORR of 30% was reached, which may be due to a higher baseline tumor burden resulting in lower exposure to treatment [30]. None of the 5 patients with CLL in the phase I part of the GAUSS trial met the criteria for response [31].

In another phase II study called GAGE (GAO4768g), which included 80 patients with previously untreated CLL and compared 2 dosing regimens of obinutuzumab monotherapy, a higher ORR was demonstrated in the 2,000 mg group (67 vs. 49%) compared with the 1,000 mg group [39].

In the GALTON trial (GAO4779g; phase Ib), 41 patients with previously untreated CLL received obinutuzumab in combination with either FC or bendamustine [40]. The data showed that these combinations have clinical activity, with grade 3/4 neutropenia and infections in 29 and 19% of the patients in the obinutuzumab plus FC arm and in 50 and 5% of the patients in the obinutuzumab plus bendamustine arm, respectively.

Table 3 provides an overview of the published trials in CLL, as discussed in this section.

The randomized controlled phase III study CLL11 (BO21004) [41] enrolled 781 patients with previously untreated CLL (median age 73 years) and relevant comorbidity indicated by a Cumulative Illness Rating Scale (CIRS) score > 6 and/or creatinine clearance of 30-69 ml/min. The study compared obinutuzumab plus chlorambucil with rituximab plus chlorambucil and chlorambucil alone. Patients received 6 28-day cycles of either chlorambucil (0.5 mg/kg on days 1 and 15), obinutuzumab (1,000 mg on days 1, 8, and 15 of cycle 1, and on day 1 of cycles 2-6) plus chlorambucil or rituximab (375 mg/m² on day 1 of cycle 1, 500 mg/m² on day 1 of cycles 2-6) plus chlorambucil. In the arms comparing obinutuzumab plus chlorambucil vs. rituximab plus chlorambucil, the administered median cumulative doses of antibodies were 8,000 mg of obinutuzumab and 5,106 mg of rituximab, respectively.

Obinutuzumab plus chlorambucil was superior to rituximab plus chlorambucil in terms of PFS (median PFS 26.7 vs. 15.2 months; hazard ratio (HR) 0.39; 95% confidence interval (CI) 0.31-0.49; p < 0.001), complete response (20.7 vs. 7.0%), and rate of negative testing for minimal residual disease, both in peripheral blood (37.7 vs. 3.3%) and bone marrow (19.5 vs. 2.6%) (fig. 2). After a median observation time of 19 months, no significant difference in overall survival was observed (HR for death 0.66; 95% CI 0.41-1.06; p = 0.08). Obinutuzumab plus chlorambucil did prolong overall survival, however, when compared to chlorambucil alone (HR for death 0.41; 95% CI 0.23-0.74; p = 0.002). A recent update with approximately 1 year of additional observation time showed an even better improvement in PFS with obinutuzumab plus chlorambucil (median PFS 29.2 vs. 15.4 months with rituximab plus chlorambucil, HR for death 0.40, 95% CI 0.33-0.50, p < 0.001) and confirmed the overall survival benefit of obinutuzumab plus chlorambucil over chlorambucil monotherapy (HR for death 0.47, 95% CI 0.29-0.76, p = 0.0014). No statistically significant overall survival benefit for obinutuzumab plus chlorambucil over rituximab plus chlorambucil could be demonstrated so far (HR for death 0.70, 95% CI 0.47-1.02, p = 0.0632), although these overall survival data are still immature [42].AE occurred more frequently with obinutuzumab plus chlorambucil than with chlorambucil alone or rituximab plus chlorambucil, primarily including IRR (the most common AE in 66% of the patients), neutropenia, and thrombocytopenia, but the risk of infection was not increased. IRR frequency was high during the first infusion of obinutuzumab (20% grade 3/4 IRR, no deaths), but considerably lower during subsequent infusions (0% grade 3 or 4 IRR). Rapid and profound B-cell depletion and faster recruitment and activation of immune effector cells by obinutuzumab with subsequent cytokine release may possibly explain the more frequent and severe IRR during the first infusion with obinutuzumab [41]. Of note, preliminary data indicate a correlation between CD20 surface expression on CLL cells as well as FcγRIII polymorphisms and the risk of developing any grade of IRR with the first infusion of rituximab or obinutuzumab [43].

Based on these data, obinutuzumab has recently been granted approval in Europe in combination with chlorambucil chemotherapy for the treatment of adult patients with previously untreated CLL and comorbidities making them unsuitable for full-dose fludarabine-based therapy.

Ongoing phase III studies investigate obinutuzumab as compared with rituximab in B-cell NHL entities other than CLL and in combination with other chemotherapy regimens. The phase III trial GOYA (BO21005, NCT01287741) compares the combination of CHOP plus obinutuzumab with CHOP plus rituximab in previously untreated DLBCL, and GALLIUM (BO21223, NCT01332968; phase III) tests obinutuzumab vs. rituximab in combination with different chemotherapies, i.e. CHOP, CVP (cyclophosphamide, vincristine, prednisone), and bendamustine, followed by maintenance with obinutuzumab/rituximab, in patients with untreated indolent NHL (FL or marginal zone lymphoma). Rituximab-refractory patients with indolent NHL were treated with obinutuzumab plus bendamustine or bendamustine alone in the phase III GADOLIN study (GAO4753g, NCT01059630). Due to superiority of the obinutuzumab/bendamustine combination at the pre-planned interim analysis, the study was terminated prior to the protocol-specified final analysis [44]. The phase IIIb safety trial GREEN (MO28543, NCT01905943) investigates obinutuzumab alone or in combination with FC, bendamustine, or chlorambucil both in previously untreated and relapsed/refractory CLL; preliminary data were in line with the known safety profile of obinutuzumab [45].

Furthermore, obinutuzumab is an especially interesting option for the assessment of chemotherapy-free combination therapies due to its optimized effector functions. Preclinical data suggest that effective combinations with targeted anticancer agents are possible and that synergistic effects can be expected, e.g. with the Bcl-2 inhibitor venetoclax (ABT-199/GDC-0199) [27], MDM2 inhibitors RG7112 and RG7388 [28], as well as the proteasome inhibitor bortezomib [46]. Preclinical data demonstrated that the kinase inhibitor ibrutinib interferes with the ADCC-related effector function of rituximab through inhibition of interleukin-2 inducible tyrosine kinase [47]. However, minimal interference of idelalisib and ibrutinib was observed with the immune effector function of obinutuzumab compared with rituximab most likely due to the direct cell death induction and stronger FcγR signaling, supporting the clinical testing of these combinations [13, 14, 48]. First safety data from a phase 1b trial investigating obinutuzumab plus venetoclax in CLL was encouraging [49]. Corresponding clinical studies exploring chemotherapy-free or -reduced treatments in CLL are currently being initiated including several trials of the German CLL Study Group (GCLLSG). A series of phase II GCLLSG studies (all-comers, i.e. both untreated and treated, and both fit and comorbid patients) will sequentially evaluate bendamustine, obinutuzumab, and either ibrutinib (BIG), venetoclax (BAG), or idelalisib (BCG). Furthermore, following the recently published phase Ib data of the GALEN study, which showed that oral lenalidomide plus obinutuzumab was well tolerated and effective in patients with relapsed or refractory FL [50], the ongoing phase II part of the study will assess obinutuzumab in combination with lenalidomide in patients with relapsed/refractory follicular and aggressive B-cell lymphomas (DLBCL and MCL).

The authors would like to thank Physicians World Europe for providing medical writing assistance for the preparation of this manuscript, supported by Roche Pharma AG, Grenzach-Wyhlen, Germany.

Valentin Goede: Roche: speaker honoraria, advisory board, research funding, travel grants; Mundipharma: speaker honoraria, advisory board, research funding, GSK: speaker honoraria. Christian Klein: Roche: employment, equity. Stephan Stilgenbauer: honoraria and research funding from AbbVie, Celgene, Genentech, Gilead, GSK, Janssen, Mundipharma, Roche.