Targeting CD20: teaching an old dog new tricks

You are asked to evaluate a 62-year-old female patient who presents with complaints of persistent adenopathy in the neck and groin. Imaging reveals extensive adenopathy, including multiple cervical, axillary, and inguinal nodes measuring 3 to 5 cm. In addition, there is matted retroperitoneal adenopathy in excess of 10 cm. Lactate dehydrogenase and b2 microglobulin are elevated. Marrow is modestly involved, but there are no cytopenias. Positron emission tomography scan shows uniform standardized uptake value of 4 to 6, and needle core biopsy of retroperitoneal nodes shows grade 1 to 2 follicular lymphoma.

Under pressure from your practice administrator to reduce infusion chair time, you initially recommend bendamustine in combination with subcutaneous rituximab; however, her insurance denies the prior authorization and states that only biosimilar rituximab products are covered by her policy. After a tense “peer to peer review,” you cannot obtain subcutaneous rituximab, but you recall recent favorable data with obinutuzumab and obtain authorization for the second generation anti-CD20 without issue. The patient receives 6 cycles of bendamustine with obinutuzumab followed by obinutuzumab maintenance with National Comprehensive Cancer Network–recommended pneumocystis and varicella prophylaxis and achieves a complete response (CR).

CD20 is a cell surface marker unique to B cells. It is a member of the tetraspanin family of proteins consisting of nearly 3 dozen different proteins in humans named for their 4 transmembrane domains resulting in 2 extracellular loops, a short intracellular loop, and both the N-terminal and C-terminal cytoplasmic domains. Originally discovered in 1980, it was among the first B cell–specific markers identified.¹ 

Tetraspanin molecules lack known ligands and do not possess catalytic activity; however, they help regulate cellular processes by organizing adhesion and signaling molecules within the plasma membrane into microdomains and coordinating cytosolic signaling molecules.²  Despite the robust clinical value of targeting CD20, many aspects of CD20 function remain unexplored. After antigen engagement of the B-cell receptor, CD20 disengages from the B-cell receptor complex and forms homotetramers, which regulate calcium mobilization necessary for certain antibody responses.³  Furthermore, engagement of CD20 results in a variety of intracellular signaling events that modulate critical signaling pathways that may influence B-cell survival.⁴ 

Direct targeting of CD20 by rituximab results in redistribution of CD20 to specialized areas of the plasma membrane known as lipid rafts and enables complement-dependent cytotoxicity, antibody-dependent cell-mediated cytotoxicity (ADCC), and antibody-dependent cell-mediated phagocytosis.⁵  Rituximab received the first regulatory approval in 1997 for the treatment of patients with relapsed indolent non-Hodgkin lymphoma (NHL). At the time of approval, it was the first monoclonal antibody used to treat a malignant condition.⁶  Since that time, rituximab has established itself as a vital component of a multitude of commonly used regimens in B-cell malignancies. Side effects of therapy are generally mild and include infusion reactions, neutropenia, hypogammaglobulinemia, and increased infections. The following review chronologically outlines efforts to build on the rituximab foundation by briefly discussing radioimmunotherapy, reviewing recent data sets addressing second generation anti-CD20 antibodies, and outlining the clinical development of biosimilar CD20 agents and subcutaneous rituximab; then, it looks at emerging combination immunotherapy approaches and novel CD20-targeting modalities.

In the years after the introduction of rituximab, radioimmunoconjugates were developed. The history of these agents provides a cautionary warning that effective therapies may still fail in the marketplace if they do not fit well within the existing infrastructure for cancer care delivery. ⁹⁰Y-ibritumomab tiuxetan (Zevalin) and ¹³¹I-tositumomab (Bexxar) gained Food and Drug Administration (FDA) approval for use based on a multitude of studies showing impressive activity and safety, yet tositumomab has been withdrawn from the market entirely, and ibritumomab use remains very limited.⁷  Ibritumomab remains available at a number of treatment centers, and long-term follow-up has demonstrated clinically meaningful single-agent activity.⁸  These therapies failed in part because of concerns regarding cost, access limitations related to the challenging logistics of administration, choice of phase 3 study design, and apprehension about long-term toxicity, including myelodysplastic syndrome/acute myeloid leukemia. These same clinical and economic risks may apply to a variety of novel immunotherapy treatment approaches currently in development.

As a greater understanding of rituximab immune effects accumulated, efforts were initiated to apply novel biologic insights to improved anti-CD20 therapies. Ofatumumab incorporated several novel features to improve on rituximab. The molecule was fully humanized rather than a murine chimeric antibody. Furthermore, it bound a different epitope on CD20 with stronger affinity, resulting in enhanced complement-dependent cytotoxicity and improved antibody-dependent cellular cytotoxicity.⁹  Initial approval was granted in chronic lymphocytic leukemia (CLL) after a single-arm phase 1/2 study in patients refractory to both fludarabine and alemtuzumab based on an overall response rate of ∼42% and a progression free survival (PFS) of 6.5 months.¹⁰  In a subsequent frontline CLL study, the addition of ofatumumab to chlorambucil improved response rates and PFS (13 vs 22 months) compared with chlorambucil alone, but owing to the absence of a rituximab-containing arm, it was unable to distinguish if ofatumumab yielded any benefit over rituximab.¹¹  Ofatumumab gained additional approvals in combination with fludarabine and cyclophosphamide in relapsed CLL¹²  and as maintenance for patients who achieved a response to salvage chemotherapy in CLL.¹³  In a pivotal study in which ofatumumab was directly compared with rituximab, patients with relapsed diffuse large B-cell lymphoma (DLBCL) received salvage therapy with either rituximab, dexamethasone, cytosine arabinoside, cis-platinum (R-DHAP) or ofatumumab, dexamethasone, cytosine arabinoside, cis-platinum (O-DHAP), and ofatumumab failed to show any advantage over standard rituximab.¹⁴  Presently, ofatumumab lacks any regulatory approval in NHL, and owing to evolving standards of care, it is infrequently used in CLL.

Obinutuzumab introduced several novel aspects to anti-CD20 therapy. Both rituximab and ofatumumab are type 1 antibodies that result in CD20 migration to specialized areas of the plasma membrane known as lipid rafts, whereas obinutuzumab does not and is, therefore, designated as a type 2 anti-CD20 antibody. Membrane clustering of type 1 antibodies bound to CD20 results in diminished CD20 surface expression via either internalization or trogocytosis,^15,16  which is not seen with type 2 obinutuzumab.¹⁶  Type 1 antibodies exhibit greater complement fixation, whereas type 2 antibodies have less complement activation and greater direct cell killing. The biologic mechanisms of direct cell killing are only partially characterized and may result from activation of intracellular signaling pathways,^4,17  leading to a direct nonapoptotic lysosome-mediated cell death¹⁸  regulated by calcium signaling and reactive oxygen species.^19,20  Another obinutuzumab modification of anti-CD20 therapy was “glycoengineering” or the removal of fucose from the crystallizable fragment (Fc) region of the antibody, which results in substantially enhanced antibody-dependent cellular cytotoxicity.^21-23  This modification is also incorporated into the type 1 anti-CD20 antibody ublituximab, which is in late-stage development for both CLL and NHL.²⁴ 

Preclinical demonstration of improved anti-CD20 effector functions requires clinical validation to impact patient care. In contrast to ofatumumab, which failed to demonstrate superiority against rituximab in any clinical setting, several pivotal studies directly compared obinutuzumab with rituximab and demonstrated improved outcomes in both CLL and follicular lymphoma (FL) but not DLBCL.

In the German CLL11 study, elderly patients with treatment-naive CLL were randomized to either obinutuzumab/chlorambucil vs rituximab/chlorambucil vs chlorambucil monotherapy. Both anti-CD20 arms demonstrated improved progression-free survival vs chlorambucil monotherapy, but there were also important differences in outcomes among the rituximab- and obinutuzumab-containing arms. Obinutuzumab resulted in deeper responses, with more patients achieving complete responses and minimal residual disease negativity compared with rituximab. Median PFS favored obinutuzumab over rituximab,²⁵  and recent updates to this study have demonstrated an overall survival benefit for the obinutuzumab arm.²⁶ 

A second large study of first-line treatment of follicular lymphoma (Gallium) compared rituximab with obinutuzumab-based chemoimmunotherapy (investigator choice of cyclophosphamide, vincristine, prednisone [CVP], cyclophosphamide, doxorubicin, vincristine, prednisone[CHOP], or bendamustine) followed by anti-CD20 maintenance. Although overall response rates were similar, 3-year PFS favored obinutuzumab at 80.0% vs 73.3%.²⁷  Rates of early progression (POD24), known to be associated with adverse outcome, demonstrated relative risk reduction of 46.0% (17.4% vs 10.1%) in the rituximab- and obinutuzumab-treated patients, respectively.²⁸  Although PFS data remain immature, modeling of outcomes demonstrates an anticipated PFS benefit in excess of 3 years for patients treated with obinutuzumab over rituximab.²⁹  In a separate study of patients with relapsed indolent lymphoma refractory to rituximab, patients received bendamustine with or without obinutuzumab (Gadolin). Median progression-free survival and overall survival favored the addition of obinutuzumab, demonstrating the ability of obinutuzumab to provide improved outcomes in the rituximab refractory setting.^30,31 

Despite improvements in CLL and indolent NHL, in a frontline study of patients with DLBCL (Goya) treated with either rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone (R-CHOP) or obinutuzumab, cyclophosphamide, doxorubicin, vincristine, prednisone (G-CHOP) (obinutuzumab), 3-year PFS did not differentiate between the 2 arms at 67% vs 70%, respectively.³²  Therefore, the advantages of obinutuzumab over rituximab seem to be disease and context dependent.

Prescription drug costs represent the single largest expenditure in health care. It is estimated that, in 2020, the total spending on cancer pharmaceuticals will reach $150 billion in the United States alone. These costs result in high rates of financial hardship among cancer patients and create barriers to access in many areas of the world. Patent protection for rituximab expired in 2018, creating the opportunity for the introduction of lower-cost biosimilar molecules.

Historically, small molecule generic drugs quickly became available after expiration of patent protection of an innovator molecule. Typical small molecule drugs have a molecular mass of <1000 Da, whereas monoclonal antibodies have a molecular mass of 150 000 Da, underscoring the difficulty of creating a biosimilar molecule. Furthermore, monoclonal antibodies are synthesized in living cells, where subtle changes in conditions may result in altered molecular structure.

To facilitate introduction of biosimilar molecules, novel abbreviated licensure pathways were introduced after the Biologics Price and Competition Act passed in 2009. According to regulatory guidance, biosimilarity indicates “that the biological product is highly similar to the reference product notwithstanding minor differences in clinically inactive components and that there are no clinically meaningful differences between the biological product and the reference product in terms of the safety, purity, and potency of the product.” Rather than following traditional pathways for approval for innovator molecules, FDA approval is based on the totality of evidence demonstrating similar structure, function, animal toxicity, human pharmacokinetics and pharmacodynamics, clinical immunogenicity, and clinical safety and effectiveness.³³  Furthermore, the clinical data used to support the application could be generalized across multiple indications of the reference product.

An initial clinical study in rheumatoid arthritis randomized 154 patients in 2:1 fashion to receive either an anti-CD20 biosimilar CT-P10 (rituximab-abbs) or reference rituximab. Efficacy, safety, immunogenicity, pharmacokinetics, and pharmacodynamic end points were all similar between the 2 molecules.³⁴  CT-P10 was subsequently studied in treatment-naive advanced-stage follicular lymphoma in which all patients received 8 cycles of CVP and were randomly assigned to either reference rituximab or CT-P10; 140 patients were randomly assigned 1:1 in a noninferiority study design. Overall response rates were 92.6% and 97.0% for rituximab and CT-P10, respectively. Treatment emergent adverse events were reported in 80% of patients receiving reference rituximab and 83% of patients receiving CT-P10.³⁵  To evaluate CT-P10 in a population of lymphoma patients without the confounding impact of cytotoxic chemotherapy, a monotherapy study was conducted in patients with low-tumor burden follicular lymphoma. Patients received weekly therapy for 4 weeks and were eligible to continue with maintenance therapy for up to 2 years if they maintained disease control. Among 258 patients, response rates at 7 months were 83% and 81% for CT-P10 and rituximab, respectively. Importantly, the 90% confidence interval for the overall response rate was entirely within the prespecified margin of 17%, consistent with guidelines for both the United States and Europe.³⁶  In December 2018, rituximab-abbs (Truxima) was the first rituximab biosimilar approved for use in the United States.

Similar development strategies have been used by additional molecules, including ABP-798³⁷  and PF-05280586.³⁸  More anti-CD20 biosimilar approvals are expected.

To extend patent protection for rituximab and eliminate prolonged infusion times, a subcutaneous delivery system for rituximab was developed. Patients first receive at least 1 IV infusion before subsequent subcutaneous injections. Subcutaneous injections contain both rituximab and hyaluronidase within the same injection. Hyaluronidase temporarily depolymerizes hyaluronan, increasing absorption of the monoclonal antibody, and the effects on subcutaneous tissues are reversible within 24 to 48 hours. Dosing differs by indication but not by body surface area, with a dose of 1400 mg in follicular lymphoma and DLBCL and 1600 mg in CLL.

In the SABRINA trial, previously untreated patients with follicular lymphoma received investigator choice of either CHOP or CVP chemotherapy, and they were randomly assigned to either IV or subcutaneous rituximab, with overall response rate as the study primary end point. Response rates were 84.9% for IV rituximab and 84.4% for subcutaneous rituximab, with similar safety except for higher rates of injection site reactions.³⁹  The MabEASE study was performed in 576 untreated patients with DLBCL. All patients received CHOP and were randomly assigned 2:1 to subcutaneous vs IV rituximab. Primary end point was CR/Cru, and secondary end points included safety and patient preference. At the end of induction, rates of complete response were 50.6% and 42.4% for subcutaneous rituximab and IV rituximab, respectively. Patients preference surveys favored subcutaneous rituximab, and time of therapy differed significantly, with median of 6 minutes for subcutaneous injection vs 2.6 to 3 hours for IV rituximab. The study was not powered to detect differences in progression-free survival or overall survival.⁴⁰  Finally, in the SAWYER trial, 176 patients with untreated CLL receiving fludarabine and cyclophosphamide were randomly assigned to IV vs subcutaneous rituximab, with the primary end point of the noninferiority study being serum drug trough concentration at cycle 5.⁴¹  Based on these studies, subcutaneous rituximab has labeled indications in FL, DLBCL when combined with CHOP, and CLL when combined with FC.

Six years after completion of maintenance obinutuzumab, the patient described above (now age 70 years old) experiences symptomatic progression. Repeat biopsy confirms that she still has low-grade disease. Aware of progress in the field of immunotherapy for the treatment of lymphoma, she asks, “is there anything I can do to boost my immune system to fight the cancer?”

The immune checkpoint inhibitors targeting PD-1, PD-L1, and anti-CTLA4 have dramatically altered treatment algorithms and prognostic outlook for many solid tumors and some hematologic malignancies. Except for primary mediastinal DLBCL, however, efficacy in NHL has not yet demonstrated enough activity to gain regulatory approval. Considering the immune effector functions enabled by anti-CD20 antibodies, it would be expected that finding the right immunologic partner for CD20-targeting agents could substantially impact function.

Lenalidomide is an immunomodulatory agent with pleiotropic effects on a variety of different immune effector cells, including the proliferation and activation of natural killer (NK) cells, which results in enhanced immune synapse formation and ADCC.⁴²  Several pivotal phase 3 trials have clarified the utility of combining lenalidomide with rituximab in indolent NHL. In the AUGMENT trial, 358 patients with relapsed follicular or marginal zone lymphoma were randomly assigned to rituximab either with or without 12 months of lenalidomide. The addition of lenalidomide resulted in substantially improved progression-free survival (14.1 vs 39.4 months, respectively) with modest increased toxicity.⁴³  FDA approval of the regimen in the relapsed setting occurred in May 2019, and the addition of lenalidomide to rituximab monotherapy should be considered standard of care. The MAGNIFY study is a second randomized study in patients with relapsed and refractory FL, marginal zone lymphoma, and mantle cell lymphoma; it compares different durations of lenalidomide therapy and will help clarify the optimal length of lenalidomide therapy.⁴⁴ 

In the RELEVANCE trial, 1030 patients with treatment-naive follicular lymphoma were randomly assigned to either standard chemoimmunotherapy (investigator choice of bendamustine, rituximab [BR], R-CHOP, or rituximab, cyclophosphamide, vincristine, prednisone [R-CVP]) or the combination of lenalidomide and rituximab. Although the trial was designed to demonstrate superiority of the experimental regimen, overall response rates and progression-free survival were highly similar between the 2 arms.⁴⁵  Because of the failure of the frontline study to reach the primary end point, approval for this indication is unlikely. For patients in whom chemoimmunotherapy is contraindicated or who favor a noncytotoxic approach, however, this combination may be considered.

Lenalidomide has also demonstrated activity in relapsed DLBCL, particularly the activated B-cell subset. Adding lenalidomide to R-CHOP in ABC-DLBCL yields results that compare favorably with historical controls. The ROBUST study directly compared R-CHOP with or without lenalidomide in frontline ABC-DLBCL.⁴⁶  Unfortunately, it has been reported that this study failed to demonstrate improved progression-free survival.⁴⁷ 

CD47 has recently emerged as a novel macrophage immune checkpoint that can be targeted with blocking antibodies, yielding potential synergistic activity when combined with anti-CD20 therapies. Signaling between CD47 and the ligand signal regulatory protein α (sirp-α) regulates macrophage-mediated phagocytosis, whereby CD47 serves as the “don’t eat me” signal expressed on the tumor cell. Hu5F9-G4 (5F9) is an immunoglobulin G4 (IgG4) antibody that blocks this inhibitor signal and potentiates the elimination of target cells. When anti-CD20 antibodies are bound to a target cell, the Fc portion of the antibody serves as a prophagocytotic signal for the macrophage.

In a population of patients with relapsed DLBCL/FL who were almost entirely rituximab refractory, the combination of 5F9 at the phase 2 dose with rituximab resulted in an overall response rate of 50%, with 36% complete response. Because the same signaling pathway helps regulate the removal of aged red blood cells, on-target anemia was as a common side effect with the agent.⁴⁸  Several other CD47 blocking antibodies are in clinical development, including 2 bispecific variants that target CD47/CD19 or CD47/CD20 to limit hemolysis or other on-target side effects outside of the tumor cells.⁴⁹ 

As the first approved therapy of the novel drug class, blinatumomab established the clinical therapeutic potential of immunoglobulin-based molecules targeting 2 different antigen binding sites: one directed at a target cell marker and the other at an immune effector cell.⁵⁰  Blinatumomab targets CD19/CD3 with a novel protein structure called a “bispecific T-cell engager” in which 2 different single-chain variable fragments (Fv) were joined by a glycine-serine linker.⁵¹ 

Recently, several new molecules targeting CD20/CD3 have entered human clinical trials with provocative results reported at American Society of Hematology 2018. REGN1979 is a full-length IgG4 bispecific antibody that allows for intermittent treatment (weekly for 12 doses and then biweekly for 12 doses) rather than the continuous infusion of blinatumomab. In a phase 1 study, patients with relapsed or refractory follicular lymphoma treated at doses ≥5 mg had overall response rate of 100%, with 80% complete response (8 of 10). Patients with relapsed/refractory DLBCL treated at doses between 18 and 40 mg had overall response of 60%, with 20% rate of complete response (n = 10). Although adverse events leading to treatment discontinuation were seen in only 2.9% of patients, infusion-related reactions and cytokine release syndrome were observed.⁵² 

Mosunetuzumab is a full-length IgG1 CD20/CD3 bispecific antibody with amino acid substitution within the Fc region to inactivate ADCC administered with “step up” dosing on days 1, 8, and 15 of cycle 1 followed by day 1 of each 21-day cycle thereafter. In a phase 1 dose escalation study, patients with relapsed/refractory follicular lymphoma had overall response rate of 69.2%, with 38.5% CR (10 of 26). Although follow-up remains brief, very few early relapses were seen among patients achieving a CR. Cytokine release was reported in ∼20% of patients, and neurologic side effects included dizziness and headache. Grade 3 neurotoxicity was very rare.⁵³ 

CD20-TCB (RG6026) is a novel “2:1” format T cell–engaging bispecific antibody in which there are 2 separate high-avidity CD20 binding sites and 1 CD3 binding region. This binding avidity offers the theoretical advantage of being able to be combined with other anti-CD20 antibodies without being blocked. In a relapsed DLBCL population, patients treated in the highest-dose cohort had an overall response rate of 55%, including 27% complete response.⁵⁴  Additional follow-up with longer observation and more patients treated will help determine the development path for this molecule.

It is premature to draw conclusions regarding relative efficacy and safety of these therapeutic approaches, but it seems that these agents have a clinical benefit profile to justify additional investigation. Furthermore, there has been substantial innovation in bispecific antibody structure and design. It is expected that additional CD20-targeting bispecific agents will move into clinical testing.

Chimeric T-cell therapy has emerged as a disruptive technological innovation in the management of B-cell hematologic malignancies, with FDA approval already established in both acute lymphoblastic leukemia and relapsed DLBCL. CD19 antigen loss may contribute to loss of efficacy for existing CD19-targeting chimeric antigen receptor T-cell (CAR-T) cells. Because of the established therapeutic efficacy of anti-CD20 agents, several different approaches have been introduced to exploit the target utilizing CAR-T technology. Human clinical trials are already underway with CAR-T cells directed at CD20 rather than CD19 (NCT 03277729 and NCT 03576807). Bispecific CD19/CD20 CAR-T products have also been developed to overcome this limitation. Other innovative approaches include CAR-T cells that express CD16, which is the NK-cell surface protein that binds to the Fc portion of therapeutic antibodies. When given in combination with anti-CD20 antibodies, these CAR-T cells target B cells for destruction.⁵⁵ 

For many years, rituximab enjoyed a dominant position in the field of anti-CD20 therapy. Despite the evidence for radioimmunotherapy, this approach has largely been abandoned, offering a variety of cautionary lessons. Eventually, second generation anti-CD20 therapies found their way into the treatment landscape either by exploiting niches of unmet medical need (ofatumumab) or by demonstrating superiority over rituximab in specific clinical situations (obinutuzumab). With the introduction of competition from biosimilar rituximab in the marketplace, subcutaneous rituximab was developed, with advantages in patient preference and ease of administration. Having been validated as a valuable target for therapeutic intervention with passive immunotherapy, anti-CD20 therapy is now at an important inflection point, being intensively studied with many of the novel immunotherapeutic approaches, including bispecific antibodies, checkpoint inhibitors, immunomodulatory drugs, and CAR-T. The treatment landscape for therapies directed at CD20 has evolved considerably in the last several years and is poised for significant innovation in the years to come.

Jeff P. Sharman, Willamette Valley Cancer Institute/US Oncology, 520 Country Club Dr, Eugene, OR 97401; e-mail: jeff.sharman@usoncology.com.