Novel immunotherapy approaches to follicular lymphoma

Balancing novel treatment options for relapsed follicular lymphoma (FL) from emerging immunotherapies against the established benefits and toxicity of autologous and allogeneic hematopoietic cell transplantation (HCT) make it challenging for clinicians to choose the optimal treatment strategies for a given patient. A risk-stratified approach is critical to this decision-making process and should consider patient age, comorbidity, and likelihood of durable disease control. Patients experiencing progression with 24 months of chemoimmunotherapy are a vulnerable population in need of novel therapeutic approaches.^1,2  Few therapies have been studied for this group, and clinical trials are recommended. Several approaches (eg, lenalidomide, phosphatidylinositol 3-kinase [PI3K] inhibitors) may have benefits in FL as a result of effects on the immune system, as well as direct effects on the tumor. Ascertaining which of these factors are a greater determinant of response in a given patient remains an important unanswered question for individuals and patient groups.

Several drugs and regimens may exert their predominant activity in FL via effects on the immune system. FL and other common subtypes of B-cell non-Hodgkin lymphoma arise from the B-cell lineage; as a result, they typically express B-lymphocyte–differentiation antigens, including CD19, CD20, CD22, and CD79. These markers provide useful targets for antibodies and cellular therapies. The widespread utilization of anti-CD20 monoclonal antibody–based therapies (rituximab [R], and, to a lesser degree, obinutuzumab, ibritumomab tiuxetan, and ¹³¹I tositumomab) alone and in combination with chemotherapy regimens have improved outcomes for patients with FL in the first-line and the relapsed settings. In the 1980s, groundbreaking research at Stanford University led by Ron Levy’s group generated patient-specific anti-idiotype monoclonal antibodies as treatment for patients with FL, demonstrating that monoclonal antibodies produced a tumor response and could be administered safely.³  The identification of antigens restricted to B cells later led to therapeutic antibodies focused against these cell surface markers; Press et al demonstrated that treatment with a murine monoclonal anti-CD20 antibody could be administered safely but had modest clinical activity.⁴  The chimeric anti-CD20 monoclonal antibody R has gained widespread use in the first-line and relapsed settings for FL. Although specific details of the mechanism of action of and the resistance to R in various B-cell malignancies are poorly understood, numerous in vitro and in vivo studies have revealed that R leads to the destruction of lymphoma cells by 3 common mechanisms: complement-dependent cytotoxicity, recruitment of effector cells and antibody-dependent cell-mediated cytotoxicity, and direct induction of apoptotic signaling.^5-9  R can be used alone or in combination with chemotherapy for patients with relapsed FL, depending on the patient’s disease burden and goals of therapy.

Some patients fail to respond to first-line treatment with single-agent R or chemoimmunotherapy, and others experience relapse after initial response to therapy. Given the reliance of R on immune-effector mechanisms, the host immunologic environment has been postulated to play a role in failure of this therapy. Possible mechanisms of resistance to R include inhibition of each of the pathways involved in its mechanism of action and loss of CD20 expression. In fact, gene-expression profiling revealed that immune signatures reflecting the T cells and another reflecting macrophages in the tumor are associated with FL survival.^10,11  Efforts to overcome resistance to R have included the development of type II anti-CD20 antibodies, such as obinutuzumab, and approaches to augment the immune milieu in the tumor microenvironment. In the relapsed setting, an open-label phase 3 trial randomly assigned patients with R-refractory FL to treatment with obinutuzumab (1000 mg on days 1, 8, and 15, cycle 1; day 1, cycles 2-6) plus bendamustine (90 mg/m²) or bendamustine alone (120 mg/m² on days 1 and 2 for all cycles).¹²  The combination was followed by obinutuzumab maintenance. The study was stopped at the interim analysis; with a median follow-up of ∼22 months, the obinutuzumab plus bendamustine group demonstrated significantly improved progression-free survival (PFS; median not reached) compared with bendamustine alone (14.9 months; P = .0001). Grade 3 to 5 adverse events occurred in 68% patients in the obinutuzumab plus bendamustine group vs 62% of patients treated with bendamustine alone. A subsequent analysis with a median follow-up ∼32 months demonstrated a median PFS of 25.8 months for patients treated with the combination and 14.1 months for patients treated with bendamustine alone (hazard ratio, 0.57; 95% confidence interval, 0.44-0.73).¹³  This follow-up analysis also demonstrated that the combination of obinutuzumab plus bendamustine, followed by obinutuzumab maintenance, was associated with improved overall survival (OS; hazard ratio, 0.67; 95% confidence interval, 0.47-0.96). With extended follow-up, grade 3 to 5 adverse events were noted in 72.5% of patients treated with obinutuzumab plus bendamustine and in 65.5% of patients treated with bendamustine alone. When comparing obinutuzumab plus bendamustine with bendamustine alone, there were some differences in the relative frequencies of adverse events noted between arms for neutropenia (34.8% vs 27.1%), thrombocytopenia (10.8% vs 15.8%), anemia (7.4% vs 10.8%), and infusion-related reactions (9.3% vs 3.4%). Together, these data supported regulatory approval of obinutuzumab plus bendamustine in this setting.

In vitro experiments suggested that adding lenalidomide to R enhanced antitumor activity via improved NK cell synapse formation, increased NK cell function, and enhanced antibody-dependent cellular cytotoxicity.¹⁴  Lenalidomide is an immunomodulatory drug approved for use in multiple myeloma and mantle cell lymphoma that provides antitumor activity in lymphoid malignancies through pleiotropic mechanisms that are predominantly immune mediated. Preclinical studies in FL demonstrated that lenalidomide enhanced T-cell immune synapses, activated CD8 T cells, skewed the frequency of T-helper (T_H) subsets (favoring T_H1 over T_H2 subsets), and reduced regulatory T cells.¹⁵  However, the clinical impact of these effects remains unclear. The RELEVANCE trial and the BIONIC trial, which address roles for lenalidomide in the first-line setting, have presented initial results, but long-term follow-up is needed to understand the impact of first-line use of lenalidomide on subsequent lines of therapy.¹⁶ 

A phase 2 study involving R-refractory FL demonstrated lenalidomide activity as a single agent and in combination with R, with an overall response rate (ORR) of 63% (n = 43) for the combination and a median PFS of 22 months for all patients (n = 50) that was significantly longer than the PFS for the preceding line of therapy (9 months).¹⁷  In a randomized trial comparing lenalidomide alone (n = 45) vs lenalidomide plus R (n = 46) in patients with relapsed FL, Leonard et al showed that of 36% of patients receiving lenalidomide and 63% of patients receiving lenalidomide and R completed 12 cycles, resulting in ORRs of 53% (20% complete response [CR]) and 76% (39% CR), respectively (P = .029), and significant differences in median time to progression: 1.1 vs 2 years, respectively (P = .0023).¹⁸  Preliminary results of the MAGNIFY trial suggest that the combination of lenalidomide and R was tolerable and produced favorable outcomes in FL patients who were refractory to both R and an alkylating agent (double-refractory) or who had early relapse < 2 years after initial diagnosis with 1-year PFS of 66% for double refractory patients and 45% for early relapsed patients.¹⁹  Based on data from these studies, lenalidomide with R provides a useful alternative to chemoimmunotherapy for relapsed FL.

PI3K enzymes are components of the intracellular PI3K/Akt/mammalian target of rapamycin signaling pathway involved in cell surface receptor signals for numerous tissue-dependent cellular functions.^20-22  The tissue distribution of PI3K isoforms helps with prediction of the expected activity and toxicity observed with inhibitors of different PI3K isoforms. PI3K α and β isoforms are expressed broadly in multiple tissues, including PI3Kα playing a role in insulin signaling, whereas PI3Kδ is predominantly expressed on leukocytes and is important in regulatory T-lymphocyte function.^23,24  The PI3Kδ inhibitor idelalisib is an oral agent that was evaluated in 72 FL patients refractory to an alkylating agent and R, demonstrating an OR of 54% and leading to accelerated approval.²⁵  Copanlisib, a pan-PI3K inhibitor with preferential activity against α and δ isoforms, demonstrated similar activity.²⁶  The most common unique adverse events (≥grade 3) for this agent were hypertension (31%) and hyperglycemia (13%) related to inhibition of PI3Kα. These effects are often transient but should be monitored closely, particularly with initiation of therapy and before and after each infusion. Duvelisib, a dual-inhibitor of PI3Kδ and PI3Kγ, and umbralisib, a selective inhibitor of PI3Kδ and casein kinase-1ε, have completed early-phase trials; additional data are needed to determine the place of these agents in the management of FL.²² 

The role of HCT in the management of relapsed FL has been controversial in the modern era, particularly due to concerns surrounding early and late toxicity and the lack of novelty associated with this approach. Limited prospective data exist to address these controversies and are fraught with selection bias, heterogeneous populations, and lack of intent-to-treat analyses, but they suggest that high-dose therapy and autologous stem cell transplant (ASCT) (early in the disease course and in the setting of remission) are associated with favorable outcomes.^27-30  Long-term follow-up from a European series of autologous HCT in FL also demonstrated a median PFS of nearly 10 years and a median OS of 21 years.³¹  Despite the limitations described above associated with these findings, these impressive outcomes may be difficult to surpass with novel agents. Another analysis compared data from the Center for International Blood and Marrow Transplant Research registry with that from the National Lymphocare Study to examine the outcomes associated with HCT for patients with FL who experienced early relapse.³²  Patients from the Center for International Blood and Marrow Transplant Research who received autologous HCT within 1 year of treatment failure had improved 5-year OS compared with those not undergoing autologous HCT from the National LymphoCare Study (73% vs 60%, P = .02). A secondary analysis of 2 German randomized trials involving early relapsed FL patients (although most patients were R naive) also showed that autologous HCT was associated with significantly higher 5-year OS (77% vs 46%).³³  Collectively, these results demonstrate that the efficacy of autologous HCT continues to make this approach a meaningful option for younger fit FL patients who relapse early after initial chemoimmunotherapy. The risks and benefits of transplantation should be considered when making treatment decisions with these patients. For patients failing autologous HCT, those with extensive bone marrow involvement, or those who lack chemosensitive disease, allogeneic HCT or consideration of a clinical trial involving chimeric antigen receptor (CAR) therapy may be appropriate.

Allogeneic HCT has been associated with lower disease relapse rates than ASCT and is potentially curative therapy for patients with relapsed FL.³⁴  A retrospective analysis compared autologous vs allogeneic HCT outcomes among FL patients who experienced early treatment failure following frontline chemoimmunotherapy.³⁵  Inherent to a retrospective analysis, the baseline characteristics were not balanced between the 2 groups. Allogeneic HCT patients were younger, more heavily pretreated, and were more likely to have advanced-stage disease at diagnosis and chemorefractory disease pretransplant, whereas ASCT patients were more likely to have grade 3 histology. The risk of relapse was lower among patients who underwent allogeneic HCT, but OS was similar, owing to the lower nonrelapse mortality associated with ASCT. The favorable 5-year OS rates in this retrospective analysis of high-risk FL patients suggest that the application of HCT should be personalized. Transplant-related mortality and morbidity associated with graft-versus-host disease (GVHD) has hampered enthusiasm surrounding this approach. With the evolution of reduced-intensity conditioning and improvements in supportive care, is there a role for allogeneic HCT for high-risk patients otherwise facing poor outcomes? Twelve-year results from a single-institution trial of nonmyeloablative allogeneic transplantation with or without ibritumomab, tiuxetan demonstrated 11-year OS and PFS rates of 78% and 72%, respectively, providing compelling evidence that FL may be curable with immunotherapy strategies.³⁶  For the appropriate candidate (eg, a young otherwise healthy fit patient experiencing early relapse with an available matched sibling or unrelated donor), consideration of allogeneic HCT early in the disease course is reasonable.

Programmed cell death protein 1 (PD-1) and programmed cell death ligand-1 (PD-L1) are cell surface target that have promoted the development and evaluation of several immune checkpoint inhibitors. Responses to PD-1–targeted therapy have been reported in FL. In a phase 1 study of relapsed or refractory hematologic malignancies that included 10 patients with relapsed FL, the PD-1 inhibitor nivolumab produced a 40% ORR, with 1 complete remission.³⁷  A phase 2 clinical trial has been designed to evaluate nivolumab monotherapy in patients with “double refractory” FL who have failed a CD20 antibody and an alkylating agent (NCT02038946), and additional studies will evaluate nivolumab in other FL patient populations (NCT03245021, NCT03015896, NCT03121677). In a single-arm study involving 23 patients, including 18 with FL, pembrolizumab (200 mg) was administered every 3 weeks until progression, excessive toxicity, or completion of 2 years of therapy.³⁸  Drug-related adverse events that were ≥grade 3 occurred in 9 patients (39%); the most common were thrombocytopenia (13%), anemia (13%), neutropenia (8%), and dyspnea (8%). Overall, 11 patients (48%) experienced stable disease, and 2 FL patients had partial responses. A phase 1b trial of patients with relapsed or refractory FL examined the PD-L1 inhibitor atezolizumab and demonstrated 1 PR (33%).³⁹  Studies combining atezolizumab with a number of agents are underway or in development. Overall, checkpoint inhibitors have demonstrated modest single-agent activity in the relapsed setting. Detailed biomarker evaluation, such as tumor T-cell infiltration and expression of PD-1 and PD-L1 by immunohistochemistry or assessment of tumor mutation burden (as performed in some solid tumors),^40-42  is needed to identify and select patients who may benefit most from this therapy.

Checkpoint inhibitors are also being studied in a variety of combinations in FL. A phase 2 study examined pembrolizumab and R in 30 patients with FL grades 1 to 3a who relapsed after ≥1 prior therapy with R-sensitive disease.⁴³  Among evaluable patients, the ORR was 64%, the CR rate was 48%, and 15 patients remain in remission with a median follow-up of 11 months (range, 3.7-21).⁴⁴  Pembrolizumab also has been combined with G100, a Toll-like receptor-4 agonist that activates the innate and adaptive arms of the immune system with a similar adverse event profile.⁴⁵  A phase 1b/2 study (NCT02596971) examined the safety and efficacy of induction with atezolizumab, obinutuzumab, and bendamustine, followed by maintenance with atezolizumab and obinutuzumab. The most common treatment-related grade 3 or 4 adverse events were neutropenia in 6 patients and thrombocytopenia in 2 patients.⁴⁶  An interesting novel immunotherapy, Hu5F9-G4 (5F9), stimulates tumor cell phagocytosis and an antitumor T-cell response by targeting CD47, a protective “don’t eat me” signal on cancer. Anti-CD47 and R are being evaluated in a phase 1b/2 study (NCT02953509), and preliminary data demonstrated that this combination is well tolerated; common treatment-related adverse events include chills (41%), headache (36%), anemia (32%), and fever (27%), with the majority of events being grade 1 or 2.⁴⁷  At present, the use of checkpoint inhibitors in FL remains experimental; use outside of a clinical trial can only be based on limited data from trials with short follow-up.

Despite the advances with pharmacological interventions and stem cell transplantation, treatments with greater efficacy are urgently needed, particularly for patients with refractory and transformed FL. CAR T cells are T cells that have been genetically engineered to recognize specific tumor-associated antigens, including some of the targets noted above, such as CD19 and CD20. Current CAR T-cell technologies also incorporate costimulatory T-cell signaling domains, such as CD28 and/or 4-1BB. CAR T cells been tested in clinical trials involving patients with FL, and more studies are ongoing. Two anti-CD19 CAR T-cell approaches are approved for use in patients with diffuse large B-cell lymphoma.^48,49  Although early results are promising, toxicity remains a concern.

In 1 trial of anti-CD19 CAR T cells, patients with FL were eligible if they had measurable progression of disease <2 years after the second line of chemoimmunotherapy.⁴⁹  Leukapheresis products were stimulated with paramagnetic beads coated with antibodies to CD3 and CD28 and were transduced with the CD19-BB-ζ transgene. After leukapheresis, patients could receive bridging therapy at the discretion of their treating physician while awaiting the manufacture of cells. Lymphodepleting regimens were chosen by the investigator based on patient-specific characteristics, and anti-CD19 CAR T cells (1 × 10⁸ to 5 × 10⁸ CTL019 cells) were infused 1 to 4 days after the completion of lymphodepleting chemotherapy. Ten of 38 patients did not receive treatment as planned because of disease progression. Eleven of 14 patients with FL experienced a response therapy, and 3 patients who had a partial response at 3 months had a CR by 6 months; 1 patient continued to have a partial response at 6 months and had progressive disease at 1 year. Persistence of anti-CD19 CAR DNA by polymerase chain reaction (range, 6-24 months) was noted in 14 of 16 patients who had a CR. The median PFS was not reached, and 70% of patients were progression-free at the median follow-up of 28.6 months; 89% of patients with FL who had a response maintained their response.

Although the efficacy of PI3K inhibitors has been promising, serious immune-mediated and infectious toxicities have been identified with the approved agents. Immune-mediated adverse effects, including diarrhea, transaminitis, and pneumonitis, are thought to occur as a result of T-cell activation and appear to be most pronounced when PI3K inhibitors have been used in the first-line setting and when used in combination regimens.^50-53  These immune-mediated toxicities also suggest that the activity of these agents may be immune mediated, at least in part. In the pivotal phase 2 trial of single-agent idelalisib, 47% of patients experienced transaminitis (13% cases were ≥grade 3), 43% experienced diarrhea (13% were ≥grade 3), and respiratory symptoms including cough, dyspnea, and pneumonia, were observed in 29%, 18% (3% were ≥grade 3), and 11% (7% ≥ grade 3) of patients, respectively.²⁵  However, a phase 2 trial of idelalisib combined with R in untreated FL and small lymphocytic lymphoma (NCT02258529) was terminated because of toxicity, and excessive transaminitis, pneumonia, rash, hypotension, sepsis syndrome, and a death related to colitis were described when idelalisib was combined with lenalidomide and R.^51,52  In another study involving the SYK inhibitor entospletinib combined with idelalisib, high rates of pneumonitis (including 2 deaths) led to early study termination.⁵⁰  Intact immune systems in younger and untreated patients are thought to be risk factors for immune-mediated events.⁵³  Based on these findings, use of approved PI3K inhibitors in combination or in the first-line setting is not recommended. Pneumocystis jirovecii pneumonia prophylaxis and monitoring for cytomegalovirus have been recommended for patients receiving idelalisib, and management of toxicities has been summarized in recent articles.^23,54  With careful consideration and management of these toxicities, approved PI3K inhibitors provide a valuable option for therapy for patients with relapsed FL who have received an alkylating agent and R. Oral agents are more likely adopted in this setting by patients and physicians, and the use of each agent should be adapted to patient groups in which expected toxicities can be reduced.

Immune-mediated adverse event profiles have been reported in studies involving PD-1 and PD-L1 inhibitors in various cancers and included autoimmune-mediated hypothyroidism, rash, diarrhea, pyrexia, and neutropenia.^55-57  The most common adverse events reported are skin rash and gastrointestinal symptoms, affecting ∼19% to 26% of patients.⁵⁸  Adverse events (≥grade 3) occurred in 6% of the patients when treated with single-agent PD-1 inhibitors. In studies involving solid tumors, patients experienced fewer severe adverse effects, including autoimmune pneumonitis, colitis, hepatitis, myocarditis, and hypophysitis.^59,60  Most of the adverse events occur during the first 12 to 14 weeks of treatment, although they can occur at other times. Most of the events reported, with the exception of endocrinopathies, resolve within a median of 5 weeks. Among patients with hematological malignancies, an increased incidence of GVHD has been observed in patients who received PD-1 inhibitors before or after allogeneic stem cell transplantation.^61,62  A retrospective study of 31 patients who received PD-1 inhibitors after allogeneic stem cell transplantation revealed that 17 patients developed treatment-emergent GVHD, including 9 patients with grade 3/4 acute or severe chronic GVHD. Additional studies are necessary to define the optimal timing and use of PD-1 inhibitors in FL. Currently, these agents are best used within the context of a clinical trial. Evaluating potential biomarkers, including expression of PD-L1 in the tumor and/or PD-1 in tumor-infiltrating lymphocytes and chromosome 9p24.1 alterations that can identify the patients likely to benefit from PD-1 therapy, remains a critical area of investigation.

The adverse effects and management of toxicities associated with CAR T-cell infusions have been well described and summarized.⁶³  A common and distinct adverse effect following infusion of CAR T cells across trials is the onset of immune activation producing an associated cytokine release syndrome (CRS), resulting in elevated inflammatory cytokines involving high fever, myalgia, anorexia, malaise, capillary leak, and tachycardia/hypotension. Mild CRS can involve these constitutional symptoms and ≤grade 2 organ toxicities, and severe CRS can involve renal impairment, cardiac dysfunction, hepatic failure, and disseminated intravascular coagulation with ≥grade 3 organ toxicity, and can be life threatening, requiring rapid and aggressive clinical intervention.⁶³ 

Neurologic toxicities, including confusion, delirium, expressive aphasia, obtundation, and seizures, also have been described in patients with lymphoma who have received anti-CD19–specific CAR T cells.⁶³  The pathophysiology of the neurologic adverse effects associated with CAR T-cell administration remains unknown. Across all lymphoma subgroups in the trial discussed above, 11 of 28 patients had neurologic toxic effects related to CAR T-cell therapy.⁴⁹  Effects ranged from mild cognitive disturbance to global encephalopathy; 3 patients had encephalopathy ≥ grade 3. One FL patient with encephalopathy had progressive neurologic deterioration that resulted in death. For most patients, neurologic symptoms were self-limiting and resolved completely within 1 week.

FL remains a lymphoma subtype that is remarkably sensitive to immunotherapy-based treatment strategies. Anti-CD20 antibody therapy administered as a single agent and in combination as a first-line treatment and at relapse continues to be the most broadly used therapy for this disease. Autologous and allogeneic stem cell transplantations provide meaningful durable remissions for patients with FL. However, identifying the most suitable patients and the optimal timing for these approaches has become increasingly challenging with the advent of novel therapies. Lenalidomide and PI3K inhibitors are emerging as agents that can be applied in the relapsed setting. Other immunotherapy approaches, including checkpoint inhibitors and CAR T cells, appear promising but remain experimental. Utilization of all forms of immunotherapy requires careful consideration of the unique toxicities associated with these agents and the means to mitigate them by selection of appropriate patients, optimal timing, and the use of supportive care.

Christopher R. Flowers, Department of Hematology and Oncology, Winship Cancer Institute, Emory University, Atlanta, GA 30322; e-mail: crflowe@emory.edu.