Editorial theme | Neurotoxicity and anti-BCMA agents

Chimeric antigen receptor (CAR) T-cell therapy (CAR-T) and bispecific antibodies represent two novel immunotherapeutic treatment modalities in the relapsed/refractory (R/R) setting of multiple myeloma (MM). While both of these modalities are changing the treatment landscape of R/R MM, knowledge and management of the specific adverse events (AEs) associated with CAR-T and bispecific antibodies is imperative.^1,2

This review provides a detailed look at the neurotoxicity associated with B-cell maturation antigen (BCMA)-directed agents, and represents Part II of an educational series on the management of AEs associated with novel agents; read Part I about cardiovascular toxicity with proteasome inhibitors, here.

Neurotoxicity

Neurotoxicity—specifically immune effector cell-associated neurotoxicity syndrome (ICANS)—is a known AE in association with CAR T-cell therapy and, to a lesser extent, bispecific antibody therapy. In some cases, the neurotoxicity observed with BCMA-directed therapy has atypical features and a prolonged timeframe compared with ICANS, though reports on this are few.³

The clinical presentation and course of ICANS varies widely, though signs and symptoms typically reported in patients with MM treated with BMCA-targeted CAR-T include confusion, delirium, transient aphasia, encephalopathy, bradyphrenia, agitation, hallucination, obtundation, seizure, mild cerebral edema, polyneuropathy, tremor, dizziness, and vertigo. Most reported cases of ICANS in anti-BCMA CAR-T have been mild and transient (unlike anti-CD-19 CAR-T, where severe neurotoxicity and death due to cerebral edema have been reported).⁴

BCMA-targeted CAR-T: ICANS vs other neurotoxicity

The neurotoxicity observed in patients in the pivotal idecabtagene vicleucel (ide-cel) trial for R/R MM (KarMMa; NCT03361748) was in line with the expected manifestations and timeline associated with ICANS: it was seen in 23 patients (18%), and had a median time to onset of 2 days—in proximity to events associated with cytokine release syndrome (CRS)—and a median duration of 3 days. The most commonly observed symptoms were confusional state and encephalopathy. In addition, tremor was reported in three patients, and hemiparesis was reported in four patients.³

In the ciltacabtagene autoleucel (cilta-cel) trial (CARTITUDE-1; NCT03548207), however, the median time to onset was 8 days, and the median duration of symptoms was 4 days. Of the 20 patients (21%) who experienced neurotoxicity, 16 were determined to have ICANS, and 12 were thought to have other types of neurotoxicity that occurred after the resolution of ICANS (eight patients experienced both types of neurotoxicity). Symptoms associated with other type of neurotoxicity included movement and/or neurocognitive changes (n = 5) and nerve palsy/peripheral motor neuropathy (n = 7). For 6 of the 12 patients experiencing this non-ICANS neurotoxicity, the manifestations did not resolve: one has ongoing neurotoxicity, one died as a result of neurotoxicity, and four had neurotoxicity at time of death due to other causes.³

BCMA-targeted bispecific antibodies

Blinatumomab is currently the only bispecific antibody approved by the U.S. Food and Drug Administration (FDA) and European Medicines Agency (EMA), although several other bispecific antibodies are being investigated in early-stage clinical trials. Of note, several cases of peripheral neuropathy have been reported in bispecifics trials, and in May 2021, the FDA placed on temporary hold the bispecific antibody elranatamab  due to peripheral neuropathy. Patient recruitment for the pivotal elranatamab trial (MagnetisMM-3; NCT04649359) was safely reinitiated in June 2021.^3,4

What is the mechanism behind neurotoxicity?

BCMA is a cell-surface receptor that is preferentially expressed by mature B lymphocytes and has been identified on the surface of almost all MM cell lines; its almost exclusive expression by hematopoietic tissue makes it an ideal therapeutic target for MM. While BCMA-targeting agents are thought to have limited off-target activity, reports of ICANS would indicate the presence of unwanted effects. Atypical neurotoxicity (that is, non–ICANS-related) has also been observed in association with these therapies, and studies of mouse and human brain tissue have suggested, but not confirmed, the expression of BCMA by neural cells, indicating that BCMA has a role in neural development.³

CAR T-cell therapy

Due to an overall lack of data, the precise pathophysiologic mechanism of ICANS is not clear, and most of the information forming these hypotheses comes from CD19-directed CAR T-cell therapy. The current working consensus is that ICANS is the result of off-target activity due to one or more of the following mechanisms:³

• Disruption of the blood brain barrier (BBB)
• Activation of by-standing monocytes
• Hyperactivation of myeloid cells
• Infiltration of the central nervous system (CNS) by T cells

In addition, severe CRS is a predisposing factor for ICANS, suggesting an association between these two toxicities; however, the differences in onset and duration indicate that these processes are complementary but ultimately pathophysiologically independent.³ Increased permeability of the BBB allows for an influx of cytokines from the bloodstream into the cerebrospinal fluid, both of which are seen in patients with severe neurotoxicity.^5,6

The presence of BCMA in neural tissue, in addition to the atypical toxicity that has been observed in BCMA-directed CAR T-cell therapy, suggests that there may also be an on-target, off-tumor mechanism(s) for neurotoxicity.³

Bispecific antibodies

While it is thought that blinatumomab does not cross the BBB, it is not clear whether or not that is the case. Until further data is collected, the mechanism by which bispecific antibodies cause neurotoxicity is uncertain.³

Incidence of neurotoxicity

CAR-T

Clinical trials of anti-BCMA CAR T-cell therapy in MM to date have been small, non-randomized, and early-phase, leading to heterogeneity in the reported incidence of AEs, including neurotoxicity. Definitions of neurotoxicity (and thus what is and is not reported) and grading also vary between studies, contributing to this heterogeneity.^5,7

Gils Roex and colleagues⁷ published a metaanalysis of 23 BCMA CAR T-cell products in a total of 640 patients and found a pooled neurotoxicity rate of 10.5%, although there was significant variation between the studies: the 57-person trial of cilta-cel reported a neurotoxicity rate of 2%, while a 10-person trial of FCARH143+GSI reported a rate of 60%. In the metaanalysis, the investigators found higher rates of neurotoxicity in studies where the median age was ≥60 years and in studies where the median number of prior lines of therapy was ≥5; there were no observed associations between neurotoxicity rates and lymphodepletion regimen (even those containing fludarabine) or the monoclonal antibody origin of the antigen-recognition.⁷

Ide-cel, which has been approved by both the FDA and EMA for the treatment of R/R MM, reported neurotoxic effects in 18% of patients in the phase II KarMMA study, as previously noted; Grade 3 events were reported in 3% of patients, and there were no Grade 4 or 5 neurotoxicities. Neurotoxicity appeared to be dose-dependent, and the use of corticosteroids for the treatment of neurotoxic effects increased as the CAR T-cell dose increased.^8-10

Cilta-cel has not yet been approved but is likely to receive approval in the near future. The phase Ib/II CARTITUDE-1 trial reported an all-grade neurotoxicity rate of 20.6% and a Grade ≥3 neurotoxicity rate of 10.3%. As noted previously, the CARTITUDE-1 investigators differentiated between ICANS and what they believed to be a distinct (and sometimes overlapping) neurotoxic entity with different symptoms, duration, and time of onset. Grade ≥3 ICANS occurred in 2.1% of patients in CARTITUDE-1, and all patients recovered.¹¹

With only one (recently) approved anti-BCMA CAR-T agent, experience with this type of therapy in MM is limited, particularly real-world experience; the management of these patients, as well as the effect of neurotoxicity on patient outcomes, is largely based on what has been reported in clinical trials.⁵ The Society for Immunotherapy of Cancer (SITC) recommends that the management of patients with neurotoxicity should be based on experience from registration trials.¹²

Bispecifics

All-grade neurotoxicity rates reported in bispecific antibody trials in MM (most but not all of which are BCMA-targeting) ranged from 5% to 28%, and Grade ≥3 rates ranged from 0% to 2%. In most cases, reported symptoms—which included headache, confusion, aphasia, cognitive disorder, and encephalopathy—were associated with CRS and resolved with treatment of CRS. Of note, AMG420, a BCMAxCD3 bispecific T-cell engager, had a high rate of neuropathy, and of the 42 patients enrolled in the trial, two patients (5%) had Grade 3 polyneuropathy (a dose-limiting toxicity); neuropathy was not reported as a major side effect in the other trials. Elranatamab, as mentioned previously, was placed on hold for 2 months due to three cases of peripheral neuropathy.²

Clinical management of anti–BCMA-related neurotoxicity

Patient selection: CAR-T

Patient selection for CAR T-cell therapy, in general, is largely dictated by access, both geographically and financially, as well as by treatment history/refractoriness. Access, however, does not guarantee that a patient will receive CAR-T, as this is not an ‘off-the-shelf’ therapy. Patients’ own T cells must be collected via leukapheresis, after which lymphodepletion is required to create a suitable environment into which the CAR T-cells are infused; many patients also require a bridging therapy during this time to prevent disease progression. It is a complex and intensive process, and the process does not end at infusion, as CAR-T is associated with unique and potentially life-threatening toxicities.^5,12 In the SITC consensus statement on immunotherapy for the treatment of MM, it is stated that patient suitability for CAR-T is also often based on the potential for toxicity, and they recommend the following as part of a patient’s baseline evaluation¹²:

• Bone marrow function
• Cardiopulmonary function
• Hepatic function
• Renal function
• Performance status
• Organ status (with respect to the patient’s ability to tolerate CRS and prolonged cytopenias)
• Comprehensive neurological assessment

Patient selection: Bispecifics

As an ‘off-the-shelf’ therapy, bispecific antibodies are not associated with some of the restrictions seen with CAR T-cell therapy; these therapies are less expensive and more accessible, though they do require repeat dosing. Currently, there are no consensus recommendations for optimal patient selection for bispecific antibody therapy in MM; however, clinical trials of bispecifics have been in the relapsed/refractory setting.¹²

AE grading

Several ICANS grading systems have been published, including the Common Terminology Criteria for Adverse Events (CTCAE) criteria, the CAR T-cell therapy 10-point neurological assessment (CARTOX-10), and the similar immune effector cell-associated encephalopathy (ICE) score. The American Society for Transplantation and Cellular Therapy (ASTCT) grading system is the most commonly used tool for the assessment of ICANS in clinical trials (Table 1).^5,13 Using the ASTCT system, the ICANS grade is determined by the most severe event that is not attributable to any other cause.¹³

Table 1. ASTCT grading system for ICANS in adults*

AE diagnosis and management

Close monitoring and supportive care are the typical management of ICANS, and patients with ICANS should ideally have their neurological status monitored at least four times per day following CAR T-cell infusion. The ICE score is a useful tool for monitoring neurological status (Table 2).

Table 2. ICE encephalopathy tool for assessment of ICANS*

It is important to rule out other causes of neurological symptoms in patients who have received CAR T-cell therapy. Many of these patients are at risk of infection (progressive multifocal leukoencephalopathy due to the JC virus) and bleeding events (cerebral hemorrhage), and imaging and spinal fluid analysis can be used to narrow the list of differential diagnoses.⁸

Seizure management is of particular importance in ICANS. Patients with suspected ICANS should be monitored via electroencephalogram (EEG), and patients who are at increased risk for ICANS (those with high tumor burden and/or extramedullary disease) should be treated with seizure prophylaxis.⁸

Corticosteroids are often used for the treatment of ICANS Grade ≥2 due to their immunosuppressive effects, and they have been associated with a rapid resolution of ICANS in trials of anti-BCMA CAR T-cell therapy in MM, though currently there is no consensus regarding the optimal dosage or duration of treatment.⁷ ASTCT recommendations for the management of ICANS are shown in Table 3.

Table 3. ASTCT management strategy for ICANS in adults*

Conclusion

CAR T-cell therapy and bispecific antibodies are highly effective treatment options for patients with R/R MM, a treatment setting that has long been a challenge, with limited responses seen to available therapies. BCMA has been identified as an ideal target for these treatments, as it is expressed almost exclusively on cells of hematopoietic origin and is highly expressed in MM. Despite this, neurotoxicities have been reported in association with both CAR-T and bispecific antibody therapy, and in some cases, these neurotoxicities have been severe. The management of neurotoxicity associated with anti-BCMA therapy is challenging; both clinical trial and real-world experience and data are limited, and clinicians must learn how to manage these patients in real-time.