Advances in the mechanisms and management of CAR T-cell toxicities

Abstract

Chimeric antigen receptor (CAR) T-cell therapy has revolutionized the treatment of hematologic malignancies and is being investigated in solid tumor malignancies. Toxicities such as cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) are now well-recognized, and improved supportive care and management with immunosuppressive agents has made CAR T-cell therapy safer and more feasible than it was at the time of the first commercial approval of tisagenlecleucel. In recent years, there has been greater recognition of previously less well-defined toxicities, including movement disorders, immune effector cell-associated hematotoxicity (ICAHT), immune effector cell-associated hemophagocytic lymphohistiocytosis-like syndrome (IEC-HS), and the substantial risk of infection that patients with persistent B-cell aplasia and hypogammaglobulinemia face. A more diverse selection of immunosuppressive and supportive care pharmacotherapies is now being utilized for toxicity management, but there remains no universal algorithm for the timing of their use. As CAR T-cell products are developed that target new antigens, more toxicities involving damage to normal tissue expressing the target antigen are a potential hurdle. Continued prospective evaluation of toxicity management strategies and the design of less toxic CAR T-cell products are both critical for continued success of the CAR T-cell field.

I. Introduction

There are currently 6 commercially available CAR T-cell therapies. Tisagenlecleucel (tisa-cel), axicabtagene ciloleucel (axi-cel), lisocabtagene maraleucel (liso-cel), and brexucabtagene autoleucel (brexu-cel) target CD19 and treat B-cell malignancies. Idecabtagene vicleucel (ide-cel) and ciltacabtagene autoleucel (cilta-cel) target B-cell maturation antigen (BCMA) and treat multiple myeloma. Anti-CD22 CAR T cells¹ and anti-GPRC5D (G protein–coupled receptor, class C, group 5, member D) CAR T cells^2,3 are being developed as rescue therapies for patients with relapsed B-cell malignancies and multiple myeloma, respectively, after prior treatment with available CAR T-cell therapies. CAR T-cell therapies are additionally being explored in solid tumor malignancies⁴.

Since the early development of CAR T-cell therapy, it was recognized that the release of cytokines and other immunologic proteins by CAR T cells and downstream myeloid-derived immune cells caused the inflammatory toxicities of cytokine release syndrome (CRS) and the now-named immune effector cell-associated neurotoxicity syndrome (ICANS)^5-8. Deaths due to severe CRS and cerebral edema were observed to be a barrier to safe and widespread use of this therapy^9,10. However, the use of the IL-6 receptor antagonist tocilizumab and corticosteroids has made CAR T-cell therapy safer and more feasible over time. Herein we will review the current clinical experience with the common inflammatory toxicities CRS and ICANS, as well as the evolving understanding of secondary hemophagocytic lymphohistiocytosis (HLH), prolonged cytopenias, infections, including covid-19, on-target toxicities, and second malignancies.

II. CRS

CRS is a well-described inflammatory syndrome resulting from secretion of cytokines and other immunologic proteins from CAR T cells and other immune cells, such as monocytes and macrophages^11,12. Classic signs and symptoms include fevers, chills, sinus tachycardia, hypotension, hypoxia, and dyspnea. CRS-related deaths, which may be due to respiratory failure, cardiac arrest, multi-organ failure, or HLH are well-reported^9,13-18. The timeline of CRS occurrence, along with other common toxicities, is summarized in Figure 1. In some cases, specific toxicities may be difficult to differentiate or may overlap: mild confusion during high-grade fevers with CRS may or may not represent early ICANS; secondary hemophagocytic lymphohistiocytosis (HLH) may be challenging to distinguish from severe CRS, and cytopenias may occur in isolation or co-occur with HLH.

Figure 1: Timeline of selected CAR T-cell toxicities.

Approximate timeline of selected CAR T-cell toxicities, including occurrence, duration and resolution. BCMA: B-cell maturation antigen. CRS: cytokine release syndrome. ICANS: immune effector cell-associated neurotoxicity syndrome. ICAHT: immune effector cell-associated hematotoxicity. IEC-HS: immune effector cell-associated hemophagocytic lymphohistiocytosis-like syndrome.

Risk factors for development of CRS include nodal and/or bone marrow burden of malignancy^13,18-21, poor performance status¹⁸, higher baseline patient inflammatory markers^13,17,22, lower baseline patient platelets counts²⁰, receipt of/requirement for bridging therapy¹³, more intensive lymphodepletion chemotherapy²⁰, higher cell doses^15,20,23, and higher blood levels of CAR T cells^{13,14,19,24,25}. Risk of severe CRS varies with CAR T-cell product and malignancy being treated (Table 1). Multiple risk stratification models have been developed to predict the occurrence of severe CRS. These models typically include pre-lymphodepletion patient laboratory values and/or early post-infusion cytokine or immunologic protein levels^20,26-29. As endothelial activation is a mechanism of CRS and ICANS^20,30, the Endothelial Activation and Stress Index (EASIX) score, which incorporates baseline creatinine, lactate dehydrogenase (LDH), and platelet counts, and modified EASIX formulas have been used for severe CRS risk stratification^27,28.

Table 1:

Frequencies of major CAR T-cell toxicities for commercially available CAR T-cell products: reports from clinical trials and standard of care use (all values in %)

Patient population and product	Grade 3- 5 CRS*	Vasopressor requirement	Hypoxia and/or supplemental oxygen	Grade 3-5 neurologic toxicity (by CTCAE v 4.03 criteria, unless otherwise specified)	Ongoing B- cell aplasia at one year in evaluable responders (anti-CD19 CAR T products only)	Prolonged grade 3-4 cytopenias (by CTCAE v4.03 criteria ≥ 28-35 days, unless otherwise specified)	Grade 3-5 infections	Treatment- related morality
Anti-CD19 CAR T-cell products
Large B-cell lymphoma
Axi-cel9,17,39,86,87,89,116	6.5-16	6-17	22-31	21-35	47-49	29-30	16-28	2.0-5.7
Tisa-cel33,120,168	4.5-17~	1.9-6	8.4-24	5.1-12%	NR	32	20	0-1.3
Liso-cel13,169	1.1-2.2	0-2.6	10	4.3-10	73	37-43	12.3-15	2.2-2.6
Follicular lymphoma
Axi-cel14	6.5	5.4#	24#	15	52	33	18#	0.81
Tisa-cel19	0	6.4	19	3.1	NR	Reported by individual lineages: neutropenia: 15.5%; thrombocytopenia: 16.5%; anemia: 3.1%	5.2	0
Mantle cell lymphoma
Brexu-cel 24,88	3.0-15~	16	34	31-36^	55	26% at more than 90 days	32	2.9-15
Acute lymphoblastic leukemia
Tisa-cel (children and young adults)32,120,170	16-47!	12-25@	17-24	9.0-13%	71	Neutropenia: 21.6% (< 500 cell/mcL); thrombocytopenia: 18% (< 20,000 cells/mcL)120	24	1.3
Brexu-cel (Adults ≥ 18)25	24	40	29	25	50% at 15 months	36	25	3.6
Anti-BCMA CAR T-cell products for multiple myeloma
Ide-cel15,18	3.1-5.5~	NR	NR	3.1-5.7^	NA	Neutropenia: 41-60; thrombocytopenia: 48-59%.	22	1.9-2.3
Cilta-cel16,79	5.2+	4.1	6.2	2.1&	NA	Neutropenia: 30%; thrombocytopenia: 41%.	20	6.2

NA: not applicable. NR: not reported. SOC: standard of care use. *Per Lee, D.W. et al. Current concepts in the diagnosis and management of cytokine release syndrome. Blood 124; 188-195, doi:10.1182/blood-2014-05-552729 (2014), unless otherwise stated. ~Lower value graded according to ASTCT criteria (Lee, D. W. et al. ASTCT Consensus Grading for Cytokine Release Syndrome and Neurologic Toxicity Associated with Immune Effector Cells. Biology of Blood and Marrow Transplantation 25; 625-638, doi:10.1016/j.bbmt.2018.12.758 (2019)) and upper value by Lee 2014 criteria. ^%Lower value per ASTCT ICANS criteria, and upper value by CTCAE v4.03. ^#Some toxicity data include a minority of patients with marginal zone lymphoma. ^Upper value per ASTCT criteria. ^!Lower value by ASTCT criteria and upper value by University of Pennsylvania criteria. ^@Upper value includes only high-dose vasopressors. ⁺Lee 2014 and ASTCT criteria. ^&Per ASTCT criteria.

Before the publication of the American Society for Transplantation and Cellular Therapy (ASTCT) guidelines in 2019³¹ (Supplemental Table 1) multiple grading systems for CRS had been devised, including the original Lee criteria in 2014¹¹, the University of Pennsylvania grading system used in early multicenter trials of tisa-cel^32,33, and with the Memorial Sloan Kettering system used in early evaluation of anti-CD19 CAR T cells for leukemia^8,21. Use of these different grading systems makes comparison of CRS toxicities between these commercial products challenging. However, more uniform application of the ASTCT scale in newer clinical trials and in reports of standard-of-care use of commercial CAR T-cell products will likely make this heterogeneity less of a concern in the future. The ASTCT system has the advantage of being simple and easy to use, though it bears noting that this system reflects only changes in temperature, blood pressure, and respiratory status. Organ toxicities such as coagulopathy, reduced ejection fraction, hepatic or renal insufficiency, and electrolyte derangements, must be captured separately for reporting. ASTCT grade 3 CRS is a broad category, including patients with transient low-dose vasopressor use, as well as patients requiring high doses of a single vasopressor, with or without the addition of vasopressin, for days. Rates of severe CRS with commercial products and grading systems used are summarized in Table 1.

The mainstay of first-line CRS management is the IL-6 receptor antagonist tocilizumab.^31,34 General recommendations for CRS management are summarized in Table 2. Suggested thresholds to give tocilizumab have evolved over time, with algorithms (Table 3) shifting to incorporate tocilizumab for grade 2 and some cases of grade 1 CRS by ASTCT criteria^12,16. Tocilizumab use for selected organ toxicities, such as acute renal or hepatic failure, cardiac ejection fraction decrease, or profound electrolyte derangements, may also be considered³⁵, though in most cases where severe organ toxicities occur, grade 2 or higher CRS is already present. Whether tocilizumab given for established CRS increases ICANS risk is unclear. Patients who receive more tocilizumab exposure are more likely to have high-grade ICANS¹⁷, but these patients also have higher grade CRS and may have underlying factors that predispose to both CRS and ICANS. Animal models have demonstrated no effect of tocilizumab on CAR T-cell efficacy³⁶. In retrospective analyses of clinical trials, no effect of tocilizumab for established CRS has been observed on blood CAR T-cell levels^24,37. Patients receiving tocilizumab for CRS appear to have response rates comparable to patients who do not receive tocilizumab^9,17,24,37. While large, randomized trials are not available, tocilizumab does not appear in retrospective analyses to decrease response durability⁹, progression-free survival (PFS)³⁸ or overall survival (OS)³⁹.

Table 2:

Prevention, monitoring, evaluation, and treatment of CAR T-cell toxicities: CRS and Neurologic Toxicities^*

Toxicity category and references	Pre-treatment assessment	Preventative measures	Post treatment monitoring	Diagnostic evaluation if toxicity occurs	Supportive care for toxicity	Pharmacologic/immunologic toxicity management
CRS 12,35,43,45,46,113,171	-Baseline CRP, ferritin, comprehensive metabolic panel (CMP) including magnesium and phosphate, CBC, PT, PTT, fibrinogen -Baseline TTE or MUGA and ECG -Formal cardiology evaluation in patients with a history of cardiovascular disease -Obtain central venous access.	-Arrange caregiver for first 30 days after cell infusion.	-outpatient: patient and caregiver counseling for temperature monitoring; admit for fever, dyspnea, other evidence of hemodynamic changes -inpatient: q 8 hr vital signs. Patients with CRS may require more frequent (e.g., q 4 hr) vital signs -continuous cardiac monitoring and pulse oximetry for patients with ≥ grade 2 CRS. -At least daily CRS scoring per ASTCT grading system. More frequent monitoring for established CRS. -CMP with magnesium and phosphate, CBC, PT, PTT, fibrinogen, daily. -CRP and ferritin at least 3 times per week in first 2 weeks. Daily if active CRS.	-Rule out infectious causes of fever: blood cultures. Depending on symptomatology: nasal swab for covid-19 and other respiratory viruses, urine cultures, sputum cultures AND 1. Treat infections if identified. 2.Empiric antibiotics for neutropenic fever -Assess ECG, troponin, BNP, and TTE for patients with grade 3-4 and consider for grade 2 CRS	-Acetaminophen and cautious IV fluid support for fevers. -Maintenance IV fluids for ≥ Gr 1 CRS -Fluid boluses for hypotension -Consider transition to vasopressors for hypotension refractory to 2 fluid boluses and anti-IL-6 therapy.	Tocilizumab: -First line for CRS. Indications per published guidelines presented in Table 3. -May be given for CRS concurrently with corticosteroids for co-occurring ICANS. Corticosteroids: -For any grade CRS with inadequate response to tocilizumab or for concurrent treatment with tocilizumab for high grade CRS. -Additional indications and dosing recommendations per published guidelines presented in Table 3. -Taper steroids as rapidly as feasible once desired response has been achieved. Third-line agents: -Consider siltuximab, anakinra, ruxolitinib, emapalumab, antithymocyte globulin, cyclophosphamide.
ICANS 12,43,45,46	-Patients with baseline neurologic disorders: neurology consult. -Consider baseline brain MRI -Perform baseline neurologic evaluation including ICE or CAPD scoring. -Baseline labs as above in the CRS row.	-Arrange caregiver for first 30 days after cell infusion. -Patients should be counseled to avoid driving and hazardous activities for 8 weeks after infusion. -Minimize sedating medications	-outpatient: patient and caregiver counseling for monitoring of neurologic changes; admit for any change from neurologic baseline. -inpatient: ICE/CAPD and neurologic exam at least twice daily. More frequent monitoring may be needed for established ICANS. -Inpatient labs as above under “CRS Monitoring.”	-In moderate to severe or quickly escalating cases, consider head CT to evaluate for hemorrhage, increased intracranial pressure, or cerebral oedema -For Gr 3-4 ICANS, assess for papilledema. -If patient in appropriate clinical condition, consider as adjuncts: --brain MRI to evaluate for cerebral edema --EEG if seizures observed or occult seizures possible and for ≥ Gr 2 ICANS. -- LP if coagulopathy absent or corrected and no evidence of increased intracranial pressure. IT hydrocortisone for ICANS management can be considered if LP performed.	-Fall precautions -modified physical therapy for recovering patients as appropriate -Consider ICU care for patients with ≥ Gr 3 ICANS or ≥ Gr 2 ICANS associated with ≥ Gr 2 CRS. -Consider repeat neuroimaging with CT or MRI every 2-3 days for persistent Gr 3-4 ICANS.	Corticosteroids: -Results of neuroimaging should not delay corticosteroid treatment. -Indications and dosing recommendations per published guidelines presented in Table 3. -After each corticosteroid dose, assess need for subsequent dosing. -Taper corticosteroids as rapidly as feasible once desired response has been achieved. -For concurrent ≥ Gr 1 CRS, give tocilizumab along with corticosteroids. Tocilizumab may be repeated q 8 hrs if needed; however, caution as per risk of worsening ICANS. Limit to 3 doses in 24-hr period and maximum of 4 doses. Second-line therapy: -Consider IT hydrocortisone, anakinra, cyclophosphamide, antithymocyte globulin
Movement disorders/delayed atypical neurologic toxicities after anti-BCMA CAR T-cell therapy 46,80	-Assess baseline malignancy burden -Baseline brain MRI and EEG -Baseline neurology consultation for patients with any neurologic comorbidities		-Monitoring for CRS and ICANS as above -Handwriting assessments with known instrument80 -Extended neurologic monitoring and patient reporting up to 1 year post infusion	-Neurology consultation -Brain PET/CT, brain MRI, and EEG -lumbar puncture to evaluate for leptomeningeal malignancy and to rule out infection -Blood PCR testing for HHV6 and HHV7 -Blood thiamine level	-Administer most effective/feasible bridging therapy to reduce malignancy burden -Corticosteroids for any grade ICANS; second-line agents such as anakinra for non-responding ICANS -Tocilizumab for any concurrent ICANS and CRS	-Treat Parkinsonism per institutional guidelines -Optimal immunosuppressive strategies unknown; 10 mg dexamethasone daily may be considered for mild symptoms. -Consider systemic cyclophosphamide for persistent, severe, or refractory symptoms. -IV Immunoglobulins, plasmapheresis, intrathecal and systemic corticosteroids, and chemotherapy have been attempted. Consider cyclophosphamide for patients with high blood CAR T cell levels.

^*Recommendations per one or more of the cited published guidelines. In cases of disagreement between guidelines, the recommendation incorporating the earliest intervention or most conservative monitoring strategy has been presented. Additional recommendations per authors’ opinion where indicated. ASTCT: American Society of Transplantation and Cellular Therapy. BCMA: B-cell maturation antigen. BNP: B-type natriuretic peptide. CAPD: Cornell Assessment of Pediatric Delirium. CMP: comprehensive metabolic panel. CRP: C-reactive protein. CRS: cytokine release syndrome graded per ASTCT criteria. CT: computed tomography. ECG: electrocardiogram. EEG: electroencephalogram. Gr: grade. HHV: human herpes viruses. Hrs: hours. ICANS: immune effector cell-associated neurotoxicity syndrome. ICE score: Immune Effector Cell-Associated Encephalopathy score. IT: intrathecal. LD: lymphodepletion. LDH: lactate dehydrogenase. LP: lumbar puncture. MUGA: multigated acquisition scan. MRI: magnetic resonance imaging. NSAIDs: nonsteroidal anti-inflammatory drugs. PET: positron emission tomography. PT: prothrombin time. PTT: partial thromboplastin time. TTE: transthoracic echocardiogram.

Table 3:

Comparisons of the ASCO, EBMT/EHA, SITC, and NCCN Toxicity Management Guidelines

	ASCO43	EBMT/EHA44	SITC45	NCCN46
CRS *
indication for first dose of tocilizumab	≥ Gr 2 CRS; Consider for Gr 1 CRS with fever > 3 days	≥ Gr 2 CRS; Consider for Gr 1 CRS with fever/symptoms > 3 days Note: repeat tocilizumab for ≥ Gr 2 CRS with absence of improvement in first 12 hrs	-Consider for Gr 2 CRS for elderly patients, those with comorbid conditions, and pediatric patients with prolonged Gr 2 CRS and/or intolerance of fever. -Administer for all Gr 3-4 CRS -Administer a second dose of tocilizumab with steroids if no improvement in CRS	≥ Gr 2 CRS and for Gr 1 CRS lasting > 3 days in patients with significant symptoms, comorbidities, and/or > 65 years old. -Consider for Gr 1 CRS after axi-cel or brexu-cel lasting > 24 hrs. -Consider for Gr 1 CRS starting < 72 hrs after liso-cel infusion +/− 10 mg dexamethasone. -Repeat in 8 hrs if no improvement. No more than 3 doses in 24 hrs, maximum 4 total doses.
indication for first dose of corticosteroids	Gr 2 CRS → hypotension persisting despite 2 IV fluid boluses AND 1-2 doses of tocilizumab OR any Gr 3-4 CRS	Gr 2 CRS refractory to 1 dose of tocilizumab OR any Gr 3-4 CRS	Any CRS refractory to 1 dose of tocilizumab (steroids administered with a second dose of tocilizumab)	Persistent refractory hypotension after 1-2 doses of anti-IL-6 therapy, or for ≥Gr 2 CRS as per product-specific recommendations below.
corticosteroid dosing recommendations*	-refractory Gr 2 CRS → dexamethasone 10 mg IV q 12 hrs -Gr 3 CRS → dexamethasone 10 mg IV q 6 hrs with taper upon improvement -Gr 4 CRS → methylprednisolone 500 mg IV q 12 hrs with taper upon improvement	refractory Gr 2 CRS OR Gr 3 CRS → dexamethasone 10 mg IV q 6 hrs for 1-3 days. -Gr 4 CRS → dexamethasone 20 mg IV q 6 hrs x 3 days, progressive tapering within 3-7 days	-Example dosing regimens are methylprednisolone 2 mg/kg/day or dexamethasone dosed as 0.5 mg/kg (max 10 mg per dose).	Gr 2 CRS → dexamethasone 10 mg IV q 12-24 hrs depending on product: -axi-cel: Consider dexamethasone 10 mg IV q 24 hrs after initial tocilizumab, regardless of tocilizumab response. -liso-cel: Consider dexamethasone 10 mg IV q 12-24 hrs if CRS occurs < 72 hours after CAR T-cell infusion. -ide-cel: dexamethasone 10 mg IV q 12-24 hrs. -Gr 3 CRS → dexamethasone 10 mg q 6-12 hrs.
dose escalation of corticosteroids	If refractory symptoms after intervention at any grade, treat per recommendations for the next highest grade	If refractory symptoms or deterioration in symptoms after intervention at any grade, treat per recommendations for the next highest grade	Not specified	-Gr 4 or refractory Gr 3 CRS → dexamethasone 10 mg IV q 6 hrs.
Recommendations for refractory high-grade CRS	-methylprednisolone 500 mg IV q 12 hrs; consider anakinra, siltuximab, ruxolitinib, cyclophosphamide, and antithymocyte globulin	-methylprednisolone 1000 mg IV daily x 3 days with taper; consider 1 additional dose of tocilizumab	-For CRS refractory to 2 doses of tocilizumab and corticosteroids, consider anakinra, siltuximab, and high-dose methylprednisolone	-Continued refractory Gr 3-4 CRS despite dexamethasone 10 mg IV q 6 hrs → consider 3 doses of IV methylprednisolone 1-2 grams/day, depending on product prescribing information; for continued refractory CRS, consider q 12 hr dosing.^ -Third line: consider anakinra -Fourth line: consider ruxolitinib, cyclophosphamide, intravenous immunoglobulins, antithymocyte globulin, intrathecal chemotherapy, or extracorporeal cytokine adsorption with continuous renal replacement therapy (CRRT). Must be balanced with risk of infection.
Notes	-Consider earlier intervention for CRS due to axi-cel and brexu-cel	none	none	Prophylactic dexamethasone 10 mg orally q 24 hrs x 3 days, with first dose prior to CAR T-cell infusion, may be considered for axi-cel, after weighing potential risks and benefits.
ICANS *
indication for first dose of corticosteroids	-Gr 2 ICANS in high-risk patients or for high-risk products -All Gr 3-4 ICANS	-Gr 2-4 ICANS	-Gr 2 ICANS due to axi-cel or brexu-cel; consider for Gr 2 ICANS due to other products -All Gr 3-4 ICANS	-≥Gr 2 ICANS. -Consider for Gr 1 ICANS occurring < 72 hrs after infusion of liso-cel or ide-cel.
corticosteroid dosing recommendations	-Gr 2 ICANS → dexamethasone 10 mg IV x 2 doses and reassess -Gr 3 ICANS → dexamethasone 10 mg IV q 6-12 hrs or methylprednisolone 1 mg/kg IV q 12 hrs -Gr 4 ICANS → methylprednisolone 1,000 mg IV 1-2 times daily x 3 days	-Gr 2-3 ICANS → dexamethasone 10 mg IV q 6 hrs x 1-3 days - Gr 4 ICANS → methylprednisolone 1000 mg IV daily x 3 days, then 250 mg twice daily for x 2 days, 125 mg twice daily for x 2 days, 60 mg twice daily x 2 days.	-Recommendations not specified. -Give at least 2 doses and taper quickly once ICANS has improved. -In agreement with axi-cel package insert, may give dexamethasone 10 mg IV q 6 hrs for Gr 2-3 ICANS; give methylprednisolone 1000 mg daily x 3 days for Gr 4 ICANS.	- Gr 1 ICANS occurring < 72 hrs after infusion of liso-cel or ide-cel → dexamethasone 10 mg q 12-24 hrs x 2 doses and reassess. -Gr 2 ICANS: 1 dose of dexamethasone 10 mg and reassess. Can repeat q 6-12 hrs if no improvement. -Gr 3 ICANS: dexamethasone 10 mg q 6 hrs or methylprednisolone 1 mg/kg IV q 12 hrs. Consider methylprednisolone 1 gram/day x 3-5 days for 3-5 days for axi-cel and brexu-cel.
Recommendations for refractory high-grade ICANS	-methylprednisolone 1,000 mg IV 2-3 times daily; consider anakinra, siltuximab, ruxolitinib, cyclophosphamide, antithymocyte globulin, or intrathecal hydrocortisone (50 mg) plus methotrexate (12 mg)	-Consider anakinra, siltuximab, intrathecal chemotherapy, or systemic chemotherapy	Not specified	-Gr 4 ICANS: High dose corticosteroids such as IV methylprednisolone 1-2 grams q 12-24 hrs.^ -corticosteroid refractory → anakinra 100 mg q 6 hrs.

^*Alternative corticosteroids at an equivalent dose may be considered.

^{^}A possible taper of methylprednisolone is methylprednisolone 1000 mg/day for 3 days, followed by rapid taper at 250 mg every 12 hours for 2 days, 125 mg every 12 hours for 2 days, and 60 mg every 12 hours for 2 days. ASCO: American Society of Clinical Oncology. EBMT: European Society for Blood and Marrow Transplantation. EHA: European Hematology Association. Gr: grade. Hrs : hours. NCCN: National Comprehensive Cancer Network. SITC: Society for the Immunotherapy of Cancer.

^*CRS and ICANS grading per American Society of Transplantation and Cellular Therapy grading system unless otherwise specified.

Standard therapy for tocilizumab-refractory CRS is corticosteroids, commonly methylprednisolone or dexamethasone in escalating doses with immediate cessation or tapering if necessary once toxicity control has been achieved. The optimal dosing and timing of corticosteroids is unknown. There were early reports of a decrease in CAR T-cell blood levels, poor CAR T-cell persistence, and malignancy relapse following high doses of corticosteroids administered to patients for management of inflammatory syndromes that would now be termed CRS^8,40,41. However, some retrospective analyses have not identified a detrimental effect of corticosteroids upon response rate or durability^9,17,24. In a retrospective analysis of patients receiving axi-cel for large B-cell lymphoma (LBCL), higher cumulative corticosteroid doses were associated with shorter PFS; additionally, earlier use, more prolonged use, and higher cumulative dose of corticosteroids were associated with shorter overall survival³⁸. These associations were maintained despite limiting the analysis to patients with high LDH as a proxy for high-burden malignancy³⁸. Similarly, a retrospective analysis of 62 patients receiving various anti-BCMA CAR T-cell therapies demonstrated no effect of any use or higher dose corticosteroids on survival outcomes, but corticosteroid duration ≥ 5 days was associated with inferior PFS⁴². However, for these analyses it is possible that patients receiving corticosteroids had other biological differences that would predispose them to both worse toxicity and poorer response with CAR T-cell therapy. Dose and duration of corticosteroid therapy might be critical factors in determining the impact of corticosteroids on CAR T cells.

Algorithms for CRS toxicity management with provisions for ordering of tocilizumab, corticosteroids, and other therapies, have been published by professional societies, specifically, the American Society of Clinical Oncology (ASCO)⁴³, the European Society of Blood and Marrow Transplantation (EBMT)⁴⁴, the Society for the Immunotherapy of Cancer (SITC)⁴⁵, and the National Comprehensive Cancer Network (NCCN)⁴⁶. US FDA package inserts for the cellular products may provide product-specific recommendations for toxicity management. These algorithms are compared in Table 3. Generally, thresholds for administering tocilizumab and corticosteroids are similar. The SITC guidelines were published first and are the most general and open-ended. The NCCN guidelines are the most specific and incorporate the most product-specific management recommendations⁴⁶.

With earlier toxicity intervention, the rates of high-grade toxicities have decreased, with rates of grade 3 or higher CRS and neurologic toxicities both decreasing during the course of ZUMA-1⁹, and with recent data from the United Kingdom indicating that with commercial use, tocilizumab and corticosteroids are being used more often, and high-grade CRS is being reported less frequently⁴⁷. Multiple preventative and early intervention strategies evaluated in the prospective setting are summarized in Supplemental Table 2 and Table 4, respectively. Prophylactic and early intervention tocilizumab and corticosteroids have demonstrated some efficacy in decreasing rates of high-grade CRS, with decreases in multiple peak inflammatory cytokines and immunologic proteins observed, and with no clear detrimental effects on response rate, remission durability, or blood CAR T-cell levels^48-51. A preliminary analysis of ZUMA-1 cohort 3, which evaluated prophylactic tocilizumab on Day +2 following axi-cel for LBCL, lead to concerns as to whether early prophylactic tocilizumab increases risk of severe neurologic toxicity⁵², though this trend was not observed in a trial of prophylactic tocilizumab on Day 0 prior to infusion of T cells expressing a CAR with a 4-1BB costimulatory domain⁴⁹ (Supplemental Table 2). In other analyses, early intervention for CRS with tocilizumab, either as monotherapy⁵³, or along with corticosteroid therapy as early intervention and/or prophylaxis^48,50,51, does not clearly increase or decrease risk of severe neurotoxicity, (Table 4 and Supplemental Table 2). Bruton’s tyrosine kinase (BTK) inhibitors^54,55 and Janus kinase (JAK) inhibitors⁵⁶ are also being investigated as CRS prophylaxis (Supplemental Table 2).

Table 4:

Prospectively evaluated early intervention strategies with immunosuppressive agents for CAR T-cell toxicity management.

Preventative strategy	Patient population	CAR Product(s)	Sample size	Detailed intervention approach	CRS and ICANS Outcomes	Anti-malignancy response outcomes
Early intervention with tocilizumab and corticosteroids to prevent severe CRS48	Pediatric/Young adult B-cell ALL	Locally manufactured investigational anti-CD19 CAR T-cell product (University of Washington)	n = 23, DLT cohort n = 20, early intervention cohort	DLT cohort: tocilizumab and/or corticosteroids for dose-limiting and life-threatening CRS or neurologic toxicity Early intervention cohort: tocilizumab for ≥ 10 hrs fever ≥ 39°C, non-fluid-responsive hypotension, requirement for oxygen supplementation; dexamethasone 5-10 mg q 6-12 hrs for persistent/recurrent fever, vasopressor requirement, escalating oxygen requirement	Non-significant decrease in severe CRS (sCRS) in the early intervention cohort. sCRS defined as requirement for vasopressors, inotropes, or intubation. Similar rates of severe neurotoxicity (≥ gr 3 by CTCAE criteria, or ≥ grade 2 seizures) between cohorts. sCRS in DLT cohort: 30% sCRS in early intervention cohort: 15% severe neurotoxicity in DLT cohort: 22% severe neurotoxicity in early intervention cohort: 25%	MRD negative CR, DLT cohort: 91% MRD negative CR, early intervention cohort: 95% No difference in leukemia-free survival or overall survival between cohorts
Early intervention with tocilizumab to prevent severe CRS in high tumor burden patients53	Pediatric/Young adult B-cell ALL	Tisa-cel	n = 70 (15 high tumor burden)	BM ALL ≥ 40%* considered high tumor burden Tocilizumab at time of at least 2 temperatures of ≥ 38.5°C in 24 hrs, at least 4 hrs apart in high tumor burden patients	Decrease in grade 4 CRS compared with historical controls. Grade 4 CRS^ in high tumor burden patients on trial: 27% Grade 4 CRS post hoc historical comparison in high tumor burden patients: 50% Neurologic toxicity& gr ≥ 2 in trial high tumor burden: 53% Neurologic toxicity gr ≥ 2 in historical control high tumor burden: 54%	No difference in CAR T-cell expansion, ORR, or 2-yr EFS between trial high tumor burden and historical control high tumor burden patients ORR trial high burden patients: 87% ORR historical control high burden patients: 85%
Early intervention with tocilizumab and corticosteroids for CRS and neurologic toxicity (ZUMA-1, cohort 4)50	Adult LBCL	Axi-cel	n = 41	Levetiracetam seizure prophylaxis starting on Day 0 Tocilizumab 8 mg/kg for ≥ gr 2 CRS, gr 1 CRS > 24 hrs, and for CRS occurring with ≥ gr 2 neurologic toxicity. Corticosteroids for ≥ gr 1 neurologic toxicity, ≥ gr 2 CRS, gr 1 CRS lasting more than 3 days. Per-protocol doses range from dexamethasone 10 mg once to methylprednisolone 1000 mg/day x 3 days, depending on toxicity grade.	Any grade CRS: 93% Low rates of both severe CRS% and neurologic toxicity&, with response rates and durability comparable to ZUMA-1 cohorts 1-2. ≥ gr 3 CRS: 2% Any neurologic toxicity: 61% ≥ gr 3 neurologic toxicity: 17% ≥ gr 3 CRS, historical comparison of ZUMA-1, cohorts 1-2: 13% ≥ gr 3 neurologic toxicity, historical comparison of ZUMA-1, cohorts 1-2: 28% Lower cumulative corticosteroid doses compared to ZUMA-1 cohorts 1-2 Propensity score matching analysis: ≥ gr 3 CRS and ≥ gr 3 neurologic toxicity still numerically lower than ZUMA-1 cohorts 1-2, when controlled for tumor burden and other patient factors	ORR: 73% CR rate: 51% 12-month estimated PFS: 57%; 12-month estimated OS: 68%. Median peak blood CAR T-cell levels were 52.9 cells/mcL (cohort 6) and 43 cells/mcL (cohorts 1-2).

BM: bone marrow. CR: complete remission. LBCL: large B-cell lymphoma. EFS: event-free survival. Gr: grade. Hrs: hours. MRD: minimal residual disease. ORR: overall response rate. *BM biopsy post-lymphodepletion and prior to CAR T-cell infusion. ^CRS graded per University of Pennsylvania Scale. ^&Neurotoxicity graded by the CTCAE v4.03 grading system. ^%CRS per 2014 Lee criteria.

The role of anti-cytokines other than tocilizumab, small molecules, and T-cell ablative agents as therapies for steroid-refractory CRS, or as steroid-sparing agents in established CRS, remains poorly defined. In absence of a randomized trial, evaluation of therapies for steroid-refractory toxicities is complicated by the fact that CRS will almost always naturally resolve with time, or with greater cumulative doses of corticosteroids, which are usually continued while rescue therapies are attempted. The IL-6 inhibitor siltuximab^13,15,57 and the IL-1 receptor antagonist anakinra^13,15,16,57 have been used at some centers for CRS. Anakinra prevents CRS in mouse models without affecting antitumor efficacy^36,58. Use of anakinra for steroid-refractory CRS has been reported in case series^59-61. The JAK inhibitor ruxolitinib may have utility in steroid-refractory CRS; however, in vitro evidence that ruxolitinib may dampen CAR T-cell proliferation is a concern⁶². More recently, the interferon-γ (IFN-γ) inhibitor emapalumab, which was developed for primary HLH, has been used anecdotally for CRS refractory to steroids and multiple other anticytokine therapies⁶³, with preclinical evidence indicating IFN-γ blockade does not inhibit CAR T-cell antitumor activity⁶⁴. Dasatinib, by inhibiting lymphocyte-specific protein tyrosine kinase, reversibly halts CAR T-cell proliferation, cytokine reduction, and cytolysis^65,66. However, limited clinical evidence exists for use of dasatinib for steroid-refractory CRS⁶⁷. T-cell ablative therapies, such as cytotoxic chemotherapy with cyclophosphamide, antithymocyte globulin, and CD52-targeted monoclonal antibodies have been incorporated into CRS treatment algorithms for refractory cases in which complete depletion of CAR T cells is an acceptable outcome³², though large-scale reports of experience with these therapies are lacking.

II. Neurologic toxicities

1. ICANS

Neurologic toxicity remains a major concern and barrier to use of CAR T-cell therapy. The diverse constellation of possible neurologic symptoms include headaches, hand tremors, confusion/encephalopathy, dysphasias, and varying decreases in level of alertness, from somnolence to obtundation³¹. Less commonly, there can be generalized or focal muscle weakness, myoclonus, and seizures³¹. Cerebral edema caused deaths on a prior clinical trial of JCAR015 in patients with B-cell acute lymphoblastic leukemia (ALL)¹⁰, and neurologic deaths have occurred with commercially available products^17,18,25. Neurologic toxicity usually occurs concurrent with or shortly after CRS, and less commonly before CRS^13,32,37,68. Delayed instances of ICANS starting more than 3 weeks after CAR T-cell infusion⁶⁹ and ICANS in absence of CRS²⁵ have been reported. ICANS occurs not only following anti-CD19 CAR T cells, but also after anti-BCMA CAR T cells^15,16,70.

CAR T cells have excellent penetration into the cerebrospinal fluid (CSF), though not all patients with detectable CSF CAR T cells experience substantial neurologic toxicity^6,37,68. Additionally, the blood-brain-barrier is highly permeable to systemic cytokines in patients experiencing severe ICANS^30,68, and in humans and in a non-human-primate model some pro-inflammatory cytokines may be more concentrated in the CSF than serum, indicating local production of inflammatory proteins in the central nervous system (CNS)^68,71. Patients with high-grade neurologic events have higher levels of cytokines and inflammatory proteins in the CSF^24,68, suggesting that these proteins are involved in the pathogenesis. Risk factors for ICANS have high overlap with risk for CRS and include high nodal and/or marrow tumor burden^{13,18,21,22,30,68}, increased baseline patient inflammatory markers^13,17,18,22, lower baseline platelets⁶⁸, receipt of/requirement for bridging therapy^13,18, higher cell dose (within-product comparison)^30,68 and peak blood CAR T-cell levels and area under the curve (AUC)^{13,14,22,24,25,68}. Patients with evidence of pre-CAR-T clonal hematopoiesis additionally have increased risk of severe ICANS⁷². This is not well understood but is possibly due to an underlying pro-inflammatory state⁷². Not surprisingly, high-grade neurologic toxicities occur more often in patients who experience high grade CRS^30,32,68. As with CRS, modified EASIX scores, or EASIX scores used with baseline C-reactive protein (CRP) and ferritin values, can assist in predicting risk of severe ICANS^27,28.

The rate of neurologic events varies with the CAR T-cell product (Table 1). In an analysis of the French DESCAR-T registry data incorporating propensity score-matching, axi-cel for treatment of LBCL had higher rates of grade 1-2 and ≥ grade 3 ICANS compared with tisa-cel for the same indication⁷³. The more severe neurotoxicity caused by axi-cel and brexu-cel is thought to be due to the CAR construct incorporating a CD28 hinge, transmembrane, and costimulatory domain, leading to more rapid in vivo CAR T-cell expansion and higher levels of cytokine and other immunologically active proteins by axi-cel^45,69.

The National Cancer Institute’s Common Terminology for Adverse Events (CTCAE) version 4.03 was used to grade neurologic toxicity for first the trials leading to US FDA approval of all now commercially available CAR T-cell products^{9,14,15,24,25,32,33} except cilta-cel, for which neurologic toxicity was graded by the ASTCT ICANS scale in the phase 2 portion of the CARTITUDE-1 clinical trial¹⁶. The ICANS grading scale (Supplemental Table 3) incorporates a cognitive score, the immune effector cell encephalopathy (ICE) score, which includes the domains of orientation, naming, following commands, writing, and attention³¹. Children less than 12 years old may be evaluated with the Cornell Assessment of Pediatric Delirium (CAPD) rather than ICE³¹. ICANS also incorporates level of alertness and presence/absence of motor defects, seizures, elevated intracranial pressure, and cerebral edema, all of which result in a grade of 3 or higher if present³¹.

Corticosteroids remain the most highly utilized first-line therapy for ICANS. The optimal timing for initiation of corticosteroids is unknown. Due to concerns of decreased CAR T-cell proliferation or persistence, corticosteroids were reserved for grade 3 and higher neurologic toxicity by CTCAE criteria in early treatment algorithms^9,74. However, like CRS, dexamethasone is now considered for grade 2 and some cases of grade 1 ICANS^12,16. Low-dose prophylactic corticosteroids and early intervention with corticosteroids for established neurotoxicity may decrease rates of severe neurologic toxicity and concurrently decrease the cumulative dose of corticosteroids required for toxicity management^50,51 (Table 4 and Supplemental Table 2). Case series suggest intrathecal corticosteroids may have a role, given that lumbar puncture can be safely performed, even in ICANS refractory to systemic corticosteroids^75,76. Use of siltuximab for first-line management of ICANS is being investigated (NCT04975555).

CRS occurring at the same time as ICANS requires careful management. This situation is addressed with specific guidance in the ASCO and NCCN consensus recommendations^43,46. Because of the possible risk of tocilizumab worsening neurologic toxicity, ICANS treatment with corticosteroids may take priority over management of low-grade CRS⁴³. This being said, tocilizumab may be administered alone for CRS overlapping with grade 1 ICANS or given along with corticosteroids to manage grade 2 or higher ICANS^43,46, though patients should be monitored for worsening neurologic toxicity, especially after repeated doses⁴³ (Table 2). Corticosteroids are generally dosed per the grade of ICANS in published guidelines^43,46, but in cases of severe CRS and mild to moderate ICANS, corticosteroids could reasonably be dosed to address the more severe toxicity. Co-occurrence of grade 2 or higher CRS and grade 2 or higher ICANS is a potential indication for intensive care monitoring^43,46.

Mouse models have shown efficacy of anakinra in prevention of neurotoxicity³⁶. Prospective evaluation indicates that anakinra may be effective as a preventative agent for ICANS⁷⁷ (Supplemental Table 2). Anakinra had limited use for active ICANS in few patients in the early trials leading to commercial use of CAR T-cell therapy^15,16, though it is now commonly used in the standard of care setting for corticosteroid-refractory ICANS^57,61. Whether it can be used as a steroid sparing agent is unclear, and one case series suggests it does not hasten steroid tapering⁷⁸. Similar to CRS, T-cell depleting therapies, such as cytotoxic chemotherapies and T-cell targeted agents, may have a role for refractory cases in which the complete ablation of CAR T cells is desirable.

2. Movement disorders after anti-BCMA CAR T cells

The CARTITUDE-1 study evaluating cilta-cel as fourth-line therapy for multiple myeloma resulted in atypical delayed neurologic syndromes in 12% of patients¹⁶. The median time of onset of these neurologic treatment-emergent adverse events was 27 days (range, 11-914)^79,80. Six percent of patients experienced movement disorders, with manifestations including Parkinsonism, micrographia, tremor, gait disturbance, cogwheel rigidity, psychomotor retardation, and flat affect^16,79. Parkinsonism was fatal in one patient¹⁶. Other delayed neurologic adverse events included altered mental status, concentration impairment, facial paralysis, diplopia, nystagmus, cranial nerve palsy, ataxia, sensory loss, and peripheral motor and sensory neuropathies⁸⁰. Onset of delayed neurologic toxicities was insidious, occurred after ICANS resolution, and was not responsive to carbidopa/levodopa, corticosteroids, or other immunosuppressive therapies⁸⁰. Micrographia is an early manifestation⁸⁰. Parkinsonism has additionally been reported following treatment with ide-cel⁸¹, indicating that movement disorders might be a toxicity of anti-BCMA CAR T-cell therapy as a class.

Risk factors identified for these toxicities after cilta-cel included two or more of the following: high baseline tumor burden, presence of ICANS, grade ≥ 2 CRS, and high and persistent blood CAR T-cell levels⁸⁰. Autopsies of two patients with movement disorders revealed focal gliosis and a T-cell infiltrate in the basal ganglia in both patients, but no changes to the substantia nigra pigmentation, as would be seen in primary Parkinson’s disease; there was additionally no BCMA expression identified in normal brain tissue⁸⁰. However, findings regarding BCMA expression in the relevant brain regions are conflicting. BCMA expression as assessed by immunohistochemistry has been detected in the caudate nucleus of the basal ganglia on a subset of neurons and astrocytes, as well as in neurons in the adjacent frontal cortex, of a patient with Parkinsonian symptoms after cilta-cel⁸². RNA expression of BCMA was also detected in the caudate nucleus in specimens from Allen Brain Atlas⁸². In a later report, immunohistochemical staining of samples of 63 normal patient cerebrum, basal ganglia, cerebellum, and brainstem samples did not detect BCMA expression beyond nonspecific antibody binding⁸³. Whether these neurologic toxicities are an “on-target” toxicity is not well established.

Once enhanced monitoring and preventive measures were implemented across the cilta-cel clinical development program, the rates of these delayed neurologic toxicities decreased to <1% in a subsequent report⁸⁰. These strategies included more aggressive bridging therapy to decrease pre-lymphodepletion disease burden, post-infusion monitoring with handwriting assessments, algorithms for early intervention on CRS and ICANS, and extended monitoring for neurologic toxicity up to a year after infusion⁸⁰ (Table 2). There is anecdotal evidence of effectiveness of high-dose systemic cyclophosphamide to reduce extremely high numbers of blood CAR T cells, which have been detected more than 2 weeks after CAR T-cell infusion^81,84. However, these toxicities remain a concern and a potential barrier to use for these agents.

III. Hematologic toxicities

1. CAR T-cell-related myelosuppression

High grade cytopenias occurring within the first several days to a month following lymphodepletion chemotherapy are very common and expected⁸⁵. Neutropenia overlapping with CRS commonly results in neutropenic fevers, which have been reported in 2-36% of patients^{9,13,15,19,32,33,86,87}.However, long-duration grade 3 or higher cytopenias occur commonly with both B-cell directed and plasma-cell directed CAR T-cell therapy, with rates as summarized in Table 1. Prolonged grade 3-4 cytopenias, predominantly neutropenia, lasting 6 months or more have been reported in a portion of patients^13,15,88,89. Prolonged cytopenias have additionally been reported following anti-CD30 CAR T-cell therapy for Hodgkin lymphoma^90,91, and even following CAR T-cell therapy for solid tumors⁹², suggesting myelosuppression is not limited to a specific target antigen. Additionally, cytopenias, particularly neutropenia, may be biphasic or intermittent with delayed cytopenias occurring weeks to months after initial apparent recovery^93-95.

The pathophysiology of prolonged post-CAR-T myelosuppression is incompletely understood. Risk varies with CAR product⁹⁶ and conditioning chemotherapy^94,97. Risk factors include higher baseline bone marrow burden of malignancy^20,94, lower baseline blood counts^96,98,99, higher baseline inflammatory markers⁹⁸, greater number of prior therapies²⁰, higher peak blood elevations of IL-6⁹⁹, and more severe post-CAR-T CRS^{20,94,96,99,100} and/or ICANS⁹⁶, suggesting that hematopoietic reserve, baseline inflammatory milieu, and post CAR-T inflammation contribute to prolonged cytopenias. In series wherein bone marrow biopsies are collected prospectively, post-therapy elevated bone marrow aspirate cytokines¹⁰¹ and CAR T-cells⁹⁴ have been associated with prolonged myelosuppression. Single-cell RNA sequencing of bone marrow samples has revealed that patients with prolonged cytopenias have persistent populations of clonally expanded CD8⁺ IFN-γ-expressing T cells; chronic IFN-γ exposure in the bone marrow may suppress hematopoietic cells¹⁰².

Post-cellular-therapy cytopenias have now been termed Immune Effector Cell-Associated Hematotoxicity (ICAHT) by the European Hematology Association (EHA) and the European Society of Blood and Marrow Transplantation (EBMT)¹⁰³. Late ICAHT is defined as neutropenia lasting beyond Day +30 after cellular infusion¹⁰⁴. A grading scale has been proposed based on severity and duration of neutropenia¹⁰⁴.

The management of ICAHT is evolving (Table 5). The CAR-HEMATOX score is a 5-item instrument that incorporates baseline hemoglobin, platelet, and neutrophil counts, as well as baseline ferritin and CRP⁹⁸. As this instrument has been shown to predict prolonged hematotoxicity in multiple myeloma and multiple lymphoma subtypes, clinicians may consider baseline bone marrow biopsy, early (~Day +2) growth factor support, and initiation of gram negative and fungal antimicrobial prophylaxis upon start of neutropenia for patients with a high CAR-HEMATOX score; whereas growth factor support and prophylactic antimicrobials may be delayed in patients in a low CAR-HEMATOX score, though this strategy has not been evaluated prospectively¹⁰⁴. Patients with cytopenias greater than 30 days in duration (late ICAHT) should undergo a stepwise evaluation to rule out other causes of myelosuppression (Table 5)¹⁰⁴, which may include bone marrow biopsy with cytogenetics and next generation sequencing, and evaluation for underlying viral etiologies, such as cytomegalovirus (CMV), Epstein-Barr virus (EBV), and parvovirus¹⁰⁴. Bone marrow biopsy can rule out other causes such as progressive B-cell/plasma cell malignancy or new myeloid malignancy. Bone marrow biopsy in cases of late ICAHT typically shows a nonspecific aplasia and hypocellularity^100,104. Red cell and platelet transfusions should be administered, tailored to patient status and institutional guidelines. Retrospective reports are conflicting as to whether growth factors worsen CRS/ICANS in the post-CAR-T setting, though, if this is the case, this risk must be balanced with infectious risks associated with prolonged neutropenia^105-107.

Table 5:

Prevention, monitoring, evaluation, and treatment of CAR T-cell toxicities: hematologic and immunologic toxicities^*

Toxicity category and references	Pre-treatment assessment	Preventative measures	Post treatment monitoring	Diagnostic evaluation if toxicity occurs	Supportive care for toxicity	Pharmacologic/immunologic management of toxicity (flat doses for adult patients only)
Hematologic toxicity/ICAHT 35,43-45,104	-Pre-treatment risk stratification with CAR-HEMATOTOX score: CBC, CRP and ferritin. Consider baseline bone marrow biopsy in high-risk patients.	-Minimize/discontinue myelosuppressive medications -Consider early growth factors and prophylactic antimicrobials for patients with a high CAR-HEMATOX score.	-CBC monitoring daily during CRS -At CRS resolution, CBC monitoring at least weekly for the first month and monthly for the first six months in all patients (author recommendation) -Monitoring may be modified based on risk.	-For grade 3-4 cytopenias with duration > 28 days: reticulocyte count, peripheral blood smear, and survey for infections. --Complete metabolic panel, ferritin, fibrinogen, triglycerides to evaluate for IEC-HS --Bone marrow aspirate and biopsy for with flow cytometry, cytogenetics +/− myeloid NGS panel --viral studies: PCR for CMV/EBV, parvovirus, HHV6	-Transfusion support with irradiated packed red cell and platelet blood products per institutional guidelines -Growth factor support is controversial; consider initiating when ANC < 500 cells/mcL. Avoid GM-CSF.	-TPO agonists +/− corticosteroids -stem cell boost (if cells available) -allogeneic stem cell transplant if feasible in truly refractory cases
IEC-HS 43-45,113	-Baseline CMP, CBC, CRP, ferritin, LDH, PT, PTT, fibrinogen, triglycerides.	-Appropriate prophylactic antimicrobial therapy as below (“Immunosuppression/infections”); consider gram negative and fungal coverage for patients receiving anti-cytokine and corticosteroids therapies. -Outpatients should arrange caretakers to assist in toxicity monitoring in the first 30 days.	-Temperature monitoring for outpatients; urgent evaluation of fevers -Inpatients with active CRS or recovering from CRS: daily CBC, PT/PTT, CMP, CRP, ferritin, fibrinogen	-Diagnostic labs include ferritin, CMP, CBC, PT, PTT, fibrinogen, triglycerides, and LDH -Consider other causes: blood cultures and other evaluation for infection -Consider bone marrow biopsy to evaluate for hemophagocytosis -Consider soluble CD25, blood lymphocyte phenotyping for NK-cell count, IL-10, IL-18 IFN-γ, CXCL9 ratio, CXCL10 to support diagnosis.	-Correct hypofibrinogenemia of < 100 mg/dL with cryoprecipitate, correct for < 150 mg/dL if bleeding -Vitamin K for INR > 1.5; FFP for INR > 2. -Transfusion and growth factor support (per above ICAHT)	-First line: anakinra +/− corticosteroids -Second line: increase dose of anakinra, then increase dose of corticosteroids -Alternative agents: ruxolitinib, low-dose etoposide, emapalumab
Tumor lysis syndrome (TLS) 45,46	-Assess for G6PD deficiency prior to treatment in patients considered high risk for TLS due to high tumor burden or aggressive malignancy histology	-Start allopurinol or rasburicase (for allopurinol intolerant patients) with LD chemotherapy in patients with high marrow or extramedullary disease burden.	-At least daily CMP (as Table 2 “CRS” row), with LDH, uric acid. For high-risk patients, more frequent lab monitoring (at least 2x per day, author recommendation).	-Monitor TLS labs q 6-12 hrs or per institutional guidelines for active TLS	-Hydration and electrolyte management per institutional guidelines	Rasburicase for patients without G6PD deficiency; febuxostat for patients with G6PD deficiency, as per institutional guidelines
Immunosuppression/infections12,44-46,124,125, except covid-19	-Pre-leukapheresis serologies for EBV, HBV, HCV, HIV, HSV, VZV, CMV -HBV PCR for viral DNA if surface antigen or core antibody positive. -Hepatitis C PCR for viral RNA if antibody positive. -Consider PCR testing for EBV, CMV, and HBV regardless of serology results. -Treatment of individuals with HIV can be made on a case-by-case basis^	-Delay lymphodepletion and/or CAR T-cell infusion if patient is febrile and has evidence of infection and give appropriate antimicrobial treatment. Resume therapy when patient is afebrile at least 48 hrs and clinical evidence infection is controlled/symptoms improved. -Delay treatment for patients with detectable HBV/HCV DNA/RNA or positive HBV surface antigen until infection is treated. -Entecavir or tenofovir prophylaxis for patients with + HBV core antibody, at least 6 months- 1 yr, consider longer for continued B-cell aplasia -All patients should receive prophylactic antimicrobials for pneumocystis and HSV/VZV x 6 months to 1 yr; consider longer if CD4 < 200 cells/mcL. -Re-vaccinate with killed/inactivated vaccines and live attenuated > 6 months and > 12 months after CAR T cells, respectively AND CD4 count > 200 cells/mcL (exceptions below).	-check immunoglobulin levels monthly -check CD4 count, B-cell count every ~3 months -HBV viral DNA monitoring in core antibody positive patients -Consider weekly CMV PCR monitoring in seropositive individuals who have received > 3 days of steroids until 1 month after last dose of corticosteroids.	-Rule out infectious causes of fever: blood cultures. Depending on symptomatology: nasal swab for covid-19 and other respiratory viruses, urine cultures, sputum cultures.	-Immunoglobulin replacement therapy for IgG < 400-600 mg/dL or recurrent infections; replacement given more routinely in children; may consider cessation of replacement in adults > 3 months after anti-CD19 CAR T-cell infusion -Consider fungal and gram-negative antibacterial prophylaxis in the following circumstances: intensive lymphodepletion regimens (e.g., containing anti-CD52), prolonged corticosteroid use, prolonged neutropenia.	- Treat identified bacterial, fungal or influenza infections per institutional guidelines for immunosuppressed individuals -Empiric antibiotics for neutropenic fever
Covid-19 infection 132,172,173	-PCR testing for SARS-CoV-2 within 2-3 days of lymphodepletion chemotherapy	-Delay lymphodepletion chemotherapy at least 14-20 days and until symptom improvement in patients with covid-19 infection -counsel masking and social distancing during periods of high community infection rate -pre-CAR-T vaccination for influenza and covid-19, if not already received - Caretakers should follow CDC guidelines for COVID-19 vaccination based on age and health status -Re-vaccination for covid-19 and influenza at ≥ 90 days after cell infusion.	-Patients and caretakers counseled to monitor for symptoms and seek care promptly	-SARS-CoV-2 PCR testing, concurrent testing for influenza and RSV in patients with respiratory symptoms -In patients with covid-19 symptoms and negative nasal PCR testing for SARS-CoV-2, consider repeat testing, chest CT imaging +/− bronchoscopy to evaluate for lower respiratory infection and other infectious causes	-Consider early hospital admission for monitoring in symptomatic patients, depending on institutional resources	- Nirmatrelvir and ritonavir or best available antiviral therapy for prevention of hospitalization and death in outpatients with covid-19 infection -Remdesivir or best available antiviral therapy for inpatients with covid-19, except in critical illness -Dexamethasone for inpatients with covid-19 requiring supplemental oxygen; anti-cytokine therapy (tocilizumab) or Jak inhibitors may be considered in worsening illness (infectious disease consultation recommended). -High titer convalescent plasma can be considered in outpatient or inpatient settings, with most benefit early in the course of illness, or in cases of protracted illness.

^*Recommendations per one or more of the cited published guidelines. In cases of disagreement between guidelines, the recommendation incorporating the earliest intervention or most conservative monitoring strategy has been presented. Additional recommendations per authors’ opinion where indicated.

^{^}Depending on processing capabilities of the manufacturer; infectious disease consultation strongly recommended. ANC: absolute neutrophil count. CBC: complete blood count with differential. CMP: complete metabolic panel, including sodium, potassium, creatinine, aspartate aminotransferase, alanine aminotransferase, bilirubin, direct bilirubin, phosphorus, and magnesium. CMV: cytomegalovirus. CRP: C-reactive protein. CRS: cytokine release syndrome. CXCL: chemokine (C-X-C motif). EBV: Epstein-Barr virus. FFP: fresh frozen plasma. GM-CSF: granulocyte macrophage colony-stimulating factor. HBV: hepatitis B virus. HCV: hepatitis C virus. HHV: human herpes virus. HIV: human immunodeficiency virus. HSV: herpes simplex viruses. ICAHT: immune effector cell-associated hematotoxicity. IEC-HS: immune effector cell-associated hemophagocytic lymphohistiocytosis-like syndrome. IFN-γ: interferon gamma. IL: interleukin. INR: international normalised ratio. LDH: lactate dehydrogenase. NGS: next generation sequencing. PCR: polymerase chain reaction. PT: Prothrombin time. PTT: Partial thromboplastin time. RSV: respiratory syncytial virus. SARS-CoV-2: severe acute respiratory syndrome coronavirus 2. TPO: thrombopoietin. VZV: varicella zoster virus.

Thrombopoietin (TPO) receptor agonists mechanistically may have utility in treatment of post-CAR-T-cell cytopenias, as they can bypass the inhibition of the TPO receptor by IFN-γ¹⁰⁸. While published data are limited in terms of its use^23,94,109, TPO agonists are increasingly used in cases of prolonged cytopenias, with, for instance, 15% of patients receiving these agents for management of cytopenias after commercial ide-cel¹⁸. Corticosteroids have also been attempted, though utility is unclear^94,110. Autologous stem cell boost for management of prolonged cytopenias, in cases in which cryopreserved cells are available, has demonstrated feasibility following both anti-CD19¹¹¹ and anti-BCMA CAR T-cell therapy¹¹² and is also increasingly used, with 5% of patients receiving commercial ide-cel also having a stem cell boost for management of cytopenias¹⁸. Allogeneic stem cell transplantation (allo-HSCT) represents the last resort in management¹⁰⁴.

2. Secondary HLH, or IEC-HS

Secondary HLH has been reported following tisa-cel for ALL, axi-cel for LBCL, ide-cel, and cilta-cel^9,18,32, with infrequent resulting deaths following axi-cel, ide-cel, and cilta-cel in ~1% or less of patientsl^9,16,18. CAR T-cell HLH is a life-threatening hyperinflammatory syndrome of macrophage activation, highly elevated ferritin, hypertriglyceridemia, cytopenias, coagulopathy, pulmonary compromise, and renal and/or hepatic dysfunction, which has appeared in some cases to represent the severe end of the spectrum of CRS. The “CARTOX” group initially proposed diagnostic criteria, a grading scale, and management guidelines for secondary HLH related to CAR T-cell therapy, including a stepwise approach of tocilizumab, escalating dose of corticosteroids, and etoposide¹¹³. Clinical trials of anti-CD22 CAR T cells revealed a pattern of HLH that occurred following improvement or resolution of CRS and appeared to have distinct pathophysiology¹. However, HLH is thought to occur rarely in absence of prior CRS^114,115, so the two disorders likely have shared mechanisms. Also, a bone marrow biopsy documenting hemophagocytosis is not usually required to define HLH, so in many cases it could be difficult to strictly separate severe CRS from HLH. Baseline patient natural killer-cell lymphopenia is associated with subsequent development of HLH, possibly due to a lack of dampening of CD8+ T-cell hyperactivation allowing unrestricted CD8+ T-cell expansion and delayed contraction of the CAR T-cell population¹¹⁴.

A consensus definition, grading system and management recommendations have been proposed by an expert panel working with the ASTCT¹¹⁵. The terminology immune effector cell-associated hemophagocytic lymphohistiocytosis-like syndrome (IEC-HS) has been suggested to describe this entity¹¹⁵. The pathophysiology shares many features with CRS, including overlapping cytokine profiles, but the underlying causes remain an area of ongoing research¹¹⁵. The definition of IEC-HS requires a serum ferritin elevated at least twice upper limit of normal or rapidly rising, but otherwise allows a broad possibility of other typical findings, such as cytopenias, transaminitis, hypofibrinogenemia, or tissue evidence of hemophagocytosis¹¹⁵. Notably, all of these findings except hemophagocytosis can also occur with CRS. Management of IEC-HS per these consensus recommendations is summarized in Table 5. Anakinra, with or without corticosteroids, appears effective for treating HLH in some cases^1,114. Anakinra and corticosteroids in increasing doses are typically used first-line, with ruxolitinib, low-dose systemic etoposide, and emapalumab being alternative agents, though management is evolving¹¹⁵.

IV. Immunologic and infectious toxicities

1. B-cell aplasia and hypogammaglobulinemia

B-cell aplasia is an important cause of immunocompromise in post anti-CD19 CAR T-cell therapy patients. Rates of B-cell recovery versus persistent aplasia vary with the anti-CD19 CAR T-cell product and with the malignancy being treated (Table 1). B-cell recovery in absence of malignancy recurrence occurs frequently following products such axi-cel or products containing the axi-cel construct for non-Hodgkin lymphoma (NHL)^14,116,117 , tisa-cel for NHL¹¹⁸, or brexu-cel for mantle cell lymphoma²⁴ or adult ALL²⁵. However, B cell recovery after tisa-cel treatment, particularly within 6–12 months infusion, is a biomarker of relapse risk in pediatric or young-adult patients with B-cell ALL ³². Long-term B-cell aplasia is an important determinant of vaccine response in patients who have received CAR T-cell therapy. For example, patients with persistent B-cell aplasia may have no humoral response to SARS-CoV-2 vaccination, although T-cell immune responses are possible¹¹⁹.

Hypogammaglobulinemia frequently occurs in patients following both anti-CD19 and anti-BCMA CAR T-cell therapy, with variable percentages of patients receiving immunoglobulin replacement^14,15,25,120. B-cell and immunoglobulin recovery can occur in the setting of LBCL slowly over a period of years¹¹⁷. However, even in the setting of ongoing B-cell aplasia, some level of humoral immunity may be retained after anti-CD19 CAR T-cell therapy, likely due to persistence of CD19-negative long-lived plasma cells, as reflected by preserved antiviral antibodies¹²¹. While long-lived plasma cells may persist after anti-CD19 CAR T cells¹²², humoral immunity to certain pathogens may decrease in the months after anti-CAR T-cell therapy¹²³. Additionally, response to vaccination may be impaired after CAR T-cell infusion, as demonstrated by lack of humoral response to pneumococcal conjugate vaccine in the first six months after anti-CD19 CAR T-cell therapy for lymphoma¹²³. Humoral immunity, on the other hand, may be compromised after anti-BCMA CAR T-cell therapy, as CD19-negative-BCMA⁺ long-lived plasma cells may be depleted, leading to higher rates of prolonged hypogammaglobulinemia in these patients^124,125. Firm evidence supporting the role of immunoglobulin replacement therapy in preventing infection is lacking. Retrospective data indicate that, in the setting of BCMA-targeting bispecific antibodies, intravenous immunoglobulin replacement (IVIG) substantially decreases rates of grade 3-5 infections¹²⁶, though the question is still understudied in the CAR T-cell setting. Guidelines developed by expert opinion suggest immunoglobulin replacement to a goal of 400 mg/dL indefinitely following anti-BCMA CAR T-cell therapy, in children who have received anti-CD19 CAR T-cell therapy (due to a less robust plasma cell repertoire), in the first 3 months after anti-CD19 CAR T-cell therapy in adults, and for any post-CAR-T-cell patient with recurrent infections¹²⁴.

2. Infections

Grade 3 or higher bacterial, viral and fungal infections are common, occurring in an estimated 5-32% of patients (Table 1) with infectious deaths being a leading cause of mortality after CAR T-cell infusion, and with infection-related deaths occurring in an estimated 1-12% of patients with in patient groups with a median of 10-16 months of follow-up^{13,15,16,24,25,32,88}. Predisposing factors for infection following CAR T-cell therapy are multifactorial, including CD4 T-cell lymphopenia (which may be prolonged more than a year)¹²⁷ induced by lymphodepletion chemotherapy, B-cell aplasia following anti-CD19 CAR T-cell therapy, plasma cell depletion following anti-BCMA CAR T-cell therapy, immunosuppressive therapies given to control CAR T-cell toxicities, and other underlying immune defects related to the hematologic malignancy and prior treatments¹²⁵. Corticosteroid and other immunosuppressive therapies may additionally suppress the patient’s ability to mount fevers or exhibit other signs of infection, making monitoring for occult infection per institutional guidelines (e.g., surveillance blood cultures) necessary⁴⁶. Pretreatment risk factors for post-CAR-T-cell infections include malignancy type (higher with ALL than NHL), higher number of prior treatment regimens, prior allo-HSCT, receipt of bridging chemotherapy, and high CAR-HEMATOX score^125,128.

Reports of whether tocilizumab increases infection risk in the setting of cellular therapy are conflicting, though likely there is not independent risk^125,127,129; however, neutropenia with ANC < 500 cells/mcL for 14 days or more and corticosteroid use at dexamethasone 10 mg/day or equivalent for 9 days or more between Days 0-21 are independent risk factors for high-grade infection in the first month post-CAR-T, though risk may be ameliorated by gram negative antibacterial prophylaxis¹²⁸.

A diversity of post-CAR-T viral infections can occur, with respiratory viruses being most common¹²⁵. CMV reactivation is especially a concern following intensive lymphodepletion with anti-CD52 agents, as has been reported in prior trials of allogeneic off-the-shelf products to prevent rejection of the allogeneic cells¹³⁰. Reactivation of polyomavirus JC and human herpes virus-6 (HHV-6) have additionally caused deaths related to viral encephalopathies in profound lymphodepletion^13,32,87. A report of fatal hepatitis B virus (HBV) reactivation occurring one year after CAR T-cell infusion suggests that prophylactic HBV antiviral therapy may need to be continued indefinitely after CAR T-cell therapy, due to persistent B-cell depletion⁸⁹. Fortunately, no cases of replication-competent retroviruses have been reported^14,25,86,131.

With the emergence of SARS-CoV-2, infectious deaths due to COVID-19 disease have occurred following both anti-CD19 and anti-BCMA CAR T-cell therapy^18,88,89. Patients who have received prior CAR T-cell therapy have high likelihood for severe disease and death due to covid-19^132,133, risk of protracted illness and persistent viral shedding¹³⁴, and poor humoral response to COVID-19 vaccines^132,135,136, though immunity can be increased with booster doses¹³⁵. T-cell immune responses to vaccines are hypothesized to offer some protection against severe COVID-19 illness¹¹⁹. Because post CAR-T-cell patients are especially vulnerable, pre- and post-CAR-T vaccination, best available prophylactic strategies to prevent severe infection, and treatment strategies for those who are actively infected are essential for this patient population¹³² (Table 5). Prophylaxis and early intervention for infectious toxicities is evolving, with substantial variation among centers (Table 5).

V. Toxicities of increasing relevance

1. Damage to normal tissue expressing the target antigen

As novel antigen targets are evaluated to increase the types of malignancies that can be treated with CAR T cells, and to provide alternative therapies for patients with hematologic malignancies with relapse post-CAR-T, the possibility of CAR T cells damaging normal tissues expressing the target antigen (so-called “on target, off tumor” effects) is increasing.

Anti-GPRC5D CAR T cells have been evaluated for multiple myeloma, including in the setting of relapse after anti-BCMA CAR T-cell therapy. GPRC5D is expressed on both normal and malignant plasma cells and additionally has low expression on skin hair follicles and hard keratinizing tissue¹³⁷. As a result, expected toxicities reported in clinical trials have included manageable rashes, nail changes, dysgeusia, and dry mouth^2,3,138. However, a serious cerebellar disorder, manifesting as dizziness, ataxia, dysarthria, and difficulties with visual fixation has been described and was thought to be due CAR T-cell targeting of GPRC5D-expressing cells in the inferior olivary nucleus in the medulla oblongata². The disorder appeared to be dose-related and potentially avoidable at lower cell doses².

Development of CAR T-cell therapy for hematologic diseases other than B-cell and plasma-cell malignancies has been challenging due to the detrimental effects of eliminating normal hematopoietic and immune cells, which may necessitate hematopoietic rescue with allogeneic stem cell transplant. Design of CAR T cells designed to target myeloid malignancies is fraught with nonideal antigen targets, such as CD33 and CD123, which are expressed on normal hematopoietic stem cells and myeloid progenitor cells, CD38, which is expressed on B, T, and NK cells, and CLL-1, which is expressed on lung and epithelial cells¹³⁹. Treatment of solid tumor malignancies with CAR T cells may be especially challenging, as few targets have expression limited to malignant cells. As an example, anti-mesothelin CAR T cells have also been demonstrated to have substantial pulmonary toxicity at high doses due to mesothelin expression on pulmonary epithelial cells¹⁴⁰.

2. Second malignancy risk

Myeloid malignancies and solid tumors

Myeloid malignancies are well-known to occur after CAR T-cell therapy but are usually attributed to patients’ prior cytotoxic chemotherapy regimens and/or autologous stem cell transplants rather than to the CAR T-cell product itself^{13,14,16,25,79,86}. A retrospective analysis of a cohort of 449 adult patients from the University of Pennsylvania who received commercial CAR T-cell therapy reported a 17% estimated risk for second malignancies, including 2.3% risk for second hematologic malignancies, over a 5-year period, which is similar to what would be expected in these patient populations receiving non-CAR-T therapies¹⁴¹. However, a boxed warning concerning “secondary hematological malignancies, including myelodysplastic syndrome and acute myeloid leukemia,” was added to the US FDA package insert for cilta-cel in December 2023, after 10 of 97 patients on CARTITUDE-1 (10.3%) developed second myeloid malignancies, demonstrating regulatory and industry concern about the rates of secondary myeloid malignancies after treatment with this product^142,143. While direct comparisons are difficult, this rate of second myeloid malignancies is numerically higher than the rate of 5-7% reported for patients with multiple myeloma¹⁴⁴. Pre-lymphodepletion clonal hematopoiesis predisposes to later development of secondary myeloid malignancies⁷². It is possible that, as CAR T-cell therapy moves earlier in hematologic malignancy treatment algorithms, risk of secondary myeloid malignancies will decrease, as patients have less exposure to prior cytotoxic chemotherapy.

T-cell malignancies

In November 2023, the US FDA published a press release stating it was investigating risk of T-cell malignancies occurring after CAR T-cell therapy, with the risk being applicable to all 6 commercially available CAR T-cell products, and recommended life-long monitoring of patients who have received CAR T-cell therapy¹⁴⁵. In total, as of December 2023, 22 cases of T-cell malignancies after CAR T cell therapy have been reported to the US FDA¹⁴⁶. T-cell malignancies have been reported following administration of 5 of the 6 commercially available products¹⁴⁶. The CAR transgene has been detected in the malignant clone in three cases of T-cell malignancies, though genetic sequencing was not available in all patients¹⁴⁶.

In terms of published case reports of T-cell malignancies, CAR T-cell derived T-cell lymphoma was observed in 2 patients who received investigational CD19-targeting, donor-derived CAR T cells genetically modified using a Piggybac transposon system; however, no integration of the transgene into known oncogenes was identified^147,148. CAR T-cell derived T-cell lymphoma has additionally been reported in a patient 5 months after receipt of cilta-cel, which utilizes a lentiviral vector, with dominant insertion of the CAR transgene into PBX2¹⁴⁹. In this case, there was evidence of a pre-existing clonal population within the leukapheresis product, and TET2 and JAK3 mutations were present prior to manufacturing¹⁴⁹. Of 449 patients who received commercial CAR T-cell therapy at the University of Pennsylvania, one patient developed both a non-small cell lung cancer and a T-cell lymphoma after axi-cel¹⁴¹. The T-cell lymphoma was determined not to express the CAR transgene but did express a JAK3 variant of uncertain significance¹⁴¹.

Overall, with more than 27,000 infusions of FDA-approved CAR T-cell products having been administered in the U.S., CAR T-cell product-derived T-cell malignancies are extremely rare¹⁴⁶, and the pathogenesis is incompletely understood. Some patients treated with CAR T cells likely have an underlying preexisting genetic predisposition to develop malignancies Patients with B-cell malignancies are predisposed to development of T cell malignancies, and vice versa¹⁵⁰; thus, some cases of new T-cell malignancies not expressing the CAR transgene may be in part related more to the patient’s underlying genetic predisposition.

In considering treatment options and counseling patients, second malignancy risk must be weighed against the potential benefits of successful treatment of the original B-cell or plasma cell malignancy and against the risks of other chemotherapy treatments for that malignancy. Notably, the rate of post-treatment second malignancies in patients with lymphoma who have received an autologous stem cell transplant is high at baseline, with an estimated cumulative 10-year risk of 20% for all second malignancies¹⁵¹ and 4-6.8% for secondary myeloid malignancies^151,152. In most cases, the potential benefits of CAR T-cell therapy are likely to outweigh the risks of second malignancies¹⁵³.

VI. Designing less toxic CAR T cells

1. Reducing CRS, ICANS, and other inflammatory toxicities

While toxicity prevention and mitigation strategies have made CAR T-cell therapy more feasible and accessible to patients, the development of less toxic CAR T-cell therapies is the ultimate goal. Selected strategies for improved design of CAR T-cell therapies are depicted in Figure 2. Modification of various parts of the CAR have been explored for toxicity prevention. AUTO1, a CD19-targeted CAR T-cell product with a single chain variable fragment (scFv) with a low antigen affinity due to rapid dissociation, resulted in no incidence of ≥ grade 3 CRS in 20 patients with ALL¹⁵⁴. While the CD28 costimulatory domain has generally been considered a driver of severe toxicity for the CAR T-cell products axi-cel and brexu-cel^73,155, modifications to other CAR features have reduced the toxicity products containing CD28 costimulatory domains. Anti-CD19 CAR constructs with a CD28 costimulatory domain and a CD8α hinge and transmembrane domain appear to lead to lower levels of peak serum cytokines and decreased rates of severe neurologic toxicity, compared to CAR constructs with a CD28 hinge, transmembrane, and costimulatory domain^69,156. Early clinical data indicate that the incorporation of an intracellular Toll-like receptor 2 domain between the CD28 costimulatory and T-cell activation domains results in reduced IFN-γ production and low rates of both CRS and ICANS¹⁵⁷. Reducing the CD3ζ immunoreceptor tyrosine-based activation motifs from 3 to 2 has resulted in low rates of CRS and ICANS in a trial of 19(T2)28z1xx CAR T cells for LBCL¹⁵⁸.

Figure 2: Selected strategies for improving CAR T-cell safety.

Strategies for improving CAR T-cell safety include structural changes to the scFv, hinge, transmembrane, costimulatory, and CD3ζ T-cell activation domains. Components which may be encoded into the viral vector include a destabilizing domain at the C-terminal end of the CAR, secreted proteins that neutralize inflammatory cytokines, and a “NOT” logic gate, which prevents destruction of normal tissue expressing the target antigen by preventing activation of the CAR if a second antigen present only on normal cells and not tumor cells (yellow circle) is present. Base editing technologies are being explored as a substitute for gene-editing technologies, as they can be used to knockdown TCRs and T-cell- expressed antigens without inducing double-strand DNA breaks and might avoid insertional mutagenesis. Note that these strategies are not meant to be employed in the same CAR T-cell but are depicted together in the figure for sake of explanation.

Additional safety components incorporated into the vector and changes to the cell culture process have been found to change inflammatory toxicity profiles of CAR-expressing cells. Suicide genes and druggable safety switches have anecdotal efficacy in eliminating most in vivo CAR T cells as a last resort for toxicity management¹⁵⁹. However, such strategies of CAR T-cell elimination have very limited human data, and overall effectiveness in toxicity management is unknown. Additionally, drug-regulatable platforms, such as drug-controlled destabilizing domains at the C-terminal of the CAR protein that can influence CAR surface expression, have the potential to reversibly inhibit and subsequently restore CAR T-cell activity⁶⁶. CAR T cells can be designed to autonomously secrete proteins that block inflammatory cytokines, such as IL-1 and IL-6; the sequences for these proteins can be incorporated into the same vector as the CAR, utilizing ribosomal skip domains¹⁶⁰. Selection of and promotion of T cells with a less differentiated phenotype in culture during the manufacturing process, which yields more effective CAR T cells, also might lead to more manageable safety profiles¹⁶¹.

2. Reducing potential for on-target damage to normal tissues and insertional mutagenesis

Engineering of CAR T cells that avoid damage to normal tissues expressing the target antigen is an area of investigation. Low antigen affinity CARs are being evaluated for the potential to prevent CAR T-cell killing of normal tissues with relatively low target antigen expression compared to the strongly antigen expressing malignant tissues¹⁶². Logic-gating techniques may also prevent targeting of normal tissues. These include incorporation of a secondary inhibitory CAR that prevents CAR T-cell activation when encountering a specific antigen that is found on normal cells but not tumor cells (“NOT-gate”)¹⁶³. Similarly, CAR T cells can be designed with both a CAR and a chemokine receptor (CCR) for a second tumor-associated antigen, such that both the CAR and the CCR must be engaged by their target antigens simultaneously for the CAR T cells to become activated^163,164.

Multiple methods are in development to reduce the need for gene-editing technology and decrease risk of insertional mutagenesis, especially in the case of healthy donor-derived CAR T cells for which multiple gene editing steps have been necessary to avoid rejection and graft versus host disease (GvHD)^130,165. Base-editing techniques may be a viable substitution for gene-editing technologies, such as CRISPR and TALEN, that generate DNA breaks¹⁶⁶. In the case of anti-CD7 CAR T cells for T-cell leukemia and lymphoma, a second step of gene editing to knock-down CD7 expression on CAR T cells to avoid fratricide may be rendered unnecessary by naturally selecting CAR T cells from culture with intracellular CD7 sequestration or CAR-mediated epitope masking¹⁶⁷.

VII. Conclusions:

With thousands of patients having been treated with CAR T-cell therapy, the management of acute toxicities such as CRS and ICANS has progressed, and cellular therapy specialists at many centers are comfortable managing these toxicities. Understanding of the pathophysiology of some of the previously less well-recognized disorders, such as ICAHT, IEC-HS, and movement disorders remains limited, and treatment strategies are evolving. Continued prospective evaluation of toxicity intervention approaches and the updating of treatment algorithms will be essential in the coming years. Development of novel CAR T-cell constructs and products with favorable toxicity profiles is hoped to make this ground-breaking therapy more accessible to patients who need it.